Fibre Reinforced Concrete: Improvements and Innovations: RILEM-fib International Symposium on FRC (BEFIB) in 2020 [1st ed.] 9783030584818, 9783030584825

Citation preview

RILEM Bookseries

Pedro Serna Aitor Llano-Torre José R. Martí-Vargas Juan Navarro-Gregori   Editors

Fibre Reinforced Concrete: Improvements and Innovations RILEM-fib International Symposium on FRC (BEFIB) in 2020

Fibre Reinforced Concrete: Improvements and Innovations

RILEM BOOKSERIES

Volume 30 RILEM, The International Union of Laboratories and Experts in Construction Materials, Systems and Structures, founded in 1947, is a non-governmental scientific association whose goal is to contribute to progress in the construction sciences, techniques and industries, essentially by means of the communication it fosters between research and practice. RILEM’s focus is on construction materials and their use in building and civil engineering structures, covering all phases of the building process from manufacture to use and recycling of materials. More information on RILEM and its previous publications can be found on www.RILEM.net. Indexed in SCOPUS, Google Scholar and SpringerLink.

More information about this series at http://www.springer.com/series/8781

Pedro Serna Aitor Llano-Torre José R. Martí-Vargas Juan Navarro-Gregori •

•

•

Editors

Fibre Reinforced Concrete: Improvements and Innovations RILEM-fib International Symposium on FRC (BEFIB) in 2020

123

Editors Pedro Serna ICITECH Universitat Politècnica de València Valencia, Spain

Aitor Llano-Torre ICITECH Universitat Politècnica de València Valencia, Spain

José R. Martí-Vargas ICITECH Universitat Politècnica de València Valencia, Spain

Juan Navarro-Gregori ICITECH Universitat Politècnica de València Valencia, Spain

ISSN 2211-0844 ISSN 2211-0852 (electronic) RILEM Bookseries ISBN 978-3-030-58481-8 ISBN 978-3-030-58482-5 (eBook) https://doi.org/10.1007/978-3-030-58482-5 © RILEM 2021 No part of this work may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilming, recording or otherwise, without written permission from the Publisher, with the exception of any material supplied specifically for the purpose of being entered and executed on a computer system, for exclusive use by the purchaser of the work. Permission for use must always be obtained from the owner of the copyright: RILEM. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Switzerland AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

Preface

2020, a weird year. I always considered that writing the preface for the BEFIB2020 Proceedings would be an interesting moment with a great responsibility to summarise in a few words what was expected from the most important symposium on Fibre Reinforced Concrete, which we can only enjoy once every four years. Unfortunately, the sudden appearance of the COVID-19 represented a drastic change in all approaches. Although we considered at the beginning that we could overcome this crisis, finally it was not possible to ensure optimal safety conditions and we decided to postpone the symposium for one year. As Organising Committee, we were aware of the effort already done in both writing and reviewing the scientific papers: a total of 153 papers were accepted from the 252 submitted abstracts. At this point, we considered that awaiting one additional year would have represented an excessive delay to maintain the scientific novelty. Therefore, to ensure the quality level of the conference and to take advantage of the great work done by the authors and the scientific committee during the peer review process, the BEFIB2020 Organising Committee decided to publish an online edition of the BEFIB2020 proceedings considering all the accepted full papers and make a new call for abstract for the next BEFIB2021 symposium edition. However, we assumed that one of the main objectives when submitting a paper for a conference is the oral presentation of the work to our colleagues and discuss about them in person. Therefore, considering the postponement, the Organising Committee gave the choice to the authors to withdraw or publish their contribution in the BEFIB2020 online publication without oral presentation. Consequently, some authors preferred to wait for another publishing opportunity to submit their work and decided to withdraw their contribution for this publication. This book entitled “Fibre Reinforced Concrete: Improvements and Innovations” is the result of the peer review process performed during 2020 and includes a total of 101 papers representing the contemporary topics of interest for the FRC researchers and practitioners. Fresh and hardened concrete properties are covered

v

vi

Preface

from rheology and early-age properties and technological aspects to the mechanical properties and quality control. But also, it can be observed an important effort on studying analytical and numerical models and a tendency to structural design, codes and standards covering interesting case studies and FRC structural applications which foreshadow a good future for the FRC. Durability and long-term properties and special FRC like UHPFRC or other smart FRC, textile concretes are still challenging topics that cover some needed advances or new concepts for the continuous progress. I would like to acknowledge the contribution of all the authors and their comprehension, as well as the great work done by the members of the scientific committee to guarantee high-quality publications. Without you this work could not have been done. I would also like to thank the support of my colleagues from the Organising Committee in the management, structuring and finalising of this publication, especially Aitor Llano-Torre for his constant and meticulous dedication and the collaboration of José R. Martí-Vargas. I would like also to acknowledge the sponsor companies KrampeHarex and SCE, FIBRAFLEX, ArcelorMittal Fibres, BASF Construction Solutions and RIMSA Metal Technology, for their support and confidence in our event. I wish we could enjoy a great even with their support next year in the BEFIB2021 edition. Finally, I want to finish these words with my most alive desire to beat the COVID-19 and to achieve the objective of our congress, which is none other than to bring together all the people who work at the FRC in the same room. Please, be careful, stay safe and work hard since I would like to welcome all of you next year in Valencia for the BEFIB 2021. I hope that the quality of this work content will cover the interest of the readers and the authors objectives. Thanks to all of you for your great contribution. Pedro Serna BEFIB2020 Chairman

Organisation

Committees Organising Committee Pedro Serna (Chairman), Spain Aitor Llano-Torre (Secretariat), Spain José R. Martí-Vargas, Spain Juan Navarro-Gregori, Spain Scientific Committee M. A. Aiello, Italy A. Aguado, Spain C. Aldea, Canada S. Al-Toubat, UAE S. Austin, UK G. L. Balázs, Hungary N. Banthia, Canada B. Barragan, France J. A. O. Barros, Portugal E. S. Bernard, Australia A. Bettencourt Ribeiro, Portugal S. Billington, USA J. Bolander, USA P. Borges, Brazil W. P. Boshoff, South Africa N. Buratti, Italy S. H. P. Cavalaro, UK J. P. Charron, Canada A. Conforti, Italy

E. Cuenca, Italy F. Dehn, Germany A. De La Fuente, Spain E. Denarie, Switzerland M. di Prisco, Italy Y. Ding, China A. Fantilli, Italy L. Ferrara, Italy A. Figueiredo, Brazil S. Foster, Australia J. Gálvez, Spain E. Garcia-Taengua, UK R. Gettu, India G. M. Giaccio, Argentina S. Grunewald, Netherlands P. Kabele, Czech Republic T. Kanda, Japan T. Kanstad, Norway I. Khan, Saudi Arabia

vii

viii

K. Kobayashi, Japan M. Konsta-Gdoutos, Greece M. Kunieda, Japan V. Li, USA Y. M. Lim, South Korea A. Llano-Torre, Spain I. Löfgren, Sweden K. Lundgren, Sweden P. Lura, Switzerland J. R. Marti-Vargas, Spain B. Massicotte, Canada C. Mazzotti, Italy A. Meda, Italy V. Mechtcherine, Germany F. Minelli, Italy B. Mobasher, USA C. Molins, Spain A. Naaman, USA J. Navarro-Gregori, Spain B. Nematollahi, Australia G. Parra-Montesinos, USA G. Plizzari, Italy D. Redaelli, Switzerland J. Resplendino, France

Organisation

K. A. Rieder, Germany M. Roig-Flores, Spain P. Rossi, France E. Schlangen, Netherlands P. Serna (Chair), Spain S. Shah, USA F. Silva, Brazil V. Slowik, Germany S. Soleimani-Dashtaki, Canada L. Sorelli, Canada H. Stang, Denmark J. Sustersic, Slovenia G. Tiberti, Italy R. D. Toledo Filho, Brazil F. Toutlemonde, France L. Vandewalle, Belgium G. Van Zijl, South Africa J. Walraven, Netherlands Y. Yao, China D.-Y. Yoo, South Korea Y.-S. Yoon, South Korea C. Zanotti, Canada R. Zerbino, Argentina

Institutions Organised by

UPV Universitat Politècnica de València

ICITECH Institute of Concrete Science and Technology

Organisation

ix

Supported By

RILEM International Union of Laboratories and Experts in Construction Materials, Systems and Structures

Sponsors Gold Sponsor

fib The International Federation for Structural Concrete

ACI Amercian Concrete Institute

x

Bronze Sponsors

Organisation

Contents

Rheology and Early-Age Properties Influence of Different Fibre Types on the Rheology of Strain Hardening Cementitious Composites . . . . . . . . . . . . . . . . . . . . . . . . . . . Hassan Baloch, Steffen Grünewald, Karel Lesage, and Stijn Matthys

3

Using Fiber Reinforced Concrete to Control Early-Age Shrinkage in Replacement Concrete Pavement . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nakin Suksawang and Daniel Yohannes

12

Early Age Shrinkage Crack Distribution in Concrete Plates Reinforced with Different Steel Fibre Types . . . . . . . . . . . . . . . . . . . . . . Sébastien Wolf, Simon Cleven, and Oldrich Vlasák

24

Technological Aspects Influence of Synthetic Fibres on Seismic Resistance of Reinforced Concrete Sections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E. Stefan Bernard Development and Mechanical Characterization of Dry Fiberreinforced Concrete for Prefabricated Prestressed Beams . . . . . . . . . . . Kamyar Bagherinejad Shahrbijari, Suman Saha, Joaquim A. O. Barros, Isabel B. Valente, Salvador Dias, and João Leite Simulation of Fibre Orientation in Self-compacting Concrete: Case Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Thomas Bauwens, Steffen Grünewald, and Geert De Schutter Mix Design and Properties of Self-compacting Fibrous Concrete . . . . . . Rafael R. Polvere, Ana R. L. Pires, Sidiclei Formagini, and Andrés B. Cheung

37

49

64 75

xi

xii

Contents

Aligned Interlayer Fibre Reinforcement for Digital Fabrication with Concrete . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Lukas Gebhard, Jaime Mata-Falcón, Tomislav Markić, and Walter Kaufmann Mixture Proportioning of Steel Fibre Reinforced Self-compacting Concrete Based on the Compressible Packaging Method: Comparison with ACI 237R-07 and RILEM TC 174-SCC Recommendations . . . . . . M. G. Cardoso, R. M. Lameiras, T. T Oliveira, F. B. Santana, and V. M. S. Capuzzo

87

99

Evaluation the Yield and Ultimate Strain of FRC in Compression . . . . 111 Salam Wtaife, Nakin Suksawang, and Ahmed Alsabbagh Electromagnetic Shielding Characteristics of High Performance Fiber Reinforced Cementitious Composites . . . . . . . . . . . . . . . . . . . . . . 123 Namkon Lee, Sungwook Kim, and Gijoon Park Mechanical Properties The Manufacture of Fiber Cement Blocks Using Chemical and Thermomechanical Pulps and Rice Husk Ash . . . . . . . . . . . . . . . . . 133 Javad Torkaman Post-Fire Flexural Tensile Strength of Macro Synthetic Fibre Reinforced Concrete . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140 Olivia Mirza, Brendan Kirkland, Kurt Bogart, and Todd Clarke Experimental Investigation on the Cyclic Behaviour of Steel Fibre Reinforced Concrete Under Bending . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 Maure De Smedt, Rutger Vrijdaghs, Els Verstrynge, Kristof De Wilder, and Lucie Vandewalle Effect of Test Setups on the Shear Transfer Capacity Across Cracks in FRC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163 Alejandro Giraldo Soto and Walter Kaufmann Bearable Local Stress of High-Strength SFRC . . . . . . . . . . . . . . . . . . . . 176 Sven Plückelmann, Rolf Breitenbücher, Mario Smarslik, and Peter Mark Impact Response of Different Classes of Fibre Reinforced Concrete . . . 189 Juan C. Vivas, Raúl L. Zerbino, María C. Torrijos, and Graciela M. Giaccio An Experimental Study on the Fatigue Failure Mechanisms of Pre–damaged Steel Fibre Reinforced Concrete at a Single Fibre Level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199 Humaira Fataar, Riaan Combrinck, and William P. Boshoff

Contents

xiii

Development of an HPFRC for Use in Flat Slabs . . . . . . . . . . . . . . . . . . 209 Julia Blazy, Sandra Nunes, Carlos Sousa, and Mário Pimentel Influence of the Steel Fibres on the Tension and Shear Resistance of Anchoring with Anchor Channels and Channel Bolts Cast in Concrete . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221 Mazen Ayoubi, Christoph Mahrenholtz, and Wilhelm Nell Fiber Reinforced Concrete After Elevated Temperatures: Techniques of Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233 Ronney Rodrigues Agra, Ramoel Serafini, and Antonio Domingues de Figueiredo Influence of the Curing Temperatures on the Mechanical Properties of Hemp Fibre-Reinforced Alkali-Activated Mortars . . . . . . . . . . . . . . . 245 Bojan Poletanovic, Gergely Nemeth, and Ildiko Merta Equivalence Between Flexural Toughness and Energy Absorption Capacity of FRC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253 Sergio Carmona and Climent Molins Alkali Resistant (AR) Glass Fibre Influence on Glass Fibre Reinforced Concrete (GRC) Flexural Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . 262 S. Guzlena and G. Sakale Fiber Reinforced Concrete Crack Opening Evaluation Using Digital Image Correlation Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 270 Kaio Cézar da Silva Oliveira, Gabriela Silva Dias, Isadora Queiroz Freire de Carvalho, Wandersson Bruno Alcides de Morais Silva, Danilo José Pereira Freitas, Christiano Augusto Ferrario Várady Filho, and Aline da Silva Ramos Barboza Effect of Distribution and Orientation of Fibers on the Post-cracking Behavior of Steel Fiber Reinforced Self-compacting Concrete in Small Thickness Elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279 Néstor Fabián Acosta Medina, Rodrigo de Melo Lameiras, Ana Carolina Parapinski dos Santos, and Fábio Luiz Willrich Ductility of the Four-Year-Old Steel Fibre Reinforced Concrete . . . . . . 290 Jakob Šušteršič, Rok Ercegovič, David Polanec, and Andrej Zajc Sensitivity of the Flexural Performance of Glass and Synthetic FRC to Fibre Dosage and Water/Cement Ratio . . . . . . . . . . . . . . . . . . . . . . . 301 Razan H. Al Marahla and Emilio Garcia-Taengua Bond Between Steel Reinforcement Bars and Fiber Reinforced Cement-Based Composites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313 Margareth S. Magalhães, Paulo José B. Teixeira, and Maria Elizabeth N. Tavares

xiv

Contents

An Experimental Study of the Influence of Moderate Temperatures on the Behavior of Macrosynthetic Fiber Reinforced Concrete . . . . . . . 322 Marta Caballero-Jorna, Marta Roig-Flores, and Pedro Serna Post-cracking Behaviour of Glass Fibre Reinforced Concrete with Recycled Aggregates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 333 Brecht Vandevyvere, Lucie Vandewalle, Els Verstrynge, and Jiabin Li Long-Term Properties A Computational Sectional Approach for the Flexural Creep Behavior of Cracked FRC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 347 Rutger Vrijdaghs, Marco di Prisco, and Lucie Vandewalle Shrinkage of Steel-Fibre-Reinforced Lightweight Concrete . . . . . . . . . . 359 Hasanain K. Al-Naimi and Ali A. Abbas Time Dependent Deflection of FRC Members Under Sustained Axial and Flexural Loading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 368 Murray Watts, Ali Amin, R. Ian Gilbert, and Walter Kaufmann Influence of the Residual Tensile Strength on the Factor for Quasi-permanent Value of a Variable Action w2 . . . . . . . . . . . . . . . 380 Darko Nakov, Goran Markovski, Toni Arangjelovski, and Peter Mark Compressive and Tensile Creep and Shrinkage of Synthetic FRC: Experimental Results and Comparison to Codes . . . . . . . . . . . . . . . . . . 392 Razan H. Al Marahla and Emilio Garcia-Taengua Creep in FRC – From Material Properties to Composite Behavior . . . . 402 Martin Hunger, Jürgen Bokern, Simon Cleven, and Rutger Vrijdaghs Durability Morphology of Corrosion of Metallic Fibers in Aggressive Media . . . . . 417 Carmen Andrade and Miguel A. Sanjuán Effects of Fibres on the Flexural Behaviour of Sound and Damaged RC Beams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 423 Raúl L. Zerbino, María C. Torrijos, Graciela M. Giaccio, and Antonio Conforti Fiber Reinforced Concrete Elements Exposed to Accelerated Corrosion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 433 Camelia Negrutiu, Ioan Sosa, Bogdan Heghes, Oana Gherman, and Horia Constantinescu Effect of Corroded Steel Fibers on Mechanical Behavior of Steel Fiber Reinforced Concrete . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 445 Minoru Kunieda, Masaki Tsutsui, and Le V. Tri

Contents

xv

Self-healing of Fibre Reinforced Concrete Containing an Expansive Agent in Different Exposure Conditions . . . . . . . . . . . . . . . . . . . . . . . . . 453 K.-S. Lauch, C. Desmettre, and J.-P. Charron Characterisation of Strain-Hardening Cementitious Composite (SHCC) Under Cyclic Loading Conditions for Self-healing Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 466 Zixuan Tang, Chrysoula Litina, and Abir Al-Tabbaa Corrosion Pattern and Mechanical Behaviour of Corroded Rebars in Cracked Plain and Fibre Reinforced Concrete . . . . . . . . . . . . . . . . . . 477 E. Chen, Carlos G. Berrocal, Ingemar Löfgren, and Karin Lundgren Evaluation of the Self-healing Capability of Ultra-High-Performance Fiber-Reinforced Concrete with Nano-Particles and Crystalline Admixtures by Means of Permeability . . . . . . . . . . . . . . . . . . . . . . . . . . 489 Hesam Doostkami, Marta Roig-Flores, Alberto Negrini, Eduardo J. Mezquida-Alcaraz, and Pedro Serna Analytical and Numerical Models Material Characterisation for Nonlinear Finite Element Analysis (NLFEA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 503 P. J. van der Aa and A. A. van den Bos Comparison Between the Cracking Process of Reinforced Concrete and Fibres Reinforced Concrete Railway Tracks by Using Non-linear Finite Elements Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 515 Jean-Louis Tailhan, Pierre Rossi, and Thierry Sedran Experimental/Computational-Based Determination of Material Parameters for Nonlinear Simulation of UHPFRC . . . . . . . . . . . . . . . . . 527 David Lehký, Martin Lipowczan, Drahomír Novák, Radomír Pukl, and Milad Hafezolghorani Mechanical Response of High Strength Fibre Reinforced Concrete Under Extreme Loads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 536 B. Luccioni, F. Isla, F. Fiengo, R. Codina, D. Ambrosini, J.C. Vivas, Raúl L. Zerbino, Graciela M. Giaccio, and María C. Torrijos Numerical Damage Modelling of Macro-synthetic Fibre Reinforced Concrete . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 548 Dayani Kahagala Hewage, Christophe Camille, Olivia Mirza, Fidelis Mashiri, Brendan Kirkland, and Todd Clarke Finite Element Analysis of Ultra High Performance Fibre Reinforced Concrete Beams Using Microplane Modelling . . . . . . . . . . . . . . . . . . . . 558 William. Wilson and Tomas O’Flaherty

xvi

Contents

Machine Learning Prediction of Flexural Behavior of UHPFRC . . . . . . 570 Joaquín Abellán-García, Jaime A. Fernández-Gómez, Nancy Torres-Castellanos, and Andrés M. Núñez-López Study of Dimensioning Aspects of FRC Based on the Beam Flexion Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 584 Iva E. Pereira Lima and Aline S. Ramos Barboza Load-Carrying Capacity of SFRC Suspended Slabs with Different Support Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 596 Olugbenga B. Soyemi and Ali A. Abbas Discrete Element Simulation of the Fresh State Steel Fiber Reinforced Self-compacting Concrete . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 610 A. Najari, A. Blanco, A. de la Fuente, and S. H. P. Cavalaro Crack Width Simulation and Nonlinear Finite Element Analysis of Bursting and Spalling Stresses in Precast FRC Tunnel Segments Under TBM Thrust Jack Forces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 621 Mehdi Bakhshi and Verya Nasri Finite Element Modelling of UHPFRC FlexuralReinforced Elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 639 Eduardo J. Mezquida-Alcaraz, Juan Navarro-Gregori, and Pedro Serna Punching Shear Response of RC Slab-Column Connections Strengthened with UHPFRC - Finite Element Investigation . . . . . . . . . . 651 Demewoz W. Menna and Aikaterini S. Genikomsou Numerical Evaluation the Effect of Specimen Thickness on Fibre Orientation in Self-consolidating Engineered Cementitious Composites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 661 Hai Tran Thanh, Jianchun Li, and Y. X. Zhang Numerical Modelling of Fiber-Reinforced Concrete Shear-Critical Beams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 670 Santiago Talavera-Sánchez, Juan Navarro-Gregori, Francisco Ortiz-Navas, and Pedro Serna Experimental Analysis of Crack Development of an UHPC Wall Element Under Shear Loading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 681 V. Příbramský, M. Kopálová, and L. Dlouhý Assessment of the Shear Behaviour of Fibre Reinforced Concrete Through Numerical Modelling of Shear-Friction Theory . . . . . . . . . . . . 693 Álvaro Picazo, Marcos G. Alberti, Alejandro Enfedaque, and Jaime C. Gálvez

Contents

xvii

Modeling the Compressive Behavior of Steel Fiber Reinforced Concrete Under High Strain Rate Loads . . . . . . . . . . . . . . . . . . . . . . . . 703 Honeyeh Ramezansefat, Mohammadali Rezazadeh, Joaquim A. O. Barros, Isabel B. Valente, and Mohammad Bakhshi Structural Design Post-cracking Strength Classification of Macro-synthetic Fibre Reinforced Concrete for Sleeper Application . . . . . . . . . . . . . . . . . . . . . 717 Christophe Camille, Dayani Kahagala Hewage, Olivia Mirza, Fidelis Mashiri, Brendan Kirkland, and Todd Clarke Structural Behaviour of Steel-Fibre-Reinforced Lightweight Concrete . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 730 Hasanain K. Al-Naimi and Ali A. Abbas Experimental Analysis of Beams Produced in Self-compacting Concrete Reinforced with Different Contents of Steel Fibres . . . . . . . . . 745 Ana R. L. Pires, Rafael R. Polvere, Sidiclei Formagini, and Andrés B. Cheung Innovation in Durable Segments for CSO Tunnels . . . . . . . . . . . . . . . . . 757 Ralf Winterberg, Michael R. Garbeth, and Brian Glynn Incorporation of Rate-Dependent Fracture Properties in the Design of Precast Concrete Tunnel Segment with Hybrid Reinforcement . . . . . 770 Stefie J. Stephen and Ravindra Gettu Codes and Standards Developments and Standardisation of Flowable Concrete Reinforced with Fibres for Structural Design, Update of fib TG 4.3 . . . . . . . . . . . . 779 Steffen Grünewald, Liberato Ferrara, and Frank Dehn Assesment of Codal Provision for SFRC Beam in Minimum Shear . . . . 791 Kranti Jain and Bichitra S. Negi Laboratory Investigations on the Installation of Fasteners in Fiber Reinforced Concrete . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 801 Panagiotis Spyridis, Lars Walter, Julia Dreier, and Dirk Biermann Quality Control Applications of Statistical Process Control in the Evaluation of QC Test Data for Residual Strength of FRC Samples of Tunnel Lining Segments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 815 Chidchanok Pleesudjai, Devansh Patel, Mehdi Bakhshi, Verya Nasri, and Barzin Mobasher

xviii

Contents

Using Decades of Data to Rethink Proportioning and Optimisation of FRC Mixes: The OptiFRC Project . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 827 Emilio Garcia-Taengua Case Studies: Structural and Industrial Applications Fire Resistance of Steel Fibre Reinforced Concrete Elevated Suspended Slabs: ISO Fire Tests and Conclusions for Design . . . . . . . . 841 Xavier Destrée, Andrejs Krasnikovs, and Sébastien Wolf Structural Behavior of a Traditional Concrete and Hollow Tiles Mixed Floor Reinforced with HPFRC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 852 D. Sirtoli, P. Riva, and P. Girardello Pedestrian Bridge over Las Vegas Avenue in Medellín. First Latin American Infrastructure in UHPFRC . . . . . . . . . . . . . . . . . . . . . . . . . . 864 Joaquín Abellán-García, Andrés M. Núñez-López, and Samuel E. Arango-Campo First Experimental Full-Scale Elevated FRSCC Slab in South America . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 873 Luis Segura-Castillo, Diego Figueredo, Iliana Rodríguez, and Nicolás García Masonry Walls Strengthened with Fiber Reinforced Concrete Subjected to Blast Load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 883 Salah Altoubat, Abdul Saboor Karzad, Moussa Leblouba, Mohamed Maalej, and Pierre Estephane Smart FRCs Interfacial Bond Quality in Functionally Graded Concretes Incorporating Steel Fibres and Recycled Aggregates . . . . . . . . . . . . . . . 897 Ricardo Chan, Isaac Galobardes, and Charles K. S. Moy Towards Rebar Substitution by Fibres – Tailored Supercritical Fibre Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 908 Katharina Look, Peter Heek, and Peter Mark A Constitutive Model for Steel-Fibre-Reinforced Lightweight Concrete . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 920 Hasanain K. Al-Naimi and Ali A. Abbas UV-C Treatment to Functionalize the Surfaces of Pet and PP Fibers for Use in Cementitious Composites. Adherence Evaluation . . . . . . . . . 938 María E. Fernández, María E. Pereira, Fernando Petrone, Claudia Chocca, and Gemma Rodríguez

Contents

xix

Potential of Using Recycled Carbon Fibers as Reinforcing Material for Fiber Concrete . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 949 Magdalena Kimm, Amna Sabir, Thomas Gries, and Piyada Suwanpinij Textile Reinforced Concrete (TRC) Development of Textile Reinforced UHPC with Reduced Steel Fiber Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 963 Mengchao Zhai, Yiming Yao, Jingquan Wang, and Barzin Mobasher Reinforcement of Concrete with Glass Multifilament Yarns: Effect of the Impregnation on the Yarn Pull-Out Behaviour . . . . . . . . . 971 A. -C. Slama, J. -L. Gallias, and B. Fiorio Influence of Fibres Impregnation on the Tensile Response of Flax Textile Reinforced Mortar Composite Systems . . . . . . . . . . . . . . . . . . . . 983 Giuseppe Ferrara, Marco Pepe, Enzo Martinelli, and Romildo D. Tolêdo Filho Experimental Investigation of Mechanical Properties of Smart Textile Reinforced Concrete Pipes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 991 Gozdem Dittel, Michelle Wangler, Bastian Maiworm, and Thomas Gries UHPFRC, SHCC and ECC Full-Scale Construction Test for Improvement of RC Void Slab Bridges Using UHPFRC – Part 1: Experimental Test Plan . . . . . . . . . . 1003 Tohru Makita, Yuji Watanabe, Shuji Yanai, and Hirokazu Kitagawa Full-Scale Construction Test for Improvement of RC Void Slab Bridges Using UHPFRC – Part 2: Test Results . . . . . . . . . . . . . . . . . . . 1012 Yuji Watanabe, Shuji Yanai, Tohru Makita, and Hirokazu Kitagawa Influence of Fiber Type on the Tensile Behavior of High-Strength Strain-Hardening Cement-Based Composites (HS-SHCC) During and After Exposure to Elevated Temperatures . . . . . . . . . . . . . . . . . . . 1022 Iurie Curosu, Sarah Burk, Marco Liebscher, and Viktor Mechtcherine Tensile and Compressive Performance of High-Strength Engineered Cementitious Composites (ECC) with Seawater and Sea-Sand . . . . . . . . 1034 Jing Yu, Bo-Tao Huang, Jia-Qi Wu, Jian-Guo Dai, and Christopher K. Y. Leung Effect of Fiber Content Variation in Plastic Hinge Region of Reinforced UHPC Flexural Members . . . . . . . . . . . . . . . . . . . . . . . . . 1042 Mandeep Pokhrel, Yi Shao, Sarah Billington, and Matthew J. Bandelt An Eco-Friendly UHPC for Structural Application: Tensile Mechanical Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1056 Amin Abrishambaf, Mário Pimentel, and Sandra Nunes

xx

Contents

Characterization of Ultra High Performance Fiber Reinforced Concrete (UHPFRC) Tensile Behaviour . . . . . . . . . . . . . . . . . . . . . . . . . 1068 Nicola Generosi, Jacopo Donnini, and Valeria Corinaldesi Slip-Hardening Bond: A Key to the Success of Ultra High Performance FRC Composites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1079 Antoine E. Naaman Testing of Thin UHPFRC Cantilever Stairs with Bolted Connections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1090 Ioan Sosa, Camelia Negrutiu, Bogdan Heghes, and Adel Todor Studying of Processing-Structure-Properties Relation of Strain Hardening Cementitious Composites (SHCC) . . . . . . . . . . . . . . . . . . . . 1100 Zhenghao Li, Jiajia Zhou, Cong Lu, and Christopher K. Y. Leung The Effect of Fiber Content on the Post-cracking Tensile Stiffness Capacity of R-UHPFRC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1112 M. Khorami, Juan Navarro-Gregori, and Pedro Serna Controlling Strength and Ductility of Strain-Hardening Cementitious Composites by Nano-Engineering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1124 Ousmane A. Hisseine and Arezki T. Hamou Comprehensive Characterization of UHPFRC Mixes for Seismic and Durability Rehabilitation of Bridge Piers . . . . . . . . . . . . . . . . . . . . 1137 C. Sevigny-Vallières, P. Marchand, B. Terrade, N. Roy, F. Toutlemonde, and A. Tagnit-Hamou Evaluation of the Splitting Tensile Strength of Ultra-High Performance Concrete . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1149 An Hoang Le Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1161

RILEM Publications

The following list is presenting the global offer of RILEM Publications, sorted by series. Each publication is available in printed version and/or in online version.

RILEM Proceedings (PRO) PRO 1: Durability of High Performance Concrete (ISBN: 2-912143-03-9; e-ISBN: 2-351580-12-5; e-ISBN: 2351580125); Ed. H. Sommer PRO 2: Chloride Penetration into Concrete (ISBN: 2-912143-00-04; e-ISBN: 2912143454); Eds. L.-O. Nilsson and J.-P. Ollivier PRO 3: Evaluation and Strengthening of Existing Masonry Structures (ISBN: 2-912143-02-0; e-ISBN: 2351580141); Eds. L. Binda and C. Modena PRO 4: Concrete: From Material to Structure (ISBN: 2-912143-04-7; e-ISBN: 2351580206); Eds. J.-P. Bournazel and Y. Malier PRO 5: The Role of Admixtures in High Performance Concrete (ISBN: 2-91214305-5; e-ISBN: 2351580214); Eds. J. G. Cabrera and R. Rivera-Villarreal PRO 6: High Performance Fiber Reinforced Cement Composites - HPFRCC 3 (ISBN: 2-912143-06-3; e-ISBN: 2351580222); Eds. H. W. Reinhardt and A. E. Naaman PRO 7: 1st International RILEM Symposium on Self-Compacting Concrete (ISBN: 2-912143-09-8; e-ISBN: 2912143721); Eds. Å. Skarendahl and Ö. Petersson PRO 8: International RILEM Symposium on Timber Engineering (ISBN: 2-912143-10-1; e-ISBN: 2351580230); Ed. L. Boström

xxi

xxii

RILEM Publications

PRO 9: 2nd International RILEM Symposium on Adhesion between Polymers and Concrete ISAP ’99 (ISBN: 2-912143-11-X; e-ISBN: 2351580249); Eds. Y. Ohama and M. Puterman PRO 10: 3rd International RILEM Symposium on Durability of Building and Construction Sealants (ISBN: 2-912143-13-6; e-ISBN: 2351580257); Eds. A. T. Wolf PRO 11: 4th International RILEM Conference on Reflective Cracking in Pavements (ISBN: 2-912143-14-4; e-ISBN: 2351580265); Eds. A. O. Abd El Halim, D. A. Taylor and El H. H. Mohamed PRO 12: International RILEM Workshop on Historic Mortars: Characteristics and Tests (ISBN: 2-912143-15-2; e-ISBN: 2351580273); Eds. P. Bartos, C. Groot and J. J. Hughes PRO 13: 2nd International RILEM Symposium on Hydration and Setting (ISBN: 2-912143-16-0; e-ISBN: 2351580281); Ed. A. Nonat PRO 14: Integrated Life-Cycle Design of Materials and Structures - ILCDES 2000 (ISBN: 951-758-408-3; e-ISBN: 235158029X); (ISSN: 0356-9403); Ed. S. Sarja PRO 15: Fifth RILEM Symposium on Fibre-Reinforced Concretes (FRC) BEFIB’2000 (ISBN: 2-912143-18-7; e-ISBN: 291214373X); Eds. P. Rossi and G. Chanvillard PRO 16: Life Prediction and Management of Concrete Structures (ISBN: 2-912143-19-5; e-ISBN: 2351580303); Ed. D. Naus PRO 17: Shrinkage of Concrete – Shrinkage 2000 (ISBN: 2-912143-20-9; e-ISBN: 2351580311); Eds. V. Baroghel-Bouny and P.-C. Aïtcin PRO 18: Measurement and Interpretation of the On-Site Corrosion Rate (ISBN: 2-912143-21-7; e-ISBN: 235158032X); Eds. C. Andrade, C. Alonso, J. Fullea, J. Polimon and J. Rodriguez PRO 19: Testing and Modelling the Chloride Ingress into Concrete (ISBN: 2-912143-22-5; e-ISBN: 2351580338); Eds. C. Andrade and J. Kropp PRO 20: 1st International RILEM Workshop on Microbial Impacts on Building Materials (CD 02) (e-ISBN 978-2-35158-013-4); Ed. M. Ribas Silva PRO 21: International RILEM Symposium on Connections between Steel and Concrete (ISBN: 2-912143-25-X; e-ISBN: 2351580346); Ed. R. Eligehausen PRO 22: International RILEM Symposium on Joints in Timber Structures (ISBN: 2-912143-28-4; e-ISBN: 2351580354); Eds. S. Aicher and H.-W. Reinhardt PRO 23: International RILEM Conference on Early Age Cracking in Cementitious Systems (ISBN: 2-912143-29-2; e-ISBN: 2351580362); Eds. K. Kovler and A. Bentur

RILEM Publications

xxiii

PRO 24: 2nd International RILEM Workshop on Frost Resistance of Concrete (ISBN: 2-912143-30-6; e-ISBN: 2351580370); Eds. M. J. Setzer, R. Auberg and H.-J. Keck PRO 25: International RILEM Workshop on Frost Damage in Concrete (ISBN: 2-912143-31-4; e-ISBN: 2351580389); Eds. D. J. Janssen, M. J. Setzer and M. B. Snyder PRO 26: International RILEM Workshop on On-Site Control and Evaluation of Masonry Structures (ISBN: 2-912143-34-9; e-ISBN: 2351580141); Eds. L. Binda and R. C. de Vekey PRO 27: International RILEM Symposium on Building Joint Sealants (CD03; e-ISBN: 235158015X); Ed. A. T. Wolf PRO 28: 6th International RILEM Symposium on Performance Testing and Evaluation of Bituminous Materials - PTEBM’03 (ISBN: 2-912143-35-7; e-ISBN: 978-2-912143-77-8); Ed. M. N. Partl PRO 29: 2nd International RILEM Workshop on Life Prediction and Ageing Management of Concrete Structures (ISBN: 2-912143-36-5; e-ISBN: 2912143780); Ed. D. J. Naus PRO 30: 4th International RILEM Workshop on High Performance Fiber Reinforced Cement Composites - HPFRCC 4 (ISBN: 2-912143-37-3; e-ISBN: 2912143799); Eds. A. E. Naaman and H. W. Reinhardt PRO 31: International RILEM Workshop on Test and Design Methods for Steel Fibre Reinforced Concrete: Background and Experiences (ISBN: 2-912143-38-1; e-ISBN: 2351580168); Eds. B. Schnütgen and L. Vandewalle PRO 32: International Conference on Advances in Concrete and Structures 2 vol. (ISBN (set): 2-912143-41-1; e-ISBN: 2351580176); Eds. Ying-shu Yuan, Surendra P. Shah and Heng-lin Lü PRO 33: 3rd International Symposium on Self-Compacting Concrete (ISBN: 2-912143-42-X; e-ISBN: 2912143713); Eds. Ó. Wallevik and I. Níelsson PRO 34: International RILEM Conference on Microbial Impact on Building Materials (ISBN: 2-912143-43-8; e-ISBN: 2351580184); Ed. M. Ribas Silva PRO 35: International RILEM TC 186-ISA on Internal Sulfate Attack and Delayed Ettringite Formation (ISBN: 2-912143-44-6; e-ISBN: 2912143802); Eds. K. Scrivener and J. Skalny PRO 36: International RILEM Symposium on Concrete Science and Engineering – A Tribute to Arnon Bentur (ISBN: 2-912143-46-2; e-ISBN: 2912143586); Eds. K. Kovler, J. Marchand, S. Mindess and J. Weiss PRO 37: 5th International RILEM Conference on Cracking in Pavements – Mitigation, Risk Assessment and Prevention (ISBN: 2-912143-47-0; e-ISBN: 2912143764); Eds. C. Petit, I. Al-Qadi and A. Millien

xxiv

RILEM Publications

PRO 38: 3rd International RILEM Workshop on Testing and Modelling the Chloride Ingress into Concrete (ISBN: 2-912143-48-9; e-ISBN: 2912143578); Eds. C. Andrade and J. Kropp PRO 39: 6th International RILEM Symposium on Fibre-Reinforced Concretes BEFIB 2004 (ISBN: 2-912143-51-9; e-ISBN: 2912143748); Eds. M. Di Prisco, R. Felicetti and G. A. Plizzari PRO 40: International RILEM Conference on the Use of Recycled Materials in Buildings and Structures (ISBN: 2-912143-52-7; e-ISBN: 2912143756); Eds. E. Vázquez, Ch. F. Hendriks and G. M. T. Janssen PRO 41: RILEM International Symposium on Environment-Conscious Materials and Systems for Sustainable Development (ISBN: 2-912143-55-1; e-ISBN: 2912143640); Eds. N. Kashino and Y. Ohama PRO 42: SCC’2005 - China: 1st International Symposium on Design, Performance and Use of Self-Consolidating Concrete (ISBN: 2-912143-61-6; e-ISBN: 2912143624); Eds. Zhiwu Yu, Caijun Shi, Kamal Henri Khayat and Youjun Xie PRO 43: International RILEM Workshop on Bonded Concrete Overlays (e-ISBN: 2-912143-83-7); Eds. J. L. Granju and J. Silfwerbrand PRO 44: 2nd International RILEM Workshop on Microbial Impacts on Building Materials (CD11) (e-ISBN: 2-912143-84-5); Ed. M. Ribas Silva PRO 45: 2nd International Symposium on Nanotechnology in Construction, Bilbao (ISBN: 2-912143-87-X; e-ISBN: 2912143888); Eds. Peter J. M. Bartos, Yolanda de Miguel and Antonio Porro PRO 46: ConcreteLife’06 - International RILEM-JCI Seminar on Concrete Durability and Service Life Planning: Curing, Crack Control, Performance in Harsh Environments (ISBN: 2-912143-89-6; e-ISBN: 291214390X); Ed. K. Kovler PRO 47: International RILEM Workshop on Performance Based Evaluation and Indicators for Concrete Durability (ISBN: 978-2-912143-95-2; e-ISBN: 9782912143969); Eds. V. Baroghel-Bouny, C. Andrade, R. Torrent and K. Scrivener PRO 48: 1st International RILEM Symposium on Advances in Concrete through Science and Engineering (e-ISBN: 2-912143-92-6); Eds. J. Weiss, K. Kovler, J. Marchand, and S. Mindess PRO 49: International RILEM Workshop on High Performance Fiber Reinforced Cementitious Composites in Structural Applications (ISBN: 2-912143-93-4; e-ISBN: 2912143942); Eds. G. Fischer and V.C. Li PRO 50: 1st International RILEM Symposium on Textile Reinforced Concrete (ISBN: 2-912143-97-7; e-ISBN: 2351580087); Eds. Josef Hegger, Wolfgang Brameshuber and Norbert Will

RILEM Publications

xxv

PRO 51: 2nd International Symposium on Advances in Concrete through Science and Engineering (ISBN: 2-35158-003-6; e-ISBN: 2-35158-002-8); Eds. J. Marchand, B. Bissonnette, R. Gagné, M. Jolin and F. Paradis PRO 52: Volume Changes of Hardening Concrete: Testing and Mitigation (ISBN: 2-35158-004-4; e-ISBN: 2-35158-005-2); Eds. O. M. Jensen, P. Lura and K. Kovler PRO 53: High Performance Fiber Reinforced Cement Composites - HPFRCC5 (ISBN: 978-2-35158-046-2; e-ISBN: 978-2-35158-089-9); Eds. H. W. Reinhardt and A. E. Naaman PRO 54: 5th International RILEM Symposium on Self-Compacting Concrete (ISBN: 978-2-35158-047-9; e-ISBN: 978-2-35158-088-2); Eds. G. De Schutter and V. Boel PRO 55: International RILEM Symposium Photocatalysis, Environment and Construction Materials (ISBN: 978-2-35158-056-1; e-ISBN: 978-2-35158-057-8); Eds. P. Baglioni and L. Cassar PRO 56: International RILEM Workshop on Integral Service Life Modelling of Concrete Structures (ISBN 978-2-35158-058-5; e-ISBN: 978-2-35158-090-5); Eds. R. M. Ferreira, J. Gulikers and C. Andrade PRO 57: RILEM Workshop on Performance of cement-based materials in aggressive aqueous environments (e-ISBN: 978-2-35158-059-2); Ed. N. De Belie PRO 58: International RILEM Symposium on Concrete Modelling - CONMOD’08 (ISBN: 978-2-35158-060-8; e-ISBN: 978-2-35158-076-9); Eds. E. Schlangen and G. De Schutter PRO 59: International RILEM Conference on On Site Assessment of Concrete, Masonry and Timber Structures - SACoMaTiS 2008 (ISBN set: 978-2-35158-0615; e-ISBN: 978-2-35158-075-2); Eds. L. Binda, M. di Prisco and R. Felicetti PRO 60: Seventh RILEM International Symposium on Fibre Reinforced Concrete: Design and Applications - BEFIB 2008 (ISBN: 978-2-35158-064-6; e-ISBN: 9782-35158-086-8); Ed. R. Gettu PRO 61: 1st International Conference on Microstructure Related Durability of Cementitious Composites 2 vol., (ISBN: 978-2-35158-065-3; e-ISBN: 978-235158-084-4); Eds. W. Sun, K. van Breugel, C. Miao, G. Ye and H. Chen PRO 62: NSF/ RILEM Workshop: In-situ Evaluation of Historic Wood and Masonry Structures (e-ISBN: 978-2-35158-068-4); Eds. B. Kasal, R. Anthony and M. Drdácký PRO 63: Concrete in Aggressive Aqueous Environments: Performance, Testing and Modelling, 2 vol., (ISBN: 978-2-35158-071-4; e-ISBN: 978-2-35158-082-0); Eds. M. G. Alexander and A. Bertron

xxvi

RILEM Publications

PRO 64: Long Term Performance of Cementitious Barriers and Reinforced Concrete in Nuclear Power Plants and Waste Management - NUCPERF 2009 (ISBN: 978-2-35158-072-1; e-ISBN: 978-2-35158-087-5); Eds. V. L’Hostis, R. Gens, C. Gallé PRO 65: Design Performance and Use of Self-consolidating Concrete - SCC’2009 (ISBN: 978-2-35158-073-8; e-ISBN: 978-2-35158-093-6); Eds. C. Shi, Z. Yu, K. H. Khayat and P. Yan PRO 66: 2nd International RILEM Workshop on Concrete Durability and Service Life Planning - ConcreteLife’09 (ISBN: 978-2-35158-074-5; ISBN: 978-2-35158074-5); Ed. K. Kovler PRO 67: Repairs Mortars for Historic Masonry (e-ISBN: 978-2-35158-083-7); Ed. C. Groot PRO 68: Proceedings of the 3rd International RILEM Symposium on ‘Rheology of Cement Suspensions such as Fresh Concrete (ISBN 978-2-35158-091-2; e-ISBN: 978-2-35158-092-9); Eds. O. H. Wallevik, S. Kubens and S. Oesterheld PRO 69: 3rd International PhD Student Workshop on ‘Modelling the Durability of Reinforced Concrete (ISBN: 978-2-35158-095-0); Eds. R. M. Ferreira, J. Gulikers and C. Andrade PRO 70: 2nd International Conference on ‘Service Life Design for Infrastructure’ (ISBN set: 978-2-35158-096-7, e-ISBN: 978-2-35158-097-4); Ed. K. van Breugel, G. Ye and Y. Yuan PRO 71: Advances in Civil Engineering Materials - The 50-year Teaching Anniversary of Prof. Sun Wei’ (ISBN: 978-2-35158-098-1; e-ISBN: 978-2-35158099-8); Eds. C. Miao, G. Ye, and H. Chen PRO 72: First International Conference on ‘Advances in Chemically-Activated Materials – CAM’2010’ (2010), 264 pp, ISBN: 978-2-35158-101-8; e-ISBN: 9782-35158-115-5, Eds. Caijun Shi and Xiaodong Shen PRO 73: 2nd International Conference on ‘Waste Engineering and Management ICWEM 2010’ (2010), 894 pp, ISBN: 978-2-35158-102-5; e-ISBN: 978-2-35158103-2, Eds. J. Zh. Xiao, Y. Zhang, M. S. Cheung and R. Chu PRO 74: International RILEM Conference on ‘Use of Superabsorsorbent Polymers and Other New Addditives in Concrete’ (2010) 374 pp., ISBN: 978-2-35158-104-9; e-ISBN: 978-2-35158-105-6; Eds. O. M. Jensen, M. T. Hasholt, and S. Laustsen PRO 75: International Conference on ‘Material Science - 2nd ICTRC - Textile Reinforced Concrete - Theme 1’ (2010) 436 pp., ISBN: 978-2-35158-106-3; e-ISBN: 978-2-35158-107-0; Ed. W. Brameshuber PRO 76: International Conference on ‘Material Science - HetMat - Modelling of Heterogeneous Materials - Theme 2’ (2010) 255 pp., ISBN: 978-2-35158-108-7; e-ISBN: 978-2-35158-109-4; Ed. W. Brameshuber

RILEM Publications

xxvii

PRO 77: International Conference on ‘Material Science - AdIPoC - Additions Improving Properties of Concrete - Theme 3’ (2010) 459 pp., ISBN: 978-2-35158110-0; e-ISBN: 978-2-35158-111-7; Ed. W. Brameshuber PRO 78: 2nd Historic Mortars Conference and RILEM TC 203-RHM Final Workshop – HMC2010 (2010) 1416 pp., e-ISBN: 978-2-35158-112-4; Eds J. Válek, C. Groot, and J. J. Hughes PRO 79: International RILEM Conference on Advances in Construction Materials Through Science and Engineering (2011) 213 pp., ISBN: 978-2-35158-116-2, e-ISBN: 978-2-35158-117-9; Eds. Christopher Leung and K. T. Wan PRO 80: 2nd International RILEM Conference on Concrete Spalling due to Fire Exposure (2011) 453 pp., ISBN: 978-2-35158-118-6, e-ISBN: 978-2-35158-119-3; Eds E.A.B. Koenders and F. Dehn PRO 81: 2nd International RILEM Conference on Strain Hardening Cementitious Composites (SHCC2-Rio) (2011) 451 pp., ISBN: 978-2-35158-120-9, e-ISBN: 978-2-35158-121-6; Eds R.D. Toledo Filho, F.A. Silva, E.A.B. Koenders and E.M.R. Fairbairn PRO 82: 2nd International RILEM Conference on Progress of Recycling in the Built Environment (2011) 507 pp., e-ISBN: 978-2-35158-122-3; Eds V.M. John, E. Vazquez, S.C. Angulo and C. Ulsen PRO 83: 2nd International Conference on Microstructural-related Durability of Cementitious Composites (2012) 250 pp., ISBN: 978-2-35158-129-2; e-ISBN: 978-2-35158-123-0; Eds G. Ye, K. van Breugel, W. Sun and C. Miao PRO 84: CONSEC13 - Seventh International Conference on Concrete under Severe Conditions – Environment and Loading (2013) 1930 pp., ISBN: 978-235158-124-7; e-ISBN: 978-2- 35158-134-6; Eds Z.J. Li, W. Sun, C.W. Miao, K. Sakai, O.E. Gjorv & N. Banthia PRO 85: RILEM-JCI International Workshop on Crack Control of Mass Concrete and Related issues concerning Early-Age of Concrete Structures – ConCrack 3 – Control of Cracking in Concrete Structures 3 (2012) 237 pp., ISBN: 978-2-35158125-4; e-ISBN: 978-2-35158-126-1; Eds F. Toutlemonde and J.-M. Torrenti PRO 86: International Symposium on Life Cycle Assessment and Construction (2012) 414 pp., ISBN: 978-2-35158-127-8, e-ISBN: 978-2-35158-128-5; Eds A. Ventura and C. de la Roche PRO 87: UHPFRC 2013 – RILEM-fib-AFGC International Symposium on UltraHigh Performance Fibre-Reinforced Concrete (2013), ISBN: 978-2-35158-130-8, e-ISBN: 978-2-35158-131-5; Eds F. Toutlemonde PRO 88: 8th RILEM International Symposium on Fibre Reinforced Concrete (2012) 344 pp., ISBN: 978-2-35158-132-2, e-ISBN: 978-2-35158-133-9; Eds Joaquim A.O. Barros

xxviii

RILEM Publications

PRO 89: RILEM International workshop on performance-based specification and control of concrete durability (2014) 678 pp, ISBN: 978-2-35158-135-3, e-ISBN: 978-2-35158-136-0; Eds. D. Bjegović, H. Beushausen and M. Serdar PRO 90: 7th RILEM International Conference on Self-Compacting Concrete and of the 1st RILEM International Conference on Rheology and Processing of Construction Materials (2013) 396 pp, ISBN: 978-2-35158-137-7, e-ISBN: 978-235158-138-4; Eds. Nicolas Roussel and Hela Bessaies-Bey PRO 91: CONMOD 2014 - RILEM International Symposium on Concrete Modelling (2014), ISBN: 978-2-35158-139-1; e-ISBN: 978-2-35158-140-7; Eds. Kefei Li, Peiyu Yan and Rongwei Yang PRO 92: CAM 2014 - 2nd International Conference on advances in chemicallyactivated materials (2014) 392 pp., ISBN: 978-2-35158-141-4; e-ISBN: 978-235158-142-1; Eds. Caijun Shi and Xiadong Shen PRO 93: SCC 2014 - 3rd International Symposium on Design, Performance and Use of Self-Consolidating Concrete (2014) 438 pp., ISBN: 978-2-35158-143-8; e-ISBN: 978-2-35158-144-5; Eds. Caijun Shi, Zhihua Ou, Kamal H. Khayat PRO 94 (online version): HPFRCC-7 - 7th RILEM conference on High performance fiber reinforced cement composites (2015), e-ISBN: 978-2-35158-146-9; Eds. H.W. Reinhardt, G.J. Parra-Montesinos, H. Garrecht PRO 95: International RILEM Conference on Application of superabsorbent polymers and other new admixtures in concrete construction (2014), ISBN: 978-235158-147-6; e-ISBN: 978-2-35158-148-3; Eds. Viktor Mechtcherine, Christof Schroefl PRO 96 (online version): XIII DBMC: XIII International Conference on Durability of Building Materials and Components(2015), e-ISBN: 978-2-35158149-0; Eds. M. Quattrone, V.M. John PRO 97: SHCC3 – 3rd International RILEM Conference on Strain Hardening Cementitious Composites (2014), ISBN: 978-2-35158-150-6; e-ISBN: 978-235158-151-3; Eds. E. Schlangen, M.G. Sierra Beltran, M. Lukovic, G. Ye PRO 98: FERRO-11 – 11th International Symposium on Ferrocement and 3rd ICTRC - International Conference.on Textile Reinforced Concrete (2015), ISBN: 978-2-35158-152-0; e-ISBN: 978-2-35158-153-7; Ed. W. Brameshuber PRO 99 (online version): ICBBM 2015 - 1st International Conference on Bio-Based Building Materials (2015), e-ISBN: 978-2-35158-154-4; Eds. S. Amziane, M. Sonebi PRO 100: SCC16 - RILEM Self-Consolidating Concrete Conference (2016), ISBN: 978-2-35158-156-8; e-ISBN: 978-2-35158-157-5; Ed. Kamal H. Kayat

RILEM Publications

xxix

PRO 101 (online version): III Progress of Recycling in the Built Environment (2015), e-ISBN: 978-2-35158-158-2; Eds I. Martins, C. Ulsen and S. C. Angulo PRO 102 (online version): RILEM Conference on Microorganisms-Cementitious Materials Interactions (2016), e-ISBN: 978-2-35158-160-5; Eds. Alexandra Bertron, Henk Jonkers, Virginie Wiktor PRO 103 (online version): ACESC’16 - Advances in Civil Engineering and Sustainable Construction (2016), e-ISBN: 978-2-35158-161-2; Eds. T.Ch. Madhavi, G. Prabhakar, Santhosh Ram and P.M. Rameshwaran PRO 104 (online version): SSCS’2015 - Numerical Modeling - Strategies for Sustainable Concrete Structures (2015), e-ISBN: 978-2-35158-162-9 PRO 105: 1st International Conference on UHPC Materials and Structures (2016), ISBN: 978-2-35158-164-3, e-ISBN: 978-2-35158-165-0 PRO 106: AFGC-ACI-fib-RILEM International Conference on Ultra-HighPerformance Fibre-Reinforced Concrete – UHPFRC 2017 (2017), ISBN: 978-235158-166-7, e-ISBN: 978-2-35158-167-4; Eds. François Toutlemonde & Jacques Resplendino PRO 107 (online version): XIV DBMC – 14th International Conference on Durability of Building Materials and Components (2017), e-ISBN: 978-2-35158159-9; Eds. Geert De Schutter, Nele De Belie, Arnold Janssens, Nathan Van Den Bossche PRO 108: MSSCE 2016 - Innovation of Teaching in Materials and Structures (2016), ISBN: 978-2-35158-178-0, e-ISBN: 978-2-35158-179-7; Ed. Per Goltermann PRO 109 (2 volumes): MSSCE 2016 - Service Life of Cement-Based Materials and Structures (2016), ISBN Vol. 1: 978-2-35158-170-4, Vol. 2: 978-2-35158-1714, Set Vol. 1&2: 978-2-35158-172-8, e-ISBN : 978-2-35158-173-5; Eds. Miguel Azenha, Ivan Gabrijel, Dirk Schlicke, Terje Kanstad and Ole Mejlhede Jensen PRO 110: MSSCE 2016 - Historical Masonry (2016), ISBN: 978-2-35158-178-0, e-ISBN: 978-2-35158-179-7; Eds. Inge Rörig-Dalgaard and Ioannis Ioannou PRO 111: MSSCE 2016 - Electrochemistry in Civil Engineering (2016), ISBN: 978-2-35158-176-6, e-ISBN: 978-2-35158-177-3; Ed. Lisbeth M. Ottosen PRO 112: MSSCE 2016 - Moisture in Materials and Structures (2016), ISBN: 9782-35158-178-0, e-ISBN: 978-2-35158-179-7; Eds. Kurt Kielsgaard Hansen, Carsten Rode and Lars-Olof Nilsson PRO 113: MSSCE 2016 - Concrete with Supplementary Cementitious Materials (2016), ISBN: 978-2-35158-178-0, e-ISBN: 978-2-35158-179-7; Eds. Ole Mejlhede Jensen, Konstantin Kovler and Nele De Belie

xxx

RILEM Publications

PRO 114: MSSCE 2016 - Frost Action in Concrete (2016), ISBN: 978-2-35158182-7, e-ISBN: 978-2-35158-183-4; Eds. Marianne Tange Hasholt, Katja Fridh and R. Doug Hooton PRO 115: MSSCE 2016 - Fresh Concrete (2016), ISBN: 978-2-35158-184-1, e-ISBN: 978-2-35158-185-8; Eds. Lars N. Thrane, Claus Pade, Oldrich Svec and Nicolas Roussel PRO 116: BEFIB 2016 – 9th RILEM International Symposium on Fiber Reinforced Concrete (2016), ISBN: 978-2-35158-187-2, e-ISBN: 978-2-35158186-5; Eds. N. Banthia, M. di Prisco and S. Soleimani-Dashtaki PRO 117: 3rd International RILEM Conference on Microstructure Related Durability of Cementitious Composites (2016), ISBN: 978-2-35158-188-9, e-ISBN: 978-2-35158-189-6; Eds. Changwen Miao, Wei Sun, Jiaping Liu, Huisu Chen, Guang Ye and Klaas van Breugel PRO 118 (4 volumes): International Conference on Advances in Construction Materials and Systems (2017), ISBN Set: 978-2-35158-190-2, Vol. 1: 978-235158-193-3, Vol. 2: 978-2-35158-194-0, Vol. 3: ISBN: 978-2-35158-195-7, Vol. 4: ISBN: 978-2-35158-196-4, e-ISBN: 978-2-35158-191-9; Eds. Manu Santhanam, Ravindra Gettu, Radhakrishna G. Pillai and Sunitha K. Nayar PRO 119 (online version): ICBBM 2017 - Second International RILEM Conference on Bio-based Building Materials, (2017), e-ISBN: 978-2-35158-192-6; Eds. Sofiane Amziane, Mohammed Sonebi PRO 120 (2 volumes): EAC-02 - 2nd International RILEM/COST Conference on Early Age Cracking and Serviceability in Cement-based Materials and Structures, (2017), Vol. 1: 978-2-35158-199-5, Vol. 2: 978-2-35158-200-8, Set: 978-2-35158197-1, e-ISBN: 978-2-35158-198-8; Eds. Stéphanie Staquet and Dimitrios Aggelis PRO 121 (2 volumes): SynerCrete18: Interdisciplinary Approaches for Cementbased Materials and Structural Concrete: Synergizing Expertise and Bridging Scales of Space and Time, (2018), Set: 978-2-35158-202-2, Vol.1: 978-235158-211-4, Vol.2: 978-2-35158-212-1, e-ISBN: 978-2-35158-203-9; Eds. Miguel Azenha, Dirk Schlicke, Farid Benboudjema, Agnieszka Knoppik PRO 122: SCC’2018 China - Fourth International Symposium on Design, Performance and Use of Self-Consolidating Concrete, (2018), ISBN: 978-2-35158204-6, e-ISBN: 978-2-35158-205-3; Eds. C. Shi, Z. Zhang, K. H. Khayat PRO 123: Final Conference of RILEM TC 253-MCI: MicroorganismsCementitious Materials Interactions (2018), Set: 978-2-35158-207-7, Vol.1: 978-235158-209-1, Vol.2: 978-2-35158-210-7, e-ISBN: 978-2-35158-206-0; Ed. Alexandra Bertron PRO 124 (online version): Fourth International Conference Progress of Recycling in the Built Environment (2018), e-ISBN: 978-2-35158-208-4; Eds. Isabel M. Martins, Carina Ulsen, Yury Villagran

RILEM Publications

xxxi

PRO 125 (online version): SLD4 - 4th International Conference on Service Life Design for Infrastructures (2018), e-ISBN: 978-2-35158-213-8; Eds. Guang Ye, Yong Yuan, Claudia Romero Rodriguez, Hongzhi Zhang, Branko Savija PRO 126: Workshop on Concrete Modelling and Material Behaviour in honor of Professor Klaas van Breugel (2018), ISBN: 978-2-35158-214-5, e-ISBN: 978-235158-215-2; Ed. Guang Ye PRO 127 (online version): CONMOD2018 - Symposium on Concrete Modelling (2018), e-ISBN: 978-2-35158-216-9; Eds. Erik Schlangen, Geert de Schutter, Branko Savija, Hongzhi Zhang, Claudia Romero Rodriguez PRO 128: SMSS2019 - International Conference on Sustainable Materials, Systems and Structures (2019), ISBN: 978-2-35158-217-6, e-ISBN: 978-2-35158218-3 PRO 129: 2nd International Conference on UHPC Materials and Structures (UHPC2018-China), ISBN: 978-2-35158-219-0, e-ISBN: 978-2-35158-220-6 PRO 130: 5th Historic Mortars Conference (2019), ISBN: 978-2-35158-221-3, e-ISBN: 978-2-35158-222-0; Eds. José Ignacio Álvarez, José María Fernández, Íñigo Navarro, Adrián Durán, Rafael Sirera PRO 131 (online version): 3rd International Conference on Bio-Based Building Materials (ICBBM2019), e-ISBN: 978-2-35158-229-9; Eds. Mohammed Sonebi, Sofiane Amziane, Jonathan Page PRO 132: IRWRMC’18 - International RILEM Workshop on Rheological Measurements of Cement-based Materials (2018), ISBN: 978-2-35158-230-5, e-ISBN: 978-2-35158-231-2; Eds. Chafika Djelal, Yannick Vanhove PRO 133 (online version): CO2STO2019 - International Workshop CO2 Storage in Concrete (2019), e-ISBN: 978-2-35158-232-9; Eds. Assia Djerbi, Othman Omikrine-Metalssi, Teddy Fen-Chong

RILEM Reports (REP) Report 19: Considerations for Use in Managing the Aging of Nuclear Power Plant Concrete Structures (ISBN: 2-912143-07-1); Ed. D. J. Naus Report 20: Engineering and Transport Properties of the Interfacial Transition Zone in Cementitious Composites (ISBN: 2-912143-08-X); Eds. M. G. Alexander, G. Arliguie, G. Ballivy, A. Bentur and J. Marchand Report 21: Durability of Building Sealants (ISBN: 2-912143-12-8); Ed. A. T. Wolf Report 22: Sustainable Raw Materials - Construction and Demolition Waste (ISBN: 2-912143-17-9); Eds. C. F. Hendriks and H. S. Pietersen Report 23: Self-Compacting Concrete state-of-the-art report (ISBN: 2-912143-233); Eds. Å. Skarendahl and Ö. Petersson

xxxii

RILEM Publications

Report 24: Workability and Rheology of Fresh Concrete: Compendium of Tests (ISBN: 2-912143-32-2); Eds. P. J. M. Bartos, M. Sonebi and A. K. Tamimi Report 25: Early Age Cracking in Cementitious Systems (ISBN: 2-912143-33-0); Ed. A. Bentur Report 26: Towards Sustainable Roofing (Joint Committee CIB/RILEM) (CD 07) (e-ISBN 978-2-912143-65-5); Eds. Thomas W. Hutchinson and Keith Roberts Report 27: Condition Assessment of Roofs (Joint Committee CIB/RILEM) (CD 08) (e-ISBN 978-2-912143-66-2); Ed. CIB W 83/RILEM TC166-RMS Report 28: Final report of RILEM TC 167-COM ‘Characterisation of Old Mortars with Respect to Their Repair (ISBN: 978-2-912143-56-3); Eds. C. Groot, G. Ashall and J. Hughes Report 29: Pavement Performance Prediction and Evaluation (PPPE): Interlaboratory Tests (e-ISBN: 2-912143-68-3); Eds. M. Partl and H. Piber Report 30: Final Report of RILEM TC 198-URM ‘Use of Recycled Materials’ (ISBN: 2-912143-82-9; e-ISBN: 2-912143-69-1); Eds. Ch. F. Hendriks, G. M. T. Janssen and E. Vázquez Report 31: Final Report of RILEM TC 185-ATC ‘Advanced testing of cementbased materials during setting and hardening’ (ISBN: 2-912143-81-0; e-ISBN: 2-912143-70-5); Eds. H. W. Reinhardt and C. U. Grosse Report 32: Probabilistic Assessment of Existing Structures. A JCSS publication (ISBN 2-912143-24-1); Ed. D. Diamantidis Report 33: State-of-the-Art Report of RILEM Technical Committee TC 184-IFE ‘Industrial Floors’ (ISBN 2-35158-006-0); Ed. P. Seidler Report 34: Report of RILEM Technical Committee TC 147-FMB ‘Fracture mechanics applications to anchorage and bond’ Tension of Reinforced Concrete Prisms – Round Robin Analysis and Tests on Bond (e-ISBN 2-912143-91-8); Eds. L. Elfgren and K. Noghabai Report 35: Final Report of RILEM Technical Committee TC 188-CSC ‘Casting of Self Compacting Concrete’ (ISBN 2-35158-001-X; e-ISBN: 2-912143-98-5); Eds. Å. Skarendahl and P. Billberg Report 36: State-of-the-Art Report of RILEM Technical Committee TC 201-TRC ‘Textile Reinforced Concrete’ (ISBN 2-912143-99-3); Ed. W. Brameshuber Report 37: State-of-the-Art Report of RILEM Technical Committee TC 192-ECM ‘Environment-conscious construction materials and systems’ (ISBN: 978-2-35158053-0); Eds. N. Kashino, D. Van Gemert and K. Imamoto Report 38: State-of-the-Art Report of RILEM Technical Committee TC 205-DSC ‘Durability of Self-Compacting Concrete’ (ISBN: 978-2-35158-048-6); Eds. G. De Schutter and K. Audenaert

RILEM Publications

xxxiii

Report 39: Final Report of RILEM Technical Committee TC 187-SOC ‘Experimental determination of the stress-crack opening curve for concrete in tension’ (ISBN 978-2-35158-049-3); Ed. J. Planas Report 40: State-of-the-Art Report of RILEM Technical Committee TC 189-NEC ‘Non-Destructive Evaluation of the Penetrability and Thickness of the Concrete Cover’ (ISBN 978-2-35158-054-7); Eds. R. Torrent and L. Fernández Luco Report 41: State-of-the-Art Report of RILEM Technical Committee TC 196-ICC ‘Internal Curing of Concrete’ (ISBN 978-2-35158-009-7); Eds. K. Kovler and O. M. Jensen Report 42: ‘Acoustic Emission and Related Non-destructive Evaluation Techniques for Crack Detection and Damage Evaluation in Concrete’ - Final Report of RILEM Technical Committee 212-ACD (e-ISBN: 978-2-35158-100-1); Ed. M. Ohtsu Report 45: Repair Mortars for Historic Masonry - State-of-the-Art Report of RILEM Technical Committee TC 203-RHM (e-ISBN: 978-2-35158-163-6); Eds. Paul Maurenbrecher and Caspar Groot Report 46: Surface delamination of concrete industrial floors and other durability related aspects guide - Report of RILEM Technical Committee TC 268-SIF (e-ISBN: 978-2-35158-201-5); Ed. Valerie Pollet

Rheology and Early-Age Properties

Influence of Different Fibre Types on the Rheology of Strain Hardening Cementitious Composites Hassan Baloch1(&), Steffen Grünewald1,2, Karel Lesage1, and Stijn Matthys1 Ghent University – Ghent, Ghent, Belgium [email protected] Delft University of Technology – Delft, Delft, The Netherlands 1

2

Abstract. Strain-hardening cementitious composites (SHCC) have a high tensile strength and display a remarkable strain-hardening behaviour. These unique characteristics make them an interesting choice for improving the strength and durability of new and existing structures. The tensile strain behaviour of SHCC is strongly influenced by its rheological properties as they determine the hardened state behaviour such as fibre-bridging strength and ultimately the degree of multiple cracking. The presence of fibres significantly affects the rheological performance of SHCC. This study aimed at investigating the relationship between rheological characteristics of SHCC mortar before and after the addition of different fibres. Polyvinyl alcohol (PVA), high modulus polyethylene (HDPE) and glass fibres were added at three different contents in order to assess their effect on the workability of SHCC. Flow tests along with rheological assessment were conducted to evaluate the fresh state behaviour of SHCC. The addition of fibres reduced the flowability of mix, especially at high dosages. A modified fibre influence factor was developed to characterize different types of fibres and was related to the viscosity and yield stress of the mix. Keywords: Fibres Flowability


Strain-hardening cementitious composites
Rheology


1 Introduction Strain-hardening cementitious composites (SHCC) are a special class of material which display an increase in stress beyond the formation of initial cracks under uniaxial tensile loading [1]. This behaviour is usually achieved by the addition of discontinuous fibres which bridge the cracks resulting in a reduced crack width and higher fracture toughness [2]. In recent years extensive research has been carried out to improve the mechanical properties of SHCCs resulting in ultra-high performance SHCC having tensile strengths up to 15 MPa and a strains capacity of up to 8% [3, 4]. These properties make SHCC an excellent choice not only as a construction material but also as repair material for existing structures. © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 3–11, 2021. https://doi.org/10.1007/978-3-030-58482-5_1

4

H. Baloch et al.

The addition of fibres comes with undesired effects on the flow and rheology of concrete and should be taken into account when designing SHCC. Short synthetic fibres are mostly used in SHCC to achieve micro-cracking behaviour. The type and content of the fibres, the mixing regime and the binder content greatly influence the rheological properties of the resulting mortar. It is important to analyse rheological properties as they can cause variation in ductility and mechanical properties of SHCC. Most work in this regard is done on rigid fibres which has shown that increasing the fibre content decreases the workability of the fibre reinforced cementitious matrix [5–7]. The influence of steel fibres is found to increase as the fibre content increases This reduction in workability was observed to become non-linear at increasing fibre content. At the end, these studies successfully demonstrated the correlation between viscosity, yield stress and flow of steel fibre reinforced matrices. However, for plastic and glass fibres these correlations are not established in detail. The objective of this study was to develop correlations for different fibre types at varying fibre contents. This study was carried out using PVA, HDPE and glass fibres at different fibre contents. Flow and rheological correlations were established for all fibre types based on the experimental results.

2 Materials and Methods 2.1

Materials

A suitable control mixture was designed with a target flow spread of 340 mm using Haegermann’s flow test and fibre contents tested were 1.0, 1,5 and 2.0 vol.% of the mixture keeping other parameters constant. Ordinary Portland Cement (CEM 1 52.5 N, Holcim, Belgium) was used as binder. Very fine silica sand (M34, Sibelco, Belgium) having an average particle size of 174 lm was used to increase the packing density of the granular skeleton and to obtain a high strength matrix. A polycarboxylate-based superplasticizer (Glenium 51, 35% con, BASF) was used to achieve the desired workability. Three different types of fibres including PVA, HDPE and glass fibres were tested having properties listed in Table 1. Table 1. Properties of fibres used Fibre type PVA HDPE Glass

2.2

Aspect ratio Diameter (lm) Length (mm) Young Modulus (Gpa) 307 39 12 42.8 600 20 12 80 857 14 12 72

Mix Formulations and Mixing Regime

A total of 10 formulations were investigated having a w/c ratio of 0.22. The amount of cement, sand and superplasticizer (SP) were kept constant in order to better understand the effect of fibres on the mixture characteristics. The mixture proportions along with

Influence of Different Fibre Types on the Rheology

5

constituent dosage per weight relative to the cement are listed in Table 2; the fibre contents are represented as volume fraction of the mixture. Table 2. Mixture formulations a

Cement Sand Water Fibre volume (%) SP Mix C 1 0.5 0.22 – 0.021 PE1 1 0.5 0.22 1.0 0.021 PE1.5 1 0.5 0.22 1.5 0.021 PE2 1 0.5 0.22 2.0 0.021 Glass1 1 0.5 0.22 1.0 0.021 Glass1.5 1 0.5 0.22 1.5 0.021 Glass2 1 0.5 0.22 2.0 0.021 PVA1 1 0.5 0.22 1.0 0.021 PVA1.5 1 0.5 0.22 1.5 0.021 PVA2 1 0.5 0.22 2.0 0.021 a Fibre content is represented as volume fraction of mix, while rest of the ingredients are weight proportion of cement

A Hobart mixer was used to prepare the mortar mixes. A total of 6-min wet mixing time was arranged to ensure a homogenous fibre dispersion throughout the matrix. Cement and sand were first dry mixed in the Hobart mixer for 60 s at 145 rpm followed by the addition of water and suerplasticizer (SP). The mixer was then operated at 145 rpm for 180 s followed by the fibre addition and 180 s of fast mixing at 285 rpm. Fibres were added in 4 about equal batches to ensure an equal distribution of fibres. 2.3

Test Methods

2.3.1 Flow Tests Flow tests of SHCCs mortars (without compacting action) were conducted using Haegermann’s cone with a height of 60 mm and diameters of 70 and 100 mm at the top and the bottom of the cone, respectively. The value of SP was adjusted to maximize the flow in the control formulation without any apparent segregation. This SP value was then kept constant for formulations containing fibres. 2.3.2 Rheological Investigation The rheological parameters were measured using the rotational type MCR 52 Rheometer produced by Anton Paar. This rheometer consists of a fixed rheometer cup having an internal diameter of 70 mm and a 39 mm rotating probe cylinder attached to the motor. SHCC mortars were prepared using the specified mixing regime and were poured in the rheometer cup. Excess material was carefully scraped off using a spatula and the filling level of material was adjusted so that it won’t flow out after probe insertion. The cup was then fixed in the machine and the rotating probe was inserted to start the measurement.

6

H. Baloch et al.

The shear rate was increased from 1 1/s to 100 1/s with 100 measuring points in between (2 s point duration). Assuming a plastic Bingham model, the average yield stress value is calculated as the axis intercept of the linear regression line of the measured points while the viscosity is calculated as the slope of the regression line. A test resulted in the shear rate-shear stress curve, further interpolations were then executed to determine the yield values and viscosities of the mixtures.

3 Results and Discussion 3.1

Flowability

3.1.1 Measurements of Flow Spread Table 3 shows the results of flow spread tests carried out. The addition of different fibres considerably reduced the flow spread of the specimens compared with the control specimen. The increase in fibre content results in a non-proportional decrease in flowability to the point where almost no flow is observed at a fibre content of 2 vol.%. This effect is visually depicted in Fig. 1 for HDPE fibres.

Table 3. Flow spread test results. Mix C PE1 G1 PVA1 PE1.5 G1.5 PVA1.5 PE2 G2 PVA2

Spread (cm) 34.0 17.3 18.5 18.8 12.5 11.5 12.8 10.7 10.3 11.0

The flowability dependent on both type and content of fibres. Glass fibres had a similar effect as the other fibre types at 1 vol.% fibre content, however they had a more pronounced effect on the workability at higher fibre contents. This outcome can be explained by the higher aspect ratio of glass fibres compared with the other fibre types. 3.1.2 Effect of the Fibre Factor The fibre factor of fibres is calculated by multiplying the volume of the fibres with the respective aspect ratio [Vf*L/D] [8]. This factor was applied to assess the flow measurements and the results are displayed in Fig. 2a. It can be seen that the fibre factor

Influence of Different Fibre Types on the Rheology

7

Fig. 1. Effect of HDPE fibre addition on the flow spread

does not corelate well when considering the different fibre types. This deviating effect can be explained by an entirely different physical and chemical nature of the fibres. As this study also was carried out with glass and plastic fibres, a more comprehensive factor was needed to take into account the different nature of the studied fibres. A generalized fibre influence factor was then developed to accurately predict the tests results obtained with the studied fibres. Considering different fibre materials, the modulus of elasticity was also included as dependent variable [9]. After incorporating all governing parameters, the fibre influence factor is stated as: F ¼ Vf La Db Ec

ð1Þ

Where a, b,c are influence coefficients. Chu et al. [10] also used a similar type of fibre influence indicator for FRC mixes without E parameter as only steel fibres were studied and E was constant. After performing several regression analyses, the authors found that a = 1, b = −0.5 and c = −0.5 results in the best correlation. The best fitting curve is plotted in Fig. 2b. It can be seen that Vf LaDb Ec seems to be a more suitable factor than the traditional Vf (L/D) which is overly simplistic in order to take into account different fibre types.

40

30 25

PE PVA Glass

20 15 10 5

y = 0,3074x2 - 5,4424x + 34,026 R² = 0,9742

35

Flow Spread (cm)

Flow Spread (cm)

40

y = 1E-05x2 - 0,0308x + 29,858 R² = 0,7322

35

30

PE

25

PVA

20

Glass

15 10 5

0 0

500

1000

Vf (L/D)

(a)

1500

2000

0 0

2

4

6

Vf Lα Dβ Eγ

(b)

Fig. 2. a. Traditional fibre factor b. Modified fibre influence factor

8

10

8

H. Baloch et al.

Another interesting observation is that the flow spread decreased at increasing fibre dosage. With increase in fibres content the decrease in flow spread is less than proportional which contradicts with other studies carried out on steel fibre types. This trend can be explained by the very stiff consistency of mixes at 2 vol.% fibre dosage as the flow spread is close to 100 mm, which is the base diameter of the flow cone itself. 3.2

Rheological Investigation

The shear stress was measured at increasing shear rate. The effect of different fibre additions is shown in Fig. 3 for different fibre types. It can be seen that because of the addition of the fibres the curves are not completely straight which is often the case with plain mortars or concretes, which are typical Bingham materials. The yield value was obtained as the y-axis intercept of each curve idealised as a straight line and the plastic viscosity was determined by calculating the slope of the linear model. 3.2.1 Relation Between Yield Stress and Flow Spread In order to study the effect of different shear-rates, the curve was assessed for five maximum shear-rate values of 20, 40, 60, 80 or 100 1/s. These values were plotted against the flow spread as shown in Fig. 4. A considerable increase in the yield stress is noted at increasing shear rate. At a shear rate of 20 1/s an almost linear increase in yield stress is noted at decreasing flow spread, however as the shear rate increased to 100 1/s, a more than proportional rise in yield stress is observed. This can be explained by the higher mechanical energy required during the initial seconds to get the mix flowing at higher shear rates. This leads to a high torque registration, which is translated in a high yield stress. 3.2.2 Relation Between Viscosity and Fibre Influence Factor The addition of more fibres increases the internal friction among the fibres. In order to characterize the relation between plastic viscosity and fibre content, the viscosity values were calculated from rheological readings and plotted against both fibre factor and the modified fibre influence factor at three different strain-rate ranges as shown in Fig. 5. R2 values along with equations were plotted for each trendline which depicts a better correlation of viscosity with the modified fibre influence factor. The plastic viscosity increased at increasing fibre factor in both cases. With a maximum shear rate of 20 1/s there is a steep increase in viscosity values at increasing fibre influence factor, however as the strain-rate increases, the viscosity values drop indicating the breakdown of the mixes. A more preferred fibre orientation during testing could be an explanation for the decrease in viscosity, which first increases at increasing fibre dosage and is about constant at a higher fibre influence factor.

Influence of Different Fibre Types on the Rheology 1400

1400

1200

1200

Shear Stress (Pa)

1000

Shear Stress (Pa)

9

800 600 PVA 1

400

PVA1.5 200

1000 800

600 PE1

400

PE 1.5 200

PE2

PVA2 0

0 0

20

40

60

80

0

100

20

40

60

80

100

She ar rate 1/s

She ar rate 1/s

(b)

(a)

1400 1200

Shear Stress (Pa)

1000 800 600 Glass1

400

Glass1.5 200 Glass2 0 0

20

40

60

80

100

She ar rate 1/s

(c) Fig. 3. a. Shear stress-shear rate curves for PVA fibres b. Shear stress-shear rate curves for PE fibres c. Shear stress-shear rate curves for glass fibres

20 1/s

20

60 1/s

16 14 12 10

y = 2E-05x2 - 0,0572x + 21,976 R² = 0,7474

8

18

Slump Flow (mm)

18

16 14

12

0

100

200

300

Yield Stress (Pa)

400

16 14 12 10

10

y = 0,0001x2 - 0,0982x + 27,109 R² = 0,8163

8

y = 0,0001x2 - 0,1033x + 31,088 R² = 0,8152

8 6

6

6

100 1/s

20

18

Slump Flow (mm)

Slump Flow (mm)

20

0

100

200

300

Yield Stress (Pa)

400

0

100

200

300

400

Yield Stress (Pa)

Fig. 4. Relationship between yield value and flow spread at different shear rates

10

H. Baloch et al. 25

20 1/s

60 1/s

15

y = -4E-06x2 + 0,0104x + 2,8298 R² = 0,3096

20

Viscosity (pa.s)

20

100 1/s

25

y = -6E-06x2 + 0,016x + 2,0343 R² = 0,3923

20

Viscosity (pa.s)

Viscosity (pa.s)

25

15

10

15

10

10

y = 5E-07x2 + 0,0045x + 11,885 R² = 0,3237

5

5

200

700

5

200

700

1200

700

1200

1700

Vf (L/D)

(a)

25

25

20 1/s

20

60 1/s

10

y = -0,358x2 + 5,1022x - 8,6393 R² = 0,7282

20

Viscosity (pa.s)

15

100 1/s

25

y = -0,4997x2 + 7,241x - 13,773 R² = 0,7853

20

Viscosity (pa.s)

Viscosity (pa.s)

200

1700

Vf (L/D)

1700

Vf (L/D)

1200

15

10

15

10

y = -0,0005x2 + 1,8979x + 5,2359 R² = 0,6544 5

5 2

4

6

Vf L1 D-0.5 E-0.5

8

10

2

4

6

8

Vf L1 D-0.5 E-0.5

10

5 2

4

6

8

10

Vf L1 D-0.5 E-0.5

(b) Fig. 5. a. Relationship between viscosity and fibre factor b. Relationship between viscosity and modified fibre influence factor

4 Conclusions This paper discusses the effect of different fibre types and contents on flow and rheological properties of SHCC. An experimental program of 10 SHCC mixes was executed with 3 different fibre types with three contents per fibre type. Based on the experimental results, the following conclusions can be drawn: • The addition of fibres significantly reduced the flow spread of the mixes. This effect is more pronounced at higher fibre dosages and should be accounted for in the mix design. • The yield stress exhibits an almost linear relation with flow spread when assessing the flow curves at a lower shear rate of 20 1/s, however the yield stress increases significantly as the shear-rate values increased. • The effect of fibres on the flowability depends not only on the fibre content but on the type of fibres as well. A modified fibre influence factor was developed to depict fibre behaviour resulted in a better prediction ability compared with the traditional fibre factor. • The plastic viscosity increased steeply with increase in fibre influence factor at lower shear rates. No significant increase in viscosity was observed at higher strain rates.

Influence of Different Fibre Types on the Rheology

11

Acknowledgements. The authors would like to acknowledge the financial support of the Higher Education Commission, Pakistan [HRDI-UESTP (BATCH-V)]. The support of Owens-Corning for free supply of glass fibres used in this research is also appreciated.

References 1. Kong, H.J., Bike, S.G., Li, V.C.: Constitutive rheological control to develop a selfconsolidating engineered cementitious composite reinforced with hydrophilic poly(vinyl alcohol) fibers. Cem. Concr. Compos. 25(3), 331–341 (2003). https://doi.org/10.1016/ S0958-9465(02)00056-2 2. Li, V.C., Mishra, D.K., Wu, H.C.: Matrix design for pseudo-strain-hardening fibre reinforced cementitious composites. Mater. Struct. 28, 586–595 (1995). https://doi.org/10. 1007/BF02473191 3. Ranade, R., Li Prof, V.C., Rushing, M.D., Roth, J., Heard, W.F.: Micromechanics of highstrength, high-ductility concrete. ACI Mater. J. 110, 4–375 (2013). https://doi.org/10.14359/ 51685784 4. Yu, K., Wang, Y., Yu, J., Xu, S.: A strain-hardening cementitious composites with the tensile capacity up to 8%. Constr. Build. Mater. 137, 410–419 (2017). https://doi.org/10. 1016/j.conbuildmat.2017.01.060 5. Grünewald, S., Walraven, J.C.: Parameter-study on the influence of steel fibers and coarse aggregate content on the fresh properties of self-compacting concrete. Cem. Concr. Res. 31 (12), 1793–1798 (2001). https://doi.org/10.1016/S0008-8846(01)00555-5 6. Stähli, P., van Mier, J.G.: Rheological properties and fracture processes of HFC’-in proceeding BEFIB (2004) 7. Grunewald, S.: Performance-based design of self-compacting fibre reinforced concrete. Delft University Press (2004) 8. Grunewald, S., Walraven, J.C.: Rheological study on the workability of fibre-reinforced mortar. In: Ozawa, K., Ouchi, M., (Eds.), Proceedings of the second international symposium on self-compacting concrete, pp. 127–136 (2001) 9. Martinie, L., Rossi, P., Roussel, N.: Rheology of fiber reinforced cementitious materials: classification and prediction. Cem. Concr. Res. 40(2), 226–234 (2001). https://doi.org/10. 1016/j.cemconres.2009.08.032 10. Chu, S.H., Li, L.G., Kwan, A.K.H.: Fibre factors governing the fresh and hardened properties of steel frc. Constr. Build. Mater. 186, 1228–1238 (2018). https://doi.org/10.1016/ j.conbuildmat.2018.08.047

Using Fiber Reinforced Concrete to Control Early-Age Shrinkage in Replacement Concrete Pavement Nakin Suksawang1(&) and Daniel Yohannes2 1

Mechanical and Civil Engineering, Florida Institute of Technology, Melbourne, USA [email protected] 2 Structural Division, SBA Communications, Boca Raton, USA

Abstract. Unlike ordinary concrete pavement, replacement concrete pavement needs to be open to traffic within 24 h (sooner in some cases). Thus, high earlystrength concrete is used; however, it frequently cracks prematurely as a result of high heat of hydration that leads the slab to develop plastic shrinkage. FRC is known to provide good resistance to plastic shrinkage. This paper explores the potential use of fiber- reinforced concrete (FRC) in concrete pavement replacement particularly in controlling plastic shrinkage. Five different fiber types, including steel, glass, basalt, nylon, and polyethylene fibers were investigated. Additionally, the effect of fiber length was also investigated for the polyethylene fiber. The fibers were added at low dosage amounts of 0.1% and 0.3% by volume. A retrained shrinkage test was conducted to assess the cracking potential of the concrete mixtures and the ability for each fiber type to resist cracking. Results indicated that both polyethylene and nylon fibers provided the best resistance to early-age shrinkage. Short fibers (1-in.) provided additional post-cracking capacity. For replacement concrete pavement, it is recommended that a short polyethylene fiber be used to eliminate uncontrolled cracking. Keywords: High early-strength concrete
Concrete pavement Polyethylene fibers
Cracking
Plastic shrinkage


FRC


1 Introduction The Florida Department of Transportation (FDOT) Design Standards [1] require a full depth replacement of concrete slab with severe distresses. The construction standards and requirements of the replacement slab are provided in Section 353 of the Standard Specifications for Road and Bridge Constructions [2]. Two of the most important acceptance criteria for the replacement slab are the plastic property, specifically the 6-h compressive strength of 2,200 psi (15.2 MPa), and the 24-h compressive strength of 3,000 psi (20.7 MPa). The 6-h compressive strength is also used as the determination point to open the slab to traffic, and therefore, it is highly emphasized in the standard specifications. In fact, if the replacement slab does not meet the plastic property © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 12–23, 2021. https://doi.org/10.1007/978-3-030-58482-5_2

Using Fiber Reinforced Concrete

13

requirements and the engineer determines that this will severely impact the replacement slab service life, the contractor would have to replace the slab at no cost to the FDOT. Thus, to ensure the replacement slab meets the plastic property requirements, low water-cement ratio concrete and concrete accelerators are used. As a consequence, the heat of hydration increases causing larger early-age shrinkage, which could potentially lead to cracking [3–7]. As a consequence, the FDOT also specified a limit on the concrete temperature not to exceed 100 °F (38 °C) and requires the contractor to cure the slab with curing compound and cover the surface with white burlap-polyethylene curing blanket immediately after the slab hardens. Furthermore, if uncontrolled cracks appear on the replacement slab during the life of the contract, the contractor will have to replace the slab free of charge. While Section 353 is comprehensive and provides many levels of protection to FDOT, replacement slabs do crack. If cracks are discovered during the contract period, then the contractor will be responsible for replacing the slab again at his expense. However, some contractors may file a lawsuit against the FDOT to avoid their contractual obligation. Regardless of the outcome, further delay to the roadway, in which the slab is constructed, opening to traffic would have a direct impact on the traveling public. Therefore, there is a need in a better solution that can prevent or at least minimize the number of cracks for replacement slabs. One proposed solution is to replace ordinary concrete used in the construction of replacement slabs with fiber reinforced concrete (FRC). The use of FRC in slab-ongrade is not new as the building industry had been benefitting from it for over 30 years [8]. However, unlike slabs-on-grade in buildings, replacement slabs on roadways are exposed to the outdoor environment and need to withstand heavy truck traffic. FRC can enhance concrete in many different ways. Concrete fracture toughness and ductility can be improved using steel and synthetic fibers at high-dosage amount (>1% by volume) [9–13]. The glass fiber can improve fracture toughness but does not provide ductility enhancement. At low-dosage amount ( 0 for compression); and k is a factor that takes into account the size effect and given by: k ¼ 1þ

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi   200=deq  2:0 deq inmm

ð3Þ

The characteristic post-cracking residual tensile strength (fFtuk) of the FRC for shear is determined at a crack opening displacement (COD) of wu= 1.5 mm and is given by: fFtuk ¼ 0:45fR1k  0:6ð0:65fR1k  0:5fR3k Þ  0

ð4Þ

where fR1k and fR3k are flexural strengths determined in accordance with MC2010. The fRi values of the SFRC beam are available in Table 4. In the present stage of the optimization of the SFRC, relatively few number of tests were carried out for the evaluation of the fRi, the average values of these parameters were adopted for determining fFtu, by avoiding a too detrimental conjugated effect of this low number of specimens and relatively large CoV on the fRi,k values. The proposed section has no stirrups, hence the equation for determining the contribution of the transverse bar reinforcement (VRd,s) is not provided here. The design shear resistance cannot be greater than the crushing capacity of concrete in the web: VRd;max ¼ kc

fck coth þ cota bw :z: 1 þ cot2 h cc

ð5Þ

where z = 0.9 deq is the effective shear depth, h is the inclination of the CDC, kc = ke ηfc, ke = 0.55 and: gfc ¼ ð30=fc Þ1=3  1:0

ð6Þ

Development and Mechanical Characterization of Dry Fiber-reinforced

61

According to the above-mentioned equations, the design shear capacity of the FRC beam is 251 KN which is much higher than the maximum shear force (52.5 KN). 4.4

Deflection Control

For the deflection calculation of the slab’s module, the DefDOCROS software was used. The DefDocros program uses the M-v relationship data derived from DOCROS and predicts the force versus deflection (F-u) response of simply supported beams by using the displacement method described elsewhere [25]. According to the Eurocode-2 [19], the maximum deflection should not exceed L/250, where L is the length of the beam. Hence, the deflection limit for the slab’s module with 12 meters span length is 48 mm. The total long-term deflection of the studied beam is 36 mm where the creep factor is 1.73 and we have a 6.72 mm upward deflection due to the prestress force. Hence, the deflection of the beam is in the safe zone.

5 Conclusions From the observations of this experimental work, it can be concluded that dry SFRC can be produced with a minimum cement content of 300 kg and the steel fibre content of 60 kg per cubic metre with 40 MPa compressive strength at 3 days. At 3 days, the average fR1 and fR3 of SFRC containing fibres 60 kg/m3 were found to be 5.98 MPa and 6.38 MPa respectively and average shear strength of 13.49 MPa was attained by the produced SFRC. Analytical results show that using prefabricated prestress SFRC beam in slab’s module in a residential building the height of slab will be 350 mm and the shear resistance in the section is less than the maximum shear force so the beam can be cast without stirrups which are economically advantageous. Further, experimental investigations on the prototype precast long-span prestressed beams will be conducted for having a better understanding of the post-cracking behaviour against the flexural and shear. A design methodology was developed capable of optimize the prestressed prefabricated SFRC beam and the slab’s module by attending the SLS (deflection) and ULS (flexural and shear capacity) design criteria. Acknowledgements. The authors acknowledge the support provided by FEDER funds through the Operational Programme for Competitiveness and Internationalization (POCI) within the scope of the project n. 33883, SlabImp- Prefabricated lightweight and multifunctional large span slabs. The first two and the last Authors would like to acknowledge the grant provided by this project.

References 1. Park, H., Kang, S., Choi, K.: Analytical model for shear strength of ordinary and prestressed concrete beams. Eng. Struct. 46, 94–103 (2013) 2. Adebar, P., Mindess, S., StPierre, D., Olund, B.: Shear tests of fiber concrete beams without stirrups. ACI Struct. J. 94(1), 68–76 (1997)

62

K. B. Shahrbijari et al.

3. Mobasher, B., Yiming, Y., Soranakom, C.: Analytical solutions for flexural design of hybrid steel fiber reinforced concrete beams. Eng. Struct. 100, 164–177 (2015) 4. Frazão, C., Camões, A., Barros, J., Gonçalves, D.: Durability of steel fiber reinforced selfcompacting concrete. Constr. Build. Mater. 80, 155–166 (2015) 5. Mo, K.H., Yap, K.K.Q., Alengaram, U.J., Jumaat, M.Z.: The effect of steel fibres on the enhancement of flexural and compressive toughness and fracture characteristics of oil palm shell concrete. Constr. Build. Mater. 55, 20–28 (2014) 6. Atiş, C.D., Karahan, O.: Properties of steel fiber reinforced fly ash concrete. Constr. Build. Mater. 23(1), 392–399 (2009) 7. Mohammadhosseini, H., Tahir, M., Sam, A.R.M.: The feasibility of improving impact resistance and strength properties of sustainable concrete composites by adding waste metalized plastic fibres. Constr. Build. Mater. 169, 223–236 (2018) 8. Thomas, J., Ramaswamy, A.: Mechanical properties of steel fiber reinforced concrete. J. Mat. Civ. Eng. 19, 385–392 (2007) 9. Mohammadi, Y., Singh, S.P., Kaushik, S.K.: Properties of steel fibrous concrete containing mixed fibres in fresh and hardened state. Constr. Build. Mater. 22(5), 956–965 (2008) 10. Padmarajaiah, S.K., Ramaswamy, A.: Flexural strength predictions of steel fiber reinforced high-strength concrete in fully/partially prestressed beam specimens. Cem. Concr. Compos. 26(4), 275–290 (2004) 11. Barros, J.A.O., Taheri, M., Salehian, H., Mendes, P.J.D.: A design model for fibre reinforced concrete beams pre-stressed with steel and FRP bars. Compos. Struct. 94(8), 2494–2512 (2012) 12. Khanlou, A., MacRae, G., Scott, A., Hicks, S., Clifton, G., et al.: Shear performance of steel fibre-reinforced concrete. In: Australasian Structural Engineering Conference 2012: The Past, Present and Future of Structural Engineering, p. 400. Engineers Australia (2012) 13. Zamanzadeh, Z., Lourenço, L., Barros, J.: Recycled steel fibre reinforced concrete failing in bending and in shear. Constr. Build. Mater. 85, 195–207 (2015) 14. Brocks, W., Cornec, A., Scheider, I.: Computational aspects of nonlinear fracture mechanics. GKSS Forschungszentrum Geesthacht GMBH-Publications-GKSS, no. 30 (2003) 15. European Committee for Standardisation (CEN), ‘Eurocode 1: Basis of Design and Actions on Structures. Part 1: Basis of Design’, CEN, Brussels, ENV 1991–1 (1994) 16. Bymaster, J.C., Dang, C.N., Floyd, R.W., Hale, W.M.: Prestress losses in pretensioned concrete beams cast with lightweight self-consolidating concrete. Structures, vol. 2, pp. 50– 57 (2015) 17. AASHTO, LRFD. Specifications for Highway Bridges (1997) 18. Robitaille, S., Bartlett, F.M., Youssef, M.A.: Evaluating Prestress Losses During Pretensioning. St. Johns Canadian Society for Civil Engineering, Canada (2009) 19. European Committee for Standardisation (CEN), ‘Eurocode 2: Design of Concrete Structures. Part 1: General Rules and Rules for Buildings’, European Prestandard, CEN, Brussels, ENV 1992–1-1 (1991) 20. Basto, C.A.A., Barros, J.A.O.: Numeric simulation of sections submitted to bending. Technical report 08–DEC/E–46, p. 73 (2008) 21. Barros, J.A.O., Taheri, M., Salehian, H., Mendes, P.J.D.: A design model for fibre reinforced concrete beams pre-stressed with steel and FRP bars. Compos. Struct. 94(8), 2494–2512 (2012) 22. Soranakom, C., Mobasher, B.: Flexural design of fiber-reinforced concrete. ACI Mater. J. 106(5), 461–469 (2009) 23. Model Code 2010. fib Model Code for Concrete Structures 2010. International Federation for Structural Concrete (fib). Ernst & Sohn, Berlin, Germany (2013)

Development and Mechanical Characterization of Dry Fiber-reinforced

63

24. Barros, J.A.O., Foster, S.J.: An integrated approach for predicting the shear capacity of fibre reinforced concrete beams. Eng. Struct. 174, 346–357 (2018) 25. Varma, R.K.: Numerical models for the simulation of the cyclic behaviour of RC structures incorporating new advanced materials (2013)

Simulation of Fibre Orientation in Self-compacting Concrete: Case Studies Thomas Bauwens1,2, Steffen Grünewald1,3(&), and Geert De Schutter1 1

3

Ghent University, Ghent, Belgium [email protected] 2 Dirk Bauwens nv, Evergem, Belgium Delft University of Technology, Delft, The Netherlands

Abstract. Recent developments in concrete technology with high potential include ultra high performance concrete and self-compacting fibre reinforced concrete, which have a flowable consistency and can transport relatively high fibre dosages. Flowability is achieved by adopted mix design and both the mix design and flow affect the distribution and orientation of the fibres, which affect the post-cracking behaviour and accordingly the structural performance. With new materials also come new manufacturing and design approaches. The prediction of fibre orientation with computational fluid dynamics (CFD) simulations can be an important instrument to predict, understand and influence fibre orientation. With better understanding the mix design and casting process can be optimized. This paper reports about a study executed to determine the applicability of the software package Autodesk Moldflow for fluid dynamics simulations of flowable fibre concrete. After a discussion of relevant literature, two reference cases address stretching and shearing flow conditions in a qualitative and quantitative way. Concrete was modelled as an incompressible Bingham material with addition of a fibre orientation model that was developed by Folgar and Tucker. A third case, a square panel, was used as a reference and structural element for flow simulations. Parameters varied were among others rotary diffusion, wallslip and duration of casting. Keywords: Self-compacting concrete simulation
CFD


Fibres
Fibre orientation
Flow

1 Introduction Flowable Concrete The use of flowable concrete (FC) facilitates the construction with concrete and very specific solutions can be realised such as remote casting, casting in very congested areas and the production of concrete surfaces with architectural appearance. FC is a cluster of types of concrete, which distinguishes itself from traditional concrete through rheological characteristics obtained by tailor-made mix design and component selection. The flow behaviour of concrete can be modelled by rheological laws with a relation between the shear stress s and the shear rate c_ . The simplest model that captures the nature of the flow of concrete is the Bingham model [1], which is shown © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 64–74, 2021. https://doi.org/10.1007/978-3-030-58482-5_6

Simulation of Fibre Orientation in Self-compacting Concrete

65

by Eq. (1). This relatively simple law can be defined by two fundamental parameters: the yield stress sy and the plastic viscosity µp. Fibres, especially at higher dosages, have a significant effect on both rheological characteristics. s ¼ sy þ lp c_

ð1Þ

Fibre Reinforced Concrete Fibres have been added to concrete to transfer forces across cracks occurring during hardening or in service. Accordingly, the most important contribution of fibres is situated in the post-cracking tensile behaviour of concrete. The contribution of the fibres depends on their location and orientation with regard to the crack surface. The fibre orientation in flowable concrete types depends on the boundary conditions of the flow; the influence of fibre orientation on the structural performance is caused by the probability that a fibre will cross a crack. Fibre orientation can be experimentally determined for example by manual counting, image analysis, computed tomography, magnetic methods and electric conductivity. The knowledge of how fibres orient and how this information can be implemented for design and execution is key for the optimisation of their application. Potentially, fibres can be used more efficiently with adequate material selection, mix design and casting operation. Fluid Dynamics Simulations with Fibres The flow of concrete is a rather complex problem to model in a simulation [2]. A first method to model concrete is computational fluid dynamics (CFD), which assumes that the concrete is a single-phase liquid [3, 4]. The problem of the flow of concrete is solved by a set of flow equations, that are defined as partial differential equations (PDE). These PDE are converted into a set of algebraic equations which are solved at discrete points in space and time. The second method is the discrete element method (DEM) [5]. This method will reduce the concrete as if it was a set of interacting particles. The interaction in itself is modelled as normal and tangential forces that act on each other using springs, dashpots and slip elements to model the flow. This method primarily focusses on the discrete character of concrete. As always, there is a research field that tries to combine the best of both worlds. More advanced simulations will require more assumptions and come at higher computational costs, but more phenomena can be modelled and explained. All fibre orientation phenomena can be explained by two main drivers for fibre orientation: shearing and stretching flows. In an extensional flow mode, the shear forces will apply a torque on the fibre to align in the direction of the extensioning flow. In a flow dominated by shear stresses, the torque exerted on the fibre reaches a minimum when the fibre is parallel to the flow direction (e.g. in the presence of a wall). Both phenomena are illustrated by Figs. 1a and 1b. In order to express the orientation distribution of fibres, first of all, the orientation of a single fibre is defined. The fibres are assumed to be rigid cylinders, uniform in length and diameter; the orientation of a single fibre can be described as a unit vector p. The theory of fibre orientation is based on the works of Einstein [6] and Jeffery [7], who described respectively the movement of a sphere and ellipsoid in a fluid. The movement of a single fibre in a fluid can be assumed similar to the movement of an ellipsoid in a fluid. Therefore, the differential

66

T. Bauwens et al.

equation of Jeffery is applicable to the modelling of fibre orientation. Folgar and Tucker [8] suggested a diffusion coefficient to take into account the interaction between fibres and they proposed the relation of Dr ¼ CI c_ , in which CI is the fibre-fibre interaction coefficient depending on the fibre geometry and volume fraction which can be obtained through experiments. c_ is the effective shear rate. From a computational point of view, the modelling of every single fibre is not feasible, since this approach requires a lot of storage and computational effort. Therefore, it was necessary to capture the nature of the fibre orientation with a simplification. This simplification was investigated and implemented in the fibre orientation model by replacing the fibre orientation distribution by a second order tensor.

a) Stretching flow: fibres align in the stretching direction

b) Shearing flow: fibres align in the shearing direction

Fig. 1. Principal mechanisms for fibre orientation a, top) stretching flow and b, right) shearing flow.

The use of fibre orientation modelling has been applied to the study of the orientation of fibres in concrete a few times in the past decade [9, 10]. With new design approaches that allow taking into account fibre orientation, this knowledge can also be used to optimize material and structural performance and the casting process. In this study, the commercial software Autodesk Moldflow was applied as a tool, that still needed verification through appropriate experimental and numerical studies. Three cases were selected for first validations and to determine the need for future research.

2 Simulation Model 2.1

Software Applied

Autodesk Moldflow is a software type of the CFD-class that was designed for simulating and optimizing injection molding processes, which has some similarities with the casting of concrete and therefore, with an adaptation of the software, the flow of concrete can be modelled. These adaptations can be done by User Defined Functions for rheological and constitutive material laws. The biggest advantage of Moldflow over other softwares is the already implemented fibre orientation model of Folgar-Tucker.

Simulation of Fibre Orientation in Self-compacting Concrete

67

The latter was the reason why this software package was selected to perform simulations. 2.2

Material Characteristics

To make Moldflow suitable for the modelling of concrete, adaptations were made with regard to the constitutive and rheological properties of the material [11]. Since concrete was not yet modelled within Moldflow Insight, some adaptations were developed on the pressure, volume, velocity-relationship and rheological laws of the model. Moldflow models the fluid default with a 2-domain Tait pvT-model and Cross-WLF viscosity model. Concrete was modelled as an incompressible fluid with a Bingham plastic rheological law. This was done by assembling a dynamic link library (.dll-file) in Visual Studio and using Moldflow Application Programming Interface (API), which allows the user to insert the desired plastic viscosity and yield stress in the model. Three parameters are necessary to characterize the implemented Bingham model: the plastic viscosity, the yield stress and the shear rate cut-off. The first two parameters are necessary to facilitate the standard Bingham model. The third parameter is necessary to reduce the computation time by limiting the viscosity for low shear rates. This simplification does not influence the final solution since fibre orientation happens at high shear rates and the cut-off is chosen low enough to not influence the flow. Potentially, computation time could be saved by increasing this criterion. For the following three cases, a plastic viscosity of 105 Pa
s, a yield stress of 37 Pa and a shear rate cut-off of 0.0001 s−1 were used. The first two parameters are the same as measured by [12]; the addition of steel fibres typically increases both characteristics compared to a reference SCC without fibres. The dosage and the type of fibres were taken into account by the rotary diffusion coefficient CI. A CI. of 0 means that no fibre-fibre interactions were taken into account and therefore the simulations were run with the differential equation that was derived by Jeffery.

3 Case 1: Radial Flow The fibre orientation model was verified by simulating flow fields which induce shearing and stretching flows (Sects. 3 and 4). First a radial flow field was simulated in which the fibres were expected to align perpendicular to the flow direction. The model consists of a cylinder with a radius of 0.50 m and a height of 0.30 m. The results of the fibre orientation at the end of the simulation are shown in Fig. 2 and they confirm qualitatively the principle of flow-induced fibre orientation by stretching. At the wall of the panel mould the fibres are mainly oriented perpendicular to the flow direction, as was also observed e.g. in experimental studies executed with panel casting [13] and on tunnel segments [14]. In the centre, the fibre orientation is relatively random. A second observation is that fibre orientation occurs very fast, such observation can be also made after the execution of the slump flow test, after which the preferred orientation of fibres at the border of the flow spread is visible. The transition from green to yellow/red colour in the simulations mainly is observed in the last third of the panel radius (orientation number larger than 0.70). The maximum fibre orientation obtained by

68

T. Bauwens et al.

panel testing was 0.866 [13], as this was an average value the results of the simulations are realistic. Orientation numbers of up to about 0.90 were found in tunnel segments [14] in a similar flow condition.

Fig. 2. Final average fibre orientation after radial flow, fibre orientation tensors in X-Y plane.

4 Case 2: Flow Through a Pipe 4.1

Simulation Without Wall Slip

The shearing flow is obtained by simulating the flow in a cylindrical pipe with a diameter of 300 mm and a length of 10 m (time to fill the pipe: 140 s). As illustrated in Fig. 3, the fibres quickly align parallel to the flow direction close to the walls as expected (location 2 m from inlet). This confirms in a qualitative manner that the fibre orientation model is correct for shearing flow fields. Martinie and Roussel [15] carried out flow simulations on the shear flow between two parallel plates; with a fibre reinforced SCC having a yield strength of 50 Pa orientation numbers of up to 0.966 were obtained close to the walls. 4.2

Simulation of Wall Slip

Furthermore, the influence of a wall slip model on the fibre orientation was investigated with Moldflow. This was done by variation of wall slip model parameters for the pipe flow simulation. From a macroscopic point of view, the physical phenomenon of wall slip can be modelled as a slip velocity. In this approach the velocity at the boundary is not taken equal to zero, as is assumed in classic no-slip boundary conditions of Newtonian fluid mechanics. When wall slip occurs, the shearing of the fluid will reduce

Simulation of Fibre Orientation in Self-compacting Concrete

69

Fig. 3. Fibre orientation within the pipe.

Fig. 4. Flow simulations a, left) velocity profile in the pipe, with wall slip and without wall slip and b, right) fibre orientation within the pipe, with wall slip.

and the influenced zone of fibre orientation due to shearing will be smaller. As a result, there is only a noticeable velocity gradient at the wall of the pipe and therefore the fibre orientation will occur slower compared to a model without wall slip. Fibre located in the centre will orientate only very slowly. The effect on the velocity profile is shown in Fig. 4a. Wall slip was observed with a flow velocity of about 8 cm/s. In the middle of the pipe the flow occurs as a plug flow. In the model without wall slip, a parabolic velocity profile is obtained. The fibre orientation was investigated at the same location (2 m away from the inlet) as shown by Fig. 3 and the orientation profile is shown in Fig. 4b. The results show that the influence of the wall slip on the fibre orientation is not negligible. It can be noticed that the fibre orientation is lower at the walls of the pipe compared to the model without wall slip and that the fibre orientation remains random in the middle of the pipe. Both phenomena can be explained by the difference

70

T. Bauwens et al.

in velocity profiles. Since shearing is equal to the gradient of the velocity profile, the shear rate will be less at the walls of the pipe and zero in the middle of the pipe. As fibre orientation is driven by shearing, it is logical that the fibre orientation is relatively lower at the walls and no fibre orientation takes place in the centre of the pipe. It can be concluded that the influence of the wall slip on the fibre orientation is not negligible and therefore further analysis is required on the subject of fibre orientation and the influence of wall slip.

5 Case 3: Casting of a Slab 5.1

Simulation Set-up

5.1.1 Material Characteristics In order to verify the accuracy of the model, experimental data from literature was chosen as a reference case. Žirgulis [12] discusses in his PhD-thesis experimental results as well as numerical simulations of the casting of fibre reinforced concrete. The experimental work was done at the Norwegian University of Science and Technology in Trondheim and the numerical simulations were carried out at the Technical University of Denmark in Lyngby. The experimental work consisted of casting square slabs with a width of 1.2 m and a thickness of 150 mm. The pouring was executed at a distance of 200 mm from the sides through a hose with a diameter of 150 mm from a height of 200 mm. The concrete contained 0.5 vol.% steel fibres. The fibres had hooked ends and were produced of cold-drawn wire, having a length of 60 mm, an aspect ratio of 80, and a minimum tensile strength of 1050 N/mm2. The rheological properties were determined 20–30 min after mixing with the 4C rheometer (plastic viscosity: 105 Pa
s; yield stress: 37 Pa). The density of the fresh concrete was 2486 kg/m3. The surfaces of the formwork were smooth. The following five general conclusions were drawn by Žirgulis: 1) The fibres align at the walls due to shearing, 2) The fibres align perpendicular to the flow in the middle of the plate due to stretching, 3) The stretching effect is more present in the top half of the plate than the bottom half and 4) The fibres shear over the bottom at the left under corner and 5) At the inlet the fibres are randomly oriented. Based on all comparisons between results of CT scans and flow simulations, it was concluded in [12] that the stretching and shearing were underestimated in the numerical simulation, but qualitatively, the results were accurate. The conclusions of Žirgulis were qualitatively assessed and confirmed in this study. 5.1.2 Geometry, Mesh and Process Settings The geometry of the slab was then reduced to the simplest form which was a simple box with a cylindrical inlet condition with a diameter of 150 mm. The final mesh contained 1,377,439 3D tetrahedral elements. These were generated by first creating a grid on the planes of the model and then meshing over the depth of the elements. A bias factor of 2 and a global mesh factor of 0.9 were used as local mesh refinements to solve the viscous effects. The viscosity function and the pvt-relation were installed by userdefined API’s. The modelled fluid was an incompressible Bingham material with

Simulation of Fibre Orientation in Self-compacting Concrete

71

properties reported in Sect. 2.2. The inlet was automatically controlled in such a way that the full slab would be filled in 300 s. The mould was underfilled (concrete kept a slope during casting) and hence the simulation was aborted when the slab had a minimum thickness of 150 mm. This was after approximately 120 s. 5.2

Simulation Results

5.2.1 Model Without Rotary Diffusion The slab model was simulated with Moldflow to show its capabilities. No rotary diffusion was included in the first simulation since it was used as a benchmark. Practically this means that fibre interactions are neglected, the stretching patterns are less visible. Some rotary diffusion is necessary to show stretching behaviour. This can be explained by the Jeffery orbitals, the orbitals that are perpendicular to the flow are less stable and therefore during simulation the numerical solution can occasionally fail. When using rotary diffusion, this disturbance is averaged over the neighbouring cells. Figure 5a shows the fibre orientation tensors at the bottom of the slab. The range of orientation numbers is indicated in Fig. 5 from dark red (0.90) to dark blue (0.333). At the casting location, shearing flow is most present and hence the fibres orientate parallel to the flow direction, which is radial. In the corner opposite to the inlet, the fibre orientation is even more pronounced due to the converging flow field. Figure 5b shows the fibre orientation tensors at the top of the slab. At the top of the slab, the shear rate is smaller and hence the stretching flow is more present. The fibres aligned more perpendicularly to the flow direction compared to the bottom of the mould. The fibre orientation simulation seems to be accurate, also when compared with the outcomes of simulations described by [12].

Fig. 5. Overview of fibre orientation in the slab a, left) at the bottom and b) at the top of the slab, fibre orientation tensors in X-Y directions

The conclusions of Žirgulis are confirmed qualitatively with the Moldflow simulations (Fig. 5): At the inlet where the concrete was cast fibres were randomly oriented. Fibres preferably oriented along walls and perpendicular to the flow direction in the

72

T. Bauwens et al.

middle of the plate. In the top half of the plate stretching was more pronounced compared to the bottom half. 5.2.2 Influence of Rotary Diffusion Figure 6 shows the effect of the rotary diffusion on the local fibre orientation at the side of the plate; the fibre-fibre interaction coefficient CI was varied. As discussed, rotary diffusion takes into account fibre-fibre interactions and increases the stability of the solution. The position of the simulation location for Fig. 6 is at half the length of the mould (side of the mould); values in the x-axis are distances from the bottom of the mould. The fibres aligned along the wall at this location. In general, the principal value of the fibre orientation is reduced at increasing interaction. It can be concluded that the rotary diffusion reduces the fibre orientation component Tyy and increases the fibre orientation in the other directions.

Fig. 6. Principal values of fibre orientation tensor for various CI-values (wall closest to inlet, half the length, close to the wall).

5.2.3 Influence of Wall Slip The influence of wall slip on the principal value of the fibre orientation tensor in the centre of the slab is shown by Fig. 7. The fibre orientation at the bottom of the slab is not greatly influenced, the fibre orientation at the top is. The fibre orientation at the top shows more a stretching fibre orientation, which resembles the experimental results better compared to the model without wall slip. 5.2.4 Influence of Duration of Casting Figure 8, the influence of a longer duration of casting (about doubled to 220 s) on the principal value of the fibre orientation tensor in the centre of the slab is shown. This reduces the flow rate, and hence less shearing will be present. It can be noticed that, at the bottom, the fibre orientation is lower when the flow velocity decreases, and fibre orientation is higher at the top of the slab. Both phenomena can be explained by the lower shearing flow condition during the casting of the slab.

Simulation of Fibre Orientation in Self-compacting Concrete

73

Fig. 7. Principal values of fibre orientation tensor with and without wall slip, in the centre of the plate.

Fig. 8. Principal value of fibre orientation tensor for longer casting duration, in the centre of the plate.

6 Conclusions This paper discusses the application of Autodesk Moldflow fluid dynamics software for the simulation of fibre orientation in specific cases with self-compacting fibre reinforced concrete. Case studies were executed to simulate flow fields which induce shearing and stretching flows; cases studied were a round panel, a pipe and a slab. Affecting parameters on the outcome of the simulation were assessed. Based on the study the following conclusions can be drawn: • The qualitative accuracy of the fibre orientation model was demonstrated through case studies, where the fibre orientation tensors aligned respectively parallel and perpendicular to the flow direction.

74

T. Bauwens et al.

• The addition of rotary diffusion in the model smears out the fibre orientation tensor. This reduces the principal values of the fibre orientation tensors and increases the stability of the solution. • With wall slip the gradient in the velocity profile decreases and therefore, the shearing of the flow is reduced. • A lower flow rate reduced the velocity and shear rate in the concrete. This reduced the principal value of fibre orientation at the walls. • A more detailed analysis of the quantitative accuracy is required for further implementation.

References 1. Wallevik, O.H.: Rheology – a scientific approach to develop self-compacting concrete. In: 3rd International Symposium on SCC, Rilem, Reykjavik, pp. 23–31 (2003) 2. TC 222-SCF, Simulation of fresh concrete flow, State-of-the-art report of the RILEM Technical Committee 222-SCF, Eds.: Roussel, N., Gram, A., ISBN: 978–94-017-8883-0 (2014) 3. Thrane, L.N.: Form Filling with Self-compacting Concrete. PhD-thesis, DTU Lyngby (2007) 4. Gram, A.: Modelling Bingham Suspensional Flow - Influence of Viscosity and Particle Properties Applicable to Cementitious Materials. PhD-thesis, KTH Royal Institute of Technology, Stockholm, ISSN 1103–4270 (2015) 5. Ferrara, L., Shyshko, S., Mechtcherine, V.: Predicting the flow-induced fiber dispersion and orientation in self-consolidating concrete by distinct element method. BEFIB 2012, ISBN: 978–2-35158-132-2, pp. 213-214 (2012) 6. Einstein, A.: Eine neue Bestimmung der Moleküldimensionen. PhD-thesis, Universität Zürich. https://doi.org/10.3929/ethz-a-000565688 (1905) 7. Jeffery, G.B.: The motion of ellipsoidal particles immersed in a viscous fluid. Proc. R. Soc. Lond. Ser. A 102(715), 161–179 (1922) 8. Folgar, F., Tucker, C.L.: Orientation behavior of fibers in concentrated suspensions. J. Reinf. Plast. Compos. 3(2), 98–119 (1984) 9. Martinie, L.: Comportement rheologique et mise en œuvre des materiaux cimentaires fibres. PhD-thesis, 2010PEST1077 (2010) 10. Svec, O., Skocek, J., Stang, H., Olesen, J.F., Poulsen, P.N.: Flow simulation of fiber reinforced self-compacting concrete using Lattice Boltzmann method. Dissemination (2011) 11. Bauwens, T.: Fibre Orientation in Self-compacting Concrete: Effect of Mix Design and Casting Process. Master-thesis, Ghent University, Ghent (2018) 12. Zirgulis, G.: Fibre Orientation in Steel-Fibre-Reinforced Concrete: Quantification methods and influence of formwork surface and reinforcement bars in structural elements. PhD-thesis, NTNU Trondheim (2015) 13. Abrishambaf, A., Barros, J.A.O., Cunha, V.M.C.F.: Relation between fibre distribution and post-cracking behaviour in steel fibre reinforced self-compacting concrete panels. Cem. Concr. Res. 51, 57–66 (2013) 14. Grünewald, S.: Performance-based design of self-compacting fibre reinforced concrete. PhDthesis, Delft University of Technology (2004) 15. Martinie, L., Roussel, N.: Fiber-reinforced cementitious materials: from intrinsic properties to fiber alignment. In: Khayat, K.H., Feys, D. (eds), Design, Production and Placement of Self Consolidating Concrete, Dordrecht, RILEM Bookseries, vol. 1, pp. 407–415 (2015)

Mix Design and Properties of Self-compacting Fibrous Concrete Rafael R. Polvere(&), Ana R. L. Pires, Sidiclei Formagini, and Andrés B. Cheung FAENG/UFMS, Federal University of Mato Grosso do Sul, Campo Grande, Brazil [email protected]

Abstract. The development of self-compacting fiber reinforced concrete (SCFRC) marks an important milestone of the Brazilian building industry, because it combines the benefits of high fluidity in the fresh state, better performance on tensile strength and the control of the cracks. To an efficient performance, it is necessary a good granular mixture proportioning. Thus, the objective of this research is the evaluation of the influence on fiber contents in self-compacting concrete of 40 MPa and its properties in fresh and hardened state. The difference of the concretes was the volume levels of the steel fibers in each mixture proportioning: SCC0F (no addition), SCC0.5F (0,5% in volume) and SCC1F (1,0% in volume). The concrete mix design was based in the compressible particle model. The results show that the insertion of steel fibers interferes diminishing the workability and fluidity of the mixtures and improve the tensile strength. Keywords: Self-compacting fiber reinforced concrete
Steel fibers
Mixtures design, compressible packing model

1 Introduction Self-compacting concrete (SCC) is characterized by its self-compacting capability without the need for additional internal or external vibration. Your development in 1988 enabled the execution of concrete structures without the need for vibration due to its flow ability and self-consolidating properties [1]. It was an important milestone in improving the efficiency of the construction industry, because it is an easy and fast to apply and requires a reduced number of people for application. The requirements properties of SCC in the fresh state are mainly resumed to the filling ability, the passing capacity, and the resistance to segregation. As concrete has low tensile strength, the introduction of fibers improves its behavior when requested to tensile and flexural bending. Fibers, when added in appropriate volumes, can overcome this deficiency by attributing ductile behavior. Steel fibers proved to have the potential to increase the post cracking energy absorption capacity of cement-based materials, enhancing the ductile properties of concrete structures behavior [2].

© RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 75–86, 2021. https://doi.org/10.1007/978-3-030-58482-5_7

76

R. R. Polvere et al.

This way steel fiber reinforced self-compacting concrete (SCFRC) combines the benefits of SCC in the fresh state reducing cracks and showing an improved performance in the hardened state compared to conventional concrete. The composition and production of SCFRC are more complicated than SCC, because it has an ideal volume fraction of steel fiber to ensure higher workability and better mechanical performance. Due to the packing effect of the fiber-aggregate solid skeleton, the self-compacting performance could not be achieved once the volume fraction of steel fiber in concrete exceeds the limited value even if the concrete mixture is a homogeneous and stable suspension [3]. The limitation, usually less than 2%, is mainly affected by the properties of raw material [4] and granular mix proportion [5]. Grunewald [6] proposed a performance-based mix design method for SCFRC. The method used the Compressible Packing Model (CPM) [7] and assumes that the steel fiber is the equivalent packing diameter [8]. The packing density of the aggregates and also the fibers in the mix of SCFRC determine the amount of cement paste that is required to fill the interstices of the granular skeleton. When reinforced in modulus and amount with fibers, concretes minimize this fragile behavior. According to Gois [9], fibers act as connecting bridges, transferring stresses from side to side of the matrix, minimizing stresses at the ends of the cracks. Naaman [10] proposed that the main advantage in adding fibers to concrete is the ability to modify the behavior of the material from brittle to ductile when concrete is evaluated the traction, because the fibers, when crossing the cracks, create bridges that make it difficult to increase their opening. Already Melo Filho [11] observed significant increases in tensile strength after the appearance of the first crack when evaluating the direct tensile tests was performed. Although the mechanical benefits of fibers added to concrete are known, few studies exist on the joint action of the hybrid effort of steel fibers on concrete. In this paper, the objective is the comparison and analysis between SCC and SCFRC, which have two different levels of steel fiber incorporation. To achieve the overall goal it’s necessary to fulfill the specific objective, which is the study of the determination of the dosage and characterization of the properties of steel fiber reinforced self-compacting concretes, in the fresh and hardened state.

2 Experimental Program Three concrete mixes were designed: one reference mix of SCC without fibers (SCC0F) and two SCFRC with steel fibers volume insertion of 0.5% (SCC0.5F) and 1.0% (SCC1F). All concrete were produced with the same dry granular mixture proportioning (except the steel fiber contents) in an inclined shaft concrete mixer, being a production of 122 L of concrete for each mix. This amount was sufficient for the fresh state characterization tests, molding of specimens to determine the properties in the hardened state.

Mix Design and Properties of Self-compacting Fibrous Concrete

2.1

77

Materials

The cementitious materials used to produce the concretes were Brazilian Portland cement type CP II E– 32 (addition varying from 5 to 32% of blast furnace slag) and silica fume (SF). Aggregates were used: coarse aggregate originating from the region’s basaltic rock crushing (G12.5), natural very fine sand (NS) and gravel sand (GS). As the natural sand is very fine crushing sand was used for grain size correction. It was chosen to work with a superplasticizer additive that provided fluidity, avoiding segregation and giving concrete more workability and water from the supply network. The steel fibers inserted in the last two mixes had hooks at the ends and a ratio (l/d) of 50, 30 mm of length and 0.60 of diameter. The materials were experimentally characterized according to the input parameters required by the packing density model [12] used for the granular mixture proportioning. The particle size distributions of aggregates (by mechanical sieving) [13] are illustrated in Fig. 1. The maximum aggregate size was limited to 12.5 mm, so that there was no discontinuity in the granular skeleton, since that the natural sand was very thin, as well as reducing the risk of segregation during the application of self-compacting concrete. The other physical properties of aggregates required for the preparation of self-compacting concrete mixtures are presented in Table 1.

Accumulated Fraction (%)

0.0 20.0

G.12,5% GS NS

40.0 60.0 80.0 100.0

0.1

1 Diameter (mm)

10

Fig. 1. Aggregate grading curve.

2.2

Concrete Mixture Proportioning

The theoretical mixer design of self-compacting concretes was performed by the computer model [12] based on the theoretical formulation using the compressible packing density model [7]. From various simulations of different granular compositions, appropriate selection of materials was carried out to produce the maximum packing density of the granular mixture, ensuring that all desired physical properties both in the fresh state and after hardening were achieved. The challenge of mixture was to associate the use of very fine natural sand with the high density coarse aggregate. For high levels of fines required for a good consistency of the self-compacting concrete, the continuity of the granular skeleton is guaranteed by the mortar, with the large aggregate

78

R. R. Polvere et al. Table 1. Aggregate properties. Property

Test Methodology Unity Material G 12.5 SN 3 1.59 Unit Mass [14] g/cm 1.68 Specific Mass [15] g/cm3 2.88 2.63 Water Absorption % 0.71 – Shape Index [16] – 2.03 – Organic Impurities [17] ppm – 300 Powdery Material [18] % 1.89 1.40 Characteristic Maximum Dimension [13] mm 12.5 0.60 Fineness Module [13] – 6.06 1.17

CS 1.90 2.85 – – – 11.0 4.75 3.14

being dispersed in smaller quantities in the mortar, which facilitates the segregation of the concrete. The three mix designs are presented in Table 2. Table 2. Mixes of the concrete produced to 1 m3. Unity Consumption by m3 of concrete SCC0F SCC0.5F SCC1F 390 390 Cement kg/m3 390 Silica Fume kg/m3 38 38 38 Natural Fine Sand kg/m3 661 661 661 Crushing Fine Sand kg/m3 301 301 301 Coarse aggregate kg/m3 966 966 966 3 Steel Fibers kg/m – 40 81 Water l/m3 190 198 198 Superplasticizer l/m3 4.80 4.90 4.50 Mortar Content % 69.7 69.7 69.7 Expected Compressive Strength, 28 days MPa 40 40 40 Ratio w/c 0.49 0.51 0.51

Material

2.3

Testing Method

Tests were performed shortly after concrete production to evaluate their behavior in the fresh state. Fresh state tests evaluate and frame self-compacting characteristics and properties according to the requirements of the technical standard of Brazilian NBR 15823 [19]. The tests were performed to determine the spreading, flow time and visual stability index, determining the passing ability, and viscosity determination. Figure 2 shows parts of the tests performed in the fresh state of the concrete: Abrams cone spreading; flow through L box; and flow through V funnel.

Mix Design and Properties of Self-compacting Fibrous Concrete

(a) Abrams cone.

(b) L box.

79

(c) Funnel V.

Fig. 2. Fresh concrete tests.

After completion of the tests in the fresh state, 20 cylindrical and 3 prismatic specimens were molded in each mixture to characterize the hardened state concrete. All cylindrical specimens were 10 cm in diameter and 20 cm high and the prismatic specimens were 15  15  50 cm3. The samples were kept in a humid chamber for curing until the tests were performed. To determine the mechanical properties of concrete, compression tests were performed [20], tensile strength [21], splitting test [22], modulus of elasticity [23], and determination of the absorption, voids index and specific mass of concrete [24].

3 Presentation and Analysis of Results 3.1

Fresh Concrete Properties

The behavior of the concrete in the fresh state was different in the mixtures produced, since the same composition of the materials was considered and only the fiber content was varied. Even with the mixture of very fine natural sand with the high density aggregate it was possible to produce SCC with satisfactory behavior in the fresh state. In general, they had good spread through the Abrams cone as shown in Fig. 3 and experimental data available in Table 3. SCC0F was homogeneous, cohesive and without evidence of segregation and exudation. SCC0.5F indicated evidence of exudation as showed at the edges of the spread concrete, not being significant to affect its stability. In SCC1F the spread showed that fresh concrete was less homogeneous, with a small accumulation of coarse aggregate in the central region. It is believed that for the granular mixture used, the 1% fiber percentage has affected the mobility of the aggregates contributing to such behavior and appearance, being a critical point for this fibers percentage and the mixture proportioning not ware adjusted for the fiber contents. The spreading values of the Abrams cone varied within the necessary range to be classified as SCC according to NBR 15823 [19]. In respect to the apparent plastic

80

R. R. Polvere et al.

a. SCC0F.

b. SCC0.5F.

c. SCC1F.

Fig. 3. Spreading and aspects of concretes. Table 3. Properties and framing of concrete in the fresh state. Test

SCC0F Result Visual Inspection Highly stable Spreading (slump flow) 750 mm Apparent viscosity t500 2 mm, the VC increases again. The different responses not only depend on the material of the filament, they are significantly influenced by both the bond mechanism and the amount of fibres.

Fig. 4. Representative impact curves (left). Variation of toughness with CMOD during standard bending tests (right).

For comparison Fig. 4 (right) represents the evolution of static toughness (calculated as the area below the load-CMOD curve) evaluated from static bending tests. As expected, the values of energy are lower than those measured during impact tests; impact loads are applied during a very short time and also part of the energy can be dissipated by other mechanisms as vibrations or friction among others; it is well known that the cracks propagation is time dependent. Although in general terms for each FRC the increments in toughness are qualitatively consistent between static and impact tests, it can be seen that for great crack openings polymer FRC behaves better during impact tests than what could have been predicted based on static tests. While S25 or S50 achieved greater toughness in static bending than P5 or P10 respectively, the contrary occurs when the impact curves are compared. Even though, this situation could vary between fibres made of a same type of material, these results demonstrate that differential behaviours can appear. This highlights the necessity of a test for impact resistance evaluation. Table 2 presents the mean values of compressive strength (fc) as well as the first crack, or limit of proportionality (fL) and the residual stresses fR1 and fR3, corresponding to CMOD of 0.5 and 2.5 mm respectively, obtained in bending tests. In addition, the mean values of the cracking and post-cracking energies (EC, EP), the total energy (ET), the initial visible crack opening (CODC) and the COD growth rate (VC) from impact tests are given. There are no significant differences in EC between all concretes, indicating that it mainly depends on matrix strength. It was also confirmed that fibre incorporation reduces

Impact Response of Different Classes of FRC

195

the initial crack size even in the case of low toughness FRC. In the case of steel fibres although more energy was applied, the initial opening was smaller in S50 than in S25; in concretes incorporating polymer or glass macro fibres CODC was greater for the higher dosages of fibres as greater impact energy was required to cause the initial crack as seen when comparing the corresponding EC values.

Table 2. Summary of results from static and impact tests. Concrete Static characterization fL fR1 fR3 fc (MPa) R 44.2 4.04 – – S25 44.5 4.75 3.27 2.93 S50 44.8 4.42 5.36 4.87 P5 47.3 4.38 1.77 1.94 P10 46.3 4.21 2.67 3.67 G6 47.1 4.51 1.81 0.90 G12 46.6 4.83 2.88 1.67

fR3/fR1 – 0.89 0.91 1.10 1.38 0.50 0.58

Impact test FRC class EC EP (J) – 103 22 3b 104 215 5c 111 670 2d 95 589 2.5e 102 1183 2a 89 180 3a 114 221

ET

CODC VC (lm) (mm/J)

125 319 780 684 1285 269 334

751 116 69 96 119 91 111

0.191 0.018 0.007 0.005 0.002 0.025 0.019

Analysing the results of VC it appears that both in steel and polymeric FRC the COD growth rate decreases as the volume of fibres increases. The post cracking energy EP is clearly associated with the COD growth VC as shown in Fig. 5.

Fig. 5. Variation of post-cracking energy (EP) with COD growth rate (VC).

Figure 6 plots the relationship between the results of impact tests and the residual stresses fR1 and fR3. Contrary to the case of EC, which is practically independent of FRC toughness, it can be seen that as the residual stresses increase the total cumulated energy increases (Fig. 6 left). At the same time, although VC decreases as FRC residual capacity

196

J. C. Vivas et al.

increases, as expected, its values depend on the fibre type, and polymer FRC showed VC results clearly lower than those obtained with other types of fibres (Fig. 6 right).

Fig. 6. Variation of ET (left) and VC (right) with the residual stresses fR1 (empty symbols) and fR3 (filled symbols) calculated from standard bending tests.

Figure 7 shows the influence of the residual stresses fR3/fR1 ratio on impact test results: the measured energies and the COD growth rate. The dashed vertical lines indicate the limits for the fR3/fR1 ratio (a, b, c, d, e) used to characterise the shape of the FRC post-cracking branch (hardening/softening) [17]. It should be noted that the FRC studied cover all of the post-peak responses established in this Code. It can be seen that as the residual stress ratio increases the cracking energy is not modified while the postcracking energy (and consequently the total energy) clearly increases (Fig. 7 left). When the values of VC are plotted against fR3/fR1 ratio (Fig. 7 right), the rate strongly decreases (R2 = 0.88). In both cases the results follow similar tendencies beyond the type of fibre used. This would appear as an additional advantage of the classification system adopted for FRC in the fib Model Code.

Fig. 7. Influence of fR3/fR1 ratio on impact test parameters. Left: effects on cracking, postcracking and total energies. Right: effects on VC.

Impact Response of Different Classes of FRC

197

4 Conclusions A drop-weight test on notched prisms was used to evaluate the contribution of different fibres on the impact resistance. Impact resistance was characterised in terms of cracking (EC), post-cracking (EP) and total (ET) cumulated energy and the COD growth rate (VC). Diverse classes of FRC obtained by incorporating different contents of steel, glass and polymer macrofibres were analysed. Main conclusions for the studied FRC are summarized as follows. • A repeated drop-weight test was implemented which easily enables to evaluate both the resistance to the first crack and the behaviour in cracked state of fibre concretes. Impact test results among specimens of a same FRC show reasonable variability. • The energy of cracking was not significantly affected by type and content of fibres. Nevertheless, fibre incorporation reduced the initial crack size even in the case of low toughness FRC. • Increase in concrete impact toughness, expressed as total cumulated energy during impact test, were mainly observed after matrix cracking, and for steel and polymeric FRC. • When comparing the post-cracking response some differences in the shape of the impact curves were found as a function of the material type of the fibres. Polymeric FRC were particularly efficient at large crack openings. • A consistent relationship between the residual stresses measured in the static bending test and the post-cracking impact test results for the same fibre type was observed. As the residual stresses increase the total cumulated energy increases and VC decreases. • The results of VC show an inverse tendency with the residual stresses ratio fR3/fR1 beyond the type of fibre incorporated in concrete, which would appear as an additional advantage of the classification system adopted for FRC in the fib Model Code. At the present, these findings correspond to the studied FRC and further studies are necessary for generalization of the conclusions. Experimental tests exploring the effects of concrete strength and other varieties of steel, polymer and glass macrofibres are in progress. Acknowledgements. The authors thank the collaboration of Eng. Francisco Hours and Pablo Bossio on the experimental works, funding from LEMIT-CIC and the projects CONICET PIP112-201501-00861 and UNLP 11/I188.

References 1. Drdlová, M., Buchar, J., Krátký, J., Řídký, R.: Blast resistance characteristics of concrete with different types of fibre reinforcement. Struct. Concr. 16, 508–517 (2015) 2. Luccioni, B., Isla, F., Codina, R., Ambrosini, D., Zerbino, R., Giaccio, G., Torrijos, M.C.: Experimental and numerical analysis of blast response of high strength fiber reinforced concrete slabs. Eng. Struct. 175, 113–122 (2018)

198

J. C. Vivas et al.

3. Luccioni, B., Isla, F., Codina, R., Ambrosini, D., Zerbino, R., Giaccio, G., Torrijos, M.C.: Effect of steel fibers on static and blast response of high strength concrete. Int. J. Impact Eng. 107, 23–37 (2017) 4. Almusallam, T.H., Siddiqui, N.A., Iqbal, R.A., Abbas, H.: Response of hybrid-fiber reinforced concrete slabs to hard projectile impact. Int. J. Impact Eng. 58, 17–30 (2013) 5. Yahaghi, J., Muda, Z.C., Beddu, S.B.: Impact resistance of oil palm shells concrete reinforced with polypropylene fibre. Constr. Build. Mater. 123, 394–403 (2016) 6. Zhu, X.C., Zhu, H., Li, H.R.: Drop-weight impact test on U-shape concrete specimens with statistical and regression analyses. Materials (Basel) 8, 5877–5890 (2015) 7. Hrynyk, T.D., Vecchio, F.J.: Behavior of steel fiber-reinforced concrete slabs under impact load. ACI Struct. J. 111, 1213–1224 (2014) 8. ACI Committee 544: Measurement of Properties of Fiber Reinforced Concrete 544.2R-89. In: ACI 544.2R (1999) 9. Banthia, N., Mindess, S., Trottier, J.: Impact resistance of steel fiber reinforced concrete. ACI Mater. J. 93, 472–479 (1996) 10. Mohee, F.M.: The effects of strain rate on concrete strength under dynamic impact load. J. Bangladesh Electron. Soc. 16, 83–90 (2016) 11. Radomski, W.: Application of the rotating impact machine for testing fibre-reinforced concrete. Int. J. Cem. Compos. Light. Concr. 3, 3–12 (1981) 12. Zhang, X.X., Ruiz, G., Yu, R.C.: A new drop weight impact machine for studying the fracture behaviour of structural concrete. WIT Trans. Built Environ. 98, 251–259 (2008) 13. Banthia, N.P., Mindess, S., Bentur, A.: Impact behaviour of concrete beams. Mater. Struct. 20, 293–302 (1987) 14. Bindiganavile, V., Banthia, N.: Polymer and steel fiber-reinforced cementitious composites under impact loading, part 1: bond-slip response. ACI Mater. J. 98, 10–16 (1998) 15. Bindiganavile, V., Banthia, N.: Polymer and steel fiber-reinforced cementitious composites under impact loading, part 2: flexural thoughness. ACI Mater. J. 98, 17–24 (1998) 16. Rahmani, T., Kiani, B., Shekarchi, M., Safari, A.: Statistical and experimental analysis on the behavior of fiber reinforced concretes subjected to drop weight test. Constr. Build. Mater. 37, 360–369 (2012) 17. International Federation for Structural Concrete (fib): Model Code, vol. 1 (2012) 18. Technical Committee CEN/TC 229: EN 14651:2005 Test Method for Metallic Fibered Concrete - Measuring the Flexural Tensile Strength (Limit of Proportionality (LOP), Residual) Méthode (2005) 19. Vivas, J., Zerbino, R.L.: Estudio de la resistencia al impacto de hormigones reforzados con fibras. In: 19 Congress International Metallurgy and Materials, CONAMET-SAM, Valdivia, Chile, pp. 140–141 (2019) 20. American Society for Testing and Materials: ASTM E436 – 03 Standard Test Method for Drop-Weight Tear Tests of Ferritic Steels. In: ASTM B. Stand. 91 (1997) 21. American Society for Testing and Materials: ASTM E208-17(2018) Standard Test Method for Conducting Drop-Weight Test to Determine Nil-Ductility Transition Temperature of Ferritic Steels. In: ASTM B. Stand. 06 (2000) 22. American Society for Testing and Materials: ASTM C 39 M:2003, Standard Test Method for Compressive Strength of Cylindrical Concrete Specimens. In: ASTM B. Stand. 03 (2003)

An Experimental Study on the Fatigue Failure Mechanisms of Pre–damaged Steel Fibre Reinforced Concrete at a Single Fibre Level Humaira Fataar1, Riaan Combrinck1, and William P. Boshoff2(&) 1 Unit for Construction Materials, Stellenbosch University, Stellenbosch, South Africa 2 University of Pretoria, Pretoria, South Africa [email protected]

Abstract. In fibre reinforced concrete (FRC), energy is dissipated in the wake of the crack tip through the actions of fibre bridging and fibre pull out. This is the main mechanism which inhibits crack growth, thus increasing the load carrying capacity of FRC by providing post–cracking ductility. Furthermore, the same mechanism is present when FRC undergoes fatigue loading. Typical applications for FRC which undergo significant fatigue loading during their service life include paving applications such as bridge decks, highways and industrial floors. The continuous exposure to cyclic loading results in a decrease in apparent stiffness of the material, which may lead to fatigue failure [1, 2]. Fatigue failures are almost always unexpected, and can have a catastrophic outcome [3, 4]. Thus, the fatigue characteristics become vital performance and design parameters [1]. In this paper, the mechanisms of fatigue failure of predamaged hooked-end steel fibre reinforced concrete (SFRC) are investigated at a single fibre level. An initial pre-damage was applied to the fibres before the cyclic loading commenced. The pre–pull out ranged from 0.6 mm to 2.5 mm, and the cyclic loading was applied at 70% and 85% of the maximum static pull out capacity of the fibre embedded in the concrete. Keywords: FRC


Fatigue
Single fibre
Steel fibres
Cyclic loading

1 Introduction As concrete technology progresses, concrete’s compressive strength is constantly being increased. However, an increase in compressive strength results in a more brittle concrete which is prone to sudden failures [5]. Since concrete is known for its weak tensile strength, steel is added in the form of reinforcing bars or fibres to contribute to the tensile capacity of the composite [6]. The use of discontinuous fibres in cementitious materials have been known to suppress crack development and slow down crack propagation, improve the composite durability as well as its fatigue life [7–13]. The fibres’ bridging activity is only activated at crack initiation, and only then does it contribute to the load carrying capacity. The load gets transferred from the concrete matrix to the fibre by shear deformation at the fibre-matrix interface [7, 14, 15]. © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 199–208, 2021. https://doi.org/10.1007/978-3-030-58482-5_18

200

H. Fataar et al.

Fatigue is defined as the process of continuous, permanent internal structural changes to a material which is subjected to cyclic loading. In concrete, the internal structural changes are associated with the progressive growth of internal micro cracks [1, 16, 17]. A continuous exposure to cyclic loading results in decreased apparent stiffness, which may lead to fatigue failure [1, 2]. The fatigue loading can be divided into two categories, low-cycle loading and high-cycle loading. The low-cycle loading consists of fewer load cycles which are applied at high stress levels. In contrast, the high-cycle fatigue loading consists of a large number of load cycles at low stress levels [1]. Generally, there are three main failure mechanisms for fibre reinforced concrete (FRC). These include fibre pullout, fibre failure, and matrix failure [14, 18]. Fibre pullout is the most common type of failure mechanism. It occurs when the bond mechanisms, which includes adhesion, friction, and mechanical interlock, have failed [19]. Fibre failure occurs when the fibre ruptures, and matrix failure occurs when the failure strain of the matrix has been reached [14]. The purpose of this research is to use single fibre pullout tests in order to understand the failure mechanisms involved with pre-damaged fatigue SFRC.

2 Experimental Method 2.1

Material Properties and Concrete Mix

The materials used for the concrete mix, excluding the steel fibres, were locally sourced. The concrete mix design is displayed in Table 1. The cement used was a CEM II 52.5 N, supplied by Pretoria Portland Cement. The coarse aggregate used in the concrete was a 13 mm crushed Greywacke stone, and a natural pit sand (locally called Malmesbury sand) was used as a fine aggregate. Municipal water was used, along with Chryso Optima 206, which was a high range water reducing superplasticiser. The steel fibres were supplied by BEKAERT and their DRAMIX 3D–65/60-BG hooked-end fibres were used in this study. Table 2 provides the properties of the steel fibres. Table 1. Concrete mix composition Material Cement Sand Stone Water Superplasticiser (1.4% by weight of cement)

kg/m3 450 880 820 194 6.3

An Experimental Study on the Fatigue Failure Mechanisms

201

Table 2. Fibre properties (DRAMIX 3D-65/60-BG) [20] Tensile strength Modulus of elasticity Length (l) Diameter (d)

2.2

1 160 MPa 200 GPa 60 mm 0.9 mm

Specimen Preparation

The specimen preparation was done in two states – fresh state and the hardened state. During the fresh state, the concrete constituents were mixed together in a pan mixer. The concrete mix was designed to be used for steel fibre reinforced beams as well as the single fibre pull out tests. The concrete was then sieved on a shaker table through a 4.75 mm sieve to collect the mortar alone. The test specimens were created by dividing 100 mm cube moulds into quarters. The moulds were divided using a wooden cross-shaped partition placed in the centre of the mould as displayed in Fig. 1. This resulted in four specimens, each measuring 40  40  100 mm3. The fresh mortar was cast into the moulds and was vibrated for 30 s on a shaker table. No fibres were added to the concrete mix, only a single fibre was embedded in each specimen after the mortar was cast. The single fibre was placed in the centre of the specimen to a fixed embedment length of l/2 (30 mm), and was vibrated once more to close any voids that could have developed when inserting the fibre. All the specimens were then placed in an undisturbed part of laboratory and was demoulded after 24 h. Once demoulded the fibre on each specimen was coated with a wax to protect it from corrosion while in the curing tank. All the specimens were placed in a curing tank to cure for an additional 26 days, completely submerged in water at 24 ± 1 °C.

Fig. 1. Specimen mould

One day prior to testing, the specimens were removed from the curing tanks and allowed to dry, after which the wax was removed from the fibre with a clean cloth. The

202

H. Fataar et al.

exposed hooked-end of the fibre was carefully removed using pliers to allow the fibre to be clamped onto the test setup. Once the specimens were dry, it was glued onto steel plates with a two-part epoxy provided by Sika called Sikadur–30. The epoxy took 24 h to set and the specimens were tested between 28 and 40 days. 2.3

Test Setup

The tests were conducted on a 50 kN servo–controlled hydraulic actuator, which was fixed onto a steel frame. An external 5 kN load cell was added to the test setup to ensure a better resolution on the load control of the test, since the maximum pullout loads were expected to be below 5 kN based on similar pullout tests conducted at Stellenbosch University [21]. Additionally, an external linear variable differential transducer (LVDT) was incorporated in the test setup to verify the crosshead displacement readings provided by the actuator. Once the specimens were ready to be tested, the steel plate with a single specimen on it was bolted down onto a box section which was fixed to two channel sections of the steel frame. The crosshead of the actuator was then lowered into position until the exposed fibre was completely inserted into the clamp. The fibre clamp was then tightened by two grub screws which held the fibre in place. The test setup can be seen in Fig. 2.

Fig. 2. Test setup

2.4

Loading Regime

Before the fatigue cyclic loading commenced, static pullout tests were conducted. This was done in order to obtain the average maximum pullout load. The static tests were

An Experimental Study on the Fatigue Failure Mechanisms

203

position-controlled and it measured the load required to completely pull a fibre embedded at l/2 out of the mortar. Once the static tests were completed, the cyclic fatigue pullout tests could commence. The fatigue loading was applied by a sinus load application as displayed in Fig. 3. The maximum load (Amax) was taken as a percentage of the average maximum static pullout load (70% or 85%). The minimum load (Amin) was specified as a small positive value (generally 20 N) in order to maintain cyclic tensile loads. The fatigue tests were loaded up to 2 million cycles at a stable frequency of 6 Hz.

Load (N) Amax

Mean Amin

0

Time (s)

Fig. 3. Sinus loading

The cyclic tests were performed in three steps: (1) pre-damage; (2) load/unload; and (3) cyclic load. The pre-damage step consisted of an initial pullout of the fibre to a certain predetermined degree. This step was position-controlled and the pullout was set by the user. The second step was the load/unload step, which required the pullout load attained in the first step to either increase or decrease in order to reach the specified mean value of the sinus load. The third step was the fatigue loading, which cycled from the mean value to the maximum and minimum amplitude until the fibre completely pulled out, or reached the 2 million cycle limit set. Both the second and third steps were load-controlled.

3 Results and Discussion 3.1

Static Tests

Before any fatigue tests could be performed, it was necessary to conduct static pullout tests. This was done in order to find the average maximum load required to completely pull out the hooked-end fibre. A total of six specimens were tested, and the results are plotted in Fig. 4. The test results followed the same trend for all the specimens. The load increased rapidly before it reached a small peak at a displacement of around 0.1 mm, followed by a relaxation in the load. The load then increased further until it

204

H. Fataar et al.

reached a global peak at a displacement of 0.9 mm, then it started to decrease slightly. It was then followed by another smaller increase in load to a less notable peak. The load then begins to decrease steadily until the fibre has completely pulled out. The first peak was possibly due to the initial fibre debonding, but further investigation is required. The global peak was then associated with further fibre debonding and the initiation of the hook straightening. The last peak was due to the fibre straightening out completely and eventually pulling out. Similar trends can be found in literature [15, 22, 23].

Fig. 4. Static pullout tests

3.2

Fatigue Tests

The fatigue tests were conducted on specimens at load levels of 70% and 85%, at a predamage of 0.6 mm, 1.2 mm, 1.8 mm, and 2.5 mm. The tests were conducted until failure, or until the 2 million cycle limit was reached. Unless otherwise specified, three specimens were tested for each load level percentage and pre-damage. Figure 5 illustrates the average number of cycles sustained for each test group. A higher load level coupled with a higher pre-damage resulted in early failure, as can be seen with the 85% load level and 2.5 mm and 1.8 mm pre-damage. The 1.2 mm and 0.6 mm predamage specimens could still withstand the 85% load level to a certain extent. However, only a single specimen did not fail. On average, the 70% load level tests were able to withstand a large number of load cycles before any type of failure occurred. Only one test batch – 70% at 1.2 mm pre–damage – sustained the full 2 million cycles without any failure. The results in Table 3 indicates the failure mechanisms for each test. Two types of failure was experienced – fibre pullout and fibre rupture. The pullout failure mechanism was mostly experienced by specimens with the higher pre-damage, since the fibre has already debonded and the hook begins to deform [23, 24]. At the lower pre-damage, the likelihood of fibre rupture increases, due to the fact that the fibre hook is still intact. It can also be noted that the 85% load level resulted in more pullout failures. The 70%

An Experimental Study on the Fatigue Failure Mechanisms

205

Load level (%)

70 0.6 mm 1.2 mm 1.8 mm 2.5 mm

85

1

100 10 000 Average cycles

1 000 000

Fig. 5. Average cycles endured by each test group

load level had an equal number of fibre rupture and fibre pullout failures, but having no failure dominate the test results. Table 3. Failure mechanism summary Load level Pre-damage (mm) Failure mechanisms 70% 2.5 2 pullout; 1 no failure 1.8 1 rupture; 2 no failure 1.2 4 no failure 0.6 1 rupture; 2 no failure 85% 2.5 3 pullout 1.8 2 pullout; 1 rupture 1.2 1 pullout; 2 rupture 0.6 1 pullout; 1 rupture; 1 no failure

3.3

Computed Tomography (CT) Scans

The CT scans provided a non-destructive method of assessing the internal structure of the specimens both before and after the fatigue testing. Initially, CT scans were taken of specimens with a pre–damage and no fatigue testing. The specimens in Fig. 6 were damaged to 0.6 mm to determine the actual measured degree of fibre displacement and compare it with the applied pre-damage. It was found that the actual measured displacement of the tip of the fibre was slightly less that the applied pre-damage. Additionally, there was the same trend in delamination observed at the curved portion of the fibre for specimens A, B and C. This was caused by the fibre debonding from the surrounding matrix as the pre-damage load was applied.

206

H. Fataar et al.

Fig. 6. CT scans for three specimens at 0.6 mm pre-damage

CT scans were also performed on two failed specimens after the fatigue loading as displayed in Fig. 7. Specimen D and Specimen E were pre-damaged to 1.2 mm before the fatigue loading of 85% load level was applied. Specimen E was found to have ruptured at the start of the hooked-end which was embedded in the matrix, whereas Specimen E ruptured at the surface of the specimen. Upon further investigation at the surface of Specimen E displayed in Fig. 8, it was found that there was a network of micro cracks. This indicated that the stress concentration at that point was high, thus resulting in crack propagation and ultimately, fibre rupture at the surface.

Fig. 7. Post-test CT scan of ruptured fibres

An Experimental Study on the Fatigue Failure Mechanisms

207

Fig. 8. Side view of Specimen E

4 Conclusions Fatigue tests at a single fibre level were conducted on pre-damaged specimens. Using hooked-end steel fibres embedded at l/2 into a mortar, the failure mechanisms were found for specimens tested at load levels of 70% and 85% of the average maximum static load. The pre-damage varied for each set of tests between 0.6 mm to 2.5 mm. The following conclusions can be made: • The main failure mechanisms were found to be fibre pullout and fibre failure in the form of fibre rupture. • At the largest pre-damage of 2.5 mm, the dominating failure mechanism was fibre pullout, due to the fact that the fibre hook anchorage has already been deformed and there is not much resistance left. • The lower pre-damage experienced a mixture of fibre pullout as well as fibre rupture occurring. • Overall, the 85% load level experienced the most failures, with only one of twelve specimens not failing. In contrast, the 70% load level experienced four specimen failures out of thirteen specimens.

References 1. Lee, M.K., Barr, B.I.G.: An overview of the fatigue behaviour of plain and fibre reinforced concrete. Cement Concr. Compos. 26(4), 299–305 (2004) 2. Kolluru, S.V., O’Neil, E.F., Popovics, J.S., Shah, S.P.: Crack propagation in flexural fatigue of concrete. J. Eng. Mech. 126(9), 891–898 (2000) 3. Bathias, C., Pineau, A.: Fatigue of Materials and Structures. ISTE , Wiley, London, Hoboken (2010) 4. Anderson, T.L.: Fracture Mechanics: Fundamentals and Applications. CRC Press, Boca Raton (2017)

208

H. Fataar et al.

5. Boulekbache, B., Hamrat, M., Chemrouk, M., Amziane, S.: Flexural behaviour of steel fibrereinforced concrete under cyclic loading. Constr. Build. Mater. 126, 253–262 (2016) 6. Hsu, T.T.C.: Fatigue of plain concrete. Am. Concr. Inst. 78(4), 292–305 (1981) 7. Namur, G.G., Alwan, J.M., Najm, H.S.: Fiber pullout and bond slip I: analytical study. J. Struct. Eng. 117(9), 2769–2790 (1992) 8. Banthia, N., Trottier, J.F.: Deformed steel fiber—cementitious matrix bond under impact. Cem. Concr. Res. 21(1), 158–168 (1991) 9. di Prisco, M., Plizzari, G., Vandewalle, L.: Fibre reinforced concrete: new design perspectives. Mater. Struct. 42(9), 1261–1281 (2009) 10. Buratti, N., Mazzotti, C., Savoia, M.: Post-cracking behaviour of steel and macro-synthetic fibre-reinforced concretes. Constr. Build. Mater. 25(5), 2713–2722 (2011) 11. American Concrete Institute: Report on fiber reinforced concrete (2002) 12. Zhang, J., Stang, H., Li, V.C.: Crack bridging model for fibre reinforced concrete under fatigue tension. Int. J. Fatigue 23(8), 655–670 (2001) 13. Yoo, D.Y., Kim, S., Park, G.J., Park, J.J., Kim, S.W.: Effects of fiber shape, aspect ratio, and volume fraction on flexural behavior of ultra-high-performance fiber-reinforced cement composites. Compos. Struct. 174, 375–388 (2017) 14. Beaudoin, J.J.: Handbook of Fiber-Reinforced Concrete : Principles Properties, Developments and Applications. Noyes Publications, Park Ridge, N.J. (1990) 15. Ghoddousi, P., Ahmadi, R., Sharifi, M.: Fiber pullout model for aligned hooked-end steel fiber. Can. J. Civ. Eng. 37(9), 1179–1188 (2010) 16. Parvez, A., Foster, S.J.: Fatigue of steel-fibre-reinforced concrete prestressed railway sleepers. Eng. Struct. 141, 241–250 (2017) 17. Keerthana, K., Chandra Kishen, J.M.: An experimental and analytical study on fatigue damage in concrete under variable amplitude loading. Int. J. Fatigue 111, 278–288 (2018) 18. Naaman, A.E.: Fiber reinforced concrete under dynamic loading. In: Fiber Reinforced Concrete International Symposium, vol. 81, pp. 169–186 (1984) 19. Naaman, A.E., Najm, H.: Bond-slip mechanisms of steel fibers in concrete. ACI Mater. J. 88 (2), 135–145 (1991) 20. Dramix® steel fiber concrete reinforcement - Bekaert.com. https://www.bekaert.com/en/ products/construction/concrete-reinforcement/dramix-steel-fiber-concrete-reinforcement. Accessed 21 Feb 2018 21. Nieuwoudt, P.D.: Time-dependent Behaviour of Cracked Steel Fibre Reinforced Concrete, March 2016 22. Nieuwoudt, P.D., Boshoff, W.P.: Time-dependent pull-out behaviour of hooked-end steel fibres in concrete. Cement Concr. Compos. 79, 133–147 (2017) 23. Marković, I.: High-performance Hybrid-fibre Concrete : Development and Utilisation. Delft University, the Netherlands (2006) 24. Cunha, V.M.C.F., Barros, J.A.O., Sena-Cruz, J.: Pullout behaviour of hooked-end steel fibres in self-compacting concrete (2007)

Development of an HPFRC for Use in Flat Slabs Julia Blazy, Sandra Nunes(&), Carlos Sousa, and Mário Pimentel CONSTRUCT-LABEST, Department of Civil Engineering, Faculty of Engineering, University of Porto, Porto, Portugal [email protected]

Abstract. Fibre-reinforced cementitious materials represent one of the most significant developments in the field of concrete technology of the last decades. The improved performance of this new class of materials (in terms of workability, compressive strength, flexural/tensile behaviour and/or durability) allows rethinking several of the existing structural solutions. This paper describes research on high-performance fibre reinforced concrete (HPFRC) to be used at the slab-column connection zones of flat slabs, in order to improve its punching shear resistance. Design of Experiments (DoE) approach was used to design HPFRC paste and aggregate particle phases. As such, a central composite design was carried out to mathematically model the influence of mixture parameters and their coupled effects on deformability, viscosity and compressive strength. After that, a numerical optimization technique was applied to the derived models to select the best mixture, which simultaneously, maximizes aggregates content and allows achieving a compressive strength of 90–120 MPa, while maintaining self-compactability (SF1 + VS2), incorporating 1% steel fibres content. Keywords: High-Performance Fibre Reinforced Concrete (HPFRC) design
Self-compacting
Flexural behaviour
Steel fibres


Mix-

1 Introduction During the last three decades, concrete has emerged from a rather simple mass construction material towards a high-performance material, which can be tailored for specific applications and according to user specifications. The rapid evolvement in terms of enhanced workability (self-compacting ability) and compressive strength was made possible by the development of superplasticizers and mix-design methods. At increasing strength, concrete becomes more brittle, which can be compensated for by the incorporation of fibres. The fibres can increase the ductility in compression as the fibres transfer stress during cracking. Figure 1 maps the most common fibre-reinforced cementitious composites with regard to compressive strength and fibre content, namely, engineered cementitious composites (ECC), ultra-high fibre performance fibre reinforced composites (UHPFRC), high-performance fibre reinforced concrete (HPFRC) and traditional fibre reinforced concrete (FRC). The potential for an increase in compressive strength due to an increase in fibre content is less significant compared to the tensile strength [1, 12]. Fibres can improve © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 209–220, 2021. https://doi.org/10.1007/978-3-030-58482-5_19

210

J. Blazy et al.

Fig. 1. Most common range of different fibre-reinforced cementitious composites, concerning compressive strength and fibre content

the tensile and flexural behaviour of FRC; its effect depends on the content, fibre type, fibre orientation and distribution, and matrix strength [1]. The use of steel fibre hybridisation has also been found to enhance the pre- and post- cracking response of FRC [2]. The performance of FRC in tension usually improves at increasing fibre content; but beyond a threshold value, the performance might even decrease due to the loss of workability and entangling of interacting fibres during pull-out (lower efficiency of the fibres). The effect of coarser aggregates in this context is also crucial concerning the performance in both the fresh and the hardened states. The addition of fibres in concrete affects its characteristics in the fresh state due to the larger surface area of fibres, which requires more paste to envelop and lubricate the fibres [12]. In particular, stiff fibres decrease the packing density of the aggregates; this effect is more pronounced the larger the aggregates are, compared to the fibre length [12]. In most cases, the FRC mix-design is a trade-off between workability requirements, desired postcracking behaviour and economical aspects. Since fibres increase the material costs, optimum fibre content has to be determined. The current study deals with the materials selection, formulation and optimization of an HPFRC mixture to be used in a localised area at the slab-column connection zones of flat slabs (the rest of the slab being cast with normal-strength concrete), in order to improve its punching shear resistance. The strategy followed in this study was to optimize the two main parts of the composite separately. Firstly, optimise the paste phase and the aggregate particle phase to achieve a target compressive strength of 90– 120 MPa and self-compacting ability. A fixed fibre content and a single fibre type were used at this stage. Secondly, combining the optimized concrete with a hybrid fibres mixture (new hooked-ends macro steel fibres + straight micro steel fibres) and assess its influence on flexural behaviour, in order to find the best combination. The current paper reports only on the first part of the study. Some literature studies on HPFRC with compressive strengths close to the target range of 90–120 MPa are reported in Table 1. Regarding the performance in the fresh

Development of an HPFRC for Use in Flat Slabs

211

state, these include conventional vibrated concretes (slump: 17.5 and 13.0 cm [6]) as well as self-compacting concretes (slump flow: 71–78 cm [2]; 76 cm [7]; 65–75 cm [11]). Compressive strength and fibre content data from these studies are also plotted in Fig. 1, identifying the zone where most HPFRC mixtures can be found. It can be observed that in most HPFRCs the fibre content ranges from 0.5% to 1.5%. Table 1. Characteristics of several HPFRC materials from different studies. Ref.

Binder

w/b

[1]

AST M T ype I cement

0.25 (w/c)

CEM I 52.5R

0.27 (w/c)

[2]

dmax (mm) 10

10 a

Vf

lf/df; geometry

0.5%; 1%; 1.5%

40/0.62; 1HE

fcm,28d (MPa) 88; 90; 93

0.5%; 1%; 1.5%

50/0.62; 1HE

87; 91; 93

0.5%; 1%; 1.5% 0.75%+0%

60/0.75; 1HE 60/0.9; 1HE+13/0.20; S

87; 92; 94 88

0.562%+0.188%

84

0.375%+ 0.375%

83

0.188%+0.562%

88

0%+0.75% 1.0%+0%

60/0.9; 2HE+13/0.20; S

0.75%+0.25%

[3] [4]

PC+SF+FA Cement CPN50

[5]

PC 42.5 T ype I+ SF

0.25 0.30 (w/c) 0.24 (w/c)

98 89 89

0.5%+0.5%

87

0.25%+0.75%

91

13 10

0%+1.0% 1.25% 0.76%

30/0.50; 1HE -; -

12.5 b

0.2%

28/0.40; 1HE

98 94 95 102

0.4%

100

0.6%

102

[6]

OPC+SF

0.295

10 c

1.0% 1%

13/0.18; S

103 103

[7] [8]

CEM+mS+LP OPC+zSF+FA

0.24 0.195

10 b 10 d

2% 0.5% 0.50%

30/0.55; C 30/0.50; 1HE

114 100 93

[9]

PC T ype II+LP

12.7

1% 1.6%

36/0.70; 1HE

108 91

[10]

CEM II/B-M (SV) 42.5 N+ SF

0.40 (w/c) 0.20

11 e

1%+0.5%

13/0.20+30/0.60

109

0.18

8e

3% (total)

13/0.20+30/0.60+50/0.1

122

0.18 11 e 0.5%+0.5%+0.5% 13/0.20+30/0.60+50/0.1 95 CEM I 42.5R 0.32 8c 0.7%+0.3% 6/- + 30/130-140 HSR LA+SF (w/c) [12] CEM I 52.5R+ 0.20 8c 2% 13/0.20; S 164 mS+LP (w/p) (O)PC: (ordinary) Portland cement; (z)SF: (zirconium) silica fume; mS: micro silica; FA: fly ash; LP: limestone powder; w/b: water to binder weight ratio; w/c: water to cement weight ratio; w/p: water to powder weight ratio; dmax: maximum diameter of coarse aggregate; Vf: fibre content; lf fibre length; df:fibre diameter fcm,28d: mean compressive strength at 28 days; 1HE: simple hooked-end; 2HE: double hooked-end; S: straight; C: crimped ends; nature of coarse aggregate: ( a) flint, ( b ) limestone, ( c) basalt, ( d ) granite, ( e) diabase [11]

212

J. Blazy et al.

2 Experimental Programme 2.1

Design of Experiments Strategy

The experiments were planned according to a Central Composite Design (CCD), which consists of three distinct sets of experimental runs: factorial (Fi), central (Ci) and axial (CCi) [13]. This approach helps to avoid the trial and error method and as a result, reduces the time necessary to develop the mixture of desired properties by collecting and statistically analysing relevant data. Additionally, it allows identifying design variables that have the most significant influence on the response variables, as well as the interactions between them. This design of experiments (DoE) technique involves the following steps: (1) choosing the mix-design variables and response variables; (2) selecting design variables’ ranges; (3) conducting experiments and collecting data; (4) statistical analysis of data and fitting the empirical model, by means of regression analysis; (5) optimizing mixture proportions. In this work, the research program was first conducted at the mortar level and then on the concrete level to minimize at the beginning the number of design variables and therefore to reduce the number of experiments required to create the empirical model. After choosing the optimum paste composition, resulting from the study at the mortar level, the research on the concrete level was conducted. Regarding the mortar stage, four key design variables were chosen: water to powder volume ratio (Vw/Vp); water to cement weight ratio (w/c); superplasticizer to powder weight ratio (Sp/p) and silica fume to cement weight ratio (SF/c) and four response variables: slump flow diameter (Dflow), V-funnel flow time (Tfunnel), compressive strength after 28 (fc,28d) and 98 days (fc,98d). It must be noted that sand to mortar volume ratio (Vs/Vm) and fibre content (Vf) were set fixed and equal to 0.475 and 1.5%, respectively. For concrete level, two key design variables were selected: solid volume (Vg/Vg,lim), sand to mortar volume ratio (Vs/Vm) and four response variables: time necessary to reach a 500 mm diameter in the slump flow test (t500), slump flow diameter (Dflow), compressive strength after 7 (fc,7d) and 98 days (fc,98d). The fibre content (Vf) was fixed and equals 1.0%. The main reason to incorporate the steel fibres at this stage of the study was to avoid brittle failures in compression and reduce the variability of test results. More information about the formulation of the mixtures can be found elsewhere [13]. The effect of design variables was established on five levels: - a, -1, 0, +1, + a. Table 2 characterizes the experimental plans used for both the mortar and concrete studies. After collecting the experimental data, the commercial software Design-Expert was used to analyse the results for each response variable namely, to examine the summary data plots, to fit 2nd order polynomial models using regression analysis and ANOVA, to validate the models by examining the residuals for trends, autocorrelation and outliers, as well as to interpret the obtained model graphically.

Development of an HPFRC for Use in Flat Slabs

213

Concrete level

Mortar level

Table 2. Experimental plans characterization

2.2

M aterials

Type of plan

Design variables

[-α; +α]

CEM I 42.5R Limestone powder Silica fume Superplasticizer standard sand Fibres (9/0.175;S) Tap water CEM I 42.5R Limestone powder Silica fume Superplasticizer Sand Coarse aggregate Fibres (13/0.2;S) Tap water

24Fi+8CCi+6Ci α= 2.0 nc = 6 nf= 16 na = 8

Vw/Vp w/c Sp/p SF/c

[0.450; 0.650] [0.250; 0.350] [1.5%; 2.5%] [7.5%; 17.5%]

22Fi+4CCi+4Ci α= 1.414 nc = 4 nf = 4

Vg/Vg,lim Vs/Vm

[0.440; 0.510] [0.365; 0.435]

Response variables Dflow Tfunnel fc,28d fc,98d

t 500 Dflow fc,7d fc,98d

na = 4

Constituent Materials

In the study at mortar level, the following materials were used: Type I Portland cement (CEM I 42.5R); limestone powder Betocarb® HP-OU (LP); silica fume Elkem 940-U (SF); standard sand; superplasticizer Sika® ViscoCrete® 20HE (Sp); straight steel fibres (lf = 9 mm and df = 0.175 mm, fy = 2100 MPa) and tap water. The specific gravity of cement, LP and SF were 3.07, 2.68 and 2.40 g/cm3, respectively. The standard sand is siliceous natural sand with particles of diameter from 0.08 mm to 2 mm and round shape. Furthermore, specific gravity equals 2.63 g/cm3 and water absorption 0.3% by mass. Regarding superplasticizers, the supplier provided the following information: specific gravity of 1.08 g/cm3 and solid content of 40%. Concerning the study at the concrete level, the same type of cement, LP, SF, Sp and tap water were used. However, standard sand was replaced by real aggregates: natural sand with a specific gravity of 2.58 g/cm3 and water absorption of 0.7%; and a crushed coarse aggregate (dmax = 8 mm; amphibolite rock) whose specific gravity equals 3.02 g/cm3 and water absorption is 0.6%. All concrete mixtures incorporated 1% of steel fibres (lf = 13 mm and df = 0.2 mm, fy = 2500 MPa), by volume.

214

2.3

J. Blazy et al.

Mixing Sequence and Testing Methods

Mortar mixes were tested in 1.40L batches at the low speed. The mixing procedure is presented in Table 3. To characterize the fresh state properties of mortar mixes (viscosity and flowability) the V- funnel and slump flow tests were carried out, respectively, according to Okamura & Ouchi proposal [14]. Firstly, the V-funnel was filled and the time that mortar needed to pass through the funnel opening was measured. Secondly, the slump flow test was conducted. At the beginning, the cone was filled and lifted with a constant speed. Then, after the flow of the mixture stopped, two diameters in perpendicular directions were measured. The final result (Dflow) is an average of four diameters because the test was repeated two times. Finally, four prisms of dimensions 40  40  160 mm3 were cast to evaluate the compressive strength after 28 and 98 days. All the samples were covered with a plastic sheet to avoid drying and water evaporation. After 24 h, specimens were demoulded and placed underwater in a climatic chamber at 20 ± 2 °C, until the testing date. In the meantime, after 7 days, all the prism samples were cut into cubes of dimensions 40  40  40 mm3. As a result, 12 samples for each mixture were obtained: 6 for testing the compressive strength after 28 days and 6 after 98 days. Afterwards, the average of these measures was calculated: fc,28d and fc,98d.

Table 3. Mortar mixing sequence Task Add sand + cement + limestone powder + silica fume + 80% of total mixing water Add 75% of superplasticizer + 20% of total mixing water Add 25% of superplasticizer Add fibers * Stop to scrape off the material adhering to the mixing walls

Duration 2.5min+ * +2.5min 2.5min + * 1.5min 2.0min+ * +1.0min

Concrete mixes were tested in 25L batches in an open pan mixer. Steps of mixing procedure are presented in Table 4. The slump flow test was carried out according to EN 12350-8. When the cone was withdrawn upwards, the time from commencing the upward movement of the cone to when the concrete had flowed to a diameter of 500 mm was measured (t500). Then, after the flow of the mixture stopped, two diameters in perpendicular directions were measured and their average was calculated (Dflow). Finally, six cubes of dimensions 150  150  150 mm3 were cast to evaluate the compressive strength after 7 and 98 days. All the samples were covered with a plastic sheet to avoid drying and water evaporation. After 24 h, the specimens were demoulded and placed underwater and kept at 20 ± 2 °C until the testing age. Three samples were used to obtain compressive strength after 7 days and three after 98 days. Afterwards, the average of these measures was calculated: fc,7d and fc,98d.

Development of an HPFRC for Use in Flat Slabs

215

Table 4. Concrete mixing sequence Task Add sand + 25% of total mixing water + coarse aggregate Waiting for absorption Add cement + limestone powder + silica fume + 75% of total mixing water + superplasticizer Add fibers * Stop to scrape off the material adhering to the mixing walls

Duration 2.5min 2.5min 5.0min + * 3.0min

3 Experimental Campaign at Mortar Level 3.1

Collected Experimental Data

In this section, the results from tests carried out on mortar level are presented and discussed. Figure 2.a shows the range of results obtained in the fresh state. In general, the mortars exhibited very good flowability (Dflow  293 mm) and relatively high flow times, which indicates a low risk of segregation. In Fig. 2.b it can be seen that almost all mixtures experienced an increase of compressive strength from 28 to 98 days, ranging from 1.0% to 8.1%. For one mixture, the compressive strength decreased by about 1.1%. This can be a result of local defects of the tested samples.

Fig. 2. Mortar test results: a) Tfunnel .vs. Dflow b) fc,98d.vs. fc,28d

3.2

Statistical Analysis of Data, Fitted Models and Optimisation

Table 5 presents the results of the fitted models, including the residual error term (e), along with the correlation coefficients. An analysis of variance showed that these models are significant when describing the effect of Vw/Vp, w/c, Sp/p and SF/c on the

216

J. Blazy et al.

modelled responses. Since R2 and R2adj values are significantly high, a large proportion of the variability of response variables is explained by the obtained regression models. The estimates of the model coefficients presented in Table 5 indicate the relative significance of the various design variables on each response. Higher values indicate greater influence of the variable in the response and, on the other hand, a negative value reflects a response decrease to an increase in the design variable. Results in Table 5 clearly show Vw/Vp and w/c are the most influencing variables on the fresh state properties, while the compressive strength is influenced mainly by w/c and SF/c. Based on the regression models presented in Table 5, the numerical optimization technique was used to determine the range of mortar mixture parameters that would lead to adequate self-compacting concrete (SCC) mixtures, with Tfunnel in between 12 and 14 s, a target compressive strength of 150 MPa and minimum cost of paste. No constraint was added concerning Dflow because all mixtures exhibited very good deformability. The selected optimised solution is presented in Table 6, in terms of the actual values. The optimised mortar mixture was also prepared and tested in the lab. The experimental results obtained compare well with the predicted values, which confirms the accuracy of the obtained fitted models. Table 5. Fitted numerical models at mortar level (coded values of parameters) Dflow Tfunnel fc,28d fc,98d (mm) (s) (M Pa) (M Pa) Independent 322.931 9.993 147.534 155.585 Vw/Vp +11.439 -3.340 -1.172 -1.594 w/c +7.915 -1.278 -5.486 -4.994 Sp/p NS -0.367 NS +0.558 SF/c -3.019 +0.115 +3.728 +3.715 Vw/Vp×w/c NS NS NS +1.144 w/c×Sp/p NS +0.236 NS -1.246 w/c×SF/c NS NS NS +1.225 Sp/p ×SF/c NS -0.251 NS NS (Vw/Vp) 2 -1.903 +0.764 NS NS (w/c)2 NS +0.367 NS NS (Sp/p)2 NS +0.264 NS -1.582 4.193 0.371 1.898 1.795 ε, std. dev.* R2 / R2adj 0.910 / 0.895 0.988 / 0.983 0.912 / 0.902 0.924/0.895 (NS) non-significant terms; (*) error term is a random and normally distributed variable with mean zero

Development of an HPFRC for Use in Flat Slabs

217

Table 6. Optimised mortar mixture and corresponding predicted and measured test results Design variables (actual values) Vw/Vp=0.536 w/c=0.273 Sp/p=1.59% SF/c=0.100

M easured Predicted 95% Prediction Interval- Low 95% Prediction Interval-High Predicted/M easured

Dflow (mm) 315.5 314.0 304.0 324.0 0.99

Tfunnel (s) 15.2 14.0 12.8 15.2 0.92

fc,28d (M Pa) 156.3 150.0 145.6 154.4 0.96

4 Experimental Campaign at Concrete Level 4.1

Collected Experimental Data

At the concrete level, concrete mixtures composition was obtained by considering the paste mixture proportions defined in Table 6 and replacing reference sand by real aggregates (fine and coarse aggregates described above). Tests on concrete are then necessary to optimize the aggregate skeleton and aggregates content. The DoE technique can also be applied at this stage to optimize concrete mixtures, considering as independent variables, only these variables related to the aggregates- Vg/Vg,lim and Vs/Vm- as detailed in Sect. 2.2. The results from the tests carried at the concrete level are presented in Fig. 3. These results show that the designed experimental plan covers a wide range of Dflow (Fig. 3. a), including mixes that can be classified from SF1 to SF3 consistency classes. For all mixtures t500 is higher than 2 s, corresponding to VS2 class, according to EN 206-9. The relation between compressive strength after 7 days and slump flow diameter is plotted in Fig. 3.b. The range of fc,7d results obtained is very narrow (about 8 MPa), meaning that the changes introduced in the aggregate skeleton do not affect significantly the compressive strength. In Fig. 4 the slump flow of two extreme mixtures is presented. In the case of F4 mixture (the least fluid) slight aggregate segregation in the middle could be observed. On the other hand, with F1 mixture (the most fluid) separation of paste from aggregates took place at the edges. Also, the reduced workability of F4 led to a reduction in compressive strength (see Fig. 3.b), probably due to increased air content in the specimens. 4.2

Statistical Analysis of Data, Fitted Models and Optimisation

Table 7 presents the results of the fitted models for Dflow and t500 responses, including the residual error term (e), along with the correlation coefficients. An analysis of variance showed that these models are significant when describing the effect of Vg/Vg, lim and Vs/Vm on the modelled responses. Since R2 and R2adj values are significantly high, a large proportion of the variability of response variables is explained by the obtained regression models. As could be expected, both design variables - Vg/Vg,lim and Vs/Vm- influence significantly the analysed fresh state properties of SCC; Vs/Vm being the most influencing parameter. An increase of Vs/Vm or Vg/Vg,lim decreases the flowability and decreases the viscosity of the mixture (lower t500).

218

J. Blazy et al.

Fig. 3. Concrete test results: a) t500 .vs. Dflow; b) fc,7d .vs. Dflow

Fig. 4. Spread flow area of tested concrete mixtures: a) F4; b) F1

Again, the fitted numerical models and numerical optimization technique were used to determine the range of design variables that would lead to an SCC belonging to SF1 + VS2 consistency classes, while minimizing the volume of paste (or maximizing the total aggregates content). The best solution found corresponds to: Vg/Vg,lim = 0.453 and Vs/Vm = 0.430. The estimated response values of this mixture are: Dflow = 600 mm; t500 = 29.8 s; fc,7d  104 MPa (the average compressive strength of tested mixtures). In the next stage of this study, the benefits of hybridization (considering fibres with lengths of 13, 35 and 60 mm) on flexural strength will be assessed, in order to achieve the highest performance while keeping a relatively low fibre content.

Development of an HPFRC for Use in Flat Slabs

219

Table 7. Fitted numerical models at concrete level (coded values of parameters)

Independent Vg/Vg,lim Vs/Vm ε, std. dev.*

Dflow (mm) 636.181 -57.454 -71.329 17.650

(1/t 5000.5) (s) 0.2315 -0.0634 -0.0858 0.026

R2 / R2adj 0.951 / 0.941 0.927 / 0.911 (*) error term is a random and normally distributed variable with mean zero

5 Conclusions Most relevant conclusions of the current study are the following: • The paste and aggregate particle phases and fibres mixture of HPFRC can be adjusted separately using DoE approach, allowing for mix-design optimisation. • Fitted numerical models revealed that the workability and compressive strength of HPFRC mortars are mainly determined by the following pairs of design variables: (Vw/Vp, w/c) and (w/c, SF/C), respectively • Total aggregates content was maximized to achieve an HPFRC exhibiting selfcompacting ability (SF1 + VS2) while incorporating 1% straight steel fibres (lf = 13 mm; df = 0.2 mm). Acknowledgements. This work was financially supported by: UID/ECI/04708/2019- CONSTRUCT - Instituto de I&D em Estruturas e Construções and the project PTDC/ECIEST/30511/2017 funded by national funds through the FCT/MCTES (PIDDAC). Collaboration and materials supply by EUROMODAL, Secil, Omya Comital, Sika, Krampharex and Dramix and is gratefully acknowledged.

References 1. Abbass, W., Khan, M.I., Mourad, S.: Evaluation of mechanical properties of steel fiber reinforced concrete with different strengths of concrete. Constr. Build. Mater. 168, 556–569 (2018) 2. Okeh, C.A.O., Begg, D.W., Barnett, S.J., Nanos, N.: Behaviour of hybrid steel fibre reinforced self compacting concrete using innovative hooked-end steel fibres under tensile stress. Constr. Build. Mater. 202, 753–761 (2019) 3. Jang, S.J., Jeong, G.Y., Yun, H.D.: Use of steel fibers as transverse reinforcement in diagonally reinforced coupling beams with normal- and high-strength concrete. Constr. Build. Mater. 187, 1020–1030 (2018) 4. Ruano, G., Isla, F., Pedraza, R.I., Sfer, D., Luccioni, B.: Shear retrofitting of reinforced concrete beams with steel fiber reinforced concrete. Constr. Build. Mater. 54, 646–658 (2014)

220

J. Blazy et al.

5. Mousavi, S.M., Ranjbar, M.M., Madandoust, R.: Combined effects of steel fibers and water to cementitious materials ratio on the fracture behavior and brittleness of high strength concrete. Eng. Fract. Mech. 216(June), 106517 (2019) 6. Gholampour, A., Ozbakkaloglu, T.: Fiber-reinforced concrete containing ultra high-strength micro steel fibers under active confinement. Constr. Build. Mater. 187, 299–306 (2018) 7. Deeb, R., Karihaloo, B.L., Kulasegaram, S.: Reorientation of short steel fibres during the flow of self-compacting concrete mix and determination of the fibre orientation factor. Cem. Concr. Res. 56, 112–120 (2014) 8. Yoo, D.Y., Shin, H.O.: Bond performance of steel rebar embedded in 80–180 MPa ultrahigh-strength concrete. Cem. Concr. Compos. 93(February), 206–217 (2018) 9. Kazemi, M.T., Golsorkhtabar, H., Beygi, M.H.A., Gholamitabar, M.: Fracture properties of steel fiber reinforced high strength concrete using work of fracture and size effect methods. Constr. Build. Mater. 142, 482–489 (2017) 10. Skazlic, M., Serdar, M., Bjegovic, D.: Influence of test specimens geometry on compressive performance concrete. In: Second International Symposium on Ultra High Performance Concrete (2008) 11. Piérard, J., Dooms, B., Cauberg, N.: Durability evaluation of different types of UHPC. In: RILEM-fib-AFGC Int. Symp. Ultra-High Perform. Fibre-Reinforced Concr. UHPFRC 2013 –, no. 1, pp. 275–284 (2013) 12. Li, P.P., Yu, Q.L., Brouwers, H.J.H.: Effect of coarse basalt aggregates on the properties of Ultra-high Performance Concrete (UHPC). Constr. Build. Mater. 170, 649–659 (2018) 13. Nunes, S.: Performance-based design of self-compacting concrete (SCC): a contribution to enchance SCC mixtures robustness, p. 357 (2008) 14. Okamura, H., Ouchi, M.: Self-compacting concrete. Adv. Concr. Technol. 1(1), 5–15 (2003)

Influence of the Steel Fibres on the Tension and Shear Resistance of Anchoring with Anchor Channels and Channel Bolts Cast in Concrete Mazen Ayoubi1(&), Christoph Mahrenholtz2, and Wilhelm Nell3 1

Frankfurt University of Applied Sciences (FRA-UAS), Frankfurt, Germany [email protected] 2 Jordahl GmbH, Berlin, Germany 3 KrampeHarex GmbH & Co. KG, Hamm, Germany

Abstract. The current design method for anchor channels with channel bolts is based on test results for fasteners installed in conventional concrete. The field of application for steel fibre concrete have been growth over the last years and recently steel fibre reinforced concrete became popular e.g. for the production of prefabricated tunnel elements. The existing design rules for fasteners including anchor channels with channel bolts do not cover steel fibre reinforced concrete. To study the load-displacement behaviour in tension and shear, exploratory tests have been carried out on anchor channel-channel bolt-systems cast in plain and steel fibre reinforced concrete. The test results demonstrate a superior performance of channel bolts installed in anchor channels which were cast-in steel fibre reinforced concrete if compared with systems cast in plain reinforced concrete. The results of the experimental investigations will be explained und discussed in this article. Keywords: Anchor channel
Channel bolt
Anchorage
Steel fibre reinforced concrete (SFRC)
Concrete
Capacity
Ductility

1 Introduction The addition of steel fibres to the concrete mix increases the tensile strain capacity and ductility as well as improves the structural properties of concrete in its hardened state such as the tensile strength, impact strength, toughness, fatigue strength and the ability to resist cracking and spalling. Due to better mechanical and physical properties became the usage of steel fibre reinforced concrete (SFRC) over the last decades a better alternative to the conventional reinforcing concrete and being widely used as a construction material. Due to addition of steel fibres, the overbridging of cracks (Fig. 1) is ensured and so the cracking resistance of concrete increases [1]. This beneficial effect leads to an increase in the concrete resistance under quasi-static, cyclic and impact loading and furthermore, to the increase of the spalling resistance [1–3].

© RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 221–232, 2021. https://doi.org/10.1007/978-3-030-58482-5_20

222

M. Ayoubi et al.

Fig. 1. Crack-bridging mechanism, a) in normal concrete and b) in SFRC from [2, 3].

Various factors and mechanisms influence the tensile behaviour of SFRC, including the pullout behaviour of individual fibres, the random distribution of fibres, and the effects of finite member dimensions [4]. Especially in members with end-hooked fibres, the tensile behaviour of mechanical anchorage is essential in addition to the frictional bond behaviour between fibres and concrete matrix [4]. The use of medium-high steel fibres contents significantly improves the post-peak behaviour in tensile for flexure, by extending the softening branch and reducing the negative slope. Steel fibres give to the concrete a sizable post-peak residual strength. This lead to an improved fracture energy and toughness of SFRC materials with high fibre content compared to ordinary concrete [4]. Anchoring in Concrete Using Anchor Channels with Channel Bolts Anchor channels combined with channel bolts allow the reliable connection of steel components to the reinforced concrete structure. To this end, T-shaped channel bolts are locked into C-shaped anchor channels (Fig. 2a) that have been cast into the reinforced concrete. Conventional anchor channels allow the transfer of tension loads (N) and shear loads perpendicular to the channel ðV? Þ. Serrated anchor channels and matching serrated channel bolts have recently been developed to also enable load transfer in the direction of the channel ðVk Þ, thus making load transfer in all directions possible (Fig. 2b). Simulated seismic load tests showed that the load bearing behaviour of the serrated connection is very robust. This is because adjacent teeth are activated when the teeth in the contact area between the head of the channel bolt and the lips of the anchor channel start to fail. This allows the qualification of the serrated channels according to the highest seismic requirements [5, 6].

Influence of the Steel Fibres on the Tension and Shear Resistance

223

Anchor Anchor channel

V| Channel Serration

V|| Channel bolt a)

b)

N

Fig. 2. a) Components of anchor channel and channel bolt; b) Serrated systems allow load transfer in all directions.

Fig. 3. Installation sequence a) attaching of anchor channel to formwork, b) pouring of concrete, c) removing of filler, d) twisting-in of channel bolt.

For installation, the anchor channel is hot glued or nailed to the formwork (Fig. 3a). Anchor channels are generally furnished with filler material to prevent concrete slurry leaking into the profile during concreting. After the concrete is set and the formwork is stripped off, the filler is removed (Fig. 3b and c). Channel bolts are then inserted and twisted in the slot of the anchor channel to allow fastening of components at any point along its length (Fig. 3d). Connections using anchor channels with channel bolts have several advantages, making anchor channels with channel bolts suitable for the connection of any kind of component in concrete, e.g. elevator guide rails, curtain wall brackets, and particularly technical equipment in tunnels [7–9]: – Quick and easy installation of the anchor channel during concreting – Compensation of building tolerances by adjusting the position of channel bolt along the length of the anchor channel – Robust load transfer due to mechanical interlock (bolt-channel, anchor-concrete)

224

M. Ayoubi et al.

– Unlike for anchor plates no on-site welding is required, thus no weld quality issues, or fire risks – Unlike for post-installed fasteners no on-site drilling required thus no hassle with positioning or cutting of reinforcing bars, and no exposition of the workforce to harmful silica dust – Later positional adjustment, or replacement of attached components is made easy at any time The design of anchor channels with channel bolts was just recently codified as the European standard EN 1992–4 (2018) [10] which requires a qualification of the system according to the European assessment guideline EAD 330008-02-0601 (2016) [11]. SFRC in Precast Segment for Tunnel Lining Due to the advantages of SFRC in performance, durability and in terms of cost reductions compare to traditionally reinforced concrete the application and the use of SFRC in precast tunnel lining design is a growing trend [12, 13] (s. Figure 4). Anchor channels with channel bolts (aka T-bolts) allow an easy and reliable connection of any kind of component for such structures, e.g. trays, railings, lights, sprinklers and equipment. While SFRC makes the post-installation of fasteners difficult, it is an ideal substrate for anchor channels-channel bolt-systems.

Fig. 4. Construction of the tunnel lining segments, from [14].

Anchoring in SFRC The proliferation of SFRC also means that anchoring of structural and non-structural components has to be carried out in this substrate. Not many tests on fasteners used for the anchoring in SFRC have been conducted yet. Walter, E. and Ammann, W. [15] studied the behaviour of post-installed undercut fasteners, adhesive fasteners, and wedge fasteners under tension and concluded that a statistically significant increase of the capacities cannot be inferred. Also Klug, Y., Holschemacher, K. et al. [16] found it difficult to predict the increase in tension capacity of post-installed fasteners and reasoned that this is due to the inhomogeneous distribution and orientation of the fibres. Kurz, C., Thiele, C. et al. [17] reported that fasteners post-installed in SFRC develop tension capacities which are at least equivalent to the

Influence of the Steel Fibres on the Tension and Shear Resistance

225

capacity if post-installed in regular concrete. To the knowledge of the authors, no shear tests on post-installed fasteners have ever been carried out yet. All studies mentioned that drilling in SFRC is difficult because the steel fibres may cause jamming and increased wear of the drilling tools. This challenge is exacerbated because high performance concrete i.e. high strength concrete is typically used for SFRC elements. In these regards, cast-in fasteners are more suitable for the anchoring in SFRC, e.g. anchor channels with channel bolts. Nilforoush, R., Nilsson, M. et al. [18] carried out tension tests on cast-in headed fasteners in plain and steel fibre reinforced normal- and highstrength concrete with compressive strengths up to 80 MPa. It was concluded that the concrete capacity (design) method (Fuchs, W., Eligehausen, R. et al. [19]) considerably underestimates the tensile breakout capacity of headed fasteners in fibre reinforced concrete and that fibres facilitate a pronounced ductile deformation at ultimate load and prevent a brittle post-peak behaviour potentially associated with high-strength concrete. Some of the first published shear tests on cast-in fasteners have recently been conducted by Lee, J.-H., Cho, B.-S. et al. [20]. The tested cast-in headed fasteners showed a pronounced correlation of ultimate shear capacity and fibre content. However, neither tension nor shear tests on channel bolts installed in anchor channels which were cast in SFRC have been carried out to date. To investigate the performance of channel boltsanchor channels-systems in SFRC, an extensive test program was launched.

2 Experimental Investigations 2.1

Materials

For all tests, the same concrete mixture with a tested compressive strength of about 95 N/mm2 was used which is representative for applications where SFRC is used. No reinforcing bars were installed since anchor channel-channel bolt-systems are generally cast in unreinforced concrete members to study their load-displacement behaviour, e.g. for product qualification. The steel fibres provided from the producer KrampeHarex were made of circular, non–alloy steel wire with end hooks, diameter 0.75 mm, length 60 mm and a nominal steel yield strength of 1900 N/mm2. The fibre mass content was 40 kg/m3, equal to about 0.5% by weight. 320 mm long anchor channels JTA W 53/34 from the producer Jordahl were cast into the concrete members. These anchor channels were made of hot–rolled profiles with two anchors at a distance of s = 250 mm. For embedment depths smaller than the standard depth of hef = 155 mm with round headed anchors riveted to the channel, I–shaped anchors made of cut I–beams were welded to the channel. Shutter and concrete works were carried out in a precast yard of the company Max Bögl. The installation of the T–bolts JB M16 prior to testing completed the test specimens. The grade 8.8 channel bolts had a nominal steel yield strength of 640 N/mm2. 2.2

Test Program

The 8 shear and 8 tension test series presented in this paper (Table 1) are part of a larger test program. The number of test repeats within each series was typically 3 for shear and 5 for tension tests in fibre reinforced concrete and 2 for shear and 3 for tension reference tests in plain concrete. The edge distance c1 and the embedment depth

226

M. Ayoubi et al.

hef varied for the shear and tension test series, respectively. In addition to the 48 tests on anchor channel-channel bolt-systems, tests to determine the performance class of the fibre reinforced concrete were carried out that are not discussed in this paper. 2.3

Test Setup

The tests were carried out at the Jordahl Test Laboratory (three point bending tests were carried out at the Frankfurt University of Applied Sciences and concrete tests were carried out at the Max Bögl Laboratory). Two different unconfined test setups were used where the support has sufficient distance to the anchoring in order to allow the development of a full concrete breakout (Fig. 5 and 6). For shear loading, the test specimens were placed on a strong floor and tied down to counteract the vertical uplift forces deriving from eccentricity effects. A support accommodated the actuator for shear loading and provided horizontal bearings at a distance of 5c1 + s. A PTFE sheet was placed on top of the anchor channel and surrounding concrete before the loading fixture was connected with a balance beam to ensure equal loading of the two channel bolts installed above the anchors. For tension loading, a support with the mounted actuator for tension loading formed with the test specimen a self–equilibrium system. The contact area had a distance of at least 2hef from the centre of anchoring. The loading fixture was connected to a channel bolt installed above an anchor. The anchor channel-channel bolt system was monotonically loaded to failure at a constant rate within 2 to 3 min. Load cells and displacement transducers recorded load Fu and displacement d at a rate of 5 Hz. Table 1. Test Program, [6] Series*

Repeats Load direction Concrete type Concrete strength Edge Embedment depth fc,test distance hef [MPa] c1° [mm] [mm]

S-p-50-155 S-f-50-155 S-p-100-155 S-f-100-155 S-p-150-155 S-f-150-155 S-p-200-155 S-f-200-155 T-p-∞-69 T-f-∞-69 T-p-∞-95 T-f-∞-95 T-p-∞-120 T-f-∞-120 T-p-∞-155 T-f-∞-155

2 3 2 3 2 3 2 3 2 5 2 5 1 5 3 5

Shear Shear Shear Shear Shear Shear Shear Shear Tension Tension Tension Tension Tension Tension Tension Tension

Plain Fibre Plain Fibre Plain Fibre Plain Fibre Plain Fibre Plain Fibre Plain Fibre Plain Fibre

96.4 98.6 96.4 98.6 96.4 98.6 96.4 98.6 94.1 98.2 94.1 98.2 94.1 98.2 92.5 92.8

50 50 100 100 150 150 200 200 ∞ ∞ ∞ ∞ ∞ ∞ ∞ ∞

155 155 155 155 155 155 155 155 69 69 95 95 120 120 155 155

* Code: Shear or Tension-plain or fibre-c1-hef; ° ∞ equals to any distance larger than 2hef

Influence of the Steel Fibres on the Tension and Shear Resistance

227

Fig. 5. Test setup for shear loading (left) and tension loading (right), [6].

Fig. 6. Pictures of the test setup for shear loading (left) and tension loading (right).

2.4

Test Results

The coefficients of variation (cv) were reasonable despite the small number of test repeats per series (Table 2): The cv of the ultimate load Fu was typically well below 15% which is the threshold commonly accepted for concrete related failure modes in fastener qualification testing. The high cv of one series (S-f-100-155) can be attributed to a test with a bias towards an outlier. The cv of the displacement at 50% of the ultimate load d(0.5Fu,m) was always below 40% which is the maximum accepted in the context of fastener qualification.

228

M. Ayoubi et al. Table 2. Test results, [6].

Series

S-p-50-155 S-f-50-155 S-p-100155 S-f-100-155 S-p-150155 S-f-150-155 S-p-200155 S-f-200-155 T-p-∞-69 T-f-∞-69 T-p-∞-95 T-f-∞-95 T-p-∞-120 T-f-∞-120 T-p-∞-155 T-f-∞-155

Ultimate load Coeff. of var. Displacement Coeff. of var. Displacement Failure mode* d(0.5Fu,m)m cv(Fu) cv(s(0.5Fu,m)) d(Fu,m)m Fu,m [%] [%] [mm] [mm] [mm] 47.6 81.0 71.8

1.6 0.8 11.2

0.71 1.23 1.42

13.46 27.8 23.3

3.00 4.66 2.67

2C 3C 3C

172.0 105.0

21.3 4.1

1.94 1.58

15.6 16.6

5.43 3.23

3C 3C

212.6 144.4

3.6 2.8

1.77 1.83

10.7 17.9

5.21 2.84

3C 3C

269.4 89.6 110.3 99.6 106.0 105.2 109.2 93.5 91.3

2.0 8.0 2.9 15.0 3.6 – 2.4 4.3 7.1

2.71 0.66 1.31 0.90 1.23 1.06 1.42 0.95 0.93

12.4 18.4 17.8 5.5 9.4 – 10.2 37.4 37.7

7.98 5.61 9.11 8.29 9.13 11.21 11.94 20.70 22.44

3 Sb 2C 1 C, 1 Sl, 3 Sb 2C 1 C, 2 Sa, 2 Sl Sl 4 Sl, 1 Sb 1 Sl, 2 Sb 4 Sl, 1 Sb

* Sb: steel failure bolt; Sl: steel failure lip; Sa: steel failure anchor; C concrete cone or edge breakout

Overall, no clear trend of the coefficients of variation with regard to the concrete type (plain or fibre) could be inferred (s. last paragraph). By trend, the recorded displacements d(0.5Fu,m) and d(Fu,m) confirmed that fibres consistently support a more ductile behaviour also of anchor channels with channel bolts. More prominent, the fibres significantly influenced the ultimate load Fu and the failure modes of the tested anchor channelchannel bolt-systems: Subjected to shear load, only the systems cast in fibre reinforced concrete with the largest tested edge distance c1 = 200 mm failed in steel due to shearing off the bolt, otherwise concrete edge breakout was decisive. Under tension load, only the systems cast in plain concrete with the embedment depth hef  95 mm failed consistently by concrete cone breakout, otherwise steel failure of bolt, lip or anchor occurred (rupture or bending). Clearly, if steel failure is the controlling failure mode, fibres have no effect. If failure occurs due to concrete breakout, the fibres increased the capacity significantly by the factor of about 1.8 for shear and 1.4 for tension. Moreover, the fibres allow the shift of the transition from concrete breakout to steel failure (s. Fig. 7, 8 and 9). The dashed lines (Fig. 9a and 9b) represent the mentioned transmission to steel failure depending on the test results (s. Table 2, failure mode).

Influence of the Steel Fibres on the Tension and Shear Resistance

229

Fig. 7. Pictures of the failure under tension loading.

Fig. 8. Pictures of the failure under shear loading.

To compare the performance of anchor channel-channel bolt-systems cast in plain and fibre reinforced concrete further, the influence of the edge distance c1 and embedment depth hef is illustrated by means of typical curves recorded during the shear and tension tests (Fig. 10): The fibres cause a substantial increase of the shear capacity where the failure mode remains concrete edge breakout if tested with an edge distance

230

M. Ayoubi et al.

Fig. 9. Capacities of anchor channels-channel bolt-systems cast in plain and fibre reinforced concrete tested a) in shear and b) in tension for different edge distances and embedment depths, respectively, [6].

Fig. 10. Load-displacement curves of anchor channels-channel bolt-systems cast in plain and fibre reinforced concrete tested in shear with a) small and b) large edge distance and tested in tension with c) small and d) large embedment depth, [6].

Influence of the Steel Fibres on the Tension and Shear Resistance

231

of c1 = 50 mm (Fig. 10a) but changes to steel failure if tested with an edge distance of c1 = 200 mm (Fig. 10b). In this case, the displacement at ultimate load is roughly tripled. The fibres also cause a change from concrete cone breakout to steel failure, accompanied by a distinct increase in tension capacity and displacement, if customized systems with an embedment depth of hef = 69 mm are tested (Fig. 10c). In contrast, no significant influence of the fibres can be determined if the standard system with an embedment depth of hef = 155 mm is tested since for this configuration steel failure is controlling already in case the anchor channel-channel bolt-system is cast into concrete without fibres (Fig. 10d). The examples demonstrate that the fibres increase the ductility of concrete breakouts and may change it to steel failure modes.

3 Summary and Conclusion Steel Fibre reinforced concrete (SFRC) gains importance for the construction of structural members, e.g. tunnel segments. The drilling in SFRC for the post-installation of fasteners is challenging, not least because of the high concrete strengths prevalent for SFRC. For this and other reasons, anchor channels with channel bolts are often a favourable solution to connect any component to structural elements made of SFRC. However, no study on the performance of anchor channel-channel bolt-systems has been published to date. To this end, a research program was launched to compare the performance of channel bolts installed in anchor channels cast in plain and fibre reinforced concrete. The results of 48 shear and tension tests presented in this paper demonstrate that anchor channel-channel bolt-systems cast in fibre reinforced concrete sustain higher ultimate loads and develop larger corresponding displacements if compared with identical systems cast in plain concrete. The increase in capacity and ductility may lead to a positive shift from rather brittle concrete breakout to more ductile steel failure modes. The views expressed in this paper are the views of the authors only and do not necessarily reflect the views of Jordahl, Max Bögl and KrampeHarex. Acknowledgements. Anchor Channels, Concrete test members and steel fibres were kindly provided by the companies Jordahl, Max Bögl and KrampeHarex, which are greatly appreciated.

References 1. Bokor, B., Tóth, M., Sharma, A.: Influence of steel fiber content on the load-bearing capacity of anchorages in concrete (2017) 2. Mechtcherine, V.: Rissbeherrschung durch Faserbewehrung, Beherrschung von Rissen in Beton. In: 7. Symposium Baustoffe und Bauwerkserhaltung Karlsruher Institut für Technologie, S. 83–94. KIT Scientific Publishing (23. März 2010) 3. Døssland, Ä.L.: Fibre Reinforcement in load carrying concrete structures, Norwegian University of Science and Technology, Thesis for the degree of philosophiae doctor, February 2008 4. Lee, S.-C., Cho, J.-Y., Vecchio, F.J.: Diverse embedment model for steel fiber-reinforced concrete in tension: model verification. ACI Mater. J. 107, 526–535 (2011)

232

M. Ayoubi et al.

5. Mahrenholtz, C., Lambton, J., Julier, F.: Suitability of anchor channels with channel bolts for use in nuclear power plants. In: Proceedings of the 24th International Conference on Structural Mechanics in Reactor Technology (SMiRT 24), Proceeding Div VI, Busan (2017) 6. Mahrenholtz, C., Ayoubi, M., Müller, S., Bachschmid, S.: Performance of anchor channels with channel bolts cast in fibre reinforced concrete (FRC). IOP Conf. Ser. Mater. Sci. Eng. 615, 728–735 (2019) 7. Gage, C.: Are we paying enough attention to elevator shaft connections? Elevator World, October Edition (2014) 8. Gage, C.: Looking behind the façade at curtain wall connections. Chinese Curtain Wall Magazin (2014b) 9. Gottschalk, B., Mahrenholtz, C.: Befestigung von Ankerschienen mit Installationskonen (Fastening anchor channels with installation cones). BFT International Betonwerk + Fertigteil-Technik, 2017, Heft 6 (2017) 10. EN 1992–4, Eurocode 2: Design of concrete structures – Part 4: Design of fastenings for use in concrete. European Committee for Standardization (CEN); EN 1992-4 (2018) 11. EAD 330008-02-0601, Anchor channels. European Assessment Document, OJEU 2016/C 248/06, European Organization for Technical Assessment (EOTA) (2016) 12. Axhimusa. R.: Investigation of steel fiber reinforced concrete (SFRC) elements with regard to their economic viability and market growth prospects. Master Thesis at the Frankfurt University of Applied Sciences, Supervisor: Prof. Dr.-Ing. M. Ayoubi, 2019 13. Rivaz, B.: Steel fiber reinforced concrete (SFRC): the use of SFRC in precast segment for tunnel lining, p. 66 (2009) 14. Carmona, S., Molins, C., Aguado, A., Mora, F.: Distribution of fibers in SFRC segments for tunnel linings. Tunn. Undergr. Space Technol. 51, 238–249 (2016) 15. Walter, E., Ammann, W.: Fastening technology in fibre reinforced concrete. In: Proceedings of the 4th RILEM International Symposium on Fibre Reinforced Concrete, Sheffield (1992) 16. Klug, Y., Holschemacher, K., Wittmann, F.: Tragverhalten von Befestigungselementen in Stahlfaserbeton (Load carrying behaviour of fastening elements in steel fibre reinforced concrete). Innovationen im Bauwesen, Beiträge aus Praxis und Wissenschaft (2002) 17. Kurz, C., Thiele, C., Schnell, J., Reuter, M., Vitt, G.: Tragverhalten von Dübeln in Stahlfaserbeton (Load carrying behaviour of fasteners in fibre reinforced concrete). Bautechnik 89. Heft 8, 545–552 (2012) 18. Nilforoush, R., Nilsson, M., Elfgren, L.: Experimental evaluation of tensile behaviour of single cast-in-place anchor bolts in plain and steel fibre-reinforced normal- and high-strength concrete. Eng. Struct. 147, 195–206 (2017) 19. Fuchs, W., Eligehausen, R., Breen, J.: Concrete Capacity Design (CCD) Approach for fastening to concrete. ACI Struct. J. 92(6), 73–94 (1995) 20. Lee, J.-H., Cho, B.-S., Kim, J.-B., Lee, K.-J., Jung, C.-Y.: Shear capacity of cast-in headed anchors in steel fiber-reinforced concrete. Eng. Struct. 171, 421–432 (2018)

Fiber Reinforced Concrete After Elevated Temperatures: Techniques of Characterization Ronney Rodrigues Agra1,2(&), Ramoel Serafini1,2, and Antonio Domingues de Figueiredo1 1

Department of Civil Construction Engineering, Polytechnic School of University of São Paulo, São Paulo, Brazil [email protected] 2 Institute of Technological Research, São Paulo, Brazil

Abstract. The mechanical properties of fiber reinforced concrete (FRC) are negatively affected when subjected to elevated temperatures. The main concern is regarding its post-crack tensile strength, which can be severely impaired at temperatures above 300 °C. In this composite, the mechanical characterization is constantly performed by means of bending tests of prismatic specimens, as recommended by EN 14651. However, due to limiting aspects, alternative methodologies have been used for the characterization of FRC, among which are the DEWS (Double Edge Wedge Splitting) and the Double Punch tests. In this context, the present study compares the methodologies for evaluating the mechanical behavior of FRC after elevated temperatures, discussing and emphasizing its advantages and limitations. The Double Punch test does not show satisfactory response as a consequence of the degradation suffered by the sample and the puncture interaction induced by the test. On the other hand, the indirect tensile DEWS test shows that it is capable of characterizing the FRC even after exposure to elevated temperatures. Although the post-crack response of the composite varies according to the method adopted, the post-crack tensile strength in the service limit state (SLS) and ultimate limit state (ULS) are considerably reduced when compared with the ambient temperature. Keywords: Fiber reinforced concrete strength
Elevated temperatures


Test method
Post-crack tensile

1 Introduction The crescent use of fiber reinforced concrete (FRC) in the construction industry can be attributed to economical and technical aspects that benefits the production of precast elements, slabs, and tunnel linings. A significant contribution has been made by the publication of the fib Model Code 2010 [1], which presented parameters for the use of FRC for structural purposes. This led to the further increase in the use of this technology for the construction of buildings and the production of segments for tunnels built with Tunnel Boring Machine (TBM) technology [2]. The mechanical properties of FRC are negatively affected when subjected to elevated temperatures [3] and the post-crack behavior varies according the type of fiber, © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 233–244, 2021. https://doi.org/10.1007/978-3-030-58482-5_21

234

R. R. Agra et al.

amount of fiber used, the duration of temperature exposure, and maximum temperature achieved. In this sense, the main concern is regarding the post-crack tensile strength, which may be severely reduced for temperatures above 300 °C [4]. The research carried out by Dehn and Hermann [5] reveals the absent literature and normative bases that allow the FRC subject to high temperatures to be adequately treated for design checks, which can put structural safety at risk. This indicates the need for further studies with regard to requirements, test methods and procedures for the design of FRC structures when subjected to fire, especially with regard to the characterization of the composite’s behavior and obtaining constitutive equations to be applied in behavior prediction models. Although the current studies on the post-crack behavior of the FRC exposed to fire concentrate mainly on the assessment of the mechanical properties for a specific target temperature [6–8], the research carried out by Serafini et al. [9] is highlighted for using the unifacial heating fire methodology. The fire simulator operated by the burning of methane to achieve a similar behavior to the hydrocarbon fire curve. In this study, the authors evaluated the effect of fire on the mechanical properties of the macro-synthetic fiber reinforced concrete (MSFRC) for tunneling applications. The results can be used as a basis for the development of design-oriented models that allow to evaluate the structural condition of the FRC after the fire exposure. The high strength concrete (HSC) is more sensitive to the occurrence of spalling due to the low permeability and porosity of the cementitious matrix [4]. It is widely accepted that the use of polymeric micro-synthetic fibers promotes a sensitive reduction in the process of explosive spalling [10] of concrete, and that the use of steel fibers is beneficial to the fire resistance of the composite. In the absence of spalling, the verification of the residual mechanical properties of the structure and a detailed inspection of the elements affected by fire is mandatory to evaluate the degree of safety of the structure and to aid the decision-making process. In most of the fire events, the structures do not collapse and the post-fire capacity must be evaluated, since rebuilding a structure or reinforcing damaged elements has a considerable monetary impact. Therefore, the knowledge regarding the residual properties of the cementitious matrix and the reinforcing mechanism is key to decide which procedures must be employed [11]. This analysis can be conducted by means of destructive and non-destructive techniques. In order to determine the FRC mechanical properties from a structural perspective, the post-crack tensile strength of the composite must be determined [1]. The characterization of the post-crack tensile strength of FRC is usually conducted based on standard bending tests in prismatic specimens. This test has already been used to evaluate macro-synthetic fiber reinforced concretes exposed to elevated temperatures [9]. However, the experimental challenges related to bending tests are considerable due to the size of the specimens and the difficulties with handling the specimens after exposure to high temperatures. More than that, the flexural tensile strength values tend to be overestimated in bending tests and questions can arise from the suitability of using converting factors to reach a more reasonable direct tensile strength value, equivalent to a mode I fracture type. Therefore, emerging methodologies have been used in substitution to the traditional bending tests to evaluate FRCs, among them the Double Punch test [12, 13] and the Double Edge Wedge Splitting (DEWS) test [14].

Fiber Reinforced Concrete After Elevated Temperatures

235

In this sense, the current study aims to present the techniques employed to determine the post-crack tensile properties of FRC after exposure to elevated temperatures, discussing and highlighting the advantages and limitations of each method. This study contributes to the limited literature regarding the comparison between tests capable of determining the post-crack behavior of FRC.

2 The Effect of Temperature on the Mechanical Behavior of FRC A temperature gradient is induced in the FRC when it is subjected to elevated temperatures, which results in different layers of dehydration of the cement paste and micro-cracks. This culminates that the mechanical properties of the composite are affected as a function of the distance to the heated surface. In this sense, any evaluation of the global mechanical properties of FRC elements exposed to fire would results in the average mechanical response of the affected layers. However, there are a very limited number of studies that study the temperature gradient in FRC and its impact on the mechanical properties of the composite [9]. The immediate consequence of exposing concrete to heat is a change in the characteristics of the composite, such as the reduction of its strength and stiffness, as well as the generation of additional deformations linked to the stress level during the first stages of heating. In addition, as the temperature distribution in the section is not uniform, these thermal deformations are also not uniform in the section, which generates effects related to spalling, thermal expansion and lateral deflections [11]. The parameters of compressive and tensile strength, stiffness and energy absorption capacity of the FRC are largely affected by the increase of temperature. Yermak et al. [4] in their studies noted that the mechanical strength of concrete with fibers (0.75 kg/m3 of PP microfibers and 60 kg/m3 of steel fibers) is greater than that of concrete without fibers at room temperature. There was an approximate difference of 15 MPa in the compressive strength and 3,5 MPa in the tensile strength, which demonstrates that the steel fibers could act in this gain. Some studies [4, 7, 8] show that in concretes with the addition of up to 70 kg/m3 of steel fibers there are reductions in compressive strength that reach 40% after heating to 400 °C and 70% after heating to 600 °C, when compared to room temperature. After exposure to high temperatures (above 300 °C) a high decrease of the compressive strength was noted because of the thermal strain mismatch between aggregate (which expand) and paste (which shrink) inducing tangential and radial cracks at the pasteaggregate interface [4, 15]. However, the use of steel fibers reduced the rate of degradation of compressive strength, although at 900 °C the presence of steel fibers was negligible for the residual strength of concrete. In addition, the decrease in the values of modulus of elasticity is more pronounced when compared to that of compressive strength, especially between the temperatures of 150 and 450 °C [4, 6, 9, 16]. Reductions are directly associated to the mismatch between aggregates and cement paste in terms of thermal expansion, which results in the formation of extensive cracks in the interfacial transition zone.

236

R. R. Agra et al.

The elastic properties of the composite are also negatively affected by the degradation of synthetic fibers, the dehydration of hydrated products present in the concrete matrix (especially portlandite and C-S-H), the breakage of bonds in the microstructure, the increase in aggregate porosity, the reduction in the specific surface area of the hydrates, and the coarsening of the cement paste pore structure. These processes result in the increase in the porosity of the cement paste and the increase in capillarity pore size [16]. Additionally, Yermak et al. [4] in their studies noted that the addition of steel fibers did not bring a meaningful contribution to the reduction in the elastic modulus with the temperature increase. The post-cracking behavior of FRC at high temperatures changes according to the type and fiber content, the exposure temperature and the duration of the exposure. The SFRC shows lower losses in the post-peak response for temperatures above 400 ° C, while the MSFRC shows greater reductions in post-crack tensile strength and toughness due to fiber degradation [17]. Yermak et al. [4] found that the post-crack tensile strength of the FRC shows a minor decrease tendency for temperatures up to 300 °C. After that temperature mark, severe damage is caused in the composite. When compared to plain concrete specimens, the use of steel fibers has shown an increase of *10% in the post-crack tensile strength for target temperatures of 300 and 600 °C. The steel fibers contributed to improve the post-peak behavior of concretes for temperatures up to 750 °C, while at around 900 °C the steel fibers did not show any capability to enhance the post-crack tensile strength of the composite. The reduction at 900 °C occurred due to the oxidation and corrosion of the steel fibers, which start to happen at 500 and 700 °C, respectively. They also noted that the shape of the load-displacement curve changes with temperature, since the peak load increases and the post-crack response is reduced at 600 °C. The mechanical properties of steel fibers and the fiber-matrix interaction are factors that considerably influence the post-cracking behavior of FRC exposed to high temperatures. The fiber-matrix interaction begins to be affected for temperature of 500 °C and above in terms of pullout load capacity [18].

3 Post-heat Evaluation of FRC by Means of Non-destructive Tests The evaluation of the capacity of concrete structures exposed to fire is a very difficult task, since traditional destructive or non-destructive techniques are not suitable for the inspection of a heterogeneous material. Possible approaches to this problem usually involve inspection of the average response of the concrete element or a point-to-point analysis of small samples extracted from different depths. The combination of several techniques is constantly applied in this analysis, which allows investigating and measuring physical, chemical and mineralogical changes, as well as the higher temperature reached in the structural element [19]. In addition, these techniques must be applied together with mechanical tests, in order to obtain a complete evaluation of the concrete after exposure to temperature. A summary of these non-destructive techniques usually employed in the FRC is shown in Table 1. The purpose of these tests is to determine the quality and integrity of the

Fiber Reinforced Concrete After Elevated Temperatures

237

materials, without compromising the load capacity of these elements. Therefore, the methodology adopted must not harm the future use of the structural element, even if commit actions considered invasive [20]. Table 1. Main non-destructive tests applied to concrete after exposure to elevated temperatures. Analysis of concrete cover Schmidt Hammer (Sclerometer) Internal fracture test Penetration resistance (Windsor Probe) CAPO test

Point-to-point analysis of small specimens Differential Thermal Analysis (DTA) Thermogravimetric Analysis (TGA) Scanning Electron Microscopy (SEM) X-Ray Diffraction

Special techniques Ultrasonic Pulse Velocity (UPV) Tomography and Resonance Modal Analysis of Surface Waves (MASW) Digital Image Correlation (DIC)

4 Post-crack Tensile Strength of FRC Exposed to Elevated Temperatures The existence of adequate methods to assess the mechanical behavior of the FRC increases the reliability of the application of this material. The adopted method must be compatible with existing design models, which is very important for the successful implementation of the composite [21]. In order to characterize the mechanical performance of the FRC, it is necessary to consider the post-peak behavior of the loaddeformation curve, with the determination of the post-crack tensile strength of the composite and the increase in toughness due to the addition of fibers [1]. Experimental verifications from the structural point of view of the FRC are suggested with the purpose of knowing its behavior under high temperature conditions, since there is still no adequate approach to fire design for the composite. In addition, technical parameters and numerical models to assess the behavior of the FRC after fire events are very limited in the current literature [5]. For this reason, it has been usual to carry out real scale (or at least representative scale) tests to verify the safety of structure made with FRC under fire loading. Thus, mesoscale tests are required to obtain the post-heat mechanical response of the composite. The mechanical characterization of the FRC is traditionally carried out by means of bending tests. However, after exposure to high temperatures, the thermal gradients induced also vary significantly depending on the dimensions of the specimen and the structural element [19]. The test recommended by EN 14651 [22] has already been used (Fig. 1) to assess the post-crack flexural tensile strength of the macro-synthetic fiber reinforced concrete (MSFRC) subjected to a real fire simulation [9, 23]. These papers contribute to the limited amount of literature data regarding the effect of real fire scenarios on the properties of MSFRC. In the study of Serafini et al. [9] prismatic specimens were subjected to single face heating under hydrocarbon fire curve

238

R. R. Agra et al.

Fig. 1. Fire exposed specimen after flexural tensile strength test [23 adapted]

and flexural results demonstrated that the composite has no significant flexural resistance after fire exposure (Fig. 2) due to reinforcement melting (170 °C) and ignition (400–500 °C) in the flexural-tensile region of the specimen.

Fig. 2. Load-CMOD curves for MSFRC before and after fire exposure [9 adapted]

Reductions in matrix tensile strength values were associated with dehydration of hydrates in cement paste, specific surface area of hydrates reduction, increase in the total pore volume, and changes in cement paste pore distribution [24]. Post-crack flexural tensile strength is greatly affected by fire in both service and ultimate limit states. These changes are related to microstructural changes in the cement paste, physical changes of the composite due to fire exposure, and the deterioration of fiber reinforcement which results in a specimen with little to no post-crack flexural tensile strength. Dynamic elastic and rigidity modulus values decrease by over 86% after fire exposure due to the influence of cracks in the sample, physicochemical changes in the matrix, and reinforcement fiber degradation. However, experimental difficulties to perform the bending test are considerable due to the relatively large dimensions of the prismatic specimens and the difficulties in the process of handling the prisms after being subjected to high temperatures. In addition, there is an overestimation of the flexural tensile strength values when compared to the

Fiber Reinforced Concrete After Elevated Temperatures

239

results obtained by direct tensile strength tests. Moreover, doubts arise in the scientific community about the relevance of using conversion factors for flexural strength values to obtain the equivalent direct tensile strength in room temperature. Therefore, new methodologies have emerged with the proposition to replace the traditional bending tests in the task of evaluating FRC. A few emerging tests can be cited, such as the Double Punch test (DPT) and Double Edge Wedge Splitting (DEWS) test [12–14]. The Double Punch Test (DPT) has a number of advantages, such as its ease of procedure, simple specimen preparation and lower variability in the measured results than those resulting from bending tests using beam specimens [25]. Particularly, the properties of FRC can be determined by using relatively small cylindrical specimens which can be molded, cut from standardized cylinders, or cores drilled from an existing structure. For application of the simplified methodology [26], accessible in most quality control laboratories, a conventional compression test frame equipped with an axial displacement control (not necessarily with a closed-loop control) is required. The DPT also has correlations with the EN 14651 [22] test, and allows the detection of changes in the characteristics of the FRC in face of the influences of parameters that determine the mechanical performance of this composite. This make the test suitable for the quality control of FRC in construction applications [27]. One of the obstacles to its use as a definitive technique for describing the FRC postcracking response is the difficulty in determining the constitutive stress-strain equations, mainly due to the unpredictability of the number of fracture planes [14]. However, the Double Punch test is one of the few tests presented in the literature that can be performed on FRC specimens drilled from real structures that have been exposed to a fire [6]. The DPT has also been applied in order to investigate the mechanical properties of FRC exposed to high temperatures, showing that the detrimental factors for the tensile behavior of FRC are more the volume fraction and the aspect ratio of the fibers, than their type [28]. However, one of the limitations of using the DPT for FRC after exposure to high temperatures is that the deterioration of the concrete matrix results in an increased frictional interaction in the region where the loads are applied, which may influence the post-crack results [6]. Rambo et al. [6] investigated the mechanical properties of concrete samples reinforced with polymeric macrofibers exposed to temperatures up to 600 °C and the applicability of the DPT to assess the post-crack tensile strength. The effect of lower temperatures was not significant, however, as a result of the degradation suffered by the sample surface, added to the puncture that results in the crushing of the deteriorated matrix, the Double Punch test did not show a conclusive response after exposure to high temperatures. This was mainly because the additional friction effect generated by the punching interaction between the concrete matrix and the piston can be confused with the bridge effect provided by the fibers. Nevertheless, the gradient of temperature stablished within the specimens may preserve part of the material (i.e.: matrix and fibers), and consequently, the composite post cracking performance. The stress-strain curves obtained from the Double Punch tests demonstrated that the MSFRC gradually loses tensile strength an energy consumption density with increasing temperature and the post-cracking response varies significantly depending on the temperature (Fig. 3). However, the effect of temperature on the mechanical behavior of

240

R. R. Agra et al.

the MSFRC showed to be very similar to that known for conventional concrete with relation to the loss of mechanical strength and elastic modulus.

Fig. 3. Stress-strain curves obtained from the results of the Double Punch test for the MSFRC specimens submitted to different temperatures [6 adapted].

The MSFRC exposed up to 200 °C maintain similar values of the post-crack tensile strength and overall ductility when compared to the MSFRC at room temperature. However, from 400 °C upwards, the bearing capacity of the material is significantly reduced and shown to be critical to the MSFRC mechanical performance.The DEWS test constitutes an indirect tensile test by applying a compression load to a specimen with wedges and notches on two opposite faces. This is a test that can be carried out on small cubic samples, extracted from structures affected by fire or from larger specimens. In addition, it can be executed in an open-loop control system, which makes its application feasible in most laboratories. A type I fracture is obtained by means of the DEWS test and, thus, the stress-strain relationship is obtained directly, which is fundamentally important for structural safety assessment based on the investigation of its resistant capacity, in accordance with the fib Model Code [1]. Agra et al. [29] showed that the DEWS test is able to characterize the SFRC in terms of post-crack tensile strength even in severe conditions, as in the case of samples subjected to the action of fire. This was possible because the evaluation occurred without prejudice to the values obtained as a response to the material, since no concrete damage was found under the contact region. The authors cite that the tensile strength

Fiber Reinforced Concrete After Elevated Temperatures

241

values of the composite after exposure to fire were 71.1% lower than the values obtained at room temperature. This is associated with dehydration of cement paste products, loss of fiber reinforcing capacity and changes in pore distribution [24]. The values of post-crack tensile strength in SLS (COD = 0.5 mm) and ULS (COD = 2.5 mm) were 64.1% and 59.8% lower than the values obtained at room temperature (Fig. 4). These effects are related to physical-chemical changes in the matrix, microstructural changes in the composite and degradation of the fibers used as reinforcement.

Fig. 4. Average stress-COD curves for DEWS test before and after hydrocarbon fire exposure [29 adapted].

Serafini et al. [30] also applied the DEWS test at SFRC exposed to 600 °C. The results obtained also show that the tensile strength of the cementitious matrix and the post-crack tensile strength in SLS and ULS are significantly affected. The composite tensile strength values after temperature exposure of 600 °C were 82.5% lower than the value reached at room temperature and the post-crack tensile strength values at SLS (COD = 0.5 mm) reduced by 74.3% and in ULS (COD = 2.5 mm) by 72.2% when compared to room temperature results. The authors also claim that no visible damage was caused by the interaction between the roller and the SFRC for temperatures up to 600 °C. In studies related to the DEWS and Double Punch tests the authors note that some limitations – such as the presence of an unstable post-peak region – should be observed, especially when the gap between the tensile strength of the matrix and the post-crack tensile strength corresponding to the SLS is high. It can be observed when a low fiber content is employed. The region of instability can also be noticed when the open-loop machine is used at room temperature. Residual loads are underestimated when open loop system is employed. Besides, the lack of rigidity of the equipment may also cause instability. It is possible to verify the region of instability more clearly in the results obtained by means of the DEWS test, which reached values greater than 0.5 mm. However, the effect of instability is mitigated after the exposure of the specimens to elevated

242

R. R. Agra et al.

temperatures, since the gap between the tensile strength of the matrix and the tensile strength in the SLS becomes smaller. It is important to state that the service state conditions are certainly compromised after a fire event and the main concern is regarding the ultimate state condition values, according to the fib Model Code 2010 [1]. Furthermore, there is no direct influence of the region of instability in determining the residual strength related to the ULS. Thus, the evaluation occurs without damage to the values obtained as a response of the material, which makes the DEWS test a viable methodology under these conditions.

5 Conclusions This paper compares the methodologies for evaluating the mechanical behavior of FRC after elevated temperatures, discussing and emphasizing its advantages and limitations. From the present study, the following conclusions can be drawn: • The parameters of compressive and tensile strength of the cement matrix, post-crack tensile strength, stiffness and energy absorption capacity of the FRC are largely affected by the increase of temperature. • There are obstacles related to the use of flexural test setups for the determination of the mechanical properties of FRC after elevated temperatures: the considerable specimen volume, the difficulties in the process of handling the prisms, the overestimation of the flexural tensile strength values, and the relevance of using conversion factors to estimate the tensile strength. • The Double Punch test may not be adequate to assess the properties of SFRC after temperature exposure. Results presented in the literature indicate that the puncture caused by the test may result in crushing of the porous matrix and induced frictional interaction. • The DEWS test is a viable technique for the assessment of the post-crack tensile properties of SFRC after temperature exposure. One of the main advantages of DEWS test is that it generates a mode I fracture that simplify the determination of constitutive equations from the results. • The effect of post-cracking instability is mitigated after the exposure of the specimens to elevated temperatures. Furthermore, there is no direct influence of the region of instability in determining the residual strength related to the ULS, even with the SLS compromised after a fire event. Acknowledgements. The authors would like to thank the Institute for Technological Research (IPT) and its foundation (FIPT) for their financial and institutional support though the New Talents Program N.01/2017 and N.01/2018. The authors would also like to thank CAPES (Coordenação de Aperfeiçoamento de Pessoal de Nível Superior). Antonio D. de Figueiredo would like to acknowledge the financial support of the National Council for Scientific and Technological Development - CNPq (Proc. Nº: 305055 / 2019-4).

Fiber Reinforced Concrete After Elevated Temperatures

243

References 1. Fédération Internationale du Béton – FIB. ‘Fib Model Code for Concrete Structures 2010’. Switzerland (2013) 2. Liao, L., et al.: Design of FRC tunnel segments considering the ductility requirements of the model code 2010. Tunn Undergr Sp Tech 47, 200–210 (2015) 3. Silva, C.L.: Projeto de estruturas de concreto em situação de incêndio conforme ABNT NBR 15200:2012 (Blucher 2012) 4. Yemark, N., et al.: Influence of steel and/or polypropylene fibers on the behavior of concrete at high temperature: spalling, transfer and mechanical properties. Constr. Build. Mater. 132, 240–250 (2017) 5. Dehn, F., Herrmann, A.: ‘Concreto reforçado com fibras de aço em situação de incêndio – requisitos normativos, pré-normativos e códigos-modelo’. Concreto & Construções, 87 (2017) 6. Rambo, D.A.S., et al.: Study of temperature effect on macro-synthetic fiber reinforced concretes by means of Barcelona tests: an approach focused on tunnels assessment. Constr. Build. Mater. 158, 443–453 (2018) 7. Tai, Y.S., et al.: Mechanical properties of steel fiber reinforced reactive powder concrete following exposure to high temperature reaching 800 & #xB0;C. Nuclear Eng. Design 241, 2416–2424 (2011) 8. Poon, C.S., et al.: ‘Compressive behavior of fiber reinforced high-performance concrete subjected to elevated temperatures’. Cement and Concrete Research, 34, 2215–2222 (2004) 9. Serafini, et al.: Influence of fire on temperature gradient and physical-mechanical properties of macro-synthetic fiber reinforced concrete for tunnel linings. Constr. Build. Mater. 214, 254–268 (2019) 10. Kalifa, P.: High temperature behaviour of HPC with polypropylene fibres: from spalling to microstructure. Cement Concrete Res. 31, 1487–1499 (2001) 11. Fédération Internationale Du Béton - fib Bulletin 46 ‘Fire design of concrete structures structural behaviour and assessment’ State-of-art report, Lausanne, Switzerland 2008 12. Choumanidis, D., et al.: Barcelona test for the evaluation of the mechanical properties of single and hybrid FRC, exposed to elevated temperature. Constr. Build. Mater. 138, 296– 305 (2017) 13. UNE 83515: 2010 ‘Hormigones con fibras. Determinación de la resistencia a fisuración, tenacidad y resistencia residual a tracción. Método Barcelona. The Spanish Association for Standardisation’, Madrid 2010 14. di Prisco, M., Ferrara, L., Lamperti, M.G.L.: Double edge wedge splitting (DEWS): an indirect tension test to identify post-cracking behaviour of fibre reinforced cementitious composites. Mater. Struct. 46(11), 1893–1918 (2013). https://doi.org/10.1617/s11527-0130028-2 15. Xing, Z., et al.: Aggregate’s influence on thermophysical concrete properties at elevated temperature. Constr. Build. Mater. 95, 18–28 (2015) 16. Zheng, W., Li, H., Wang, Y.: Compressive stress-strain relationship of steel fiber-reinforced reactive powder concrete after exposure to elevated temperatures. Constr. Build. Mater. 35, 931–940 (2012) 17. Sukontasukkul, P., et al.: Post-crack (or post-peak) flexural response and toughness of fiber reinforced concrete after exposure to high temperature. Constr. Build. Mater. 24, 1967–1974 (2010) 18. Abdallah, S., Fan, M., Cashell, K.A.: Pull-out behaviour of straight and hooked-end steel fibres under elevated temperatures. Cem. Concr. Res. 95, 132–140 (2017)

244

R. R. Agra et al.

19. Fernandes, B., Gil, A.M., Bolina, F.L., Tutikian, B.F.: Microstructure of concrete subjected to elevated temperatures: physico-chemical changes and analysis techniques. Ibracon Struct. Mater. J. 10, 838–863 (2017) 20. Helal, J., Sofi, M., Mendis, P.: Non-destructive testing of concrete: a review of methods. Electr. J. Struct. Eng. 14, 97–105 (2015) 21. Monte, R., Toaldo, G.S., Figueiredo, A.D.: Avaliação da tenacidade de concretos reforçados com fibras através de ensaios com sistema aberto. Revista Matéria 19, 132–149 (2014) 22. Standardization, E.C.F.: EN 14651: Test method for metallic fiber-reinforced concrete – Measuring the flexural tensile strength (limit of proportionality (LOP), residual, p. 15p. CEN, London (2007) 23. Serafini, R., et al.: Influence of fire exposure on the flexural behavior of macro-synthetic fiber reinforced concrete. In: 9th international conference on concrete under severe conditionsenvironment & loading, Porto Alegre 2019 24. Ma, Q., et al.: Mechanical properties of concrete at high temperature-a review. Constr. Build. Mater. 93, 371–383 (2015) 25. Molins, C., Aguado, A., Saludes, S.: Double Punch Test to control the energy dissipation in tension of FRC (Barcelona test). Mater. Struct. 42, 415–425 (2008) 26. Pujadas, P.: Caracterización y diseño del hormigón reforzado con fibras plásticas. Tesis Doctoral - Universitat Politècnica de Catalunya, Barcelona (2013) 27. Silva, C.L.: ‘Proposta de metodologia alternativa para controle de qualidade da aplicação estrutural do concreto projetado reforçado com fibras de aço’. Dissertação de Mestrado, São Paulo 2017 28. Kim, J., Lee, G.P., Moon, D.Y.: Evaluation of mechanical properties of steel-fibre reinforced concrete exposed to high temperatures by double-punch test. Constr. Build. Mater. 79, 182– 191 (2015) 29. Agra, R.R., et al.: ‘Avaliação dos efeitos do fogo na resistencia à tração residual do concreto reforçado com fibras de aço por meio do ensaio DEWS (Double Edge Wedge Splitting). In: 5th Iberian-Latin-American Congress On Fire Safety, Porto 2019 30. Serafini, R., et al.: Double edge wedge splitting test to characterize the post-cracking design parameters of fiber reinforced concrete subjected to high temperatures. J. Mater. Civ. Eng. (forthcoming). https://doi.org/10.1061/(ASCE)MT.1943-5533.0003701

Influence of the Curing Temperatures on the Mechanical Properties of Hemp FibreReinforced Alkali-Activated Mortars Bojan Poletanovic(&), Gergely Nemeth, and Ildiko Merta Institute of Material Technology, Building Physics, and Building Ecology, Faculty of Civil Engineering, TU Wien, Vienna, Austria [email protected]

Abstract. The aim of this research was to investigate how the curing temperature influences the mechanical properties of fibre reinforced alkali activated matrices. The mortar matrix contained fly ash as a binder, sodium water glass as an activator and sand with a particle size between 0.4 and 0.8 mm. As reinforcement hemp fibres of 10 mm in length and of volume ratio 1% of the composite were used. The prism mortars had dimension of 40  40  160 mm3 and were cured under three different temperatures: 20 °C, 60 °C and 80 °C. The results showed that after adding the fibres specimens cured at room temperature (20 °C) increased their compressive strength, flexural strength and energy absorption for 4%, 57% and 711% respectively. At the 60 °C and 80 °C curing temperatures the compressive strength of the specimens decreased for 8% and 27% respectively, whereas the flexural strength decreased for 14% in case of 60 °C curing temperature but remained the same after the curing temperature of 80 °C. The energy consumption of the specimens increased for 331% and 153% when specimens were cured at the 60 °C and 80 °C respectively. Keywords: Hemp fibres
Fly ash
Activated materials
Mortar
Mechanical properties

1 Introduction The demand for cement is increasing constantly in the last decades and will exponentially increase in the near future [1, 2]. The production of Portland Cement is one of the main causes for concrete’s vast negative footprint on the environment. Approximately one ton of Portland cement production may release up to one ton of carbon dioxide (CO2) in the atmosphere [3]. A possible way to reduce concrete’s footprint is to reduce or completely replace the cement within it. A very effective way with maintaining its physical and mechanical properties is to use pozzolanic materials, such as fly ash, ground granulated blast furnace slag, metakaolin, etc. [4–7]. When pozzolanic materials react with alkaline solutions they form a so-called alkali-activated material (AAM) [4]. These materials not only reduce the CO2 release in comparison to the traditional cement-based materials, but effectively re-use industrial by-products (such as fly ash coming from thermal power plant coal combustion) that are so far mainly © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 245–252, 2021. https://doi.org/10.1007/978-3-030-58482-5_22

246

B. Poletanovic et al.

landfilled. So far, AAMs are the most commonly researched alternative to cementbased materials. Similar to cement-based materials, AAMs have high compressive strengths, but very low resistance to crack propagation under flexural stresses and low energy absorption capacity. To improve these deficiencies the material should be reinforced with short, randomly dispersed fibres. The traditional fibres used for this purpose are steel, glass, synthetics, etc. [8–10]. Due to the growing environmental awareness natural fibres such as hemp, flax, coir, sisal, cotton, etc. have been considered as an environmentally friendly alternative to traditional fibres, since their mechanical characteristics are in the same range [11, 12]. Additionally, natural fibres are worldwide locally available, biodegradable, from renewable resources and need much less energy for production than traditional ones. Limited research work is published dealing with AAMs reinforced with natural fibres [13–18] and the existing ones focus solely on matrices based on alkali-activated pastes instead of mortars (with fine aggregates).

2 Materials and Experimental Methods The mixtures used for the mortar contain fly ash as a solid prime material, water glass as an alkali activator and sand. The fly ash is Micorsit20 produced by the NewChem company. All particles in the fly ash are smaller than 20 lm. The chemical composition of the fly ash is listed in Table 1. As the alkali activator water glass with the chemical composition of 16.72% Na2O, 25.08% SiO2 and 58.2% H2O by mass was used. To make a mortar, sand (with grain size 0.4-0.8 mm) produced by the Scherf GmbH & Co KG company was used. Table 1. Chemical composition of major oxides in fly ash. Material Oxides (wt%) SiO2 Al2O3 Fe2O3 CaO Fly ash 52 25 7 5

As fibre reinforcement short, randomly dispersed primary bast hemp fibres (Cannabis sativa L) in the range of micro fibres (diameter from 8–60 lm) were used. The morphology of a typical hemp fibre is presented in Fig. 1. The density of fibres was 1500 kg/m3 and according to the literature the tensile strength lies between 300– 1100 MPa, the Elastic modulus 23,5–90 GPa and the elongation at failure 1–3.5% [19]. The fibres consist of two cell walls and one inner part (the lumen). The primary wall’s role is to protect the secondary wall which is responsible for the tensile strength of the fibres [20]. Hemp fibres consist mainly of cellulose (74.4%), hemicellulose (17.9%), lignin (3.7%), pectin (0.9%) and wax (0.8%) [20]. The length of the fibres was 10 mm and the dosage 1vol.% of the total mixture. Generally, all plant-based fibres are hydrophilic, therefore the hemp fibres were added to the mixture in a water saturated dry surface condition.

Influence of the Curing Temperatures

247

Fig. 1. Morphology of hemp fibres

The fly ash to sand mass ratio was set to 1:3. The activator to fly ash ratio was defined that the mass of Na2O from the activator is equal to 10% of the fly ash mass. All mortar samples were prepared according to the European Standard EN 196-1 [21]. Six identical prism specimens were cast from each mortar group. The prism specimens had dimensions of 40  40  160 mm3. After the mixture is poured into the molds and left for one hour at laboratory conditions (20 °C temperature and 50% relative humidity), the specimens were covered with a plastic foil and cured under different conditions. Three different curing conditions were applied, i.e.: i) 28 days in a climate chamber (20 °C temperature and 65% relative humidity); ii) 24 h in an oven under the temperature of 60 °C and then additional 27 days in a climate chamber; iii) 24 h in an oven under the temperature of 80 °C and then additional 27 days in a climate chamber. All specimens were tested at the age of 28 days. Three-point bending test (3PBT) and compressive test were conducted and the mortars flexural strength, energy absorption capacity (toughness) under flexure and compressive strength were calculated. The tests were conducted on all specimen groups according to ÖNORM EN 1015-11 [22]. 3PBT was carried out on a mechanical testing machine Zwick/Roell Z250 with a load capacity of 200 kN, rigidity of 8  10−3 mm/kN under a deflection rate of 400 m/min at a room temperature of 21 °C and relative humidity of 50%. The load was applied at the middle of the specimens with a span length of 100 mm. The flexural strength was calculated from the maximal force from the force-span deflection curve of the 3PBT and the energy absorption as the area under the force-span deflection until the deflection of 6 mm, divided with a cross section of the specimens (Fig. 2). Flexural strength and energy absorption are calculated for six specimens of each group and the mean value (with standard deviation) is used as a representative. The compressive tests were conducted on three halves (of three different) specimens tested on 3PBT with applying the force on the area of 40  40 mm2 calculating the mean value with standard deviation as a representative.

248

B. Poletanovic et al.

Fig. 2. Force-displacement curve from a fibre reinforced mortar group which is cured at 80 °C

3 Results and Discussion The compressive strengths of non-reinforced and fibre reinforced mortar groups cured under different conditions are presented in Fig. 3. Generally, it can be seen that when cured at the higher temperature, the composites’ compressive strength increased. The reason for that is the fact that higher temperatures speed up the chemical reactions. At increased curing temperature the fly ash reacts faster with the water glass, forming more minerals and increasing the compressive strength of the composite. However, the faster chemical reactions and faster forming of a final structure of the mortar disturbs the uniform development of adhesion of fibres and matrix which has in turn a negative influence on the fibre reinforced mortars mechanical properties. With increasing curing temperature the strength difference between plain and fibre reinforced mortars became more pronounced.

Fig. 3. Compressive strength of the mortars

In case of cement-based mortars, the addition of fibres decreases the density of a plain mortar and consequently also its compression strength [12]. When cured at higher temperatures, fibre reinforced alkali-activated mortars resulted in lower compressive

Influence of the Curing Temperatures

249

strength than their corresponding plain mortar. The strength decreased for 8% and 27% in case of 60 °C and 80 °C curing temperature respectively. But contrary, when cured at room temperature (20 °C) the addition of fibres increased the plain mortar’s strength for 4%. The reason could be that the lower temperature provides moderate chemical reactions within the matrix, which resulted in the slower minerals formation and consequently in better fibre-matrix bond. Generally, fibres in the matrix bridge cracks and transfer the stresses which can increase the ultimate strength of the composite [14]. The flexural strength of mortars is presented in Fig. 4. The addition of short fibres in composites generally does not have a significant influence on their flexural strength since the flexural strength is mostly determined by the strength of the matrix itself. The fibres’s main role is to carry stresses after the failure of the matrix [12]. However, at lower curing temperature (20 °C) the flexural strength of the matrix increased for 57% after the addition of the fibres. The reason for the increase could be similarly as in case of the compressive strength, i.e. the moderate chemical reactions and mineral formation within the matrix provide better adhesion between fibres and the matrix. The better fibre-matrix bond contributes to better transfer of stresses by bridging the cracks. Under 60 °C curing temperature the flexural strength of mortars decreased for 14% when reinforced with hemp fibres whereas at 80 °C the flexural strength remained almost unchanged.

Fig. 4. Flexural strength of the mortars

The fibres main relevant contribution in a matrix is to increase the energy absorption capacity of the composite. When the maximal flexural strength is reached in a plain matrix a sharp drop occurs in the post peak region, indicating no deformationand energy absorption capacity of the material. On the other hand, when reinforced with fibres the composite prolongs its deformation capacity and is able to additionally carry a certain amount of stresses, increasing in such a way the overall energy absorption capacity of the composite. In the Fig. 5. the energy consumption of the non-reinforced and fibre reinforced mortars cured at different temperatures are shown. It can be seen that the most significant energy consumption improvement of the non-reinforced mortars is at room curing temperature (20 °C). The energy consumption capacity of the mortar increased

250

B. Poletanovic et al.

for 711%. Under 60 °C and 80 °C curing temperature the energy consumption increased for 331% and 153% respectively, when compared to their non-reinforced counterparts. The highest energy consumption increase (at 20 °C) could be due to the moderate matrix final forming, which results in an optimal attachment of fibres to the matrix and have a proper bond. The too rapid growth of the minerals could result in inappropriate fibres attachment to the matrix with a weaker fibre-matrix bond resulting in lower energy consumption capacity of the composite.

Fig. 5. Energy absorption of the mortars

4 Conclusions In the research plain and hemp fibre reinforced alkali activated mortars cured under three different temperature were examined. The prism mortars with the dimensions of 40  40  160 mm3 were prepared using fly ash as solid prime material, water glass as alkali activator, sand of particle size between 0.4 and 0.8 mm and short (10 mm), randomly dispersed hemp fibres. Regarding curing regime, three different temperature were used: 20 °C, 60 °C and 80 °C. The results showed that with reinforcing mortars with hemp fibres: • solely at curing temperature of 20 °C increased the composites compressive strength (for 4%). At curing temperatures of 60 °C and 80 °C it decreased for 8% and 27% respectively. • the flexural strength of the mortars increased for 57% at the curing temperature of 20 °C and decreased for 14% at 60 °C. At 80 °C the flexural strength remained almost unchanged. • all mortars increased their energy consumption capacity. The most significant increase (711%) was observed at the lowest curing temperature (20 °C), whereas at 60 °C and 80 °C a significantly lower increase of 331% and 153% respectively, was observed. Acknowledgements. The research is supported by the Stiftung Aktion Österreich-Ungarn in the frame of a bilateral research cooperation project Nr. 101ou10 and this activity has received

Influence of the Curing Temperatures

251

funding from the European Institute of Innovation and Technology (EIT), a body of the European Union, under the Horizon 2020, the EU Framework Program for Research and Innovation from the project EIT RM RIS-ALiCE: Al-rich industrial residues for mineral binders in ESEE region, no. 18258. The authors greatly acknowledge the support of the NewChem company for providing the fly ash for the research.

References 1. Cohen, B.: Urbanization, city growth, and the new united nations development agenda role in achieving the least cost path to a sustainable. Cornerstone 3(2), 4–7 (2015) 2. Activity Report 2017 of The European Cement Association CEMBUREAU, (2017) 3. Bilodeau, A., Malhotra, V.M.: High-volume fly ash system: concrete solution for sustainable development. Mater. J. 97, 41–48 (2000) 4. Pacheco-Torgal, F., Labrincha, J., Leonelli, C., Palomo, A., Chindaprasit, P.: Handbook of Alkali-Activated Cements. Woodhead Publishing, Mortars and Concretes (2014) 5. Provis, J.L., Palomo, A., Shi, C.: Advances in understanding alkali-activated materials. Cem. Concr. Res. 78, 110–125 (2015) 6. Provis, J.L.: Alkali-activated materials. Cem. Concr. Res. 114, 40–48 (2018) 7. Poletanovic, B., Kopecsko, K., Merta, I.: Durability of Hemp Fibre Reinforced Cementitious Mortars by Means of Fibre Protection and Cement Substitution with Metakaolin. In: International Conference on Interdisciplinary Approaches for Cement-based Materials and Structural Concrete, Madeira Island, Funchal, Portugal; 24.10.2018–27.10.2018; in: Cement-based Materials and Structural Concrete, p. 957–962 (2018) 8. Noushini, A., Hastings, M., Castel, A., Aslani, F.: Mechanical and flexural performance of synthetic fibre reinforced geopolymer concrete. Constr. Build. Mater. 186(20), 454–475 (2018) 9. Tung, T., Tran, T.M., Hong Hao, P.: Experimental and analytical investigation on flexural behaviour of ambient cured geopolymer concrete beams reinforced with steel fibers. Eng. Struct. 200(1), 109707 (2019) 10. Sathanandam, T., Awoyera, P.O., Vijayan, V., Sathishkumar, K.: Low carbon building: Experimental insight on the use of fly ash and glass fibre for making geopolymer concrete. Sustain. Envir. Res. 27(3), 146–153 (2017) 11. Merta, I., Mladenovič, A,. Turk, J., Šajna, A., Pranjić, M.: Life cycle assessment of natural fibre reinforced cementitious composites. In: 6th International Conference on NonTraditional Cement and Concrete, Brno, Czech Republic. 19.06.2017–22.06.2017 12. Merta, I., Šajna, A., Poletanović, B., Mladenović, A.: Influence of natural fibres on mechanical properties and durability of cementitious mortars. In: CoMS - 1st International Conference on Construction Materials for Sustainable Future, Zadar; 19.04.2017– 21.04.2017 13. Amalia, F., Akihaf, N., Nurfadilla, S.: Development of coconut trunk fiber geopolymer hybrid composite for structural engineering materials. In: Materials Science and Engineering (2017) 14. Alomayri, T., Low, I.M.: Synthesis and characterization of mechanical properties in cotton fiber-reinforced geopolymer composites. J. Asian Ceram. Soc. 1, 30–34 (2013) 15. Alomayri, T., Shaikh, F.U.A., Low, I.M.: Characterisation of cotton fibre-reinforced geopolymer composites. Compos. B 50, 1–6 (2013)

252

B. Poletanovic et al.

16. Chen, R., Ahmari, S., Zhang, L.: Utilization of sweet sorghum fiber to reinforce fly ashbased geopolymer. J. Mater. Sci. 49(6), 2548–2558 (2013). https://doi.org/10.1007/s10853013-7950-0 17. Sá Ribeiro, R.A., Sá Ribeiro, M.G., Sankar, K., Kriven, W.M.: Geopolymer-bamboo composite – A novel sustainable construction material. Constr. Build. Mater. 123, 501–507 (2016) 18. Assaedi, H., Shaikh, F.U.A., Low, I.M.: Characterizations of flax fabric reinforced nanoclaygeopolymer composites. Compos. B 95, 412–422 (2016) 19. Yan, L., Kasal, B., Huang, L.: A review of recent research on the use of cellulosic fibres, their fibre fabric reinforced cementitious, geo-polymer and polymer composites in civil engineering. Compos. B 92, 94–132 (2016) 20. Merta, I., Poletanovic, B., Kopecsko, K.: Durability of natural fibres within cement-based materials - review”, concrete structures. J. Hungarian Group fib (Federation International de Beton) 18, 10–16 (2017) 21. ÖNORM EN 196–1, Methods of Testing Cement - Part 1: Determination of Strength, 2016 10 15 22. ÖNORM EN 1015–11: 2018 01 01, Methods of test for mortar for masonry - Part 11: Determination of flexural and compressive strength of hardened mortar (2018)

Equivalence Between Flexural Toughness and Energy Absorption Capacity of FRC Sergio Carmona1(&) and Climent Molins2 1

Civil Engineering Department, Federico Santa María Technical University, Valparaíso, Chile [email protected] 2 Civil and Environmental Department, Barcelona Tech, Barcelona, Spain

Abstract. The increase of toughness and residual strengths are the great benefits of incorporating fibres in concrete. However, there is still no agreement regarding the most suitable experimental procedure for its determination and the most accepted and used tests, the bending tests, are complex to execute, require of sophisticated equipment with closed loop control and are unsuitable for the systematic quality control in works. Due to the latter, some authors have proposed equivalences between the different tests with the aim of simplifying the control. This paper presents the first stage of an experimental research in which a linear correlation between the flexural toughness determined by mean of the three-point bending test given in the European standard EN 14651 and the energy absorption capacity determined by square panel test according to EFNARC recommendation, was obtained, with differences less than 13% between the experimental results and predicted values. This achievement will allow to define a flexural equivalent resistance, based on the energy absorption capacity for controlling fibre reinforced concrete properties in the construction of tunnel linings. Keywords: Toughness panel test


Energy absorption capacity
Beam test
Squared

1 Introduction In recent years, fibres have begun to be widely used to reinforce concrete in the construction of tunnel linings in different Latin American countries, such as Bolivia, Chile, Colombia and Peru. However, a recurring problem, both for designers and contractors, is the characterization and control of fibre reinforced concrete, because most laboratories do not have the equipment and experience to perform the three-point bending test (3PB) according to EN-14651 standard [1], which requires a closed loop control (CLC) system for its execution. In contrast, due to its greater simplicity and stability when concrete cracks, many laboratories are able to perform the square panel test, as shown in Fig. 1, obtaining satisfactory results. Nevertheless, this test does not allow determining the residual strengths of fibre reinforced concrete, the parameters that, according to the MC 2010 © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 253–261, 2021. https://doi.org/10.1007/978-3-030-58482-5_23

254

S. Carmona and C. Molins

[2], characterize the post cracking resistance of FRC. Then arises the need to establish an equivalence between both tests, which will allow to control fibre reinforced concrete at works by mean of square panel test, and thus verify compliance with the project specifications. In the particular case of Fig. 1, all data was recorded manually, dial indicator measures were taken while testing and the load was applied by a manual jack.

Fig. 1 Square panel test performs in non-standard testing system in a laboratory in Bolivia.

In order to produce a robust relationship between the square panel and the 3PB tests, the state-of-the-art on comparing test methods for the mechanical characterization of fibre reinforced concrete was reviewed. Many authors establish correlations among standard tests based on the crack width [3, 4. 5, 6]. Nevertheless, the crack width in the square panel test is not measured as is in the 3PB test, and the panels exhibit many cracks in the final state, at the end of the test. Even though the failure mechanisms between both test might be different, this should not pose a problem to obtain a correlation between both tests. In fact, several codes and studies from the literature propose correlations between the results of test methods with completely different cracking mechanisms [7]. Using this idea, a code-type expression has been proposed in which the energy absorption capacity, determined by testing squared panels, and energy dissipated by cylinders subjected to double punching test or Barcelona test are correlated [8] and [9]. With that correlation, a good fit between the absorbed energy at a deflection of 25 mm in square panel with the dissipated energy at a total crack opening displacement of 6 mm. Taking in account those results, in this research, it was decided to correlate the energy measurement in both tests. However, in standardized 3PB test there is no definition of toughness at all. Then, a new definition of a toughness parameter was necessary for that. The aim of this paper is to present the results of the first stage of an experimental research, design to develop a correlation between the energy absorption capacity by the square panel test and flexural toughness obtained in 3PB bending test.

Equivalence Between Flexural Toughness and Energy Absorption

255

2 Square Panel Test According to the EFNARC recommendations [10], this test is conducted on a square panel of 600 mm side and 100 mm thick supported on its four edges, which is loaded at the centre by a contact surface of 100  100 mm, as can be seen in Fig. 2. Load is applied under deformation control at a rate of midspan deflection of 1.5 mm/min. This test is similar to the square panel test defined in EN 14488-5 standard [11]. The load – deflection curve shall be continuously recorded until a displacement of 25 mm is achieved at the centre point of the panel. Using this response, the energy absorption capacity, (d), can be calculated as: E25 ¼

25 Z

PðdÞdd

ð1Þ

0

where PðdÞ is the load as function of deflection.

Fig. 2 EFNARC square panel test setup [10].

3 Three-Point Bending Test (3PB) As is well known, according to EN-14651 standard [1], the characterization of the properties of the FRCs is carried out by testing a beam with a 25 mm deep central notch, loaded at the midspan (3PB), as shown in Fig. 3. The test should be carried out in a closed loop control system under crack mouth opening displacement (CMOD) control, at a rate of 0.05 mm/min, and with a sufficiently high rigidity of the loading system to avoid instability in the transition between the pre and post cracking stage. During test, the load and CMOD must be recorded continuously at a rate of 5 Hz. Using the P  CMOD response obtained during test, the EN-14651 standard defines the residual strengths, FR;j , as: FR;j ¼

3Pi l 2bh2sp

ð2Þ

256

S. Carmona and C. Molins

where l = 500 mm, b = 150 mm and hsp = 125 mm are the span, width and height of the specimen, respectively, shown in Fig. 2, y Pi , with i = 1, 2, 3, 4, is load at CMOD1 = 0.5 mm, CMOD2 = 1.5 mm, CMOD3 = 2.5 mm and CMOD4 = 3.5 mm, respectively. Based on the toughness definition given by ASTM C–1609 standard [12] and considering the linear relationship between the CMOD and the deflection of the beam, which is validated in the EN-14651 standard, in this research the flexural toughness has been defined, T3:5 , as the area under the curve P  CMOD up to CMOD4 = 3.5 mm, calculated as: T3:5 ¼

3:5 Z

PðCMODÞdCMOD

ð3Þ

0

This value of flexural toughness will be used in this research to establish an experimental correlation with the energy absorption capacity of the FRC.

Fig. 3. Test setup as EN – 14651 standard [1].

4 Experimental Research All the experimental research was developed in the Laboratory at Federico Santa Maria Technical University in Valparaiso, Chile. 4.1

Tested FRC

The concretes were prepared with a Chilean pozzolanic cement of Type IP [13] and crushed river sand; the mix proportions are presented in Table 1. Considering that Barchip 48 (BC – 48) is one of the most widely fibres used in Latin American project, in this research eight different contents of those fibres were used to reinforce concretes. The fibres features are given in Table 2 and are shown in Fig. 4.

Equivalence Between Flexural Toughness and Energy Absorption

257

Table 1. Features and properties of tested concretes. Material (kg/m3) Cement type IP Sand 0/10 Superplasticizer admixture Water reducer admixture Active admixture Free water Fiber content Beams Number of Square panel specimens Cylinders FRS properties Compressive strength, (MPa) Slump (mm) (%)

1.5 2.0 420 1955 2.10 2.10 2.94 156 1.5 2.0 10 10 3

2.5

3.0

FRS 4.0

6.0

8.0

12.0

2.5

3.0

4.0

6.0

8.0

12.0

45.3 255 0.16

39.9 245 0.27

43.0 250 0.33

46.7 235 0.44

44.5 210 0.66

43.7 220 0.88

41.2 180 1.32

42.1 260 0.22

The concretes were prepared at laboratory using a conventional paddle mixer of 200 litres’ capacity. For each FRC 600  600  100 mm square panels, standardized beams and three compression cylindrical specimens were cast. Fibre content, volumetric substitution (Vf ), slump and compressive strength (fc ) are given in Table 1. Table 2. Synthetic fibres properties (manufacturer’s data). fst df kf E Fibers/kg Designation lf (mm) (mm) lf/df (MPa) (GPa) (N°) BC – 48 48 0.70* 68.6 640 12 59500 (*) Equivalent diameter determined with the manufacturer’s information.

Fig. 4. View of fibres Barchip – 48 used in this research.

258

4.2

S. Carmona and C. Molins

Tests and Results

All the tests were conducted in a hydraulic closed-loop control system of 100 kN capacity, under deformation control. In the panel tests, the deflection was measured with a LVDT of 50 mm range, placed in the centre bottom face of the specimen. The 3PB tests were conducted according to standard EN–14651 under CMOD control, which was measured using a clip-gauge extensometer with a total range of 5.0 mm, at a rate of 0.05 mm/min until reaching a CMOD = 0.1 mm. Then, the speed was increased to 0.2 mm/min until the end of the test. Before to the bending tests, a 25 mm deep notch was cut at midspan of each beam. In both tests, the load and the deformation were recorded continuously by the testing system at a rate of 5 data/s. The mean curves obtained with each tested concrete are shown in Fig. 5, in which can be clearly seen the effect of the fibre content in the post peak response of FRC. As can be observed in Fig. 5a, for concretes reinforced with low fibres content, i.e. less and equal to 4.0 kg/m3, the curves exhibit two peaks related with flexural cracks. The loads corresponding to the first and second cracks do not seem to depend on the amount of reinforcing fibre. After the second peak, the curves show a softening behaviour, with a decreasing slop as fibre content increases. However, in the panels reinforced with medium and high amounts (FRS–6.0, FRS–8.0 and FRS–12.0), a flexural crack was initially open, which caused the first peaks that can be observed in the P  d curves. However, due to the higher fibres content, the testing system had to increase the applied load to maintain the established deformation rate, which in addition to the friction force developed in the supported, gave rise to a punching failure, which is reflected in the formation of cracks around the loaded section. At the same time, in the P – CMOD curves show in Fig. 5b, it can be seen that the first peak does not depend on fibre content, nevertheless the post peak behaviour varies from a deep softening for concrete reinforced with low fibre content to strong hardening for high fibre content.

(a)

(b)

Fig. 5. Mean curves load–deformation obtained with (a) square panel tests, and (b) 3PB tests.

Equivalence Between Flexural Toughness and Energy Absorption

259

5 Experimental Correlation Proposed With the results of panel and beam tests, the energy absorption capacity and flexural toughness were calculated using the Eqs. (1) and (3), respectively. The diagrams showing the energy absorption–deflection and flexural toughness–CMOD are shown in Fig. 6.

(a)

(b)

Fig. 6. (a) Curves energy absorption – deflection of panels and (b) flexural toughness obtained with 3PB tests.

In order to obtain an experimental correlation based on standardized parameters that show the FRC behaviour in advanced cracking states, the values of energy absorption capacity of square panels at a deflection of 25 mm, denoted as E25 , and the flexural toughness of beams at a CMOD = 3.5 mm, denoted as T3:5 , were used. The values of E25 and T3:5 are summarized in Table 3, along with the coefficient of variation in bracket. As can be seen, the results scatter trends to decrease when fibres content increases and, in all cases, the scatter is larger in the 3PB tests. Table 3. Mean results of conducted tests. Concrete FRS FRS FRS FRS FRS FRS FRS FRS

– – – – – – – –

Fibre content (kg/m3) 1.5 1.5 2.0 2.0 2.5 2.5 3.0 3.0 4.0 4.0 6.0 6.0 8.0 8.0 12.0 12.0

E25 (J) 400 (18.6) 482 (15.4) 553 (16.6) 637 (15.6) 669 (14.8) 859 (13.3) 1098 (9.5) 1744 (6.1)

T3:5 (J) 9.3 12.3 18.3 19.4 25.2 39.5 48.7 83.8

(19.0) (21.9) (22.5) (20.5) (17.0) (16.0) (15.1) (14.3)

Equation (4) Difference (J) (%) 8.5 −5.3 13.1 6.9 17.1 −6.6 21.8 12.7 23.6 −6.1 34.3 1.3 47.7 −2.1 83.9 0.3

260

S. Carmona and C. Molins

The values of E25 and T3:5 are plotted in Fig. 7, where it can be seen there is a linear relationship between both parameters, given by the expression: T3:5 ðE25 Þ ¼ 0:0561  E25  13:92

ð4Þ

where T3:5 ðE25 Þ is the predicted flexural toughness as function of energy absorption capacity, E25 . This expression fits well with the experimental data, with a correlation coefficient r 2 ¼ 0:9972, and differences between experimental and predicted values less than 12.7%, as can be seen in the Table 3.

Fig. 7. Linear correlation between E25 and T3:5 obtained for BC - 48.

The high value of the correlation coefficient ðr 2 Þ of this linear relationship shows that there is a direct proportionality between the flexural toughness at a CMOD = 3.5 mm, equivalent to a beam deflection of 3.02 mm [1], and the energy absorption capacity determined at a deflection d = 25 mm, for a wide range of fibre content, which will allow establishing a correlation between equivalent strengths determined using the flexural toughness and the energy absorbed by the square panel.

6 Conclusions The conclusions of this research are the following: • In recent years, the use of fibre reinforced shotcrete has replaced steel mesh reinforced shotcrete in tunnel linings construction in Latin America. However, the lack of laboratories to characterize and control of fibre reinforced concrete properties is a very common limitation. • To overcome that limitation, this paper proposes to establish an empirical equivalence between the square panel test and the notched beam test given in standard EN 14651, since the former is easier to perform with simple laboratory equipment. Then, a flexural toughness has been defined as the area under the load-CMOD curve, until a CMOD = 3.5 mm.

Equivalence Between Flexural Toughness and Energy Absorption

261

• To work out the correlation, an experimental research was carried out using concretes reinforced with different synthetic fibre contents, obtaining for this particular fibre, a linear relationship between the energy absorption capacity and the flexural toughness, with differences less than 13% between the experimental results and values predicted using the proposed linear correlation. • This achievement will allow to define a flexural equivalent resistance, based on the energy absorption capacity for controlling fibre reinforced concrete properties in the construction of tunnel linings. Acknowledgments. This research was supported by Fondecyt Project “Use of the Generalized Barcelona Test for Characterization and Quality Control of Fibre Reinforced Shotcretes in Underground Mining Works”, N°1150881.

References 1. European Committee for Standardization, EN 14651: Test method for metallic fibre concrete–measuring the flexural tensile strength (Limit of Proportionality (LOP), Residual) (2007) 2. CEB-FIP: Model code – first complete draft. FIB Bull. 55, 1–318 (2010) 3. Alberti, M.G., Enfedaque, A., Gálvez, J.C.: On the mechanical properties and fracture behavior of polyolefinfiber-reinforced self-compacting concrete. Const. Build. Mater. 55, 274–288 (2014) 4. Conforti, A., Minelli, F., Plizzari, G.A., Tiberti, G.: Comparing test methods for the mechanical characterization of fiber reinforced concrete. Struct. Concr. 19, 656–669 (2018). https://doi.org/10.1002/suco.201700057 5. Carmona, S., Molins, C., Aguado, A.: Correlation between bending test and Barcelona tests to determine FRC properties. Const. Build. Mater. 181, 673–686 (2018) 6. Carmona, S., Molins, C.: Use of BCN test for controlling tension capacity of fibre reinforced shotcrete in mining works. Const. Build. Mater. 198, 399–410 (2019) 7. Galeote, E., Blanco, A., Cavalaro, S.H.P., De la Fuente, A.: Correlation between the Barcelona test and the bending test in fibre reinforced concrete. Const. Build. Mater. 152, 529–538 (2017) 8. Carmona, S., Molins, C.: Application of BCN test for controlling fibre reinforced shotcrete in tunnelling works in Chile. IOP Conf. Ser. Mater. Sci. Eng. 246, 012010 (2017) 9. Carmona, S., Molins, C., García, S.: Application of Barcelona test for controlling energy absorption capacity of FRS in underground mining works. Const. Build. Mater. 246, 118458 (2020). https://doi.org/10.1016/j.conbuildmat.2020.118458 10. European Federation of national associations of specialist contractors and material suppliers for the construction industry, European Specification for Sprayed Concrete (1996) 11. European Committee for Standardization: EN 14488-5:2006 Testing sprayed concrete. determination of energy absorption capacity of fibre reinforced slab specimens (2006) 12. ASTM ‘C1609/C1609M-19: Standard test method for flexural performance of fibrereinforced concrete (Using Beam With Third-Point Loading)’, ASTM International, West Conshohocken, PA (2019). www.astm.org 13. ASTM: C595/C595M-18, Standard specification for blended hydraulic cements. ASTM International, West Conshohocken, PA (2018). www.astm.org

Alkali Resistant (AR) Glass Fibre Influence on Glass Fibre Reinforced Concrete (GRC) Flexural Properties S. Guzlena(&) and G. Sakale Institute of Technical Physics, Faculty of Material Science and Applied Chemistry, Riga Technical University, P. Valdena 7, Riga, Latvia [email protected]

Abstract. Glass fibre reinforced concrete (GRC) is lightweight material mostly used for façade panels and decorative elements. GRC can be made using two methods – spraying and premixing. Glass fibre in both cases has main influence on material flexural properties and ductility. Historically ordinary E type glass fibre has been used, but during concrete aging and alkaline medium fibres become fragile (weight and diameter loses). New type, alkali resistant (AR) glass fibres have been developed. In this research AR glass fibre amount and length influence on GRC flexural properties is investigated. Fibre length was changed from 6 mm till 41 mm for different samples and cut during spraying process. Fibre amount was changed from 0–7%. Samples were analysed using SEM-EDX to evaluate AR glass fibre and concrete matrix bond. GRC mechanical properties was evaluated using four-point bending tests and characterised by level of proportionality (LOP) and modulus of rupture (MOR). . Keywords: GRC


AR glass fibre
Flexural strength

1 Introduction Concrete is most widely used engineering material in construction all over the world due to its strength, durability and low cost as compared to other construction materials. The major drawback of concrete is its low tensile strength. Crack appearance due to stress during loading reduces concrete strength, durability and makes concrete more vulnerable to deleterious outside environment [1]. To increase tensile strength different fibres, like organic, polymer, basalt, steel and glass fibres can be incorporated in matrix. Glass fibre reinforced concrete (GRC) is a composite which consist of concrete and glass fibres. Glass fibre has low water absorption, high modulus of elasticity and tensile strength [2]. Fenu et al. have even concluded that basalt fibre reinforcement was shown to be less performing than glass-fibre reinforcement when used under dynamic conditions [3]. This material is used for non-structural parts of buildings like facade panels for over than 30 years. GRC is known for its high tensile and impact strength which is due to glass fibres. These properties allow make GRC panels 12 mm thin [4–6]. There are two types of techniques used to produce GRC elements. First of them is premixing, © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 262–269, 2021. https://doi.org/10.1007/978-3-030-58482-5_24

Alkali Resistant (AR) Glass Fibre Influence on Glass Fibre Reinforced Concrete (GRC)

263

when concrete is mixed with chopped glass fibres and then casted in formwork. Fibre length usually is around 12 mm and it is added 3% of mortar mass. If the fibre length and amount is increased workability of the mix is reduced. Barluenga et al. have noted that using short fibres in premix the cracked area of concrete surface can be reduced. If crack appears perpendicular to fibre it limits the size of the crack, but if the crack appears parallel to fibre, reinforcement does not work and crack size grows freely [7]. For spray technique fibre length can be changed, but usually it is 25–40 mm and 5% by mortar weight. During spraying process uniform well, mixed composite is achieved comparing to premix method. This type of material is with high durability, impact and tensile strength [5, 8]. Glass fibres used in material must be alkali resistant due to concretes high alkalinity. In the past, ordinary glass fibres, like E glass fibres were used, but during the time they showed low durability. Fibres lose their durability due to hydration reaction in concrete. Portlandite (Ca(OH)2) is produced during hydration process and increases alkalinity in cementitious matrix to high alkalinity (pH 12). Hydroxyl ions (OH-), which is at high concentration in cementitious matrix due to alkaline pore solution, brakes Si-O-Si bonds in glass fibres Eq. 1 [9, 10] and causes weight and diameter loses making fibres fragile [6, 8, 9, 11]. in solution

ð1Þ

To improve glass fibre chemical stability in concrete ZrO2 have been added. In the EN 15422:2008 is mentioned AR glass fibre must contain minimum 16% of ZrO2, to use it in concrete. But ZrO2 addition do not solve his problem completely, as researches showed decrease in aged GRC flexural properties comparing to young GRC [12–14]. There exist different methods to reduce fibre embrittlement. Admixtures like fume, metakaolin, nano silica can be added to a concrete mixture. This permits pozzolanic reactions and transforms portlandite to C-S-H in this way decreasing alkalinity of the concrete [2, 6, 9]. Use of low alkalinity cement, like sulphoaluminate cement could solve the problem, because Portlandite (Ca(OH)2) is not produced during hydration process, but pH value in pore solution still is high, around pH 10. To densify the interface between fibers and cement matrix with polymers, like PVA, AC and others. This technique reduces lime diffusion into fibres [9, 10]. Due to thickness and common application of GRC tensile test is used to evaluate GRC properties. Level of proportionality (LOP) and modulus of rupture (MOR) is used to describe sample tensile strength. LOP value describes sample maximal linear elastic deformation, which means every load that is put below LOP will not harm material, no cracks will be seen on surface and material will return to its present state. Sample elastic region, below LOP, can be described by Hookes law stress is proportional to the strain and the slope is Young’s modulus. Basically, LOP value demonstrates matrix maximum flexural strength. After load is increased over LOP value material exhibits plastic behaviour, it deforms irreversibly. In this moment multiple cracks can appear on GRC surface, but

264

S. Guzlena and G. Sakale

material still have load bearing properties. Increasing load elongation increases of GRC sample due to fibres connection in concrete matrix until crack is too large and fibres pull out or brake at fracture point. MOR values describe samples ultimate strength before failure [2, 8, 13]. LOP and MOR values are influenced by many factors as: • • • •

fibre amount, fibre length, sprayed GRC or premix is used, compaction.

In this research impact of amount of glass fibres in composition, length of fibre on sprayed GRC properties have been evaluated.

2 Materials and Methods GRC samples were made using spraying method. Basic components of GRC was cement CEM I 52,5R, fine quartz sand, superplasticizer and acrylic polymer. AR-glass fibre roving was used. Fibre length from 6 mm till 41 mm was changed for different samples and cut during spraying process. Fibre amount was changed from 0–7% by weight of GRC material (concrete matrix) in the uncured, green state. Sample flexural test was carried according to BS EN 1170 [15]. Samples were cut from test board in size 275  50 mm and tested with 4-point bending test after 28 days. GRC samples and AR glass fibres were investigated using Phenom ProX Desktop SEM.

3 Results 3.1

SEM – EDX Analysis

For all tests have been used AR glass fibre roving which was cut during manufacturing process. In SEM image (Fig. 1) can be seen bundle of AR glass fibres. Fibre surface is smooth, and the shape is regular. EDX analysis was done on glass fibre fracture zone, marked in Fig. 1. In Table 1 is shown element analysis. Base oxides in fibre is Si, Na and Zr. With EDX analysis is not possible to obtain precise amount of ZrO2 amount in fibre, but this method can be used to indicate Zr in fibre and to know that this is AR glass fibre. In Fig. 2 calcium carbonate crystals [16] are grown on surface of AR glass fibre which means that glass fibre surface has good bonding abilities with concrete matrix. Fenu et al. have also noted that fibres in fracture are almost completely covered by a thin surface layer formed by the products of reaction between the outer alkali resistant glass of the fibre and the cementitious matrix [3]. In Fig. 3 is shown fibre bundles incorporated in GRC matrix. GRC performance in strength depends on mortar bonding with fibre filaments, fibre bundles and concrete matrix strength [8]. As can be concluded from the SEM image fibre bundles have good incorporation in matrix, no air voids or cracks can be seen around interface.

Alkali Resistant (AR) Glass Fibre Influence on Glass Fibre Reinforced Concrete (GRC)

265

Fig. 1. SEM image of AR-glass fibre used in GRC samples. EDX analysis was done at marked location. Table 1. AR-glass fibre element analysis Element symbol Weight concentration, % O 46.92 Si 25.32 Na 12.72 Zr 9.38 Ti 2.91 C 2.75

Fig. 2. AR-glass fibre in GRC sample

266

S. Guzlena and G. Sakale

Fig. 3. AR-glass fibre bundles in GRC sample cross section

3.2

Mechanical Properties

Concrete as itself is brittle material in tensile strength. Fibre addition increase material elastoplastic behaviour. In this section GRC property dependence from fibre amount and length are presented. AR glass fibre addition to concrete matrix has significant impact on sample ultimate load bearing capacity. As shown in Fig. 4, if 41 mm long fibres are used MOR values are increased with increase of fibre amount in the concrete, meanwhile LOP values do not change significantly. LOP value is more affected by the strength of concrete matrix than fibre addition [2]. Concrete as material is nonelastic, brittle material, as shown in Fig. 4, LOP and MOR values are almost the same if no AR glass fibres are added to concrete matrix. In Fig. 5 is shown that test samples without fibres are with very low plasticity and brittle fracture point. When the amount of fibres is increased from 0% to 5% and 7%, material elasticity is increased. By increasing fibre amount from 5% to 7% increases the loadbearing capacity of the material, but as the fibre length has not change the deflection of material does not change. Fibre length was changed from 6 mm to 41 mm, but amount during these experiments was 5%. In Fig. 6 is shown LOP/MOR dependence from fibre length. Fibre length do not considerably change LOP values. As we concluded before the LOP values are more dependent on the matrix properties of the concrete. MOR values increases by increasing fibre length from 6 mm to 41 mm. As shown in Fig. 7 by adding longer fibres plastic behaviour of material increases. H. Kasagani et al. have also noted that samples with longer fibres have higher deformation capacity comparing to samples with shorter fibres [17]. Longer fibres have higher chance to defom in the bundle structure improving post crack deformation and

Alkali Resistant (AR) Glass Fibre Influence on Glass Fibre Reinforced Concrete (GRC)

267

35

Strenght, MPa

30 25 20 15 10 5 0 0% LOP (7d)

5% MOR (7d)

LOP (28d)

7% MOR (28d)

Fig. 4. Influence of fibre amount on LOP/MOR values

Fig. 5. Influence of fibre amount on deflection

still keep increment of force applied to GRC. Analysing collapse of samples can be seen that using longer fibres deflection stays constant for higher ΔForce. The filaments slowly pulls out of the bundle, not brittle when applied under load, crack size (deflection) do not change.

268

S. Guzlena and G. Sakale

35

Strenght, MPa

30 25 20 15 10 5 0 6 mm LOP (7d)

12.5 mm MOR (7d)

LOP (28d)

41 mm MOR (28d)

Fig. 6. Influence of fibre length on LOP/MOR values

Fig. 7. Influence of fibre length on deflection of material

4 Conclusions Using AR-glass fibres is possible to change GRC flexural properties: • by increasing fibre amount from 5% to 7% increase the load-bearing capacity of the material, but as the fibre length of all samples have been the same (41 mm) the deflection of material does not change. • LOP value, which describes material elastic deformation, is dominated by concrete matrix flexural strength, addition of fibres has inessential effect.

Alkali Resistant (AR) Glass Fibre Influence on Glass Fibre Reinforced Concrete (GRC)

269

• by increasing fibre length from 10 mm till 41 mm, MOR values increases considerably due to composite plastic behaviour. Acknowledgements. Authors of the article are grateful to LTD Skonto Concrete Cladding for a materials and opportunity to use spraying equipment.

References 1. Khaliq, W., Ehsan, M.B.: Crack healing in concrete using various bio influenced self-healing techniques. Constr. Build. Mater. 102, 349–357 (2016) 2. Madhkhan, M., Katirai, R.: Effect of pozzolanic materials on mechanical properties and aging of glass fiber reinforced concrete. Constr. Build. Mater. 225, 146–158 (2019) 3. Fenu, L., Forni, D., Cadoni, E.: Dynamic behaviour of cement mortars reinforced with glass and basalt fibres. Compos. Part B Eng. 92, 142–150 (2016) 4. Branco, F.A., Ferreira, J., Brito, J.D.E., Santos, J.R.: Building structures with GRC Fernando A. Branco, João Ferreira, Jorge De Brito and José R. Santos, no. April, pp. 1–11 (2001) 5. Ferreira, J.G., Branco, F.A.: GRC mechanical properties for structural applications, vol. 1, no. 1, pp. 1–20 6. Correia, J.R., Ferreira, J., Branco, F.A.: A rehabilitation study of sandwich GRC facade panels. Constr. Build. Mater. 20(8), 554–561 (2006) 7. Barluenga, G., Hernández-Olivares, F.: Cracking control of concretes modified with short AR-glass fibers at early age. Experimental results on standard concrete and SCC. Cem. Concr. Res. 37(12), 1624–1638 (2007) 8. Bartos, P.J.M.: Glassfibre reinforced concrete: a review. IOP Conf. Ser. Mater. Sci. Eng. 246 (1) (2017) 9. Arabi, N., Molez, L., Rangeard, D.: Durability of alkali-resistant glass fibers reinforced cement composite: microstructural observations of degradation. Period. Polytech. Civ. Eng. 62(3), 1–7 (2018). SE-Research Article 10. Scheffler, C., et al.: Interphase modification of alkali-resistant glass fibres and carbon fibres for textile reinforced concrete I: fibre properties and durability. Compos. Sci. Technol. 69(3– 4), 531–538 (2009) 11. Genovés, V., Gosálbez, J., Miralles, R., Bonilla, M., Payá, J.: Ultrasonic characterization of GRC with high percentage of fly ash substitution. Ultrasonics 60, 88–95 (2015) 12. Moceikis, R., Kičaite, A., Keturakis, E.: Workability of glass reinforced concrete (GRC) with granite and silica sand aggregates. IOP Conf. Ser. Mater. Sci. Eng. 251(1) (2017) 13. Enfedaque, A., Cendón, D., Gálvez, F., Sánchez-Gálvez, V.: Analysis of glass fiber reinforced cement (GRC) fracture surfaces. Constr. Build. Mater. 24(7), 1302–1308 (2010) 14. Holubova, B.: Corrosion of Glass Fibres in Ultra High Performance Concrete and Normal Strength Concrete. Ceram. Silikaty 61(4), 1–9 (2017) 15. “EN 1170-5.pdf.” 16. Davies, R., et al.: Multi-scale cementitious self-healing systems and their application in concrete structures. In: 9th International Concrete Conference 2016: Environment, Efficiency and Economic Challenges for Concrete, 4–6 July 2016 (2016) 17. Kasagani, H., Rao, C.B.K.: Effect of graded fibers on stress strain behaviour of Glass Fiber Reinforced Concrete in tension. Constr. Build. Mater. 183, 592–604 (2018)

Fiber Reinforced Concrete Crack Opening Evaluation Using Digital Image Correlation Techniques Kaio Cézar da Silva Oliveira(&), Gabriela Silva Dias, Isadora Queiroz Freire de Carvalho, Wandersson Bruno Alcides de Morais Silva, Danilo José Pereira Freitas, Christiano Augusto Ferrario Várady Filho, and Aline da Silva Ramos Barboza Technology Center, Structure and Materials Laboratory, Federal University of Alagoas - UFAL, Maceió, Brazil [email protected] Abstract. The analysis of mechanical properties in fiber reinforced concrete (FRC) elements is basically done through destructive tests since the results obtained by these methods are already well established in normative codes. One of them is the 3-point flexural test normalized by EN 14651 using a notched beam to measure the crack width (CMOD). Within the context of the mechanical properties evaluation that do not require the production or extraction of specimens, and can be applied in fully functioning structures, the digital image correlation (DIC) is a technique which has been proposed. This non-destructive test analyzes a group of images correlating one with each other, evaluating the changes that occurred during the load has been applied. It is a non-invasive test, capable of results with acceptable precision and a considerable low cost, proving to be a promising technique in the field of behavior analysis. Thus, this study compares the cracking results obtained through the extensometry technique (Linear Variable Differential Transformer – LVDT) and the digital image correlation. The samples were made using steel fibers. The results obtained using the DIC technique were validated by the data obtained through the LVDTs, and the absolute error was considerably low. Keywords: Concrete


Correlation
Fiber
Image

1 Introduction Concrete is one of the most used materials in civil construction, however, when submitted to tensile efforts, it shows a low resistance and deformation capacity. In view of the problem presented, the addition of staple fibers (steel, glass, polypropylene and natural) represents an improvement alternative for such behavior, since the main contribution of the fibers occurs in the post-cracking stage. The fibers act as stress transfer bridges between the cracks, redistributing the efforts in the cementitious matrix, and ensuring that even after the cracking the composite presents resistive capacity, restricting a sudden rupture [1]. © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 270–278, 2021. https://doi.org/10.1007/978-3-030-58482-5_25

Fiber Reinforced Concrete Crack Opening Evaluation

271

In order to expand the use of Fiber Reinforced Concrete (FRC) in elements for structural purposes, a working group called Special Activity Group (SAG 5) of the Federation Internationale du Beton (FIB), included in the fib Model Code 2010 (2013) sections addressing material capacity and recommendations for sizing and behavior analysis. Configuring itself as an important step for the SFRC, since it defined the minimum performance requirements in view of the safety limit states Ultimate Limit State (ELU) and Service Limit State (ELS). Although not a normative character, the fib Model Code 2010 (2013) is a document that aims to subsidize future regulations on the SFRC [2]. The fib Model Code 2010 (2013) establishes the test standardized by EN 14651 (2007) for the characterization of the FRC for structural purposes. It consists of a beam test under three-point bending carried out in a closed system with displacement speed control, and through it, it is possible to determine the post-crack strength, the SFRC toughness and the crack opening induced by a notch at the lower end of the specimen [3]. The code states that alternative tests can be used to determine the same mechanical performance parameters if the results found are correlated with the reference test. From this context follows the proposal of the present work, which aims to use the correlation of digital images. The behavior assessment method using digital image correlation is suitable as a Non-Destructive Test (NDT), since there is no need to impose any type of deformation or destructive process on the structure, it is interesting to use the technique for structures already in full operation, being a technique focused on the monitoring of structural health [4–5].

2 Digital Image Correlation DIC is a technique that allows studying qualitatively, and quantitatively, the mechanical behaviour of materials when submitted to different types of requests. Initially proposed by Sutton et al. (1983), the technique is based on the comparison of pairs of digital images, captured at different times of the loading application. Each of the images is composed of pixels, and each pixel has a grayscale value, which varies from 0 to 255, according to the amount of light reflected by the surface of the specimen (Fig. 1).

Fig. 1. Grayscale intensity matrix and its digital image correlation.

272

K. C. da Silva Oliveira et al.

As there may be pixels with the same gray intensity value, a region or set of pixels, called a subset, is defined and used for the correlation method. After defining the subset in the image captured before the deformation, a computational tool is used to identify the post-deformation image, for that same set of pixels. The result obtained for the central part of the subset is the average of the displacements suffered by each of the pixels that compose it (Fig. 2) [6–11].

Fig. 2. Displacement of a subset during specimen deformation.

A displacement vector is generated for each subset, and at the end of the analysis of all defined subsets, a displacement field results. It is important to highlight that the unique pattern of subsets is guaranteed when the region studied has no repetitions, isotropic and has high contrast patterns. Obtaining significant results through the method is directly related to the lighting conditions and standardization of the surface captured by images [12].

3 Dic and Crack Opening Evaluation One of the main parameters of the behaviour of steel fiber reinforced concrete (SFRC) obtained through the test recommended by EN14651 (2007), is the opening of the notch in the lower face of the beam, located in the middle of the span, called CMOD. This parameter is recorded during the experiment using mechanical equipment such as clip-gage or displacement transducer. For the application of digital image correlation, the displacement values were recorded in real-time through a library of the ITOM ® software, which was created and has been updated by researchers from the Mechanical Research Group on Advanced Structures and Materials from Federal University of Alagoas, to assist in the process of acquiring the data generated by the LVDTs (Linear Variable Differential

Fiber Reinforced Concrete Crack Opening Evaluation

273

Transformers), shown in Fig. 3, and to automate the capture of the images in a certain time interval. Still, with the help of this library, image processing, crack opening analysis, deformation and displacement fields are carried out.

Fig. 3. LVDT positioning details.

The LVDTs were connected to a system for the acquisition of analog signals, the Spider 8®. The test apparatus is shown in Fig. 4. The choice of blue LEDs for lighting the specimen was made from previous tests, which proved that this color allowed greater contrast for the captured surface. The camera used for the tests has a CMOS sensor with a resolution of 24.5 MP. To capture the images, it was necessary to carry out a camera calibration process, with which a conversion factor from pixels to millimetres is found. This conversion factor is obtained with the aid of an ITOM® tool, which receives the distance between the center of two pixels (pixel pitch) as input data. Then, about 10 (ten) photographs are taken of a checkered mesh generated on the computer screen, as shown in Fig. 5, without changing the focus or zoom of the camera, nor the positioning of the tripod, and they are transferred to the test computer and loaded into ITOM® for a calibration whose reference was pixel pitch. In this calibration, ITOM® determines the conversion factor. The image analysis process starts with reading at time t = 0 s, using ITOM®. For this reading, the software uses a tool called linecut and generates a graph (Fig. 6), in which the X-axis refers to the number of pixels contained in the linecut and the Y-axis to the intensity of gray tones (luminosity data). To avoid measurement errors, the linecut is standardized for all test images, aligned to the LVDTs. Analyzing the graph generated on the Fig. 6, it is noticed that when the line cuts a region with shades of gray that tend to black, the Y values fall dramatically. In turn, when the linecut cuts pixels with tones closer to white, the Y values increase. With this, when analyzing the graph of Fig. 6, the crack opening will be given by the number of pixels between the red and green points called width, which indicates the beginning

274

K. C. da Silva Oliveira et al.

Fig. 4. Test apparatus.

Fig. 5. Checkerboard used in camera calibration.

Fiber Reinforced Concrete Crack Opening Evaluation

275

Fig. 6. Detail of crack opening obtained through ITOM.

and end of the crack, respectively. With the crack opening values (width), conversion to millimeters is performed, multiplying the values obtained by the conversion factor described above.

4 Results To validate the image correlation process in obtaining the crack opening of beams tested under three-point bending, a comparison of the values obtained by each method was performed and then a percentage difference between the opening read by the LVDT and the opening was calculated. Starting with the analysis of the results, it was found that they were satisfactory since the average percentage difference found between the values obtained through the LVDT’s and the images was approximately 7.5%. It is worth mentioning that in previous studies the percentage differences were around 35%. The average values obtained demonstrate the method’s ability to obtain crack opening values, with most of the values fluctuating around 5%. Tables 1, 2, 3 and 4 shows the comparative averages of the crack opening values obtained through the

276

K. C. da Silva Oliveira et al.

LVDTs and the images. The general average of variation between the two methods is also presented, which in this case was approximately 7.4%. Table 1. Beam 1 and 2 results. TIME (s) 1300 1400 1500 1600 1700 1800 1900 2000 2100 2200 2300 2400 2500 2600

BEAM 1 LVDT (10−3 m) 1.7655 2.1242 2.4982 2.8654 3.2325 3.5963 3.9628 4.3145 4.6641 5.0219 5.393 5.7322 6.0772 6.4204

DIC (10−3 m) 1.9439 2.2679 2.6567 2.9159 3.1103 3.4342 3.7582 3.8878 4.3414 4.6006 4.8598 5.1838 5.3134 5.6373

Percent Difference (%) −10.1048 −6.7635 −6.3439 −1.7596 3.7801 4.5051 5.1631 9.8887 6.9183 8.3892 9.8874 9.5681 12.5686 12.1964

BEAM 2 LVDT (10−3 m) 2.6933 3.0385 3.3871 3.7542 4.1119 4.4636 4.8221 5.1846 5.5285 5.8698 6.197 6.5228 6.8545 7.1893

DIC (10−3 m) 2.6567 2.9807 3.3047 3.6286 3.823 4.147 4.5358 4.9246 5.1838 5.443 5.7021 5.9613 6.2853 6.4797

Percent Difference (%) 1.3612 1.9048 2.434 3.344 7.0253 7.0936 5.9369 5.0157 6.2362 7.2724 7.9861 8.6073 8.3041 9.8696

Table 2. Beam 1 and 2 results. BEAM 1 Mean 7.7026 Standard Deviation (SD) 3.2090 Mean + SD 10.9116 Mean − SD 4.4937 Variation 7.7026

BEAM 2 5.8851 2.6861 8.5712 3.1990 5.8851

When performing an analysis of the results obtained, a certain homogeneity was found in the values found. Initially, it was observed that in the pre-cracking period, that is, corresponding to the time interval that goes from 0 to 1300 s, the image analysis demonstrated a certain limitation. This fact is easily explained due to the high level of sensitivity of the LVDT, with small displacements being captured that are in the thousandths of a millimetres, which justifies the distortion found in the comparison of the results between the image and the LVDT.

Fiber Reinforced Concrete Crack Opening Evaluation

277

Table 3. Beam 3 and 4 results. TIME (s) 1300 1400 1500 1600 1700 1800 1900 2000 2100 2200 2300 2400 2500 2600

BEAM 3 LVDT (10−3 m) 2.8159 3.2284 3.5939 3.9712 4.3371 4.6828 5.0356 5.3773 5.7266 6.0834 6.4259 6.7648 7.1004 7.4279

DIC (10−3 m) 2.6567 3.0455 3.3047 3.6286 3.8878 4.2118 4.5358 4.7302 4.9894 5.2486 5.5078 5.7021 6.0261 6.2853

Percent Difference (%) 5.6535 5.6664 8.0473 8.6257 10.3596 10.0571 9.9245 12.0348 12.8744 13.7238 14.2880 15.7082 15.1300 15.3824

BEAM 4 LVDT (10−3 m) 1.6794 2.0838 2.4338 2.7894 3.1424 3.5044 3.8468 4.2150 4.5769 4.9182 5.2925 5.6654 6.0345 6.4108

DIC (10−3 m) 1.8791 2.2031 2.6567 3.0455 3.2399 3.4990 3.8230 4.1470 4.7950 4.8598 5.1838 5.3134 5.7021 6.0261

Percent Difference (%) −11.8890 −5.7249 −9.1586 −9.1786 −3.1027 0.1542 0.6177 1.6120 −4.7656 1.1870 2.0550 6.2137 5.5071 6.0008

Table 4. Beam 1 and 2 results. BEAM 1 Mean 11.2483 Standard Deviation (SD) 3.4443 Mean + SD 14.6926 Mean - SD 7.8039 Variation 11.2483

BEAM 2 4.7976 3.5726 8.3703 1.2250 4.7976

5 Conclusions The application of the digital image correlation (DIC) method proposed by Sutton et al. (1983), the main focus of this work, was able to quantify and monitor the development of the crack formation process in SFRC beams under loading and to quantify their opening. Comparing the values obtained with the images, with values obtained utilizing usual reference devices, such as displacement transducers, an average error of around 7,4% was obtained, which represents a good approximation for using the method in the context of Engineering. The standardization of the experimental process was carried out to allow reproducibility without the need to perform the tests of all parts at once. In this standardization, we can mention the positioning of the LVDT’s in the image plane to avoid reading errors during the correlation process; the positioning of the camera, being at a fixed distance from the specimen; the use of the same test material (LVDT’s, Spider, camera, LEDs, cables), to avoid any measurement errors.

278

K. C. da Silva Oliveira et al.

The technology developed in the study is promising because it is a non-destructive and non-invasive test, and can become suitable for monitoring structures already implemented, detecting cracks throughout the useful life of the structural element.

References 1. Bentur, A., Mindess, S.: Fibre Reinforced Cementitious Composites. Crc Press, London (2007) 2. FÉDÉRATION INTERNATIONALE DU BÉTON–FIB. Fib model Code for Concrete Structures 2010. p. 402, Switzerland (2013) 3. EUROPEAN COMMITTEE FOR STANDARDIZATION. EN 14651: Test method for metallic fiber-reinforced concrete – Measuring the flexural tensile strength (limit of proportionality (LOP), residual), p. 15. CEN, London (2007) 4. Yoneyama, S., Ogawa, T., Kobayashi, Y.: Evaluating mixed-mode stress intensity factors from full-field displacement obtained by optical methods. Eng. Fract. Mech. 74, 1399–1412 (2007) 5. Yoneyama, S., Murasawa, G.: Digital image correlation, in experimental mechanics. Encyclopedia of Life Support Systems (EOLSS) (2009) 6. McComirck, N., Lord, J.: Digital image correlation. Mater Today, pp. 52–54 (2010) 7. McNeill, S.R., Peters, W.H., Sutton, M.A.: Estimation of stress intensity factor by digital image correlation. Eng. Fract. Mech. 28, 101–112 (1987) 8. Chu, T.C., Ranson, W.F., Sutton, M.A.: Applications of digital-image-correlation techniques to experimental mechanics. Exp. Mech. 25, 232–244 (1985) 9. Kuntz, M., Jolin, M., Bastien, J., Perez, F., Hild, F.: Digital image correlation analysis of crack behavior in a reinforced concrete beam during a load test. Can. J. Civ. Eng. 33(11), 1418–1425 (2006) 10. Pan, B.: Digital image correlation for surface deformation measurement: historical developments, recent advances and future goals. Meas. Sci. Technol. 29(8), 082001 (2018) 11. Peters, W.H., Ranson, W.F.: Digital imaging techniques in experimental stress analysis. Opt. Eng. 21, 427–432 (1982) 12. Pan, B., Qian, K.M., Xie, H.M., Asundi, A.: Two-dimensional digital image correlation for in-plane displacement and strain measurement: a review. Meas. Sci. Technol. 20, 062001 (2009)

Effect of Distribution and Orientation of Fibers on the Post-cracking Behavior of Steel Fiber Reinforced Self-compacting Concrete in Small Thickness Elements Néstor Fabián Acosta Medina1(&), Rodrigo de Melo Lameiras2, Ana Carolina Parapinski dos Santos1, and Fábio Luiz Willrich3 1

3

Federal University of Latin America Integration, UNILA, Foz do Iguaçu, Brazil [email protected] 2 University of Brasília, UnB, Brasília, Brazil Laboratório de Tecnologia do Concreto, Itaipu Binacional, Foz do Iguaçu, Brazil

Abstract. In this work, an experimental investigation focused on the distribution and orientation of fibers on the post-cracking behavior of Steel Fiber Reinforced Self-Compacting Concrete (SFRSCC) to cast structural small thickness elements was assessed. To achieve this purpose, two SFRSCC panels with 45 and 60 mm of thickness were cast from their center point. From each panel, cylindrical specimens were extracted and notched either parallel or perpendicular to the SFRSCC flow direction. The post-cracking behavior was determined by means of the Modified Splitting Tensile Test. The fiber distribution was evaluated by counting the number of effective fibers crossing the fractured surfaces. Moreover, the orientation of the fiber was verified using X-ray method. Notched specimens loaded in the parallel direction of the SFRSCC flux lines presented higher post-cracking strength when compared with notched specimens loaded in the perpendicular direction. Likewise, it was also determined that smaller thickness of the structural element, represents greater residual stresses and energy absorption, in consequence of the wall effect. keywords: Steel Fiber Reinforced Self-Compacting Concrete behavior
Thickness
Fiber distribution


Post-cracking

1 Introduction Currently, Fiber Reinforced Concrete (FRC) is mainly applied to industrial floors [1, 2], prefabricated elements and in tunnels [3, 4]. When used in large structural elements, FRC can contribute cost reduction, since the labour required to place the conventional reinforcement is reduced (because the density of the reinforcement is lower when compared to conventional reinforced concrete) or even eliminated if the steel fibers replace all conventional reinforcement [5]. The advantages associated with the addition of steel fibers in concrete mixtures as mentioned above, can be enhanced by the use of a concrete with self-compacting parameters [6], resulting in steel fiber reinforced self-compacting concrete (SFRSCC). © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 279–289, 2021. https://doi.org/10.1007/978-3-030-58482-5_26

280

N. F. A. Medina et al.

SFRSCC can lead to a decrease or even eliminate the need for conventional reinforcement used in a structure. In addition, SFRSCC is easy to pour, increasing efficiency in the construction process [7]. Considering that, conventional concrete presents a brittle behavior when pulled; the addition of fibers allows the modification of this behavior, benefiting in two aspects. First, in the post-cracking resistance, since they act as a means of transferring stresses and loads between the cracks; and the second in the increase in the toughness of the material, providing energy absorption mechanisms, which are related to the processes of detachment and pullout of fibers that make the bridge through the cracks [8, 9]. Although there are several fibers of different natures, steel fibers are the most used as a reinforcement element due to the excellent compatibility between steel and concrete, and because they have a high modulus of elasticity. After the crack appears, the fibers allow the transfer of loads from the cementitious matrix to the fibers, in such a way as to provide increases in the strength of the concrete [10]. As the fiber reinforcement mechanisms are activated mainly after the cracking of the concrete matrix, the fibers have little influence on the behavior of the non-cracked elements. Thus, the resistance to compression and traction of the FRC is related to the strength of the matrix and not by the influence of fibers. In this context, as the residual tensile strength (post-cracking) represents an important design parameter for FRC structures, it effectively constitutes the mechanical property most influenced by fiber reinforcement [11]. The post-cracking behavior of the FRC is influenced by the dispersion and orientation of the fibers. The dispersion and orientation of the fibers in the hardened state are the product of a series of steps that the steel fiber reinforced concrete (SFRC) passes from the mixture to the hardened state. Namely, the main factors that indirectly influence the dispersion and orientation of the fibers are: the properties of the fresh state after mixing, the conditions of release in the form, the flow characteristics, the vibration and the wall effect produced by the form [12, 13].

2 Experimental Study 2.1

Materials and Mixture Proportions

The constituent materials used in the composition of the SFRSCC were: Portland cement CP V (C), water (W), superplasticizer of third generation (SP) based on polycarboxylates (GLENIUM 51), limestone filler, fine river sand and coarse sand, coarse aggregate and hooked end steel fibers (length, lf, of 50 mm; diameter, df, of 0.55 mm; aspect ratio, lf/df, of 67 and a yield stress greater than 1100 MPa). The adopted mix proportions are shown in Table 1, where W/C is the water/cement ratio. The content of steel fibers in all SFRSCC is kept constant and equal to 40 kg/m3. The test program was conducted on samples whose mix design has followed the recommendation presented in Melo [14] and Gomes [15]. Firstly, the proportions of the constituent materials of the paste were defined, then the proportions of each aggregate on the final skeleton were determined, and finally the paste and the solid skeleton were mixed in different proportions (36, 37, 38 and 40% of paste) until self-compacting

Effect of Distribution and Orientation of Fibers on the Post-cracking Behavior

281

Table 1. Mix proportions of SFRSCC per m3. Casting Cement (kg)

Water (kg)

A

148.69 0.4

399.95

W/C (−)

SP Limestone (kg) filler (kg) 4.79 199,97

Fine sand (kg) 471.85

Course sand (kg) 459.46

Course aggregate (kg) 900

requirements are assured in terms of spread ability and correct flow velocity. The final mix proportions are shown in Table 1. To evaluate the properties of SFRSCC in the fresh state, the slump flow test was performed according to EFNARC recommendations [16]. 2.2

Specimens

According to Barnett et al. [17], when casting a slab out from its centre, a better mechanical behavior is assures in comparison with other casting methods. Therefore, three SFRSCC panels were produced following the selected direction of casting. The dimensions of the panels were 1500 x 1500 mm2, where the panels thickness varied (45, 60 and 75 mm), according to the scheme represented in Fig. 1. The specimens that were used in this work were the number 4, 5, 8 and 13 for the 45 mm thickness plate, and the specimens 4, 5, 12 and 13 for the 60 mm thickness plate.

Fig. 1. Schematic representation of the core-extracting specimens (units in millimeters).

The test carried out on specimens extracted from these panels will therefore allow to take conclusions about the flow induced orientation and dispersion of fibers on the post-cracking behavior. The influence of fiber dispersion and orientation within the panel was assessed by means of the Modified Splitting Tensile Test, improved by Abrishambaf et al. [18] and Lameiras et al. [19]. In Fig. 1, the pale dash line represent the supposed concrete flow direction.

282

N. F. A. Medina et al.

No specimens were extracted from the vicinity of the edge of the formwork, in order to reduce the interference of the wall effect from the lateral formwork of the results. The drilling operation was performed when the panel were in their hardenmature phase. The hatch cores were extracted from a distance of 300 mm and 600 mm from its centre, all of them with 200 mm in diameter. The extracted hatch cores were subjected to a loading coincident with the notched plane performed on the specimens. In order to obtain information on the influence of the plate thickness on the fiber orientation, the post-cracking behavior of the specimens was assessed in two distinct directions. By considering h as the angle between the notched plane and the direction of the concrete flow, the notch plane is designated parallel for h = 0° or perpendicular for h = 90°. According to Abrishambaf et al. [18], in order to localize the specimen’s fracture surface along the notched plane, a 5 mm deep notches parallel to the loading direction was executed (see Notch 2 in Fig. 2). Furthermore, two additionally notches in a V-shaped groove with 45° inclination has been cut at the extremities, in both directions of the notched pane, as shown in notch 1 of Fig. 2. This V-shaped groove configuration, as reported by Di Prisco et al. [20], induced a stress field corresponding to an almost pure fracture mode I in the notched plane at the loading. Else, Di Prisco et al. [20] states that this configuration deviate the compressive stresses field from the notched plane, creating, an uniaxial tensile stress field in the notched plane. Moreover, following what was implemented by Di Prisco et al. [20], two more 5 mm deep straight notches at the V groove vertices were executed. This force the crack to open at the reduced section and move the crack tip away from the load application zones, where high stress concentrations generally arise (see notch 3 in Fig. 2).

Fig. 2. Sequence of implementation of the notches made in the specimens (units in millimetres; tf: total height; tr: notched height).

2.3

Test Setup and Procedure

2.3.1 Splitting Tensile Test The test were carried out under displacement control, through a universal servocontrolled electro-hydraulic testing machine, with a 1000 kN load cell and closed-loop configuration. The load was applied by using two steel rollers of 20 mm diameter that were accommodated into the V-shaped grooves. The test setup is depicted in Fig. 3. In each specimen five linear variable differential transducers (LVDTs) of 5 mm stroke

Effect of Distribution and Orientation of Fibers on the Post-cracking Behavior

283

were used to measure the crack opening displacement, as shown in Fig. 3(a) and (b). Three LVDT were applied at the side of the specimen corresponding to bottom side during casting (corresponding to the side in contact with the metallic formwork), and the two remaining at the other side of specimen.

Fig. 3. Experimental setup of the specimen (a) detail of the bottom side during casting; (b) upper side during casting and (c) positions of the LVDT’s (units in millimeters).

The test was conducted using the following displacement rates: 0.06 mm/min up to the displacement of 2.0 mm; 0.12 mm/min up to 2.5 mm; 0.24 mm/min until the end of the test. The exact position of each LVDT is schematically represented in Fig. 3(c). 2.3.2 Assessment of Fiber Distribution and Orientation After testing, the fiber distribution in the specimens was determined. Each of the two faces of the fracture surface of splitting test was divided in four equal regions (see Fig. 4a). This stage made possible to quantify the fiber density in the fracture surface. The fiber distribution was evaluated by counting the number of effective fibers crossing the fractured surfaces (see Fig. 4b). A fiber was considered effective when it was broken or when its visible length was, at least, twice the length of the hooked part of the fiber. However, this procedure allows to draw conclusions about their orientation and distribution in the structural element. Furthermore, the great challenge to understand better the distribution and orientation of fibers within a specimen, is to correlate images using non-destructive methods with the results obtained from the mechanical test performed. Then, the orientation of the fibers was verified using X-ray method. In some specimens, a photographic representation of the projection of the fibers was obtained.

284

N. F. A. Medina et al.

Fig. 4. Assessment of the fiber density in the fractured zone of specimens. (a) Schematic representation of the fiber counting regions; (b) overview of fiber counting for splitting test specimens.

3 Results and Discussion 3.1

Splitting Test

The Fig. 5 and Fig. 6 depict the splitting tensile stress (r) versus crack width curves, where r is determined from the following equation: r¼

2
P p
tr
h

4.0

4.0 Envelope Average L.I.95

3.5

3.0

L.S.95

2.5

Envelope Average L.I.95

3.5

Stress (MPa)

Stress (MPa)

3.0

2.0 1.5 1.0 0.5 0.0 0.0

ð1Þ

L.S.95

2.5 2.0 1.5 1.0 0.5

0.5

1.0

Crack width (mm)

(a)

1.5

0.0 0.0

0.5

1.0

1.5

Crack width (mm)

(b)

Fig. 5. Splitting tensile stress versus crack width relationship for the specimens obtained from the panel with 45 mm thickness, and with loading direction (a) parallel, and (b) perpendicular, to the SFRSCC flux lines.

Effect of Distribution and Orientation of Fibers on the Post-cracking Behavior 4.0

4.0 Envelope Average L.I.95

3.5

3.0

L.S.95

2.5

Envelope Average L.I.95

3.5

Stress (MPa)

Stress (MPa)

3.0

2.0 1.5 1.0 0.5 0.0 0.0

285

L.S.95

2.5 2.0 1.5 1.0 0.5

0.5

1.0

Crack width (mm)

(a)

1.5

0.0 0.0

0.5

1.0

1.5

Crack width (mm)

(b)

Fig. 6. Splitting tensile stress versus crack width relationship for the specimens obtained from the panel with 60 mm thickness, and with loading direction (a) parallel, and (b) perpendicular, to the SFRSCC flux lines.

where P is the compressive load applied in the specimen, tr is the height of the SFRSCC cylinder and h is the diameter of the remaining SFRSCC cylinder after the notches are executed in the specimen, as shown in Fig. 2. The values for tr and h in the equation were measured in each specimen. The crack width of the abscissa axes of Fig. 5 and Fig. 6 corresponds to the average of the values measured in the five LVDTs in the specimen. The average, envelope and Lower Bound (L.B.) and Upper Bound (U. B.) characteristics curves corresponding to a confidence level equal to 95% are also presented in Fig. 5 and Fig. 6. The splitting tensile stress (r) versus crack width responses are almost linear up to the stress at crack initiation in all specimens. Note that, the peak stress was equal to the stress at crack initiation. Once the peak load was attained, a strain-softening branch is verified on specimens. Despite the high dispersion of results observed in specimens, this behavior is generally high in SFRSCC elements, even with specimens of the same casting and executing under the same test condition. This is due to the high dependence on post-cracking behavior in relation to the orientation and distribution of the fibers, as already shown by Abrishambaf et al. [18]. Furthermore, it is quite evident that the stress at crack initiation, and mainly the post-cracking tensile strength, were higher in specimens loaded in the direction of the SFRSCC flux lines than in the orthogonal direction, due to the tendency of fibers to orientate themselves orthogonally to the SFRSCC flux lines, as demonstrated by Abrishambaf et al. [18], and verified by Lameiras et al. [19]. Other conclusions can also be drawn from residual stresses parameters (rw*) and energy absorption (GFw*) during the fracture process, as presented in Table 2 and Table 3. The results were presented separating the samples according to the casting, thickness and loading direction according to the flow lines. It is noted that the value attributed to the subscript “w*”, represents the crack width at which rw and GF are evaluated.

286

N. F. A. Medina et al.

Table 2. Maximum stress, residual stress and dissipated energy for the specimens and with load applied parallel to the SFRSCC flux lines. Specimen

Direction rmax

A4–45 mm 0° A5–45 mm Avg. CoV A4–60 mm 0° A5–60 mm Avg. CoV

r0.5

r1.0

(MPa) (MPa)

r0.3

(MPa)

(MPa) (MPa)

(N/mm) (N/mm) (N/mm) (N/mm)

2,72 2,63 2,68 2,54% 2,16 2,45 2,68 2,54%

1,56 1,95 1,75 15,80% 1,89 1,37 1,75 15,80%

1,71 1,68 1,69 1,16% 1,61 1,28 1,69 1,16%

0,41 0,54 0,47 19,36% 0,35 0,31 0,47 19,36%

1,49 2,04 1,76 22,12% 2,08 1,42 1,76 22,12%

r1.5

GF0.3

1,41 1,68 1,55 12,16% 1,41 1,26 1,55 12,16%

GF0.5 0,65 0,87 0,76 21,11% 0,59 0,46 0,76 21,11%

GF1.0 1,28 1,63 1,46 16,99% 1,12 0,82 1,46 16,99%

GF1.5 1,88 2,33 2,10 15,15% 1,59 1,17 2,10 15,15%

Table 3. Maximum stress, residual stress and dissipated energy for the specimens and with load applied perpendicular to the SFRSCC flux lines. Specimen

Direction rmax

A8–45 mm 90° A13–45 mm Avg. CoV A12–60 mm 90° A13–60 mm Avg. CoV

r0.3

r0.5

r1.0

r1.5

GF0.3

(MPa)

(MPa)

(MPa)

(MPa)

(MPa)

(N/mm) (N/mm) (N/mm) (N/mm)

3,19 2,07 2,63 30,29% 1,99 2,30 2,14 10,23%

1,54 0,90 1,22 37,16% 1,69 0,82 1,26 49,20%

1,48 0,95 1,22 30,85% 1,56 0,77 1,16 48,17%

1,48 0,94 1,21 31,07% 1,04 0,76 0,90 21,62%

1,42 0,83 1,12 36,79% 0,97 0,83 0,90 10,54%

0,50 0,29 0,40 38,21% 0,31 0,33 0,32 2,91%

GF0.5 0,75 0,42 0,59 39,44% 0,52 0,42 0,47 15,21%

GF1.0 1,37 0,77 1,07 39,86% 0,91 0,64 0,78 24,14%

GF1.5 1,98 1,10 1,54 40,18% 1,22 0,89 1,05 22,39%

From the results of Table 2 and Table 3, was verified the influence of the load application in the parallel and perpendicular direction to the flow lines. When the load was applied parallel to the flux lines, the specimens of the two thicknesses evaluated presented higher residual tensile stresses and also higher energy absorptions. This when compared to specimens loaded perpendicular to the flux lines. Therefore, can be verified that the results obtained through the variation of thickness of the SFRSCC specimens, the average residual tensile stresses and the average absorbed energies increased, while decreasing thickness. However, it is presented that the specimens with 45 mm thickness showed greater energy absorption, both parallel and perpendicular direction to the flux lines. 3.2

Assessment of the Number of Effective Fibers

The preferential alignment of fibers are corroborated by the results of effective fiber counting at the fractures surface of splitting test, as shown in Table 4 for the specimens obtained from the plate with two different thickness. The values presented in Table 4 highlights the correlation between the post-cracking response of the specimens and the effective fiber counting at the fracture plane. Hence, when the values of Table 4 are

Effect of Distribution and Orientation of Fibers on the Post-cracking Behavior

287

examined separately by load direction, it is verified a greater number of effective fibers in specimens loaded parallel to the SFRSCC flux lines, when compared with the average number of fibers counted in the specimens with the load applied perpendicularly to the SFRSCC flux lines. Thus, the values determined are consistent with the highest post-cracking parameters obtained in specimens with the fracture plane parallel to the flux lines, as shown by Abrishambaf et al. [18] and Di Prisco et al. [20]. Table 4. Fiber counting at fractured surface of specimens. Thickness (mm) 45

60

Loading direction Specimen Average number of effective fibers (fibers/cm2) Top Bottom Total Per Top Bottom thickness Parallel to flow A4–45 1,06 1,45 1,25 1,49 1.33 1.41 A5–45 2,14 1,32 1,73 Perpendicular to A8–45 1,32 1,98 1,65 1,25 flow A13–45 0,81 0,89 0,85 Parallel to flow A4–60 1,94 1,83 1,88 1,42 1.18 1.55 A5–60 0,78 1,11 0,95 Perpendicular to A12–60 1,34 2,31 1,82 1,31 flow A13–60 0,65 0,95 0,80

Furthermore, it is verified the existence of segregation when compared the values in Table 4 according to its position (top and bottom) in the plate. For the 45 mm thickness plate, 5.4% more fibers per square centimeter can be observed when compared to the top region. In the same way, for the 60 mm thickness plate, 31.5% more fibers were found in the bottom side when compared to the top side. 3.3

Visualization of Fiber Distribution and Orientation

To view the fiber orientation, were selected for the paper only the specimen corresponding to the position 5 (see Fig. 1), with load applied in the parallel direction of the SFRSCC flow, and the specimen corresponding to the position 13 (see Fig. 1), with load applied in the perpendicular direction of the SFRSCC flow. In Fig. 7, can be observed a tendency of the fibers to align perpendicularly to SFRSCC flux lines. The specimen with loading application in the parallel SFRSCC flux lines presented a larger amount of fibers crossing the fracture plane. This verification are in accordance with the values of maximum residual stresses (rmax) and energy absorption (GF), obtained in Table 2 and Table 3. In addition, it was verified in the same direction of the loading application, a greater horizontal distribution of the fibers, matching fewer fibers crossing the fracture section and lower values of maximum residual stresses and energy absorption, according to the values obtained in Table 3.

288

N. F. A. Medina et al.

Load application

Flux direction

(a)

Load application

Flux direction

(b)

Fig. 7. X-ray images of specimens extracted from the 45 mm thickness plate: (a) A5 specimen and (b) A13 specimen.

4 Conclusions • For the materials and casting condition used in the investigation, it was found that fibers have a tendency to be oriented perpendicularly to the concrete flow, namely, the radial flow; as already noted by some researchers. • The specimens presented better behaviors in the post-cracking strength when loaded in the direction parallel to the flux lines of the SFRSCC, when compared with specimens loaded in the perpendicular direction. Likewise, it was also found that smaller thickness of the structural element, represents greater residual stresses and energy absorption, in consequence of the wall effect. • The X-ray tests carried out on the specimens proposed in the study, allowed to verify this preferential tendency of the fibers in relation to the concrete flow.

5 Acknowledgments The authors wish to express their thanks for the financial support of Fundação Araucária. The first author also acknowledge the CNPq and the Itaipu Binacional for the support.

References 1. Frazier, P.: Steel fibrous concrete for airport pavement applications. Technical Report S-7412. US Army Engineer Waterways Experiment Station, Vicksburg, MS (1974) 2. Suksawang, N., Mirmiran, A., Yohannes, D.: Use of f reinforced concrete for concrete pavement slab replacement. Final Report, Florida Department of Transportation, Tallahassee, Florida (2014)

Effect of Distribution and Orientation of Fibers on the Post-cracking Behavior

289

3. Burgers, R., Walraven, J., Plizzari, G.A., Tiberti, G.: Structural behavior of SFRC tunnel segments during TBM operations. In: Barták, J., Hrdina, I., Romancov, G., Zlámal, J. (eds.) Underground Space – the 4th Dimension of Metropolises, Prague, pp. 1461–1467. CRC Press, Prague (2007) 4. Gettu, R., Barragán, B., Garcia, T., Gonzalo, R., Fernández, C., Oliver, R.: Steel fiber reinforced concrete for the barcelona metro line 9 tunnel lining. In: International Rilem Symposium on Fibre Reinforced Concretes, Proceedings, Varenna. RILEM Publications, Varenna, 2004, pp. 141–156 (2004) 5. Soutsos, M., Le, T., Lampropoulos, A.P.: Flexural performance of fibre reinforced concrete made with steel and synthetic fibres. Constr. Build. Mater. 36, 704–710 (2012) 6. Okamura, H.: Self-compacting high-performance concrete. Concr. Int. 19(7) (1997) 7. Tojal, T.L.: Contribution to the study of the adhesion of steel bars on self-compacting concrete reinforced with steel fibers. Master Thesis (Master in Structure) – Graduate Program in Civil Engineering, Federal University of Alagoas, Maceió (2011). 116 p. 8. Bentur, A., Mindess, S.: Fibre Reinforced Cementitious Composites, 2nd edn. Taylor & Francis, New York (2007) 9. Rambo, D.A.S.: Self-compacting concrete reinforced with hybrid steel fibers: material and structural aspects. Master Thesis (Master in Civil Engineering) - Graduate Program in Civil Engineering, Federal University of Rio de Janeiro, Rio de Janeiro (2012) 10. Velasco, R.V.: Self-compacting concrete reinforced with high volumetric fractions of steel fibers: rheological, physical, mechanical and thermal properties. PhD Thesis (PhD in Civil Engineering) - Graduate Program in Civil Engineering, Federal University of Rio de Janeiro, Rio de Janeiro (2008), 388 p. 11. Di Prisco, M., Plizzari, G., Vandewalle, L.: Fibre reinforced concrete: new design perspectives. Mater. Struct. 42(9), 1261–1281 (2009) 12. Laranjeira, F.: Design-oriented constitutive model for steel fiber reinforced concrete. Thesis (PhD) - Department d´Enginyeria de la Construcció, Universitat Politècnica de Catalunya, Barcelona (2010). 318 p. 13. Martinie, L., Rossi, P., Roussel, N.: Rheology of fiber reinforced cementitious materials: classification and prediction. Cem. Concr. Res. 40(2), 226–234 (2010) 14. Melo, K.A.: Contribution to the self-compacting concrete composition with the addition of limestone filler. Master Thesis (Master in Civil Engineering) - Graduate Program in Civil Engineering, Federal University of Santa Catarina, Florianópolis, (2005). 183 p. 15. Gomes, P.C.C.: Optimization and characterization of high-strength self-compacting concrete. Thesis (Ph.D.) - Departament D’Enginyeria de la Construcció, Universitat Politècnica de Catalunya, Barcelona (2002). 150 p. 16. EFNARC: The European Guidelines for Self-Compacting Concrete. United Kingdom (2005) 17. Barnett, S., Lataste, J.F., Parry, T., Millard, S.G., Soutsos, M.N.: Assessment of fibre orientation in ultra high performance fibre reinforced concrete and its effect on flexural strength. Mater. Struct. 43, 1009–1023 (2010) 18. Abrishambaf, A., Barros, J.A.O., Cunha, V.M.C.F.: Relation between fibre distribution and post-cracking behaviour in steel fibre reinforced self-compacting concrete panels. Cem. Concr. Res. 51, 57–66 (2013) 19. Lameiras, R.M., Barros, J.A.O., Azenha, M.: Influence of casting condition on the anisotropy of the fracture properties of Steel Fibre Reinforced Self-Compacting Concrete (SFRSCC). Cement Concr. Compos. 59, 60–76 (2015) 20. di Prisco, M., Ferrara, L., Lamperti, M.G.L.: Double edge wedge splitting (DEWS): an indirect tension test to identify post-cracking behaviour of fibre reinforced cementitious composites. Mater. Struct. 46(11), 1893–1918 (2013)

Ductility of the Four-Year-Old Steel Fibre Reinforced Concrete Jakob Šušteršič(&), Rok Ercegovič, David Polanec, and Andrej Zajc IRMA Institute for Research in Materials and Applications, Ljubljana, Slovenia [email protected]

Abstract. The paper deals with the results of an experimental investigation into the ductility of Steel Fibre Reinforced Concrete (SFRC) with a steel fibre content of between 0,5% and 2% by volume, and that of a comparable concrete without fibres. These investigations are part of a large-scale research project on the SFRC that lasted 4 years. At their age of 4 years high compressive strength has been achieved, with average values in the range of approximately 60 to 115 MPa. These concretes can be divided into two groups: 1st group - concrete with a maximum nominal grain size (Dmax) of 16 mm with compressive strength of 90 to 115 MPa and 2nd group - concrete with Dmax = 4 and 8 mm with a compressive strength of 60 to 80 MPa. The ductile behavior of SFRC was evaluated by the ductility factor 1/B. 1/B is a parameter for evaluating the behavior of FRC, which takes into account the entire surface under the load – CMOD curve. In this way it is possible to evaluate the behavior of the FRC with only one parameter. Wedge Spit Test (WST) method was used to obtain load – CMOD curves. In 1st group, a large influence of the fiber aspect ratio (lf/df) on the increase of 1/B and the lower impact of the aggregate type is visible. In 2nd group, the influence of polymer, Dmax in addition to fibers on ductile behavior, can be noticed. Keywords: Steel Fibre Reinforced Concrete
Ductility
Ductility factor 1/B
Absorption energy
Fiber aspect ratio
Wedge split test

1 Introduction As concrete ages, strength increases, but also the brittleness. By adding fibres, the toughness and ductility of concrete can be increased. In research, we are most interested in how concrete behaves at older ages, when the hydration of cement and the formation of cement stone are almost completely completed. Cement stone binds individual aggregate particles and fibres into a monolithic structure (but heterogeneous in composition). This increases the strength of the composite as well as the efficiency of the added fibres, thereby increasing the toughness and ductility of the composite. Thus, the ductility of SFRC at high strengths is investigated. Based on these properties, we estimated which SFRC at high strengths still exhibit ductile behaviour. Concretes with a low value of w/c ratio have a higher strength than the strength of normal concrete (compressive strength > 60 MPa). The behaviour of such concrete during fracture can be “explosive”, leading to rapid demolition of the structural element. The crack © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 290–300, 2021. https://doi.org/10.1007/978-3-030-58482-5_27

Ductility of the Four-Year-Old Steel FRC

291

propagation in concrete can be very fast. When fibres are added in concrete, they prevent rapid crack propagation and concentration into one devastating crack [8–10]. In practice, we tend to use fibres that will least affect the workability of concrete. This means that we want to work with fibres with as little aspect ratio (lf/df) as possible. The smaller the lf/df, the easier the fibres are to mix in the fresh concrete mass and the easier it is to place the SFRC. Most commonly used are steel fibres with an aspect ratio lf/df = 50 to 75. In our study, we used steel fibres with lf/df = 66 (lf = 33 mm, df = 0.50 mm). The other types of steel fibres used in our study are more effective in hardened concrete but less used in practice. Their aspect ratio is lf/df = 87 (lf = 33 mm, df = 0,38 mm). The added quantity of fibres has a major impact on the workability, and especially on the SFRC price. From a technological point of view, steel fibres of up to 2% by volume can be used in practice, which is approximately 160 kg/m3 of placed concrete. SFRC with a steel fibre content of approximately 0,5 to 1,0% by volume (40 to 80 kg/m3) are mostly prepared in our construction practice [11]. Due to the practical interest in our study, we have prepared SFRC with the aforementioned steel fibres in quantities of 0,5 to 2,0% by volume in various modifications of the mix-proportion.

2 Experimental Details 2.1

Concrete Mix-Proportions

Seven subgroups of the mix-proportions listed in Table 1 are investigated. Basic principles for choosing the concrete mix-proportions from Table 1: • The values of water/binder (w/b) ratio are low, in order to obtain the highest quality cement stone (its structure and adhesion with the aggregate particles, especially fibres). • The adhesion of cement stone and fibre is increased by the addition of SF or SBL (polymer). • Two characteristic values of the steel fibre aspect ratio were selected and used: – lf/df = 66 - (lf = 33 mm, df = 0.5 mm), – lf/df = 87 - (lf = 33 mm, df = 0.38 mm). • The amount of fibres varies from 0,5 to 2,0% by volume, that is, from the most used amount in practice (approximately 40 kg/m3 of placed concrete) to the amount of fibre that can still be manipulated by conventional technological methods of preparation and placement of concrete (160 kg/m3 of placed concrete). For some subgroups, concrete mixtures without fibres have also been prepared to evaluate the effect of added fibres on individual properties and behaviour under different loads. • Superplasticizer (SPL) improves the workability of concrete, which reduces the amount of water required, thereby reducing the value of w/b ratio. At a constant value of w/b ratio, the use of superplasticizer also reduces the amount of cement, which favourably affects the rheology of the concrete.

292

J. Šušteršič et al. Table 1. Mix-proportions (per 1 m3 of concrete mixture).

CEM* SF** Steel Fibers SPL*** AE**** SBL+ AGG# Dmax l/d amount _ % by (% of (% of (% of (mm) – (kg) (% volume CEM) CEM) CEM) of CEM) I/0 0.35 320 9.4 – – 2 0.2 – 16 I/0.5 0.35 330 9.4 66 0.5 2 0.15 – 16 I/1 0.35 348 9.4 66 1.0 2 0.3 – 16 I/1.5 0.35 375 9.4 66 1.5 2 0.2 – 16 I/2 0.35 393 9.4 66 2.0 2 0.2 – 16 II/0 0.35 439 9.4 – – 2 0.07 – 16 II/1 0.35 457 9.4 66 1.0 2 0.15 – 16 II/1.5 0.35 466 9.4 66 1.5 2 0.15 – 16 III/0.5 0.35 330 9.4 87 0.5 2 0.15 – 16 III/1.5 0.35 375 9.4 87 1.5 2 0.2 – 16 IV/1 0.35 457 9.4 87 1.0 2 0.15 – 16 IV/1.5 0.35 466 9.4 87 1.5 2 0.15 – 16 V/1 0.30 580 – 66 1.0 3 – – 16 VI/0 0.42 650 – – – – – – 4 VI/0p 0.30 660 – – – – – 10 4 VI/0.5p 0.30 657 – 66 0.5 – – 10 4 VI/1 0.41 645 – 66 1.0 – – – 4 VI/1p 0.30 655 – 66 1.0 – – 10 4 VII/1 0.40 657 – 66 1.0 – – – 8 VIIp/1 0.41 665 – 66 1.0 – – 10 8 * ** *** **** CEM - cement, SF - silica fume, SPL - superplasticiter, AE - air-entraining admixture, + SBL - solid particles of styrene-butadiene copolymer latex, #AGG – aggregate; aggregate types: fine gravel aggregate + coarse eruptive aggregate in concrete from subgroups I and III; slag aggregate in concrete from subgroups II and IV; gravel aggregate in concrete from subgroups V, VI and VII. Designation w/b of concrete ratio

• The use of air-entraining admixture (AE) introduces fine closed air bubbles into the concrete which increase the resistance of the concrete to the freezing/thawing effects of the concrete without or in the presence of de-icing salt. Introduced air bubbles can influence the better distribution of added fibres in fresh concrete. The resistance of concrete to the effects of freezing/thawing is also increased by reducing the value of w/b ratio or by adding SBL, thereby closing open capillary pores and reducing their number. • The distribution of fibres in fresh concrete and their efficiency in hardened concrete may also be affected by the largest aggregate grain (Dmax). At constant fibre length (lf = 33 mm) the ratio lf/Dmax has the values: 2, 4 and 8.

Ductility of the Four-Year-Old Steel FRC

293

• The use of high-density aggregate with high strength also enables the preparation of high strength concrete. Of course, these concretes must have good quality cement stone with low w/b ratio values. • Slag aggregate, as a secondary raw material, can replace natural eruptive aggregate. The pozzolanic properties of the slag can also be counted on, which increases the adhesion of certain aggregate grains and cement stone, thereby increasing the compactness of concrete, which in turn increases the strength and other resistance of these concretes. When designing the mix-proportions of concrete from subgroups I to IV, the selected value of w/b ratio was kept constant. Other parameters of the mix-proportions (especially the amount of cement) were chosen so as to obtain the best workability of fresh concrete. Concrete mix-proportions of subgroups VI and VII are designed for constant workability of fresh concrete and constant ratio: cement mass/aggregate mass = 1/2. Because of this, the values of w/b ratio have changed so that for polymer modified concrete with SBL the obtained value of w/b ratio is less than w/b ratio of concrete without SBL. 2.2

Characteristic Properties of the Materials Used for Concrete

• Steel fibres: Both types of fibres used (lf/df = 66 and 87) are of the same shape, given in Fig. 1. End hooks allow better anchoring of fibres in concrete, while adhesion with cement stone increases with increasing quality of cement paste (low value of w/b ratio and use of admixtures). The fibres with lf/df = 66 are produced from wire with an average maximum tensile strength Rm = 841 N/mm2, and the fibres with lf/df = 87 from wire with Rm  2000 N/mm2.

Fig. 1. Characteristic shape of the steel fibres used.

• Cement: Portland cement CEM I 42,5 was used to prepare the concrete from subgroups I to V, and Portland cement with slag PC II/A-S 42,5 for the concrete from subgroups VI and VII. • Silica fume: fine amorphous dust in the form of balls of average diameter 0,1 lm; SiO2 content is 94,0%. • Superplasticizer: modified polycarboxylate; density of SPL is 1,2 kg/dm3.

294

J. Šušteršič et al.

• Air-entraining admixture: it is made on the basis of vinyl resin; density of AE is 1,1 kg/dm3. • Styrene-butadiene copolymer latex: with the following characteristics: solid particles (dried at 105° C) = 46,0%, Cl content = 1.45%, pH value = 8,3, density = 1,01 kg/dm3. • Gravel aggregate: fraction of fine aggregate = 0-4 mm, fractions of coarse aggregate = 4-8 and 8-16 mm. • Eruptive aggregate: fractions of crushed amphibolite 4-8, 8-11 and 11-16 mm; density slightly greater than 2900 kg/m3. • Slag aggregate: Slag generated in the production of ferrochrome by the carbothermic process. Its structure is similar to the porphyry structure of magmatic rocks. The aggregate, which is obtained by crushing the slag, is one of the heavy aggregates for concrete because it has a density greater than 3000 kg/m3. For the preparation of concrete from subgroups II and IV, fractions of 0-4, 4-8 and 816 mm of slag aggregate were used. 2.3

Program of Investigations

The main purpose of adding fibres to concrete is to increase its ductility, toughness and resistance to crack propagation. Several types of tests were performed. This paper will only provide some of the results of the wedge spit test (WST). WST is a test method to perform stable fracture mechanics tests on concrete and concrete-like materials. It was proposed by Brühwiler and Wittman [1], and Linsbauer and Tschegg [2]. The method proposed by last authors [2] was used and discussed in previous investigations [3–5] in order to obtain load – CMOD curves. When a concrete element is loaded, small individual cracks begin to appear inside the concrete at a certain load. Those cracks combine into a continual crack, which can be seen on the surface of concrete element. At this point of the load - CMOD curve, the slope of the curve increases significantly. Load and CMOD at this point are denoted as first crack load, first crack CMOD, respectively. There are difficulties relating to precise determination of the location of the first crack. ASTM C 1018 defines first crack as the point on the load – deflection curve at which the form of the curve first becomes nonlinear. Determination of the point of first crack has been proposed [6, 7] as the point at which the slope of the curve departs from linearity by more than 5% and lasts for an interval of more than 0,01 mm. At our Institute (IRMA), computer program has been developed, which works in graphical form, for automatically drawing load-CMOD curves, for calculation of parameters for evaluation of concrete behaviour, and for determination of the point of first crack (FC) [4]. At the moment when the point FC is reached, the crack width begins to propagate with further loading. From the point FC, the fracture zone of the concrete begins to form. In the fracture zone, all further fracture processes proceed until the final separation of the test specimen. Absorbed energy of fracture zone (WFZ) represents the absorbed energy required to completely separate the test specimen. Absorbed energy is expressed as the product of load and crack mouth opening displacement (CMOD).

Ductility of the Four-Year-Old Steel FRC

295

When the load at first crack (FFC) and the absorbed energy of fracture zone (WFZ) are determined from the load-CMOD diagram, the characteristic CMOD of fracture zone (DCMOD) can be calculated according to Eq. (1): DCMOD ¼

WFZ FFC

ð1Þ

DCMOD is a property of a concrete, on the basis of which we can estimate the ductility of a concrete with respect to CMOD at the first crack (CMODFC). Concrete is ductile if DCMOD > CMODFC and vice versa concrete is brittle if DCMOD < CMODFC. Ductility can be expressed by the dimensionless ductility factor 1/B, which could be considered as a favourable parameter for the evaluation of the behaviour of fibre reinforced concrete (FRC). The ductility factor 1/B is expressed by Eq. (2): 1=B ¼

DCMOD CMODFC

ð2Þ

The ductility factor 1/B is higher and thus the ductility of the tested FRC, if the characteristic crack mouth opening displacement of the fracture zone (DCMOD) is larger, which means that the FRC must achieve as much large absorbed energy of the fracture zone (WFZ) at the corresponding load at the first crack (FFC). From prisms with dimensions of 10  10  40 cm, at the age of concrete 4 years, test bodies were cut - cubes with a 10 cm edge with a notch in the middle to a depth of 5 cm and a necessary slot for wedge, which is pushed vertically during the experiment. This widens the notch. The crack opening velocity was 0,04 mm/min.

3 Results and Discussion The ductility of concrete at the age of 4 years was tested when high compressive strengths were reached, with average values ranging from approximately 60 to over 115 MPa (Fig. 2). Thus, all the subgroups in Table 1 can be joined into two groups: 1st group: concrete with Dmax = 16 mm (concrete from subgroups I to V) with compressive strength of 90 to 115 MPa and 2nd group: concretes with Dmax = 8 mm (concretes from subgroup VII) and concretes with Dmax = 4 mm (concretes from subgroup VI) with compressive strengths of 60 to 80 MPa. At such high strengths, the question first arises about the form of behaviour of these concretes in loading, in our case with wedge splitting. As an example, the characteristic load-CMOD diagrams of individual concrete are visible from Fig. 3, in which concrete diagrams from subgroups I and III are given. Figure 3 shows the influence of aspect ratio of steel fibre on the ductile behaviour of SFRC. The visual evaluation of the ductile behaviour of concrete from the diagrams in Fig. 3 may very well be replaced by a numerical evaluation using a ductility factor1/B (Fig. 4 and 5).

J. Šušteršič et al.

296

Fig. 2. Average compressive strength of 4 years old concretes.

2,2 2,0

load (kN)

1,8 1,6 1,4 1,2 1,0 0,8 III/1,5 0,6

I/2

I/1,5

III/0,5

0,4 0,2 0,0 0,0

I/0

I/1 I/0,5 0,5

1,0 CMOD (mm)

1,5

2,0

Fig. 3. Characteristic diagrams load - CMOD of concrete from subgroups I and III.

First, a relative comparison can be made from the dependence of the ductility factor 1/B on the amount of steel fibres in the individual concrete from both groups.

Ductility of the Four-Year-Old Steel FRC

297

60

ductility factor 1/B

50

IV

40 30

III

20

II

10

I

0 0

0.5

1

1.5

steel fibres (% by volume) Fig. 4. The dependence of the ductility factor 1/B on the amount of steel fibres in concretes from subgroups I to IV.

30

VIIp

ductility factor 1/B

25

VII

20 15

VI

VIp

10 5 0 0

0.5 steel fibres (% by volume)

1

Fig. 5. The dependence of the ductility factor 1/B on the amount of steel fibres in concretes from subgroups VI and VII.

In 1st group, a large influence of the fibre aspect ratio (lf/df) on the increase of 1/B and a smaller influence of the type of aggregate are seen (Fig. 4). The ductility factor 1/B of SFRC with 1,5% by volume of steel fibres with lf/df = 87 is greater than 1/B of

J. Šušteršič et al.

298

SFRC with 2,0% by volume of steel fibres with lf/df = 66. In concrete without fibres and with slag, the total absorbed energy is higher than in concrete without fibres and with an eruptive aggregate. However, in the case of concrete with slag, the absorbed energy of the elastic region is higher, which is why the ductility factor 1/B for this concrete is reduced compared to the concrete with eruptive aggregate. Due to the large amount of absorbed energy of the elastic area of the concrete with the slag aggregate, the absorbed energy of the fracture zone WFZ is relatively smaller but still higher than the absorbed energy of the fracture zone of concrete with eruptive aggregate. By adding fibres to the concrete with slag aggregate, especially fibres with lf/df = 87, the absorbed energy of the fracture zone is greatly increased. Adhesion increases and thus the efficiency of fibres. The behaviour of these SFRC is similar to the behaviour of high performance SFRC. Similar observations are valid for concretes from the second group. Only a polymer (SBL) without fibres in concrete with Dmax = 4 mm increases the ductile behaviour, as can be seen from the diagram in Fig. 5 and the increase in the ductility factor. The addition of fibres increases the absorbed energy and 1/B, most of all in SFRC with 1% by volume of steel fibres and higher in absolute values in SFRC with Dmax = 4 mm than in SFRC with Dmax = 8 mm. But the ductility factor 1/B of a polymer-modified SFRC with 1% by volume is less than 1/B of SFRC with 1% by volume without polymer although the total absorbed energy of the first SFRC is greater than the second. 1/B of polymer modified SFRC with 1% by volume is reduced due to the increase in the load at the first crack FFC and the expansion of the CMODFC. The added polymer (SBL) affects the increase of the absorbed energy of SFRC, so that the behaviour of SFRC VI/1p and SFRC VII/1p is similar to the behaviour of high performance SFRC. The correlations of fibre quantity and 1/B are similar to the correlations of fibre quantity and parameters DCMOD and WFZ with which 1/B is calculated. 60

ductility factor 1/B

50 40 30 20 10 VII/1p

VI/1p

VII/1

VI/1

VI/0p

VI/0.5p

VI/0

V/1

IV/1.5

IV/1

III/1.5

III/0.5

II/1.5

II/1

I/2

II/0

I/1.5

I/1

I/0.5

I/0

0

type of concrete Fig. 6. Average values of ductility factor 1/B of the concrete from the 1st and 2nd group.

Ductility of the Four-Year-Old Steel FRC

299

A comparison of the average values of 1/B of all concrete, given in Fig. 6, indicates a significant increase in 1/B of SFRC with slag and 1,5% by volume of steel fibres with lf/df = 87. An increase of 1/B of SFRC with Dmax = 4 and 8 mm and 1% by volume of steel fibres without or with SBL is also observed. This indicates that by increasing ratio of the fibre length and the largest grain diameter of the aggregate lf/Dmax, the ductility factor 1/B also increases. Both types of fibres used are 32 mm long, so by reducing the Dmax from 16 to 8 and 4 mm, the lf/Dmax ratio increases from 2 to 4 and 8.

4 Conclusions Based on the results of four-year-old SFRCs tested by the WST method and having high compressive strengths, the following main conclusions can be drawn: • ductility factor 1/B of SFRC increases if the absorbed energy of the fracture zone WFZ is increased sufficiently with respect to the increase in the absorbed energy of the elastic area or area up to the first crack WFC; • an increase in the ductility factor1/B is achieved by using increased quantities of efficient steel fibres (fibres with a larger aspect ratio lf/df), increasing the adhesion of the fibres and the matrix, and increasing the lf/Dmax ratio (ratio of the fibre length and the largest grain diameter of the aggregate); • behaviour of SFRC with slag aggregate and especially fibres with a larger aspect ratio lf/df as well as behaviour of polymer modified SFRC are similar to the behaviour of high performance SFRC.

References 1. Brühwiler, E., Wittman, F.H.: The Wedge splitting test, a new method of performing stable fracture mechanics tests. In: Rossmanith, H.P. (ed.) Fracture and Damage of Concrete and Rock, pp. 117 – 125. Pergamon Press (1988) 2. Linsbauer, H., Tschegg, E.K.: ‘Die Bestimmung der Bruchenegie an Würfelproben’ (Fracture Energy Determination of Concrete with Cube-Shaped Specimens). Zement und Beton, 31(1), 38 – 40 (1986) 3. Šušteršič, J., Kolenc, M., Zajc, A., Riček, F., Zajc, P.M.: High-performance fibre reinforced concrete for mine roadway support panels. In: Proceedings, Second CANMET/ACI International Conference Gramado, RS, Brazil,. SP-186, pp. 101 – 112 (1999) 4. Šušteršič, J., Ukrainczyk, V., Zajc, A., Šajna, A.: Evaluation of Crack Opening Resistance 8) of SFRC. In: Banthia, N., Sakai, K., Gjørv, O.E.: Concrete Under Severe Conditions. Vol. 2. Vancouver, pp. 1594 – 1601 (2001) 5. Šušteršič, J., Zajc, A., Leskovar, I., Dobnikar, V.: Improvement in the crack opening resistance of FRC with low content of short fibres. In: Dhir, Ravindra, K. (ed.) Role of concrete in sustainable development: proceedings of the International Symposium dedicated to professor Surendra Shah, Northwestern University, USA held on 3–4 September 2003 at the University of Dundee, Scotland, UK. London: ˝Thomas Telford˝, pp. 167-174 (2003)

300

J. Šušteršič et al.

6. Chen, L., Mindess, S., Morgan, D.R.: Toughness evaluation of steel fibre reinforced concrete. In: Proceedings of 3rd Canadian Symposium on Cement and Concrete, Ottawa, pp. 16 – 29 (1993) 7. Morgan, D.R., Mindess S., Chen L.: ‘Testing and Specifying Toughness for Fibre Reinforced Concrete and Shotcrete’. Fibre Reinforced Concrete. Modern Developments. Eds.: N. Banthia and S. Mindess. The University of British Columbia, Vancouver, pp. 29 – 50 (1995) 8. Banthia, N.: Advanced Fiber Reinforced Cement-based Composites for New Generation Safe, Sustainable and Smart Infrastructure. 25th Slovenian Colloquium on Concrete: Concrete with Improved Properties. Ljubljana, IRMA, pp. 3–8 (2018) 9. Plizzari, G.: Structural applications of Fibre Reinforced Concrete. In: 25th Slovenian Colloquium on Concrete: Concrete with Improved Properties. Ljubljana, IRMA, pp. 9–18 (2018) 10. Silfwerbrand, J.: Fibre Concrete – Swedish Research & Experience. 25th Slovenian Colloquium on Concrete: Concrete with Improved Properties. Ljubljana, IRMA, pp. 19–28 (2018) 11. Šušteršič, J., Zajc, A.: Review of Certain Applications of Fibre Reinforced Concrete in Slovenia. Fibre Concrete 2015. Prague (2015)

Sensitivity of the Flexural Performance of Glass and Synthetic FRC to Fibre Dosage and Water/Cement Ratio Razan H. Al Marahla and Emilio Garcia-Taengua(&) School of Civil Engineering, University of Leeds, Leeds, UK [email protected]

Abstract. A comparative analysis of the flexural performance of FRC mixes with either glass or synthetic fibers is presented in this paper. The data used for such analysis were obtained from an experimental programme which comprised 42 notched prismatic specimens, produced and tested to EN 14651 at the age of 28 days. Different fibre dosages up to 15 kg/m3 were considered in two series of mixes with water/cement ratios of 0.26 and 0.39, which yielded average compressive strength values of 65 MPa and 50 MPa respectively. A direct correlation between fibre content and residual flexural strength was confirmed. However, statistically significant differences were observed between the two fibre types considered. FRC mixes with glass fibre contents up to 5 kg/m3 failed immediately after the first crack and showed no residual flexural strength. In general, specimens reinforced with synthetic fibres showed better levels of residual flexural strength and toughness than their glass fibre counterparts. The ratio between the residual flexural strength and the limit of proportionality provides a good illustration of such observations: it was 61% on average for FRC mixes with synthetic fibers at 10 kg/m3, whilst it was only 33% for FRC mixes with the same amount of glass fibers. Keywords: Synthetic fibres


Glass fibres
Residual flexural strength

1 Introduction Fibre reinforced concrete (FRC) is well suited for a variety of structural as well as nonstructural applications, as concrete performance in the hardened state is generally enhanced by the fibres role in bridging cracks, which results in improved crack control and enhanced properties in the cracked state [1, 2]. For applications where the structural contribution of fibres is intended to be significant, steel fibres are generally preferred [3, 4]. Fibres made with other materials have lower tensile strength and elastic modulus than steel fibres, and they are typically considered for applications where their main contribution is concerned with restraining the plastic cracks, and shrinkage cracking control. However, there is no clear-cut separation that restricts non-steel fibres to non-structural applications, and this, together with current trends favouring materials with lower carbon footprint, has motivated increasing interest in the mechanical performance of FRC with synthetic or glass fibres. © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 301–312, 2021. https://doi.org/10.1007/978-3-030-58482-5_28

302

R. H. Al Marahla and E. Garcia-Taengua

Fibres partially counterbalance the brittleness which is intrinsic to concrete as a material [5, 6], but their positive influence on flexural toughness varies greatly depending on the type of fibre material, its dimensions, and the fibre dosage. A number of studies [7–9] have investigated the flexural performance and toughness of FRC made with different types of fibres, and there is clear consensus that the addition of increasing fibre contents improves residual flexural strength and toughness. To characterise, specify and design with FRC, the residual flexural strength parameters or toughness as obtained from the bending test are used as reference, in addition to the compressive strength at 28 days. In this study, the flexural response of different FRC mixes with either glass or synthetic fibres was characterised by testing notched prismatic specimens under three point bending test conditions, following the standard EN 14651 [10]. Different fibre dosages were considered. Also, to evaluate the influence of water-to-cement (w/c) ratio on flexural toughness, and how this modifies the contribution of fibres to residual flexural strength, two reference concrete mix designs were considered.

2 Experimental Programme 2.1

Variables Considered

The factors considered in this study were: w/c ratio, type of fibre, and fibre content. Cement type CEM I 52.5 N was used, and the superplasticiser was Sika Viscocrete 25MP. Two different values were considered for the w/c ratio: 0.26 and 0.39, leading to the definition of two reference mix designs, which were adjusted for a slump value of 120-150 mm so they could incorporate different fibre contents without further adjustments. These reference mix designs are summarised in Table 1, and were intended to be representative of a range of specified compressive strengths between 4560 MPa approximately. Table 1. Reference mix designs (kg/m3) Cement Water Fine aggregate Coarse aggregate (10 mm) Coarse aggregate (20 mm) Superplasticiser

w/c = 0.39 w/c = 0.26 440 510 175 130 825 950 637 580 317 300 8 11

The two types of fibres considered in this study were: synthetic fibrillated macrofibres and alkali-resistant glass macro-fibres, and they are shown in Fig. 1. The synthetic fibres were high-modulus 54-mm long polymeric fibres and had a tensile strength of 600 MPa. The glass fibres, on the other hand, were 36 mm long and their tensile

Sensitivity of the Flexural Performance of Glass

303

strength was 1700 MPa. Synthetic fibres were used in dosages of 5, 7.5, and 10 kg/m3, whilst the dosages considered for the glass fibres were 2.5, 5, and 15 kg/m3.

Fig. 1. Samples of the glass fibres (left) and synthetic fibres (right) used in this study.

2.2

Production of FRC Mixes

Taking as reference the two mix designs presented in Table 1, either glass or synthetic fibres were incorporated in different dosages, leading to the combinations summarised in Table 2. Table 2. Summary of the FRC mixes considered in this study. Fibre type w/c ratio Fibre content (kg/m3) (None) 0.26 0.0 0.39 0.0 Glass 0.26 2.5 5.0 15.0 0.39 2.5 5.0 15.0 Synthetic 0.26 5.0 7.5 10.0 0.39 5.0 7.5 10.0

The same mixing sequence was followed in the production of all mixes. In preparation before the mixing of every batch, the total amount of water to be added was separated in two buckets: one containing 80% of the water, and the other containing the mixture of the remaining 20% of the water and the required amount of superplasticiser. First, cement and all aggregates were all poured into the mixer and dry-mixed for 2 min. After that, 80% of the water was added and mixed with the cement and

304

R. H. Al Marahla and E. Garcia-Taengua

aggregates for 3 min. During this time, the fibres were poured gradually into the mixer. Finally, the remaining 20% of the water with the superplasticiser predispersed in it was added, and the mixing continued for 4 min. In all cases, a uniform distribution of the fibres in the mix was observed. 2.3

Characterisation of FRC Mixes

For each of the combinations presented in Table 2, one batch of 70 litres was produced, and this material was used to cast 3 cubes, 3 cylinders, and 3 prismatic specimens. All specimens were tested at the age of 28 days. Prior to this, all specimens were kept in a fog room with a controlled temperature of 20C and a relative humidity of 90%. The cubes were 100-mm side and were used to determine the compressive strength, whilst the 150  300 mm cylinders were used to determine the splitting tensile strength following the standard EN 12390-6:2009 [11]. The prismatic specimens were produced and tested in flexure, according to the standard EN 14651 [10], and were 150 mm side and 600 mm long. They were all notched with a notch depth of 25 mm and 5 mm width using a wet sawing machine, and cured for a minimum of 3 days after sawing according to EN 12390-2 [12] until the age of testing. For the flexural strength test, a three point bending scheme was used, an image of which is shown in Fig. 2. The tests were carried by imposing a constant rate of 0.05 mm/min for the increasing CMOD until the CMOD reached 0.1 mm, after which the rate was increased to 0.2 mm/min. CMOD values were monitored by means of LVDTs placed right under the notch, and a K7500 service controller was used for the data acquisition and control signal.

Fig. 2. Set up for tested specimen

Sensitivity of the Flexural Performance of Glass

305

3 Test Results and Discussion 3.1

Compressive and Splitting Tensile Strength

The results of the compressive strength and splitting tensile strength tests are presented in Table 3. The average compressive strength was 67 MPa and 50.4 MPa for the reference mixes without fibres and w/c ratios of 0.26 and 0.39 respectively. Table 3. Compressive and splitting tensile strength results. Fibre type w/c ratio Fibre content (kg/m3) Compr. strength (MPa) Average Std. dev. (None) 0.26 0.0 67.0 1.0 0.39 0.0 50.4 0.6 Glass 0.26 2.5 66.6 1.5 5.0 66.8 0.6 15.0 67.5 0.7 0.39 2.5 50.8 1.9 5.0 51.3 1.4 15.0 52.2 0.8 Synthetic 0.26 5.0 67.3 0.9 7.5 67.5 0.6 10.0 67.6 0.3 0.39 5.0 51.3 1.4 7.5 52.4 0.8 10.0 52.2 0.8

Split. tensile strength (MPa) Average Std. dev. 4.2 0.2 3.8 0.6 3.9 0.6 4.0 0.6 4.0 0.4 3.6 0.5 3.7 0.8 3.8 0.5 4.1 0.4 4.1 0.3 4.0 0.3 3.7 0.4 3.7 0.7 3.6 0.4

The addition of glass or synthetic fibres, at the dosages considered in this study, did not introduce significant variations in terms of average compressive strength, and the same can be said in relation to splitting tensile strength. Although fibres have been reported to sometimes increase the compressive strength of concrete by up to 15%, ACI 544 [13], these results are in agreement with those other studies that reported no significant improvements in compressive strength due to the presence of fibres [14, 15]. In terms of variability, low fibre contents were observed to increase the standard deviation of compressive strength values with respect to the reference mixes without fibres. However, increasing fibre contents were associated with decreasing standard deviation values, for both types of fibres and w/c ratios. 3.2

Bending Test Results

For each of the FRC mixes as per the combinations listed in Table 2, three prismatic specimens were tested under flexure to EN 14651 [10], and the corresponding loadCMOD curves were obtained. From these curves, the equivalent stress corresponding

306

R. H. Al Marahla and E. Garcia-Taengua

to the limit of proportionality, fL, and the residual strength parameters fR1, fR2, fR3, and fR4 (corresponding to CMOD values of 0.5, 1.5, 2.5, and 3.5 mm, respectively) were obtained. Their average values are given in Table 4. Table 4. Bending test results: average values, expressed in MPa. Fibre type Fibre content (kg/m3) w/c ratio Glass 2.5 0.26 0.39 5 0.26 0.39 15 0.26 0.39 Synthetic 5 0.26 0.39 7.5 0.26 0.39 10 0.26 0.39

fL 5.68 5.69 6.96 6.31 5.41 5.67 7.45 6.24 7.48 6.79 6.68 7.43

fR1 0.00 0.00 0.00 0.00 2.49 2.61 2.66 2.12 3.10 2.53 3.17 2.78

fR2 0.00 0.00 0.00 0.00 2.00 1.74 2.56 2.08 2.84 2.50 3.10 2.60

fR3 0.00 0.00 0.00 0.00 2.00 1.61 2.53 2.14 2.79 2.44 3.09 2.61

fR4 0.00 0.00 0.00 0.00 1.93 1.57 2.42 2.18 2.75 2.39 3.08 2.58

All specimens produced with FRC mixes containing glass fibres at dosages of 2.5 and 5.0 kg/m3 failed without exhibiting any residual load-bearing capacity in flexure, regardless of the w/c ratio. A softening response was observed in specimens where the glass fibres dosage was 15 kg/m3. A graphical comparison of the fL values corresponding to the FRC mixes with glass fibres, and the stress-CMOD curves corresponding to specimens with a glass fibre dosage of 15 kg/m3 are shown in Fig. 3.

Fig. 3. Flexural test results for glass FRC mixes: limit of proportionality values (left) and stressCMOD curves for 15 kg/m3 fibre content.

Sensitivity of the Flexural Performance of Glass

307

On the other hand, specimens with synthetic fibres presented comparatively better performance than their glass fibres counterparts, as residual flexural capacity was developed for all fibre dosages and w/c ratios considered. Figure 4 shows the average stress-CMOD curves for the different dosages of synthetic fibres and w/c ratios considered. Furthermore, comparing the results corresponding to those mixes where fibres were added at the maximum dosages considered in this study (10 kg/m3 for synthetic fibres, 15 kg/m3 for glass fibres), synthetic fibres did better than glass fibres in terms of the level of residual flexural strength achieved. The incorporation of synthetic fibres at a dosage of 10 kg/m3 led to fR1/fL ratios between 0.42 and 0.61, for w/c ratios of 0.39 and 0.26 respectively, as opposed to 0.26 and 0.33, corresponding to specimens with 15 kg/m3 of glass fibres.

Fig. 4. Flexural test results for glass FRC mixes with w/c = 0.39 (left) and w/c = 0.26 (right).

3.3

Residual Flexural Strength

In order to better quantify the differences in the flexural response of FRC introduced by changes in the variables considered in this study, a regression analysis was done and equations for the residual flexural strength parameters were obtained. In the regression analysis, the interactions between fibre content, fibre type and w/c ratio were also considered, with the purpose of determining whether these synergies were statistically significant. The equations obtained are presented in Table 5, where Cf is the fibre content in kg/m3. The R-squared values ranged between 0.91 and 0.97, which indicated a very accurate fit with the experimental results. Unsurprisingly, fibre content and the fibre type were found to have a statistically significant effect on all residual flexural strength parameters. In addition to that, the interaction between fibre type and fibre content was found to be statistically significant. That is, the effect that a certain increase in the fibre content had on residual flexural strength was modified depending on the type of fibre considered. Regarding the effect of the w/c ratio, it was found that it had a statistically significant effect on fR1 and fR2. However, the regression analysis showed that varying the w/c ratio did not introduce statistically significant variations in fR3 and fR4.

308

R. H. Al Marahla and E. Garcia-Taengua Table 5. Regression equations obtained for the residual flexural strength parameters. FRC with glass fibres FRC with synthetic fibres Lnðf R1 Þ ¼ 3:845 þ 0:380Cf  3:31w=c Lnðf R1 Þ ¼ 1:733 þ 0:044Cf  3:31w=c Lnðf R2 Þ ¼ 3:557 þ 0:314Cf  1:89w=c Lnðf R2 Þ ¼ 1:258 þ 0:041Cf  1:89w=c Lnðf R3 Þ ¼ 4:097 þ 0:305Cf Lnðf R3 Þ ¼ 0:647 þ 0:040Cf Lnðf R4 Þ ¼ 4:650 þ 0:335Cf Lnðf R4 Þ ¼ 0:628 þ 0:041Cf

The equations presented in Table 5 were useful in visualising the sensitivity of the different residual flexural strength parameters to changes in w/c ratio, fibre type, and fibre content. Figure 5 shows the contour plots for fR1 values corresponding to mixes reinforced with either glass fibres (left) or synthetic fibres (right), as a function of w/c ratio and the fibre content. For the mixes with glass fibres, a theoretical minimum for the fibre content could be identified in order for the material to present residual flexural capacity. This minimum glass fibre content was in the range of 10 to 12 kg/m3 and was slightly dependent on the w/c ratio.

Fig. 5. Contour plots for fR1 (in MPa) as a function of w/c ratio, type of fibre, and fibre content.

The effect of w/c ratio was much more marked in mixes with synthetic fibres. As the contour plot in Fig. 5 (right) shows, reducing the w/c ratio could lead to improvements in fR1 comparable to those which would be achieved by increasing the amount of synthetic fibres in the mix. Contour plots for fR2 are not shown in this paper because they were very similar to those obtained for fR1 and led to similar conclusions. Figure 6 shows the trend followed by fR3 and fR4 values with respect to increasing fibre contents, for the two types of fibres considered. As mentioned before, fR3 and fR4 were not sensitive to variations in the w/c ratio within the range considered in this study. It can be seen that, for the mixes with synthetic fibres, the relationship between fR3 or fR4 values and the fibre content was practically linear. Also, in all cases considered in this study there was almost no difference between fR3 and fR4 values.

Sensitivity of the Flexural Performance of Glass

309

Fig. 6. Regression lines for fR3 and fR4 as a function of fibre content.

3.4

Flexural Toughness

In order to obtain an indication of the flexural toughness from the bending test results, the areas under the stress-CMOD curves up to a CMOD value of 3.5 mm were calculated. The average toughness values obtained for each case considered in this study are summarised in Table 6. Values for glass fibre contents lower than 15 kg/m3 are not presented, as those cases exhibited a brittle failure. Table 6. Toughness (mm.MPa) Fibre type Fibre content (kg/m3) w/c = 0.39 w/c = 0.26 Glass 15.0 6.95 8.03 Synthetic 5.0 7.78 9.52 7.5 9.26 10.57 10.0 10.52 11.51

The toughness values corresponding to synthetic FRC mixes were analysed by means of an analysis of variance, and this showed that the interaction between the fibre content and the w/c ratio had a statistically significant effect. The response surface shown in Fig. 7 was obtained, and the significance of such interaction is clearly noticeable: the trend followed by toughness values with respect to the fibre content changes depending on the w/c ratio, and vice versa. Reducing the w/c ratio from 0.39 to 0.26 increases toughness values by 22% if a fibre content of 5 kg/m3 is considered; however, this increase is 14% or 9.5% if the fibre content is 7.5 kg/m3 or 10 kg/m3, respectively. In consequence, it was concluded that mixes with higher synthetic fibre contents were less sensitive to changes in the w/c ratio, in terms of their flexural toughness. Conversely, increasing the synthetic fibre content led to increased toughness values, but the magnitude of such an increase was found to depend on the w/c ratio. For instance, for a w/c ratio of 0.39, toughness values were increased by 35% as a result of increasing the fibre content from 5 kg/m3 to

310

R. H. Al Marahla and E. Garcia-Taengua

Fig. 7. Contour plot for toughness values (area under stress-CMOD curve, in mmMPa).

10 kg/m3; but this increase was 21% instead of 35% if the w/c ratio was 0.26 instead of 0.39.

4 Conclusions Different FRC mixes were produced and tested in order to evaluate the sensitivity of their mechanical properties to changes in the following parameters: w/c ratio (0.26 or 0.39), type of fibre (glass or synthetic macrofibres), and fibre content (up to 10 kg/m3 or 15 kg/m3 for synthetic or glass fibres, respectively). The following conclusions were obtained: • The addition of the glass or synthetic fibres considered in this study, irrespective of their dosage, did not cause statistically significant changes to the average compressive strength or splitting tensile strength. However, increasing the fibre content was found to reduce the standard deviation of compressive strength values. • FRC specimens with glass fibres in dosages of 2.5 kg/m3 and 5 kg/m3 presented brittle failure in flexure, and only those with 15 kg/m3 of glass fibres showed residual flexural capacity. Based on the results obtained, it was estimated that glass fibres need to be added in contents of at least 10–12 kg/m3 in order to achieve residual flexural capacity. • No cases of brittle failure in bending were observed amongst the FRC specimens with synthetic fibres. They all presented a softening post-peak behaviour, irrespective of the w/c ratio and fibre content. • When glass and synthetic fibres are compared at the maximum contents considered in this study, synthetic FRC specimens clearly outperformed their glass FRC counterparts. The level of residual flexural strength, represented by the ratio fR1/fL, varied between 0.42-0.61 for synthetic fibres, as opposed to 0.26-0.33 for glass fibres. • The residual flexural strength parameters fR1 and fR2 were found to be sensitive to changes in w/c ratio, fibre type and fibre content. It was concluded that reducing the

Sensitivity of the Flexural Performance of Glass

311

w/c ratio could lead to improvements in fR1 and fR2 comparable to those achieved by increasing the fibre content. • Flexural toughness was evaluated through the area under the stress-CMOD curves up to a crack opening of 3.5 mm. Increasing the synthetic fibre content was found to decrease the sensitivity of this parameter to variations in the w/c ratio. Acknowledgements. The authors wish to acknowledge the contribution of Oscrete Construction Products, part of Christeyns UK Ltd, and Sika Ltd (UK), which very kindly provided some of the materials used in this study, as well as the support and assistance provided by the technical staff of the School of Civil Engineering, University of Leeds. The authors are also thankful to AlZaytoonah University of Jordan for the financial support granted to Ms Al Marahla in undertaking her PhD studies at the University of Leeds.

References 1. Biolzi, L., Cattaneo, S., Guerrini, G.L.: Fracture of plain and fiber-reinforced high strength mortar slabs with EA and ESPI monitoring. Appl. Composite Mater. 7(1), 1–12 (2000) 2. Kazemi, M.T., Golsorkhtabar, H., Beygi, M.H.A., Gholamitabar, M.: Fracture properties of steel fiber reinforced high strength concrete using work of fracture and size effect methods. Constr. Build. Mater. 142, 482–489 (2017) 3. Soutsos, M.N., Le, T.T., Lampropoulos, A.P.: Flexural performance of fibre reinforced concrete made with steel and synthetic fibres. Constr. Build. Mater. 36, 704–710 (2012) 4. Cho, B., Lee, J.H., Back, S.Y.: Comparative study on the flexural performance of concrete reinforced with polypropylene and steel fibers. J. Korean Soc. Civil Eng. 34(6), 1677–1685 (2014) 5. Olivito, R.S., Zuccarello, F.A.: An experimental study on the tensile strength of steel fiber reinforced concrete. Composites Part B: Eng. 41(3), 246–255 (2010) 6. Thomas, J., Ramaswamy, A.: Mechanical properties of steel fiber-reinforced concrete. J. Mater. civil Eng. 19(5), 385–392 (2007) 7. Simoes, T., Costa, H., Dias-da-Costa, D., Júlio, E.N.B.S.: Influence of fibres on the mechanical behaviour of fibre reinforced concrete matrixes. Constr. Build. Mater. 137, 548– 556 (2017) 8. Lee, J.H.: Influence of concrete strength combined with fiber content in the residual flexural strengths of fiber reinforced concrete. Composite Struct. 168, 216–225 (2017) 9. Buratti, N., Mazzotti, C., Savoia, M.: Post-cracking behaviour of steel and macro-synthetic fibre-reinforced concretes. Constr. Build. Mater. 25(5), 2713–2722 (2011) 10. BS EN 14651, Test Method for Metallic Fibre Concrete-Measuring the Flexural Tensile Strength, British Standard Institute, UK, pp. 1–20 (2007) 11. BS EN 12390–6, Testing Hardened Concrete. Tensile Splitting Strength of Test Specimens. British Standard Institution, London (2009) 12. BS EN 12390–2, Testing hardened concrete-Part 2: Making and curing specimens for strength tests, European Committee for Standardization, Brussels (2009) 13. ACI committee, ACI 544.1 R-96, State-of-the-art report on fiber reinforced concreteTechnical report, ACI Farmington Hills, Michigan (2003)

312

R. H. Al Marahla and E. Garcia-Taengua

14. Mazaheripour, H., Ghanbarpour, S., Mirmoradi, S.H., Hosseinpour, I.: The effect of polypropylene fibers on the properties of fresh and hardened lightweight self-compacting concrete. Constr. Build. Mater. 25(1), 351–358 (2011) 15. Cifuentes, H., García, F., Maeso, O., Medina, F.: Influence of the properties of polypropylene fibres on the fracture behaviour of low, normal and high-strength FRC. Constr. Build. Mater. 45, 130–137 (2013)

Bond Between Steel Reinforcement Bars and Fiber Reinforced Cement-Based Composites Margareth S. Magalhães(&), Paulo José B. Teixeira, and Maria Elizabeth N. Tavares Civil Engineering Postgraduate Program, State University of Rio de Janeiro, UERJ, Rio de Janeiro , Brazil [email protected]

Abstract. This paper deals with the bond between steel reinforcement and strain hardening cement-based composites (SHCC). Pull-out tests were carried out according to RILEM standard on specimens made with three mixtures, characterized by different fiber content (1% and 2%) beyond the control plain mortar. Ribbed steel reinforcement bars with 8 mm and 10 mm diameter were used to observe the influence of steel bar diameter. Experimental bond-slip relationships were analysed, and results show enhanced bond resistance when fiber is used in the mixture. SHCC specimens (composite with fiber content equal to 2%) presented the best bond performance in terms of bond strength and stiffness retention capacity, as well as damage control ability. Keywords: Strain hardening cement-based composites behavior
Pull-out test
Steel reinforcement


PVA fiber
Bond

1 Introduction Strain hardening cement-based composite (SHCC) is a kind of composite material that display a ductile behavior when subjected to tensile loading, different to other fiber reinforced cementitious composites and ordinary concrete. SHCC can resist the full tensile load with strain capacity up to 5.0%, leading to a high energy absorption capacity. During the increase of strain under tensile loading, a strain hardening effect is found and many, closely spaced, micro cracks (less than 70 µm in width) are formed in the material [1, 2]. The compressive behavior of SHCC is also more ductile compared to ordinary concrete, but the real benefit is attained when SHCC is used to resist a force or imposed tensile strain. Lately, SHCC have been widely utilized in civil engineering due to the remarkable mechanical properties, especially in flexural and shear dominated members, such as coupling beams, low-rise walls and beam-column joints [3–8]. Findings suggested that SHCC can effectively improve bond efficiency of reinforcement by reducing slippage of column bars [5], enhance the shear strength, energy dissipation and damage tolerance of members, and consequently improve the performance of reinforced SHCC structures. The experiences of current applications of SHCC are of paramount importance. © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 313–321, 2021. https://doi.org/10.1007/978-3-030-58482-5_29

314

M. S. Magalhães et al.

However, review on the existing literature related to SHCC indicates that the research on the bond behavior between steel reinforcement and SHCC is limited [8, 10–15]. Bond strength is one of the most important parameters in reinforced elements design, both at the ultimate and serviceability limit state. Poor bond property weakens the bearing capacity of a component and then leads to structural failure. Therefore, effective bonding between steel bars and SHCC is crucial for the components to function in a stable and collaborative manner. According to Fischer and Li [9], by using SHCC, it is possible to reduce the slip of steel bars and improve interfacial bonding between steel bar - SHCC. Chao et al. [8] observed that the composite type plays a vital role in bond behavior of steel bar - composites. According Toshiyuki and Hiroshi [10], by using SHCC is possible to reduce the cover thickness and bar spacing of the main bars in concrete members. Lee et al. [11] proposed a constitutive model applicable to the bond-slip behavior between the ECC, a kind of SHCC, and steel bar and Li et al. [12] observed that the bond strength decreased as the diameter and embedded length of steel bars increased. Other studies were conducted on bond between fiber (basalt or glass) reinforced polymer bar and ECC [16, 17]. Wang et al. [16], based on the pull-out test of basalt fiber reinforced polymer (BFRP) bar embedded in ECC, pointed out that adding PVA fiber can enhance the damage tolerance of matrix and thus improve the bond behavior. Kim and Lee [17] investigated the bond behavior between glass fiber reinforced polymer (GFRP) bar and ECC. Test results showed that energy absorption capacity of specimens was enhanced due to the addition of PVA fiber and the ductility of bond stress-slip curves significantly increased as fiber volume content increased from 1% to 2%. The purpose of the research described in this paper is to gain a better understanding on the bond behavior between steel reinforcement and SHCC. Pull-out tests were carried out on specimens made with three different mixtures, characterized by different fiber content. Ribbed steel reinforcement bar with two different diameters were used to observe the influence of steel bar diameter.

2 Experimental Program 2.1

Materials and Composite Manufacturing

The raw materials used on the preparation of the SHCC specimens, named C02 mixture, were Portland cement CPII F-32, defined by the Brazilian standard [18], composed of filler (in mass: 6–10%) with a 28 days-compressive strength of 32 MPa; fly ash with a density of 2.26 g/cm3; silica sand with a maximum diameter of 0.30 mm; a superplasticizer and water. The fiber used was a polyvinyl alcohol (PVA) fibre, manufactured by kuraray company, Japan, with a diameter of 40 µm, density of 1.31 g/cm3, elastic modulus of 40 GPa and tensile strength of 1600 MPa [19]. To study the influence of the PVA fiber content on the bond behavior of steel barcomposites, other two mixtures were produced: one mixture without fiber (CM) and another mixture with 1% PVA fiber (C01). The mixtures proportion used in this study are shown in Table 1.

Bond Between Steel Reinforcement Bars

315

Table 1. Mixture design (kg/m3) of composite. CM C01 Cement 505 505 Fly ash 605 605 Sand 404 404 Water 404 404 PVA fiber – 13 Superplasticizer/(cement + fly ash) (%) – 0.12 Fly ash/cement 1.20 1.20 Water/(cement + fly ash) 0.36 0.36

C02 505 605 404 404 26 0.13 1.20 0.36

To produce the mixtures, all dry raw materials, except fibers, were mixed for 3 min in a mechanical mixer with a 20 L capacity. Water and superplasticizer were added to form the basic matrix. The mixture was stirred for another 8 min to allow appropriate workability of the matrix. In the last step, when used, fibers were added manually to the cementitious matrix and the mixture was stirred for another 5 min. The formulation of the C02 mixture, given in Table 1, was developed by Magalhães and co-workers in earlier studies [2]. The composite exhibits a strain-hardening behavior with average tensile strain capacity of approximately 3% and average crack width around of 68 µm (see Fig. 1).

Fig. 1. Typical tensile stress – strain and average crack width - strain curves of SHCC [2].

2.2

Test Procedure

Compressive strength tests at 28 days of mixes were carried out on 3 cylindrical specimens of 50  100 mm (diameter  height). Ten pull-out specimens per each mixture were casted with a central horizontal bar. A ribbed steel rebar with a nominal yield tensile strength of 500 MPa was used. The

316

M. S. Magalhães et al.

casting direction was perpendicular to the longitudinal axis of the bar. Cubes’ side (100 mm) was 10 times the diameter of the bar (10 mm), and the embedded length (50 mm) was 5 times the diameter of the bar, following RILEM guidelines [20]. A plastic sleeve was used to limit the non-adherent zone (50 mm), situated at the loading face. The same procedure was used to all tests with the three mixes and repeated for another test set with C02 mix and bar diameter equal to 8 mm. Pull out tests were carried out at 28 days after casting: the test setup is shown in Fig. 2. A universal testing machine with a capacity of 600 kN was used to perform the tests, and displacement-control was applied to capture the post-peak behavior. The lower surface of the cube was restrained by a stiff 15 mm steel plate, with a hole of 32 mm diameter in the center. Between the specimen and the plate, a thin layer of rubber of 5 mm was placed, to ensure that a uniform contact was realized and to minimize friction effects.

Fig. 2. Setup for pull-out tests.

3 Experimental Results and Discussion Test results are given in Table 2 with regards to compressive strength (fc), ultimate bond strength (su), corresponding to the peak bond stress, and the corresponding slip (su), residual bond strength (sr) and the failure modes. The residual bond strength was obtained at a slip of 7 mm. The bond stress between anchorage steel and composite was calculated as follows: s¼

F pdla

ð1Þ

where s is the bond stress, F is the applied load, d is the bar diameter and la is the anchorage length. Two different types of failures, such as pull-out failure and slitting failure, were observed in the tested specimens. Pull-out mode generally occurs in confined specimen

Bond Between Steel Reinforcement Bars

317

Table 2. Summary of results from compressive and pull-out test (standard deviation in parenthesis). Mix CM-10 C01-10 C02-10 C02-8

fc (MPa) 39.83(0.58) 38.99 (0.20 36.27 (0.15) 36.27 (0.15)

su (MPa) 11.28 (0.43) 11.42 (0.68) 11.86 (1.76) 9.23 (1.72)

Su (mm) 0.41 (0.11) 1.14 (0.61) 1.30 (0.18) 1.09 (0.11)

sr (MPa)

Failure mode Splitting 2.99 (0.65) pull-out 3.06 (0.68) pull-out 2.86 (0.51) pull-out

in which the bond failure is due to the pull-out of the bars. In this research, specimens with 1% and 2% fiber showed appreciable damage tolerance and maintained their integrity throughout the test due to the fiber bridging effect, which prevented internal cracks from opening widely. A clean bar was pulled out in most specimens and no cracks (visible to the naked eye) were observed on the surface of the specimens, as shown in Figs. 3b, c, d, e. In contrast, the confinement was not sufficient to prevent splitting of mortar cover and a longitudinal cracking failure occurred in all mortar specimens without fibers (CM-10), which exhibited an obvious brittle feature (see Fig. 3a).

(a)

(b)

(c)

(d)

Fig. 3. Typical specimens after tests: (a) CM-10, (b) C01-10, (c) C02-10 and (d) C02-8.

The concrete-steel bars bond is considered to stem from three main factors: chemical bond, friction force and the mechanical thread strengths, which is the mainly responsible by the adhesion in ribbed steel bars. Figure 4 shows the bond stress versus slip typical curves (on the left) and a zoom of the pre-peak zone (on the right) to clearly highlight the ascending branch, for the specimens with 10 mm diameter bars. At the initial loading portion, up to approximately 2–3 MPa, the curves showed a relatively sharp slope for all specimens, indicating the load transfer by the chemical adhesion. After, the ascending portion of the curves became nonlinear, up to the applied load reached the maximum value. From Table 2, it can be observed that mortar mixture (CM–10) displayed insignificant reduction on the peak stress as compared with the composite specimens (C01–10 and C02–10) due to splitting of mortar cover. The difference observed on the ultimate bond strength between mortar (CM–10) and composites (C01–10 and C02–10) was up to –5.2%.

318

M. S. Magalhães et al.

(a)

(b)

Fig. 4. Bond stress-slip curves: (a) effect of fiber content and (b) a zoom of the pre-peak zone.

Comparing the slip values, corresponding to the ultimate bond stress, a significant increase of this bond characteristic is obtained for the C01–10 (1.14 mm) and C02–10 (1.30 mm) mixes, if compared to the other mix without fiber, CM–10 (0.41), that are instead characterized by lower value (0.41 mm). This result indicates that, the confinement effect in the mortar was enhanced due to fiber addition. The fiber has effectively controlled the crack opening and widening and then improving the bond behavior of mortar specimens that failed due to splitting cracks occurrence. Regarding to the fiber content (mixes C01–10 and C02–10), less differences was observed in the ultimate bond strength values of the specimens. This can be assigned to the fact that pull-out strength is mainly governed by composite compressive strength, which is very similar for the specimens (C01: 38.99 MPa and C02: 36.27 MPa), being mostly affected by the similarity of the cementitious matrix and little effect of increasing fiber content. The contribution of PVA fibre subsequent to matrix cracking was more apparent in the post-peak behavior, mainly in the SHCC specimen (C02–10 mix). After ultimate bond strength, the bearing capacity of CM–10 specimen has a sudden loss. As contrast, the curve of specimens with fibers declines slowly which exhibits better ductility performance. With the further increase in slippage, the bond stress of C02–10 gradually tends to be steadier than C01–10 due to higher fiber content. Indeed, the average residual bond strength of C02–10 is slightly higher than the value of C01–10 (+2.4%). The contribution of the fibers resulted in the improvement of the toughness compared to the specimen with no fibers. The properties and content of fibers plays a very important role in controlling the development of cracking. Therefore, the contribution of fibers to interfacial bond property during the crack development and propagation can be summarized as the totally dissipated bond energy and it can be evaluated by determining and evaluating the toughness at different slips value. The toughness of the composites was calculated from the area under the bond stress–slip curves up to slips values of Su (Tu), 2 mm (T2), 4 mm (T4), 6 mm (T6) and 7 mm (T7) and the average values are presented in Table 3. In this aspect, the PVA fibers with volume fraction of 2% in total volume (C02–10) showed a good result on the improvement of the

Bond Between Steel Reinforcement Bars

319

Table 3. Average toughness from pull-out test results (standard deviation in parenthesis). Toughness (KJ/m2) T2 T4 Tu C01-10 12.70 (6.73) 20.73 (1.34) 38.77 (0.82) C02-10 13.35 (2.11) 22.95 (3.89) 42.17 (5.97) C02-8 7.88 (2.38) 14.76 (4.46) 25.40 (8.33)

Mix

T6 49.66 (1.32) 52.63 (7.63) 30.61 (10.58)

T7 53.04 (1.91) 55.86 (8.18) 32.76 (11.38)

toughness at any slip level and maintenance of the residual load (see Table 2) for 10 mm bar. The Fig. 5 indicate that the basic shape of the bond stress–slip curve of SHCC specimens is not affected by the bar diameter, however the bond stress of the C02–10 specimen was always higher than that of the specimen C02–8, at any given slip, due to an increase of effective bonding area in bars with higher diameters, which leads to an increase in bond strength. This finding is contrary to those obtained in other studies [12, 15]. The results, in Table 2, also showed that the slip at peak load and residual of C02 specimen decreased, respectively, 21.6% and 6.5% with a decrease of the diameter of the steel bar. Furthermore, the toughness values, as seen in Table 3, were also reduced up to approximately 42%.

Fig. 5. Bond stress-slip curves: (a) effect of fiber content, (b) effect of steel bar diameter.

4 Conclusions The purpose of the research described in this paper was to gain a better understanding on the bond behavior between steel reinforcement and composite. Most relevant conclusions of the scientific research are: • Two different types of failures were observed in the tested specimens: ductile failure with steel bar pull-out in composite specimens and brittle failure in matrix specimens. • The increase in the content of PVA fiber don’t affect the peak stress of the composites, however the slip values, corresponding to the ultimate bond stress, increase.

320

M. S. Magalhães et al.

• The fiber has effectively controlled the crack opening and widening and then improving the bond behavior of mortar specimens that failed due to splitting cracks occurrence. • Bond strength, residual bond strength and toughness increase as the diameter of steel bars increase. Acknowledgements. The authors acknowledge the Foundation for Research Support of the State of Rio de Janeiro (FAPERJ) for the financial support and Pozo Fly for supplying the fly ash.

References 1. Li, V.C., Wu, C., Wang, S., Ogawa, A., Saito, T.: Interface tailoring for strain hardening PVA-ECC. ACI Mater. J. 99(5), 463–472 (2002) 2. Magalhães, M.S., Toledo Filho, R.D., Fairbairn, E.M.R.: Influence of local raw materials on the mechanical behaviour and fracture process of PVA-SHCC. Mater. Res. J. 17(1), 146–156 (2014) 3. Kunieda, M., Rokugo, K.: Recent progress of SHCC in Japan – required performance and applications. J. Adv. Concr. Technol. 4(1), 19–33 (2006) 4. Lim, Y.M., Li, V.C.: Durable repair of aged infrastructures using trapping mechanism of engineered cementitious composites. Cem. Concr. Compos. 19(4), 373–385 (1997) 5. Parra-Montesinos, G.J.: High-performance fiber-reinforced cement composites: an alternative for seismic design of structures. ACI Struct. J. 102(5), 668–675 (2005) 6. Deng, M.K., Zhang, Y.X.: Cyclic loading tests of RC columns strengthened with high ductile fiber reinforced concrete jacket. Constr. Build. Mater. 153, 986–995 (2017) 7. Fischer, G., Li, V.C.: Deformation behavior of fiber-reinforced polymer reinforced Engineered Cementitious Composite (ECC) flexural members under reversed cyclic loading conditions. ACI Struct J. 100(1), 25–35 (2003) 8. Chao, S.H., Naaman, A.E., Parra-Montesinos, G.J.: Bond behavior of reinforcing bars in tensile strain-hardening fiber-reinforced cement composites. ACI Struct. J. 106(6), 897–906 (2009) 9. Fischer, G., Li, V.C.: Effect of matrix ductility on deformation capacity behavior of steelreinforced ECC flexural members under reversed cyclic loading conditions. ACI Struct J. 99 (6), 781–790 (2002) 10. Toshiyuki, K., Hiroshi, H.: Bond-splitting strength of reinforced strain-hardening cement composite elements with small bar spacing. ACI Struct J. 112(2), 189–198 (2015) 11. Lee, S.W., Kang, S.B., Tan, K.H., Yang, E.H.: Experimental and analytical investigation on bond-slip behaviour of deformed bars embedded in engineered cementitious composites. Constr. Build. Mater. 127, 494–503 (2016) 12. Li, X.L., Bao, Y., Xue, N., et al.: Bond strength of steel bars embedded in high performance fiber-reinforced cementitious composite before and after exposure to elevated temperatures. Fire Saf. J. 92, 98–106 (2017) 13. Bandelt, M.J., Frank, T.E., Lepech, M.D., et al.: Bond behavior and interface modeling of reinforced high-performance fiber-reinforced cementitious composites. Cem. Concr. Compos. 83, 188–201 (2017) 14. Deng, M., Pan, J., Sun, H.: Bond behavior of steel bar embedded in Engineered Cementitious Composites under pullout load. Constr. Build. Mater. 168, 705–714 (2018)

Bond Between Steel Reinforcement Bars

321

15. Asano, K., Kanakubo, T.: Study on size effect in bond splitting behavior of ECC. High Perform. Fiber Reinforced Cem. Compos. 6, 137–144 (2012) 16. Wang, H., Sun, X., Peng, G., et al.: Experimental study on bond behaviour between BFRP bar and engineered cementitious composite. Constr. Build. Mater. 95(1), 448–456 (2015) 17. Kim, B., Lee, J.Y.: Polyvinyl Alcohol Engineered Cementitious Composite (PVAECC) for the interfacial bond behaviour of Glass Fibre Reinforced Polymer Bars (GFRP). Polym. Polym. Compos. 20(6), 545–558 (2012) 18. Brazilian Standard NBR 11578. Cimento Portland Composto. Associação Brasileira de Normas Técnicas (ABNT) (1991) 19. Magalhães, M.S., Toledo Filho, R.D., Fairbairn, E.M.R., et al.: Durability under thermal loads of polyvinyl alcohol fibers. Matéria (Rio J.) 18(4), 1587–1595 (2013) 20. RILEM/CEB/FIP. Recommendations on reinforcement steel for reinforced concrete. Revised edition of: RC6 bond test for reinforcement steel: Pull-out test, CEB News 73, Lausanne, Switzerland (1983)

An Experimental Study of the Influence of Moderate Temperatures on the Behavior of Macrosynthetic Fiber Reinforced Concrete Marta Caballero-Jorna(&), Marta Roig-Flores, and Pedro Serna Instituto de Ciencia y Tecnología del Hormigón, Universitat Politècnica de València, Valencia, Spain [email protected]

Abstract. This research shows an experimental campaign performed aiming to understand the influence of moderate temperatures (5 °C, 20 °C and 50 °C) on the mechanical behavior of macrosynthetic fiber reinforced concrete (MSFRC). To do this, a 35 MPa concrete was selected. Two types of polypropylene fibers were added in a relatively high content (10 kg/m3). Steel fiber was used in an equivalent volume as a reference fiber. Considering that there is no standardized protocol to analyze the temperature effect on MSFRCs, a testing methodology based on UNE-EN 14651:2007 + A1:2008 is proposed. This modified procedure includes a system of insulating thermal covering which allows to maintain the temperature during the tests. Additionally, this study compares the effect of temperature on the pre-cracked and uncracked sections of MSFRCs. For this purpose, MSFRC and SFRC specimens were stored in each conservation temperature. Some of these samples were pre-cracked at the age of 28 days and then, they were reassigned in their respective conservation temperature until testing at 60 days. The results show that the analyzed temperatures have some influence on the mechanical properties of MSFRCs in terms of their residual flexural strength but to a limited extent, being suitable for environments with equivalent temperatures. Keywords: Aging
Fiber reinforced concrete
Macrosynthetic fibers Moderate temperature
Residual flexural strength




1 Introduction The application of fibers in concrete emerged in the sixties as a result of the constant search of potential enhancements in this construction material related to its mechanical performance and sustainability. Nowadays, there are numerous fiber types on the market which can be used to reinforce concretes. Steel fibers are the most commonly employed but in the last few decades, macrosynthetic fibers have appeared as an alternative in several applications due to the enhancements they offer [1]. Despite the enhancement they can provide to concrete, there is still some uncertainty regarding the durability of this type of fibers and its influence in the behavior of concrete. One of the most critical conditions to which macrosynthetic fibers can be subjected is temperature. The deterioration of their properties may occur since they are © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 322–332, 2021. https://doi.org/10.1007/978-3-030-58482-5_30

An Experimental Study of the Influence of Moderate Temperatures

323

thermally dependent. Macrosynthetic fibers can be found dimensionally unstable or more brittle; they can present a loss of bond or even be melted at moderate temperatures. As a result, the mechanical behavior of macrosynthetic fiber reinforced concretes (MSFRCs) may be affected. An analysis of the literature has revealed that many of the studies are focused on the behavior of polymeric fibers reinforced concretes at high temperature, mainly aiming to assess fire conditions where it is well proven that micro synthetic fibers are suitable in order to reduce spalling, since the fibers produce channels that reduce pore pressure [2–4]. It is noteworthy that adding synthetic fibers into concrete to improve spalling resistance is also recommended by Eurocode 2 [5]. However, only scarce information can be found on MSFRCs behavior at no standard temperatures under serviceability conditions, and the existing studies show discrepancies. Richardson and Ovington [6] investigated the effect of temperature variation (between −20 °C and 60 °C) on concrete properties with steel and polymeric fibers. The results showed that all beams exposed to a temperature of 60 °C experienced a flexural strength reduction when comparing to ambient temperature. However, a lesser decrease in peak flexural strength was observed when polypropylene fibers (12%) were compared to steel fibers (49%) and to plain concrete (35%). For all beams tested at cold temperature, flexural strength increased when compared to those performed at ambient temperature. The plain beam showed an average 99% increase in flexural strength, the polypropylene fiber beam showed an average a 150% increase in flexural strength and the steel fiber beam showed an average a 79% increase in flexural strength. This outcome concurs with Mirzazadeh et al. [7] who stated that concrete beams tested at −25 °C demonstrated an increase in strength and ductility of 13% and 34% respectively, compared to those tested at room temperature. Strauss Rambo et al. [8] showed the influence of temperature on the mechanical behavior of the MSFRCs were very similar to that known for plain concrete with relation to the loss of mechanical strength and elastic modulus. The residual tensile strength and elastic modulus of the macrosynthetic fibers were not affected by the temperature up to 100 °C remaining resistant capacity may be enough to ensure safety in service depending on the temperature reached. In contrast, Buratti and Mazzotti [9] found that temperatures ranging from 20 °C to 40 °C reduced the short-term residual strength of some of the MSFRCs analyzed. They concluded that temperature should be considered as factor when designing structures with macrosynthetic reinforcement, given that the effect of temperature is more relevant for MSFRCs than for SFRCs. This is backed up by [10], in which residual strength of the MSFRC decreased (about 10%) for crack mouth opening displacement (CMOD) higher than 1.5 mm since the temperature increased. Additionally, looking at the experimental procedure to determinate the thermomechanical behavior of MSFRCs, only a few investigations have examined mechanical characterization at the targeted temperatures [11–13] since most of the investigations have carried out residual tests at room temperature after heating/cooling. The different conclusions reached by the aforementioned studies shows that the temperature effect on the properties of MSFRCs is not completely clear yet. This context has motivated the present research, which aims to evaluate the mechanical behavior of macrosynthetic fibers reinforced concrete exposed to cold and hot moderate

324

M. Caballero-Jorna et al.

temperatures in order to analyze whether macrosynthetic fibers may be a viable alternative in environments and applications where they have a currently limited use and whenever MSFRC elements should be designed considering temperature-induced mechanical degradation.

2 Materials and Methods 2.1

Concrete and Fiber Properties

This study assesses the effect of moderate temperatures on the structural properties of macrosynthetic fiber reinforced concretes when compared steel fiber reinforced concrete control samples with focus on the behavior in cracked state. To this purpose, a base concrete was selected, with a compression strength around 35 MPa and high workability, common in the precast industry. The mix contains Ordinary Portland Cement CEM II 42.5R, two coarse aggregates (4-6 mm and 6-10 mm), river sand (0– 4 mm) and limestone filler. A polycarboxylate-based superplasticizer (SikaViscoCrete5980) was employed to guarantee the workability of the mixtures, being its dosage slightly adjusted for each batch. The effective w/b ratio of the mix was 0.55. The mix designs are displayed in Table 1. Three types of fibers were added into the base concrete: two polymeric fibers and one steel fibers. Fiber contents used were equivalent in a volume to achieve near equal flexural performance for both types of fibers, 10 kg/m3 for polymeric fibers and 30 kg/m3 for steel fibers. These quantities were selected due to their relatively frequent use. The most relevant properties of the three types of fibers used are summarized in Table 2. Table 1. MSFRC and SFRC dosages used in the present study. Dosage MFRC Dosage SFRC (kg/m3) (kg/m3) CEM II 42,5R 325 325 Gravel 6-10 mm 430 430 Gravel 4–6 mm 580 580 Sand 0–4 mm 835 835 Limestone Filler 80 80 Water 178.75 178.75 Fibers 10 30

An Experimental Study of the Influence of Moderate Temperatures

325

Table 2. Properties of the fibers used in the present study.

2.2

Experimental Program

The experimental campaign was developed to study the effect of moderate temperatures (5, 20 and 50 °C) on the mechanical behavior of MSFRCs, in particular their residual flexural strengths. Specimens were cast in eight concrete batches to produce prisms of 100  100 400 mm to perform three-point bending tests at target temperatures. Additionally, to characterize the batches, slump test was performed and cylindrical samples of Ø 150  300 mm were casted for compression tests at 28 days. To date, there is no standardized protocol to evaluate the influence of temperatures on FRCs during testing. Thus, a methodology based on the three-point bending test according to standard EN 14651 (2/3 scale) was implemented. 2.2.1 Preparation and Conservation Conditions of Samples All concrete mixes were manufactured in a DIEM WERKE model DZ 180 V mixer by the same sequence to ensure reproducibility of results. First, coarse aggregates and sand were added and pre-mixed for two minutes. Then, cement and filler were added and mixed for two minutes. Afterwards, water was added and then mixing for one minute. Finally, following the gradual addition of the fibers, superplasticizer was added, and a final additional mixing was applied for about five minutes. After mixing, the slump was measured according to standard EN 12350-2:2009. Then, the concrete was taken directly from the mixer and was poured into the molds according to the UNE EN 12350-1. On the one hand, 3 cylindrical specimens were cast for compression test for each batch, testing in total 24 specimens. The guideline for determining the compression strength in specimens of concrete in the hardened state was standard UNE-EN 123903:2009.

326

M. Caballero-Jorna et al.

On the other hand, prisms for three-point bending tests were cast and compacted over a vibrating table, being in total 48 prisms. They were demolded after 24 h of casting and kept at 20 ± 2 °C and 95% of relative humidity for 3 days. After that, they were located to their respective condition of conservation for two months until testing time. In this study, three different conservation temperatures were investigated, representing possible situations during service life in a Spain’s Mediterranean climate: • 5 °C, using an industrial fridge, this condition represents moderately cool weather. • 20 °C and RH 95%, using a controlled temperature room, to represent warm and humid weather. • 50 °C and saturated, using a thermal bath in water at target temperature, to represent situations such as in concrete heaters or industrial containers.

Temperature ( C)

Different measurements of the temperature in each defined case were conducted in order to verify the conservation conditions throughout the time. Figure 1 presents that the temperatures of 5 °C, 20 °C and 50 °C were quite stable during the exposure conditions. All the mixes were studied in pre-cracked conditions. One mix with polymeric fibers was studied in uncracked and cracked conditions. The objective of this process was to evaluate if the exposure of the open crack to moderate temperatures affects the residual flexural strength in MSFRCs. The specimens that were studied in cracked conditions were taken out of their conservation conditions and pre-cracked up to CMOD of 0.5 mm, at 28 days of conservation. Then, the specimens were returned to their corresponding conservation conditions until final testing at 60 days. The rest of the samples were tested without being subjected to the pre-cracked phase after two months.

50 40 30 20 10 0 0

7

14

21

28

35 42 49 56 Conservation (Days)

63

70

77

84

91

Fig. 1. Control of temperature during the conservation period for the three conditions studied.

2.2.2 Methodology of Flexural Test at Target Temperatures This test methodology was designed to determine how the mechanical properties of fiber reinforced concrete vary considering temperature, from 5 °C to 50 °C. To ensure the temperature was constant during the test, as novelty in this procedure, a cover system was used to insulate the specimens during testing when the specimens were removed from their storage areas.

An Experimental Study of the Influence of Moderate Temperatures

327

This cover system consisted on reusable cold and hot gel that provide and maintain the required temperature, while thermal sacks were used for insulation. The blue gel is not toxic and is 99% biodegradable, which makes the product environmental-friendly. For the flexural tests, the adopted configuration was scaled with a factor 2/3 from the standardized test according to UNE-EN 14651:2007 + A1:2008. Specimens were notched at the center of the prisms and LVDTs were placed to register the displacements. Figure 2 shows two pictures of the setup decided to maintain temperature inside concrete specimen during the test. The process was run at a load velocity of 2 mm/min.

Fig. 2. System of insulating thermal covering in the sample during testing.

Finally, to calculate the residual flexural strength, Eq. (1) was employed: fR;j ¼

3 Fj l
2 bh2sp

ð1Þ

where Fj is the axial load recorded during the test; l, the distance between supports (330 mm); b, the width of the sample cross section (100 mm) and hsp, the distance between top of the notch and top of cross section (85 mm).

3 Results and Discussion The results of the characterization tests and residual strength are collected and analyzed, to determine the uniformity and quality of the different concrete mixes and to show the temperature effect on MSFRCs and SFRCs, respectively. 3.1

Characterization Tests

Workability of fresh concrete was evaluated by slump test. Table 3 sums up thereby the data related to the workability test for each batch. As is shown, the slump of concrete is affected by the fiber type, its properties and its geometrical structure. This effect may be verified in all cases, according to Bolat [14]. Moreover, it is of importance to emphasize that this influence is quite significative due to the high fiber contents added in the concrete mixes.

328

M. Caballero-Jorna et al.

All in all, polypropylene fibers reduced more the workability of concrete than steel fibers. The higher workability values were obtained for SFRC, without segregation, while a clear loss of workability was experimented by mixes with P2 fibers. It could be explained due to the aspect ratio of this kind of fiber and the number of fibers on the cross section. However, this effect can be overcome using of effective method of compaction (vibrating tables). In the case of P1 fibers, slump values obtained are also lower than those obtained for S fibers, although this does not imply a reduction in the workability. Table 3 also summarizes the mean results of the compression performed tests for each batch and their coefficients of variation (COV) calculated from 3 specimens per batch. All compressive strengths results were around 35 MPa, which were initially defined. From a general point of view, a slight decrease in compressive strength is observed for the SFRC batches. However, the results indicate that there is no significant deviation between the results.

Table 3. Results of complementary tests. Code P1 N-05 N-20 P1 N-50 P1 P-05 P-20 P1 P-50 P2 P-05 P-20 P2 P-50 S P-05 S P-20 P-50

3.2

Consistency class Compression fcm (N/mm2) COV (%) S2 35.05 1.89 S2 S2

37.26 35.68

2.49 2.39

S2 S1

35.42 35.20

1.28 1.22

S1 S3 S3

36.47 31.60 32.73

2.53 5.80 1.46

Three-Point Bending Tests

To reach the main purpose of this work, three-point bending tests were performed at different temperatures. The results of the residual flexural strengths corresponding with the limit of proportionality (fLm) and the residual flexural strength fR,jm corresponding to CMOD = CMODj (j = 1, 3) where CMOD1 = 0.5 mm, CMOD3 = 2.5 mm, and their corresponding coefficients of variation are collected in Table 4 for each type of fiber and conservation condition. Herein, it can be observed that the variability of the results is influential within this study, which may complicate to obtain strong conclusions. This variability can be produced by the small size of the specimens used, the number of the samples and the scattering of the obtained outcomes between several specimens from the same production for the three-point bending tests.

An Experimental Study of the Influence of Moderate Temperatures

329

Table 4. Results of the three-point bending tests at moderate temperatures on the studied prisms. Code Temperature (°C)

fLm (MPa)

COV (%)

fR,1m (MPa)

COV (%)

P1

3.98 4.36 4.24 3.91 4.36 5.10 4.43 4.27 4.36

10.87 11.31 10.28 8.65 6.85 6.59 6.73 3.43 2.64

2.03 1.42 1.83 1.26 1.79 1.84 2.31 2.95 2.38

11.46 24.67 30.28 10.34 38.69 25.93 4.74 18.96 17.20

P2

S

05 20 50 05 20 50 05 20 50

Non pre-cracked COV fR,3m (MPa) (%) 3.00 12.35 2.18 13.94 4.04 38.84 – – – – – – – – – – – –

Pre-cracked fR,3m COV (MPa) (%) 3.52 6.53 3.14 18.36 2.23 13.57 1.86 10.64 3.14 41.36 2.33 29.93 2.35 26.43 3.51 26.94 2.69 18.84

To visualize the comparation between these outcomes, the residual flexural strengths are depicted in Figs. 3, 4 and 5. For a better understanding, a representative color was defined for each temperature. The blue color represents the coolest temperature (5 °C) in the bar charts. The temperature of 20 °C is identified by the purple color. The hot color depicts 50 °C. In addition, the mean value of all the data for every fiber type, without considering the influence of temperature, is also showed, which is the green bar. Firstly, as can be seen in Fig. 3, similar values in terms of residual flexural strengths at peak-load are shown for each type of fiber, between 4 and 5 MPa. These values are thought to be mainly dependent of the quality of the concrete matrix. Moreover, it is possible to observe that steel and macrosynthetic fibers behaved in a different way. For S samples, the increase or descent of the temperature had no significant effect on the fLm. For P1 and P2 specimens, the effect of temperature on fL was negligible, although a slight rising trend may be observed for P2 specimens.

6

5°C

20°C

50°C

AVG

fLm (MPa)

5 4 3 2 1 0 P1

P2

S

Fig. 3. Results of fLm for each type of fiber at different temperatures.

330

M. Caballero-Jorna et al.

6

5°C

20°C

50°C

AVG

fR,1m (MPa)

5 4 3 2 1 0 P1

P2

S

Fig. 4. Results of fR,1m for each type of fiber at different temperatures.

Figure 4 shows the residual flexural behaviour at CMOD1 mm for the tested fibres at different target temperatures. P1 and P2 samples obtained lower strengths than those obtained by S specimens. It is noteworthy that the coefficients of variation were higher than in the limit of proportionality for all fibers (see Table 4), indicating a greater variability within a same batch. S specimens presented a decrease on fR,1m for cold and hot temperature compared with the reference temperature (20 °C). For P2 elements, it is possible to affirm that a decrease in temperature may cause a reduction in fR,1m. However, the obtained residual flexural strengths at CMOD1 at 20 °C and 50 °C were similar. In the case of fR,1m for P1 samples does not seem to be affected by temperature since all values presented for residual flexural strength at CMOD1 are very similar. Figure 5 shows the temperature effect on the residual flexural strength fR,3 at CMOD3 for all types of fibers and all temperatures. This parameter is the one that suffered the highest variation levels. Since this test was performed at a different stage after pre-cracking and storage in the indicated conditions, this variation may be produced by handling during the displacement of the prisms from the storage to the testing press. SFRC specimens (S) exhibited lower values of fR,3m for targeted temperatures of 5 ° C and 50 °C than for the standard temperature. As in the case of CMOD1, the cold and hot temperature could influence the residual flexural strength of SFRCs more than the standard temperature. For P2 elements, the temperature can cause changes in the values of fR,3m, but they are not relevant. The best results for this type of fiber were obtained at 20 °C. In general, with this fiber, at CMOD3 the residual strengths displayed a reduction when a variation (increase or decrease) of temperature occurs. For the specimens reinforced with the macrosynthetic fiber P1 in the pre-cracked condition (same as in P2 and S), the best behavior occurred at low temperatures, but the temperature factor does not seem to influence on fR,3m. So, from a general point of view, it is not possible to observe any trend within results of fR,3m. for those specimens that had pre-cracking. Regarding the effect of the pre-cracked condition, there is a negative influence on the residual flexural strength for the group pre-cracked (P1-P) when temperature

An Experimental Study of the Influence of Moderate Temperatures

331

increases, and on the contrary, this increase of temperature has a positive effect for the uncracked group (P1-N). Hence, the interaction between the factors pre-cracking condition and the temperature can be potentially significant in some cases and is worth an in-depth study.

6

5°C

20°C

50°C

AVG

fR,3m (MPa)

5 4 3 2 1 0 P1-N

P1-P

P2-P

S-P

Fig. 5. Results of fR,3m for each type of fiber at different temperatures.

4 Conclusions In this research, an experimental campaign in order to obtain a better understanding of the influence of moderate temperatures on the mechanical behavior of macrosynthetic fiber reinforced concretes was carried out. From the obtained results, the conclusions drawn are: • No clear trend was obtained for the flexural strength in specimens that were precracked, indicating that both MSFRC and SFRC specimens, despite their differences, did not suffer noticeable degradation at the target temperatures of 5 °C – 50 °C after 60 days of exposure. Thus, MSFRCs have potential to be used at moderate temperatures. • Regarding the responses obtained using the two polypropylene fibers, P1 fibers obtained more efficient flexural performance than P2 fibers at the tested temperatures. • The interaction of pre-cracking condition with moderate temperatures can be influential in some cases but need a specific study to determine its influence. • The proposed experimental procedure, based on the standard UNE-EN 14651:2007 + A1:2008 (2/3 scale), allows to maintain the temperature during the tests. This procedure includes a system of insulating thermal covering which manages to set up the influence of moderate temperature as a factor on the mechanical behavior of reinforced concretes. However, the reduced scale is a factor that may increase the dispersion.

332

M. Caballero-Jorna et al.

Acknowledgements. The authors would like to express their gratitude to the Spanish Ministry of Science, Innovation and Universities for funding received under the FPU Program [FPU18/06145].

References 1. Yin, S., Tuladhar, R., Shi, F., Combe, M., Collister, T., Sivakugan, N.: Use of macro plastic fibres in concrete: A review. Constr. Build. Mater. 93, 180–188 (2015). https://doi.org/10. 1016/j.conbuildmat.2015.05.105 2. Kalifa, P., Chéné, G., Gallé, C.: High-temperature behaviour of HPC with polypropylene fibres. Cem. Concr. Res. 31, 1487–1499 (2001). https://doi.org/10.1016/s0008-8846(01) 00596-8 3. Sideris, K.K., Manita, P., Chaniotakis, E.: Performance of thermally damaged fibre reinforced concretes. Constr. Build. Mater. 23, 1232–1239 (2009). https://doi.org/10.1016/j. conbuildmat.2008.08.009 4. Han, C.G., Hwang, Y.S., Yang, S.H., Gowripalan, N.: Performance of spalling resistance of high performance concrete with polypropylene fiber contents and lateral confinement. Cem. Concr. Res. 35, 1747–1753 (2005). https://doi.org/10.1016/j.cemconres.2004.11.013 5. European Committee for Standarisation, Eurocode 2: Design of concrete structures - Part 1– 2: General rules - Sturctural fire design, AENOR, Madrid (2011) 6. Richardson, A., Ovington, R.: Temperature related steel and synthetic fibre concrete performance. Constr. Build. Mater. 153, 616–621 (2017). https://doi.org/10.1016/j. conbuildmat.2017.07.101 7. Mirzazadeh, M.M., Noel, M., Green, M.F.: Effect of low temperature on the shear-fatigue performance of reinforced concrete beams, Proceedings. Annu. Conf. Can. Soc. Civ. Eng. 4, 2682–2691 (2016) 8. Rambo, D.A.S., Blanco, A., de Figueiredo, A.D., dos Santos, E.R.F., Toledo, R.D., Gomes, O.d.F.M.: Study of temperature effect on macro-synthetic fiber reinforced concretes by means of Barcelona tests: An approach focused on tunnels assessment. Constr. Build. Mater. 158, 443–453 (2018). https://doi.org/10.1016/j.conbuildmat.2017.10.046 9. Buratti, N., Mazzotti, C.: Experimental tests on the effect of temperature on the long-term behaviour of macrosynthetic Fibre Reinforced Concretes. Constr. Build. Mater. 95, 133–142 (2015). https://doi.org/10.1016/j.conbuildmat.2015.07.073 10. Buratti, N., Mazzotti, C.: Temperature effect on the long term behaviour of macro-syntheticand-steel–fibre reinforced concrete. In: 8th RILEM Int. Symp. Fibre Reinf. Concr. Challenges Oppor. (2012) 11. Castillo, A.J., Durrani, C.: Effect of transient high temperature on high strength concrete. ACI Mater. J. 87, 47–53(1987) 12. Cifuentes, H., Ríos, J.D., Leiva, C., Medina, F.: Influencia del tiempo de exposición a altas temperaturas en el comportamiento en fractura de hormigones autocompactantes reforzados con fibras. An. Mecánica Fract, pp. 1–6 (2016) 13. Cheng, F.P., Kodur, V.K.R., Wang, T.C.: Stress-strain curves for high strength concrete at elevated temperatures. J. Mater. Civ. Eng. 16, 84–90 (2004). https://doi.org/10.1061/(ASCE) 0899-1561(2004)16:1(84 14. Bolat, H., Şimşek, O., Çullu, M., Durmuş, G., Can, Ö.: The effects of macro synthetic fiber reinforcement use on physical and mechanical properties of concrete. Compos. Part B Eng. 61, 191–198 (2014). https://doi.org/10.1016/j.compositesb.2014.01.043

Post-cracking Behaviour of Glass Fibre Reinforced Concrete with Recycled Aggregates Brecht Vandevyvere1, Lucie Vandewalle2, Els Verstrynge2, and Jiabin Li1(&) 1

2

Research Group RecyCon, Department of Civil Engineering, KU Leuven Campus Bruges, 8200 Brugge, Belgium [email protected] Department of Civil Engineering, KU Leuven, 3001 Leuven, Belgium

Abstract. In the past two decades, numerous studies have demonstrated the benefit of adding fibres in structural concrete. The addition of fibres in concrete limits the crack opening after concrete cracking. For conventional concrete with natural aggregates, the post-cracking behaviour is mainly determined by the type and amount of the fibres. Due to the increased shortage of raw materials, the use of recycled aggregates (RAs) in structural concrete has gained more and more interests recently. Nevertheless, the influence of fibres on the post-cracking behaviour of concrete with RAs has not well been understood. This paper presents an extensive experimental study on the behaviour of concrete with RAs reinforced with alkali resistant (AR) glass fibres. To better understand the effect of the quality of the RAs, three types of RAs were used in the present work, which include a high grade recycled concrete aggregate (RCA+), a normal quality RCA (nRCA) as well as a mixed recycled aggregate (MA). In addition, a natural aggregate (limestone) is also used as the reference aggregate. A total of 24 fibre reinforced notched beams are tested under three-point bending load according to EN 14651. In all fibre reinforced concrete mixtures, the CEM-FIL MinibarsTM was used with a fibre content of 0.75 V%. The test results indicate that the quality of the RAs can have a significant influence on the post-cracking behaviour of the beams, especially in the Crack Mouth Opening Displacement (CMOD) range of 0.5–1.2 mm. Keywords: Fibre reinforced concrete
Recycled aggregates (RAs) resistant (AR) glass fibres
Three-point bending
Post-cracking


Alkali

1 Introduction Recycling of construction and demolition waste into recycled aggregates (RAs) saves land space, decreases the transportation cost and reduces the consumption of natural resources as well as CO2 emissions. Therefore numerous research activities have been carried out to investigate the mechanical and durability properties of concrete containing RAs [1]. The test results generally indicated a decrease in the mechanical and durability properties. This reduction is mainly caused by the increased porosity of RAs in comparison to natural aggregates due to the attached mortar. The attached mortar influences the interfacial transition zone (ITZ) and consequently the concrete © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 333–343, 2021. https://doi.org/10.1007/978-3-030-58482-5_31

334

B. Vandevyvere et al.

performance [2]. In spite of this disadvantage, a proper mix design can still meet the design requirements [3–5]. In structural concrete elements, fibres are commonly added to improve the post-cracking behaviour of concrete. The most commonly used fibres are steel fibres, polypropylene (PP) fibres and glass fibres [6]. Previous studies [7, 8] showed that those fibres have a limited influence on the compressive strength and modulus of elasticity of concrete. The biggest improvement is related to the flexural strength, fatigue and abrasion resistance, deformation capability and toughness [6]. Due to this beneficial effect, a number of research activities have been conducted in the last decades on fibre reinforced concrete with natural aggregates. Those studies revealed that the fibre amount and dimensions are very important to obtain a certain postcracking strength [9]. For concrete with RAs, another kind of behaviour occurs. Many researchers [10, 11] reported that the strength of RAs is commonly lower than that of natural aggregates. Therefore the fracture process and consequently the fibre activation in concrete with RAs can be different from that of conventional concrete with natural aggregates. To fully understand the fracture process in fibre reinforced concrete containing RAs, the following research is carried out to get a better insight in the fracture behaviour of fibre reinforced concrete with different types of RAs.

2 Experimental Program To investigate the post-cracking behaviour of glass fibre reinforced concrete with different types of RAs, a total of 8 concrete mixtures are manufactured, which include two conventional and six recycled concrete mixtures. The mix composition was obtained based on the target strength class C40/50 and the slump class S4 (>16 cm) for the plain natural concrete mixture. The mixture comprised ordinary Portland cement CEM I 52.5R/HES (400 kg/m3), fine and coarse aggregates, water (W/C = 0.48) and super-plasticizer (SP). In Table 2, the detailed concrete compositions for all the concrete mixtures are given. In the concrete mixtures, the used aggregates are sand 0–2 mm (fineness modulus: 2.25), limestone 2–6.3 mm and coarse natural and recycled aggregates 6.3–14 mm. In total, three types of recycled aggregates are used: high quality recycled concrete aggregates (RCAs+), normal recycled concrete aggregates (nRCAs) and mixed aggregates (MAs). Those aggregates are produced by a local recycling plant in Flanders, Belgium. In Table 1, the physical and mechanical properties of the aggregates are given. The physical properties, including the oven-dry density (qrd), the saturated surface-dried (SSD) particle density (qssd), the apparent particle density (qa) as well as the water absorption after 24 h (WA24) are obtained according to the standard EN 1097-6. In addition, the resistance to fragmentation (LA) of the aggregates is determined according EN 1097-2. In the fibre reinforced concrete mixtures, RA glass fibres with a content of 0.75 V% are used. These fibres are the CEM-FIL MinibarsTM produced by OWENS CORNING. The glass fibres have a fibre length of 43 mm, diameter of 0.70 mm and a Young’s modulus of 42 GPa.

Post-cracking Behaviour of Glass Fibre Reinforced Concrete

335

Table 1. Physical and mechanical properties of the aggregates qrd (kg/m3) Sand 0-2 2580 Limestone 2-6.3 2680 Limestone 6.3-14 2630 2380 RCA+ 6.3-14 nRCA 6.3-14 2230 MA 6.3-14 1890

qa (kg/m3) 2600 2710 2680 2660 2680 2420

qssd (kg/m3) WA24 (%) LA (−) 2590 0.40 – 2690 0.45 – 2650 0.81 22 2500 4.42 24 2650 6.06 34 2190 11.52 46

Table 2. Concrete composition. Aggregates (kg/m3) Sand Limestone 0-2 2-6.3 NAC – ref. 759 199 199 RAC+ – ref. 759 nRAC – ref. 759 199 MAC – ref. 759 199 NAC – 759 199 0.75 V% RAC+ – 759 199 0.75 V% nRAC – 759 199 0.75 V% MAC – 759 199 0.75 V% Concrete mixture

+

Fibre content (V%)

Limestone 6.3-14 777 – – – 777

RCA 6.3-14 – 703 – – –

nRCA 6.3-14 – – 659 – –

MA 6.3-14 – – – 560 –

0 0 0 0 0.75

–

703

–

–

0.75

–

–

659

–

0.75

–

–

–

560

0.75

The strength properties of each concrete mixture are examined on 12 cubes (150  150  150 mm3) and 6 cylinders (height: 300 mm – diameter: 150 mm). These specimens are cured in water for 28 days until test. The compressive strength test is performed according the NBN EN 12390-3:200 - Testing hardened concrete. Compressive strength of test specimens. The splitting tensile strength and E-modulus are determined according to the NBN EN 12390-6 Testing hardened concrete – Part 6: Tensile splitting strength of test specimens and the NBN B 15-203 Concrete testing – Statical modulus of elasticity with compression, respectively. The flexural behaviour of the different concrete mixtures is examined on beam specimens (600  150  150 mm3). In total, 24 fibre reinforced concrete beams as well as 24 unreinforced beams were casted and tested according to the standard EN 14651: 2005 + A1:2007 - Test method for metallic fibre concrete: Measuring the flexural tensile strength (limit of proportionality (LOP), residual). The test set-up is shown in Fig. 1.

336

B. Vandevyvere et al.

Fig. 1. Load-CMOD curves of the concrete mixtures

During the displacement controlled three-point bending test, the concrete beam is subjected to a concentrated force F at the mid-span. According to the standard, a notch in the beam is required to locate the crack plane under the bending load. In the present test, the notch in the concrete beams is made through a wet saw. During the bending test, the applied force F is measured as a function of the crack mouth opening displacement (CMOD). Each test is stopped as the CMOD reaches a value of 4.5 mm. During the test, not only the CMOD is measured (LVDT CMOD), but also the deflection (LVDT deflection) of the beams is measured. This is also indicated in Fig. 1. From the load-CMOD curve, the limit of proportionality (LOP) and residual flexural tensile strengths can be determined. The LOP represents the maximal tensile stress at the top of the notch. By considering a linear stress distribution, the following equation is valued: f fct;L ¼

3FL l 2bh2sp

ð1Þ

where: f fct;L = the limit of proportionality (LOP) (MPa); FL = the load corresponding to the LOP (N); l = span length (mm); b = width of the specimen (mm); hsp = distance between the tip of the notch and the top of the specimen (mm). The residual flexural tensile strengths corresponding to different CMODj values can be computed according to Eq. (2). f R;j ¼

3Fj l 2bh2sp

ð2Þ

where: fR,j = the residual flexural tensile strength corresponding with CMODj (MPa); Fj = the load corresponding with CMODj (N). The purpose of adding fibres is to increase the residual flexural tensile strength. The fibres enhance the concrete toughness by their energy absorption capacity. In the obtained P-CMOD curve, two parts are distinguished: a pre-crack part and a post-crack

Post-cracking Behaviour of Glass Fibre Reinforced Concrete

337

Fig. 2. Load-CMOD curves of the concrete mixtures.

part. The pre-crack part occurs before the LOP, while the post-crack part occurs after the LOP. Hereby, the area under the pre-cracking curve and the post-cracking curve represents the pre-cracking energy (Epr) and the post-cracking energy (Epost). Generally, it is known that the fibres mainly influence the post-crack energy.

3 Test Results and Discussion 3.1

Workability

In Table 3, an overview of the workability of the different concrete mixtures is given. For all the unreinforced concrete mixtures, the amount of superplasticizer (SP) was determined to obtain a S4 slump class. It seems that for the unreinforced NAC, RAC+ and nRAC 0.77 g/l is required, while for the MAC-ref. mixture an amount of 0.33 g/l is sufficient. This might be due to the fact that the additional added water (to compensate the high water absorption of the MAs) is not yet completed absorbed. When 0.75 V% CEM-FIL MinibarsTM is added, the obtained slump class is reduced to S3. Only for the MAC-0.75 V%, a S4 slump class can still be obtained. Generally, it seems that the used amount of glass fibres has an important impact on the concrete workability.

338

B. Vandevyvere et al. Table 3. Workability of the concrete mixtures. NAC RAC+ nRAC MAC ref. 0.75 V% ref. 0.75 V% ref. 0.75 V% ref. 0.75 V% Amount of SP (g/l) 0.77 0.77 0.77 0.33 Slump (cm) 18 10 16 11 19 13 20 16 Slump class S4 S3 S4 S3 S4 S3 S4 S4

3.2

Strength and Flexural Behaviour

The mechanical properties of the reference and the fibre reinforced concrete mixtures are summarized in Table 4. Earlier research [7, 8] reported that the fibre addition has only a small influence on the mechanical properties of conventional concrete. For the recycled concrete mixtures similar results are found. By adding 0.75 V% glass fibres to the recycled mixtures, a small increase of the compressive strength is observed for the concrete mixtures with 100 V% RCAs+ and nRCAs. This improvement is resulted from the bridging effect of the fibres when micro-cracks occur. Only the fibre reinforced mixture containing MAs shows a small decrease in compressive strength. Despite this decrease, the highest (splitting) tensile strength improvement is observed for the MAC-mixture. The splitting strength increases from 2.2 (MAC-ref.) to 3.3 MPa (MAC-0.75 V%). In addition, it is observed that all fibre reinforced concrete cubes do not collapse at the maximum load. This indicates an enhanced ductile behaviour of the concrete due to the glass fibres. Despite the small increase of the mechanical concrete properties, a significant improvement is Fig. 3. ANOVA-test observed in the post-cracking behaviour of the fibre reinforced concrete mixtures with 100 V% nRCAs and MAs, compared to NAC-0.75 V%. In Fig. 2, the individual P-CMOD curve of the four fibre reinforced concrete mixtures is shown. It seems that there is no real difference in the flexural tensile strength (f fct;L ) for the fibre reinforced concrete mixtures with natural aggregates, RCAs+ and nRCAs (see also Table 5). Only for the MAC-0.75 V% mixture, the flexural tensile strength is 3.3 MPa. From the obtained P-CMOD curves, it is clearly that a higher post-cracking behaviour is obtained when aggregates with a higher LA-value are used. The statistical parametric mapping method (SPM) [12] in which an one-way ANOVA test was carried out, shows that there is a significant difference in the load in a CMOD-range of 0.5–1.2 mm (see Fig. 3) for the four fibre reinforced concrete mixtures. In this test, a significance level (a) of 0.05 is used.

Post-cracking Behaviour of Glass Fibre Reinforced Concrete

339

Table 4. Mechanical concrete properties. Cubical compressive strength (MPa) [st. dev] NAC-ref. 59.6 [2.7] NAC-0.75% 59.3 [1.9] 54.6 [1.0] RAC+-ref. RAC+-0.75 V% 59.8 [2.3] nRAC-ref. 52.4 [1.4] nRAC-0.75 V% 55.9 [1.6] MAC-ref. 45.2 [0.8] MAC-0.75 V% 44.3 [1.0]

Splitting tensile strength (MPa) [st. dev] 3.6 [0.2] 3.9 [0.2] 3.6 [0.2] 3.8 [0.4] 2.9 [0.6] 3.5 [0.2] 2.2 [0.4] 3.3 [0.3]

Modulus of elasticity (GPa) [st. dev] 38.6 [2.3] 36.6 [2.6] 33.9 [1.3] 36.0 [1.5] 33.4 [1.5] 35.9 [3.6] 32.9 [0.4] 34.1 [0.5]

Table 5. (Residual) Flexural tensile strength according EN 14651:2005. f fct;L (MPa) [st. dev] fR,j at prescribed CMODj values (MPa) [st. dev] 0.5 mm 1.5 mm 2.5 mm 3.5 NAC-0.75 V% 4.2 [0.6] 3.4 [0.4] 3.3 [0.3] 2.7 [0.3] 2.1 + 3.2 [0.4] 3.3 [0.4] 2.8 [0.3] 2.2 RAC -0.75 V% 4.1 [0.3] nRAC-0.75 V% 4.2 [0.2] 3.8 [0.2] 3.8 [0.3] 3.0 [0.3] 2.3 MAC-0.75 V% 3.3 [0.3] 3.9 [0.3] 3.9 [0.3] 3.0 [0.1] 2.3

Energy (J) mm [0.2] [0.2] [0.3] [0.0]

38.7 39.9 43.3 43.3

[3.5] [4.0] [3.5] [2.2]

The higher P-CMOD curves of the nRAC-0.75 V% and MAC-0.75 V% mixture definitely results in a higher amount of absorption energy. The total energy (Epre + Epost) increases from 38.7 J to 43.3 J for the MAC-0.75 V% mixture. According to the FIB Model Code 2010 [13], traditional rebars can partly be replaced if Eq. 3 and 4 are satisfied: f R1k [ 0:4 f Lk

ð3Þ

f R3k [ 0:5 f R1k

ð4Þ

in which fLk represents the characteristic flexural tensile strength at LOP, while fR1k and fR3k are the characteristic residual strength at 0.5 mm CMOD and 2.5 mm CMOD, respectively. Due to the higher post-cracking energy in the range of 0.5–1.2 mm CMOD, a higher residual flexural tensile strength is observed for Serviceability Limit State (SLS), which corresponds to a CMOD-value of 0.5 mm. This increases from 3.4 to 3.9 MPa for the fibre reinforced concrete mixture with MAs. At ULS (CMOD = 2.5 mm), only a small difference in the obtained residual tensile strength is observed (see Table 5). Therefore, it is important to notice that Eq. (4) becomes more critical for the recycled

340

B. Vandevyvere et al.

concrete mixtures. By considering a lognormal distribution, the characteristic strength values can be determined for all parameters (f fct;L and fRj), according to Eqs. (5)–(7).
 f k ¼ exp f ln;m  kn
rln

ð5Þ

1 Xn f ln;m ¼
ln f i i n rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Xn
2 1
rln ¼ ln f  f i ln;m i n1

ð6Þ ð7Þ

In the above mentioned formulas, the kn-value depends on the number of samples and the prior knowledge on the coefficient of variation. Those kn-values are summarized in table D1 from the standard EN 1990. For five and six specimens the value is 2.33 and 2.18, respectively. In Table 6, the corresponding characteristic flexural tensile strengths are given. It seems that although a higher residual tensile strength, Eq. 3 and Eq. 4 are still satisfied. Therefore the traditional rebars can still (partly) be replaced when recycled aggregates are used. This is even the case for the MAC-mixtures.

Table 6. Characteristic (residual) flexural tensile strength according EN 14651:2005. f fct;Lk (MPa) NAC0.75 V% RAC+0.75 V% nRAC0.75 V% MAC0.75 V%

3.2

fR,jk at prescribed CMODj values Equation 3 Equation 4 FRC (MPa) class 0.5 mm 1.5 mm 2.5 mm 3.5 mm 2.8 2.1 1.7 1.3 ✓ ✓ 3a

3.4

2.4

2.5

2.1

1.7

✓

✓

3b

3.8

3.4

3.0

2.3

1.6

✓

✓

3a

2.8

3.1

3.1

2.6

2.1

✓

✓

2.5b

The LA-values, shown in Table 1, indicate that the aggregate strength of the nRCAs and MAs are significantly lower compared the natural and high quality recycled aggregates. Due to this, it is expected that the crack propagation, and therefore the fibre activation is also influenced by the aggregate strength. By considering the linear asymptotic superposition assumption, Tada et al. [14] proposed formula (8) and (9) to calculate the equivalent-elastic crack length: CMOD ¼

6PSa VI ðaÞ D2 BE

ð8Þ

Post-cracking Behaviour of Glass Fibre Reinforced Concrete

341

Where: a = (a + H0)/(D + H0); P = the load at specific CMOD-value; S = specimen loading span (500 mm); a = crack length (mm); D = beam depth (150 mm); B = beam width (150 mm); E = modulus of elasticity (GPa); H0 = thickness of the clip gauge holder (= 18 mm). For S/D = 4, the geometrical function VI(a) is given as follows: VI ¼ 0:76  2:28a þ 3:87a22 :04 a3 þ

0:66 ð1  aÞ2

ð9Þ

With the above mentioned equations and the obtained P-CMOD curves, the crack length propagation for the different concrete mixtures can be determined. This is visualized by the blue line in Fig. 4. In this figure, the crack length at LOP (PLOP) is shown in red, while the crack length at the initial cracking load (Pcrack) and postcracking peak (PMOR) is visualized in green and yellow. The used P-CMOD values at initial cracking and ultimate cracking are also given in Table 7. The initial cracking point is defined as the point where the first cracking occurs. This is commonly reached earlier than the flexural strength of the concrete matrix. The initial cracking point is where non-linearity in the load-deflection curve becomes obvious or the inclination of P-CMOD curve reduces significantly. In Fig. 4, the mean value of six beams is given, while the colored zone indicates the range [l − r, l + r]. It seems that for the high

Fig. 4. Crack length at initial cracking, ultimate and post peak load.

342

B. Vandevyvere et al. Table 7. Load-CMOD values at initial and ultimate cracking point Initial cracking point Load (kN) CMOD (mm) NAC-0.75 V% 7.4 [1.6] 0.010 [0.002] RAC+-0.75 V% 6.8 [2.6] 0.009 [0.002] nRAC-0.75 V% 6.8 [0.6] 0.008 [0.001] MAC-0.75 V% 4.1 [0.8] 0.007 [0.001]

Ultimate cracking point Load (kN) CMOD (mm) 13.1 [1.8] 0.049 [0.028] 12.6 [1.0] 0.067 [0.051] 12.9 [0.8] 0.034 [0.007] 10.3 [1.0] 0.081 [0.007]

quality recycled concrete mixtures a longer crack length (and scattering) is obtained at the different load stages. Due to the fact that the crack length of the natural and recycled mixtures does not differ much, no real difference is observed in the P-CMOD curves. By using recycled mixed aggregates, a crack length of 78.8 mm is obtained at the LOPvalue. This is caused by the lower aggregate strength which results in a long crack length. Therefore, many fibres are directly activated after concrete cracking and no force-drop is observed in the P-CMOD curve (see Fig. 3). Also for the nRAC mixture, a higher post-cracking strength is observed; although only a crack length of 49.4 mm is noticed at LOP. This can be caused by the higher ductile behaviour of the nRAC0.75 V% mixture [15].

4 Conclusions In this study, the mechanical behaviour of glass fibre reinforced concrete with different types of RAs is studied. In addition to the tests on compressive strength, splitting tensile strength and modulus of elasticity, three-point bending tests on notched beam specimens are also carried out according to the EN 14651 to determine the postcracking behaviour of the beams. Within the scope of this study, the following conclusions can be drawn: • The CEM-FIL minibarsTM AR glass fibres have a minor, but beneficial influence on the mechanical properties of the recycled concrete mixtures. The most significant improvement is observed for the splitting tensile test of the concrete with mixed recycled aggregates. The splitting strength increases from 2.2 to 3.3 MPa when the fibre amount increases from 0 to 0.75 V%. • The three point bending tests on the fibre reinforced concrete specimens show that there is a significant difference in the obtained P-CMOD curves, especially in the CMOD-range between 0.5–1.2 mm. • The fibre reinforced natural mixtures, as well as the fibre reinforced recycled concrete mixtures show that they meet the FIB Model code requirements to (partly) replace the traditional reinforcement. • The crack length (at LOP) for the RAC+-0.75 V% and MAC-0.75 V% mixtures are higher compared to the natural aggregate fibre reinforced mixture. For the fibre reinforced MAC-mixture a crack length of 78.8 mm was obtained while the NAC0.75 V% mixture has a crack length of only 57.4 mm at LOP.

Post-cracking Behaviour of Glass Fibre Reinforced Concrete

343

• Despite smaller crack length of the nRAC-0.75 V% (49.4 mm), a higher P-CMOD curve is obtained. This can be caused by the higher ductility of those concrete mixture. Acknowledgements. This research work has been supported by a start-up funding at KU Leuven – University of Leuven, Belgium (No. STG/16/011), and the Lvnong Chair in Construction Waste Recycling at KU Leuven, Belgium. The financial supports are gratefully acknowledged. The authors would also like to express their thanks to the technical staff Bernd Salaets and technical manager Kurt Scherpereel at the laboratory of Department of Civil Engineering, KU Leuven, as well as Owens Corning for providing the glass fibres, O.B.B.C nv for delivering the RAs and Willy Naessens Group for supplying the natural aggregates.

References 1. Safiuddin, M., Alengaram, U.J., Rahman, M.M., Salam, M.A., Jumaat, M.Z.: Use of recycled concrete aggregate in concrete: a review. J. Civ. Eng. Manag. 19, 796–810 (2013) 2. Otsuki, N., Miyazato, S., Yodsudjai, W.: Influence of recycled aggregate on interfacial transition zone, strength, chloride penetration and carbonation of concrete. J. Mater. Civ. Eng. 15, 443–451 (2003) 3. Levy, S.M., Helene, P.: Durability of recycled aggregates concrete: a safe way to sustainable development. Cem. Concr. Res. 34, 1975–1980 (2004) 4. Malešev, M., Radonjanin, V., Marinković, S.: Recycled concrete as aggregate for structural concrete production. Sustainability 2, 1204–1225 (2010) 5. Tu, T.-Y., Chen, Y.-Y., Hwang, C.-L.: Properties of HPC with recycled aggregates. Cem. Concr. Res. 36, 943–950 (2006) 6. ACI Committee 544: ACI 544.1R-96: Report on Reinforced Concrete (Reapproved). ACI Struct. J. 96, 1–66 (2002) 7. Thomas, J., Ramaswamy, A.: Mechanical properties of steel fiber-reinforced concrete. J. Mater. Civ. Eng. 19(5), 385–392 (2007) 8. Alhozaimy, A.M., Soroushian, P., Mirza, F.: Mechanical properties of polypropylene fiber reinforced concrete and the effects of pozzolanic materials. Cem. Concr. Compos. 18, 85–92 (1996) 9. Frank, H.P.: Fibre Cements and Fibre Concretes, D. J. Hannant. Wiley-Interscience, New York, 219 pp. (1978). J. Polym. Sci. Polym. Lett. Ed. 17, 464–465 (1979) 10. López-Gayarre, F., Serna, P., Domingo-Cabo, A., Serrano-López, M.A., López-Colina, C.: Influence of recycled aggregate quality and proportioning criteria on recycled concrete properties. Waste Manag. 29, 3022–3028 (2009) 11. Silva, R.V., De Brito, J., Dhir, R.K.: Properties and composition of recycled aggregates from construction and demolition waste suitable for concrete production. Constr. Build. Mater. 65, 201–217 (2014) 12. Friston, K.J., Karl, J., Ashburner, J., Kiebel, S., Nichols, T., Penny, W.D.: Statistical Parametric Mapping: The Analysis of Funtional Brain Images. Elsevier/Academic Press (2007) 13. FIB 2011 Model Code 2010 FIB Model Code Concr. Struct. 114–148 (2010) 14. Tada, H., Paris, P.C., Irwin, G.R.: The Stress Analysis of Cracks Handbook (1985) 15. Zhang, J., Li, V.C.: Simulation of crack propagation in fiber-reinforced concrete by fracture mechanics Cem. Concr. Res. 34, 333–339 (2004)

Long-Term Properties

A Computational Sectional Approach for the Flexural Creep Behavior of Cracked FRC Rutger Vrijdaghs1(&), Marco di Prisco2, and Lucie Vandewalle1 1

2

Department of Civil Engineering, KU Leuven, Leuven, Belgium [email protected] Department of Structural Engineering, Politecnico di Milano, Milan, Italy

Abstract. This paper presents a computational model to calculate and predict the flexural creep behavior in a cracked fiber reinforced concrete (FRC) section. The proposed model is based on uniaxial creep data and consists of three steps. In the first step, an inverse analysis algorithm is presented to model the monotonic bending behavior of a notched FRC beam in accordance with EN 14651. A simplified and numerically optimized method is compared to experimental data and a good agreement is found. In a second step, the unloading behavior of the beam is taken into account. Calibrated on experimental data, the model is able to accurately and precisely predict the unloading behavior. Further validation comes from the location of the neutral axis, and the deformation profile. In a third step, the flexural creep behavior is predicted based on the results in the second step. The creep data is supplied in uniaxial form, which allows greater applicability across various FRC mixtures. The proposed approach is able to take into account stress redistribution following fiber fracture. Furthermore, the time-dependent effects of the stress redistributions are also accounted for. As such, the model is able to predict tertiary creep and structural failure under sustained loading. Keywords: Sectional analysis
Creep of FRC
Polymeric FRC
Tensile and bending creep

1 Introduction Fiber reinforced concrete (FRC) is a composite material in which fibers are added to the fresh concrete mix [1, 2]. These fibers are used to improve the properties of the concrete in the fresh or hardened state. In structural applications, fibers can partially or totally replace the traditional reinforcement. For these purposes, the fibers provide an enhanced post-cracking tensile strength in the hardened state by bridging crack faces [3]. Commercially available fibers can be made from a number of different materials: steel, glass, synthetic and natural being the most common types [4]. Until recently, the research effort has been focusing on the properties of FRC in the fresh and hardened state [5, 6]. It has been found that the inclusion of fibers decreases plastic shrinkage cracking [7] and flowability [8] in the fresh state. In the hardened © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 347–358, 2021. https://doi.org/10.1007/978-3-030-58482-5_32

348

R. Vrijdaghs et al.

state, fibers provide a post-cracking strength to the matrix [9–11]. However, investigations on the long-term structural properties of FRC have only recently become a research focal point. The recently published Model Code 2010 [12] acknowledges that time-dependent effects may significantly alter the behavior of structural elements, both in the serviceability limit state and in the ultimate limit state. In the former, creep may cause unwanted and unacceptable deflections of FRC beams [13, 14], whilst in the latter, the residual load-bearing capacity of elements may be significantly lower than the design value based on the short-term strength assessment. However, the Model Code (MC10) does not provide design rules to take long-term behavior into account. It is clear that further research on the long-term structural properties of FRC is required. One of these time-dependent phenomena is creep. While the creep of cracked FRC has been under investigation for over two decades [15–18], much uncertainty still remains on how to model the behavior. In this paper, the creep deformations of a cracked, notched FRC beam are predicted based on uniaxial tensile creep data. A sectional approach is presented, consisting of three steps, which gradually increase the complexity and capabilities of the model. In the first step, a monotonic bending test is simulated using uniaxial constitutive laws. Secondly, unloading of the sample is taken into account, and in the third step, uniaxial creep data is used to predict timedependent flexural crack opening. The approach is validated with experimental data in each step.

2 Experimental Program and Results The experimental program focusses on a normal strength concrete (fcm,cube = 43 MPa) with 1 V% of polypropylene fibers (diameter: 0.7 mm, length: 55 mm) and consists of a series of short-term characterization and long-term creep tests. Three different tests are performed: monotonic bending according to EN 14651 [19], unloading from various points in the bending tests and uniaxial tensile creep tests, refer to Table 1. Each test setup and result is discussed in this section. Table 1. Number of specimens per test type Test type Number of specimens [-] Monotonic bending acc. to EN 14651 12 Unloading-reloading in bending 6 Uniaxial tensile creep 14

2.1

Monotonic Bending Test

The monotonic bending test is performed according to EN 14651 [19]. The test setup is shown in Fig. 1, and the reader is referred to the European Standard for all details concerning the test speed, specimen geometry and strength calculations. The results are shown in a stress-CMOD curve on the right, with indicating of the average,

A Computational Sectional Approach for the Flexural Creep Behavior

349

characteristic and experimental scatter values. According to MC10, the FRC is classified as 2d, and can be used in structural applications. 6

F

[MPa]

5

150 LVDT deflection

mm

4

Stress

3

LVDT CMOD

2

Exp. scatter Exp. mean

500 mm

f

1

f

650 mm

0

0

1

0.5

1.5

2

R,i,m R,i,k

2.5

3

3.5

4

CMOD [mm]

Fig. 1. (left) bending setup (right) bending results

2.2

Unloading-Reloading Bending Test

The unloading-reloading test uses the same specimens and test geometry as the monotonic bending tests, but the imposed CMOD rate is – both for the unloading as well as the reloading phase – set equal to 0.2 mm/min. Unloading cycles are imposed at CMOD = 0.1, 0.2, 0.3, 0.4, and 0.5 mm. After the 5 cycles, a monotonic test speed is imposed. Furthermore, the mid-span deflection is no longer measured but 5 LVDTs are attached to the side of the FRC prism to measure the deformation profile over the height. The test setup is shown in Fig. 2 (right), with the global and zoomed experimental results shown in the (middle) and (right) respectively. It is noted that the average monotonic response yields generally higher (post-peak) strengths, as it is expected that the cycles induce additional damage in the specimen. Global response

4.5

Zoom on cyclic response

4.5

Cyclic scatter Cyclic mean

[MPa] Stress

Stress

[MPa]

Monotonic mean

3

1.5

3

1.5

Cyclic scatter Cyclic mean Monotonic mean

0

0 0

0.5

1

1.5

2

CMOD [mm]

2.5

3

3.5

4

0

0.1

0.2

0.3

0.4

0.5

CMOD [mm]

Fig. 2. (left) cyclic bending setup with LVDT (middle) global results (right) zoom on cycles

2.3

Uniaxial Tensile Creep

In order to assess the uniaxial tensile creep deformations, a finite element model was built in which the fibers are modelled separately. The aim of the numerical model is to describe the creep deformations for a time period of 50 years. Further details about the construction and analysis of the model can be found in literature [20], but for the

350

R. Vrijdaghs et al.

purpose of this paper, it is sufficient to show the creep coefficient evolution in Fig. 3. This figure represents the uniaxial tensile creep coefficient determined from 25 different numerical analyses. 1.4 1.35 1.3

basic

[-]

1.25 1.2 1.15 1.1 1.05 1 0

3

6

9

12

Log time [s] 10

Fig. 3. Uniaxial tensile creep coefficient, determined from 25 FEM simulations

3 Sectional Analysis: Monotonic Behavior The sectional analysis considers a discretization of the notched sections: the cracked section is divided into 125 1 mm high elements, each assumed to operate in uniaxial compression or tension. As such, the approach can be considered a fictitious crack model, where the cracked section functions as a plastic hinge between two rigid bodies. All deformations are smeared over a length equal to the height of the section, such that the strain e can be related to the deformation d: e = d/125. In the sectional approach, the deformation and stress profile, d(y) and r(y) respectively, are calculated such that the horizontal and bending moment equilibria are fulfilled, given a certain constitutive law in compression and tension. 3.1

Constitutive Laws in Compression and Tension

In compression, the stress-strain curve of FRC is described by a Thorenfeldt model, which is based on the Young’s modulus and the compressive strength. In tension, two different approaches are implemented and compared. Firstly, the crack width-stress relation is based on the MC10 model, extended with the model proposed in [21]. In this model, the constitutive law can be directly computed from the compressive strength and the post-cracking parameters, measured in the EN 14651 test (fR1 and fR3). Secondly, a multi-linear crack width-stress relation is proposed, and in a numerical optimization scheme, the coordinates of each point of this curve are varied such that the resulting bending response corresponds as good as possible with the experimentally obtained result, as determined by the minimization of the sum of the absolute error between the measured and calculated bending moment, refer to Eq. (1).    X X  ð1Þ error ¼ ðyj
rj;i Þ  Mexp;i  b
h
  i j

A Computational Sectional Approach for the Flexural Creep Behavior

351

In Eq. (1), rj,i is the stress in the jth element corresponding at the ith CMOD value in the bending test. The height on an element h and the width of the section b are known, i.e. 1 mm and 150 mm respectively. In the optimization scheme, the location of the neutral axis yNA is determined such that the horizontal force error DH is less than 2 N, in accordance with Eq. (2). DHi ¼ b
h


X

rj;i

ð2Þ

j

3.2

Results of the Monotonic Bending

6

6

5

5

4

4

[MPa]

[MPa]

The results of the sectional approach for the monotonic bending test are shown in Fig. 4, with a comparison of the experimental result (as shown before) and the MC10 model as well as the numerical optimized model, for the global response (CMOD = 4 mm, left) and a zoomed in part up to CMOD = 0.5 mm (right). A very good agreement is found between the experimental and calculated results, especially with the optimized constitutive tensile law. This is expected as the latter has more degrees of freedom, i.e. post-peak points in the stress-crack width curve, to achieve a better fit.

3

Stress

Stress

3

2

Scatter

2

Scatter

Mean

1

Mean

1

MC10/SC

MC10/SC

Numerical optimization

0

0

0.5

1

1.5

2

2.5

3

Numerical optimization

3.5

4

0

0

CMOD [mm]

0.1

0.2

0.3

0.4

0.5

CMOD [mm]

Fig. 4. Comparison of the experimental and predicted monotonic bending behavior (left) global (right) zoomed

4 Sectional Analysis: Unloading Behavior The next step is the analysis of the unloading behavior. The aim is, as before, to calculate the stress and deformation profile, rLR(y) and dLR(y) respectively, as a function of the height y and of the applied load ratio LR, i.e. the load w.r.t. the maximum bearable force. To ensure residual deformation after complete unloading, the Euler-Bernoulli beam theory is abandoned and a bilinear deformation profile, rather than a linear, is assumed. Additionally, the uniaxial constitutive laws are extended to include unloading and damage laws.

352

4.1

R. Vrijdaghs et al.

Extension of the Constitutive Laws

The uniaxial compressive behavior is assumed linear-elastic, as the compressive stress remains below 45% of the strength for CMOD < 0.5 mm. It is assumed in MC10 that for these low stress levels, concrete remains elastic, such that no modifications are needed for the compressive constitutive law. In tension, especially in the cracked zone of the section, unloading incurs irreversible deformations, which needs to be taken into account. Therefore, the unloading stiffness in uniaxial tension cannot be equal to the Young’s modulus E0. The unloading stiffness E(e) depends on the scalar damage evolution D(e) in accordance with Eq. (3). E ðeÞ ¼ E 0
ð1  D ðeÞÞ

ð3Þ

The damage function is a monotonically increasing function, bound between 0 and 1, representing zero and complete loss of stiffness. In this approach, the damage function is further limited to ensure that tensile elements remain in tension during the unloading phase. 4.2

Implementation of the Inverse Analysis

The inverse analysis aims to find the bilinear deformation profile such that the horizontal and bending equilibrium is fulfilled, whilst simultaneously finding the optimal damage function. The input for the algorithm is every (CMOD, M)i point from the experimental unloading branches and the constitutive laws in compression and tension. The total error DEi is minimized per unloading point i, as defined in Eqs. (4–6). DEi ¼ DH i ¼

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2 2 DH i þ DM i b
h


X

ð4Þ

! rj;i =Fi

ð5Þ

j

DM i ¼

b
h


X

! ðyj
rj;i Þ  Mexp;i =Mi

ð6Þ

j

4.3

Results of the Unloading Analysis

The error defined in Eqs. (4–6) is calculated for all points on the unloading branches in Fig. 2, and is expressed as a percentage. The analysis shows that 99% of all points have a relative horizontal and bending moment error within ± 1% and ± 5%, with an average error of 0.0009% and 0.0073%, respectively, indicating a very good close agreement with the experimental data. In Fig. 5, the damage evolution in terms of the tensile deformation is shown. A clear and steep increase of the damage can be seen, even at very low crack widths: at a crack width of 0.1 mm, over 90% of the stiffness is already lost.

A Computational Sectional Approach for the Flexural Creep Behavior

353

1

0.8

Damage [-]

0.6

10%/90% quantile Min/max Median

0.4

0.2

0

-0.4

-0.3

-0.2

Deformation

-0.1

0

0.1

[mm]

Fig. 5. Evolution of the scalar damage parameter

As the unloading branches are input for the model, they cannot be used for the validation. However, the 5 LVDTs on the side of the FRC prism can be used to validate the calculated deformation profile, as well as the location of the neutral axis. In Fig. 6, the calculated deformation profile (solid lines) at full load (LR = 100%, black) and half load (LR = 50%, gray) is compared against the experimentally measured deformations at various y heights (circle markers). It should be noted that the measured deformations are not used in the calibration of the model, indicating that a very good agreement can be reached. Finally, the unloading deformation profile is assumed bilinear, as to allow for residual deformations, but at full load, the bilinear curve converges to a linear curve, representing an Euler-Bernoulli beam. As such, the unloading sectional analysis can be understood as an extension of the classical beam theory. CMOD = 0.1 mm

CMOD = 0.2 mm

CMOD = 0.3 mm

150

150

150

100

100

100

50

50

50

0 -0.1

0 -0.05

0

0.05

0

-0.2

CMOD = 0.4 mm

-0.1

0

-0.3

-0.2

-0.1

0

CMOD = 0.5 mm

150

150 LR = 100 % model LR = 100 % exp.

100

100

50

50

0

0

-0.4

-0.2

0

LR = 50 % model LR = 50 % exp.

-0.4

-0.2

0

Fig. 6. Comparison of the measured and predicted deformation profiles during unloading

354

R. Vrijdaghs et al.

5 Sectional Analysis: Creep Deformations The final step is the extension of the approach to include creep effects. The start point for the time-dependent analysis is the stress and deformation profile at a certain LR, which was obtained in the previous step of the analysis and is kept constant in time. The main assumption of the creep analysis is that the total deformation d(t, y) as a function of time t and height y can be written as the linear superposition of instantaneous deformations Dd multiplied by their creep coefficient /, as shown in Eq. (7). dðt; yÞ ¼

X

/k ðt; y; tk Þ
Ddk ðtk ; yÞ

ð7Þ

k

5.1

Extension of the Constitutive Laws

In Eq. (7), k refers to a certain stress redistribution at the moment tk incurring an instantaneous deformation Ddk. The stress redistribution might be necessary due to two reasons: (1) a certain element in the discretized section cannot take any further force due to excessive straining (2) a time-varying load is applied. In this approach, the load is assumed constant, so that (2) is not allowed to happen. On the other hand, the effect of excessive straining is investigated by (artificially) limiting the uniaxial tensile deformation capacity of the cracked concrete. Note that such an effect is not experimentally or numerically encountered, and the analysis is purely from illustrative purposes to highlight the capabilities of the model. Additionally, excessive compressive straining is neglected. Each instantaneous deformation Ddk can be related to an instantaneous stress change in each element, depending on whether this element is in compression or tension. After each stress redistribution, a new equilibrium is sought. Equations (8–9) show the incremental constitutive laws in compression and tension.     Drj ti ; yj ¼ E0
Dej ti ; yj

ð8Þ

        Drj ti ; yj ¼ E0
1  D dj ti1 ; yj
Dej ti ; yj

ð9Þ

The compressed zone is assumed visco-elastic, while the previously defined scalar damage function is included in the cracked tensile law. Note that this incremental description of the constitutive laws does not take into account the change in deformation due to creeping, and that the assumption of a bilinear deformation profile is abandoned. This is a significant simplification in the model, and further work is underway to include the interaction between creeping and continuous stress changes in the cracked section. Finally, the creep coefficients are determined for the compressive and tensile zone separately. For the former, the evolution according to MC10 is adopted for undamaged linear visco-elastic concrete. As shown previously, the creep evolution of the cracked tensile zone is determined from a two-phased numerical analysis. Both creep coefficients are shown in Fig. 7.

A Computational Sectional Approach for the Flexural Creep Behavior

355

1.4

4

1.35 1.3

3 [-]

1.25

2.5

basic

Creep coefficient

[-]

3.5

1.2 1.15

2 1.1

1.5 1 1 s.

1.05

1 m.

1 h.

1 m.

1 d.

1

1 y.

50 y.

0

3

Time

6

Log

10

9

12

time [s]

Fig. 7. Evolution of the creep coefficient in (left) compression (right) cracked tension

5.2

Results of the Creep Analysis

The sectional analysis takes thus input from the constitutive laws, an initial deformation and stress profile and the creep coefficient evolution to determine flexural creep in a cracked section from uniaxial creep descriptions. 1.4 Min/max

1.2

10%/90% quantile Mean

CMOD [mm]

1

0.8

0.6

0.4

0.2 1 s.

1 m.

1 h.

1 d.

1 m.

1 y.

50 y.

Time

Fig. 8. CMOD evolution of a cracked section

The global behavior of a flexural creep test can be expressed as the CMOD evolution as a function of time. As previously highlighted, the effect of excessive tensile straining is investigated by limiting the allowable tensile strain in the cracked zone. This value is (arbitrarily) chosen as the maximum calculated tensile strain in the cracked section before unloading, but the model allows for any limit value to be imposed (or none at all), based on experimental data or numerical analysis. The evolution of the CMOD for a 50 year sustained load at LR = 50% is shown in Fig. 8 for 150 different simulations.

356

R. Vrijdaghs et al. 0.9

50 Min/max

40

Min/max

0.8

10%/90% quantile

10%/90% quantile Mean

Mean

0.7

[MPa]

0.6

T

C

[MPa]

30

20

0.5 10

0 1 s.

0.4

1 m.

1 h.

1 d.

Time

1 m.

1 y.

50 y.

0.3 1 s.

1 m.

1 h.

1 d.

1 m.

1 y.

50 y.

Time

Fig. 9. Evolution of the maximum (left) compressive and (right) tensile stress in the section

While most simulations indicate a relatively modest CMOD growth, the majority exceeds the initial crack width of 0.5 mm after 50 years under simulated loading. Since the model also accounts for stress redistributions when the limit strain is encountered, the maximum compressive and tensile stress in the cracked zone can be evaluated as well. The evolution of both is shown in Fig. 9. The compressive stress evolution (left) can exceed the compressive strength of the material for some simulations, similar to the tensile stress (right). The exceedance of the strength values indicates structural failure. This is an important result of the model, as it indicates that it is capable not only to predict the crack width growth (even under various simplifications), but also to predict structural failure under sustained loading. Further work is still ongoing on removing the various simplifications of the threestepped model and to extend its capabilities.

6 Conclusions In this paper, the results of a sectional approach to model flexural creep based on uniaxial creep data is presented. The sectional approach considers the cracked section of the EN 14651 beam in a plastic hinge formulation and consists of three steps: monotonic loading, unloading, and creep. The model is validated with experimental data, both on a global scale (stress-CMOD curves) as well as on a local scale (deformation profiles). The main conclusions are summarized as follows: • A fast algorithm is implemented to calculate and predict the monotonic bending behavior based on simplified uniaxial constitutive laws. Furthermore, a uniaxial tensile law optimization scheme is implemented to extend the MC10 model, which further improves the agreement between the predicted and experimental result. • The proposed sectional approach can also predict the unloading behavior of the cracked sections, provided the uniaxial laws are extended to include unloading. This is implemented through the use of a scalar damage function and a simplified bilinear deformation profile under unloading. The proposed algorithm achieves low errors in the equilibrium equations and the deformation profile is validated against measured deformations along the height of the beam.

A Computational Sectional Approach for the Flexural Creep Behavior

357

• Finally, the uniaxial creep behavior is included in the analysis, to predict flexural creep from uniaxial creep data. In compression, the creep deformations are based on the viscoelastic concrete behavior assumed in MC10. In tension, a two-phased numerical model predicts the time-dependent crack opening input for the analysis. The results indicate a slow CMOD increase under a 50 year sustained load. More fundamentally, the model is also capable to predict structural failure. Further works remains in removing as much as possible the various assumptions in the different steps, but the proposed model highlights important structural aspects of cracked FRC under sustained loading.

References 1. Balaguru, P., Shah, S.P.: Fiber-Reinforced Cement Composites. Mcgraw-Hill, Texas (1992) 2. Bentur, A., Mindess, S.: Fibre Reinforced Cementitious Composites, 449 p. Elsevier Science Publishers Ltd, London (1990) 3. di Prisco, M., Plizzari, G., Vandewalle, L.: Fibre reinforced concrete: new design perspectives. Mater. Struct. 42(9), 1261–1281 (2009) 4. Daniel, J.I.: ACI Committee 544. State-of-the-Art Report on Fiber Reinforced Concrete. American Concrete Institute (2002) 5. di Prisco, M., Felicetti, R., Plizzari, G.A. : Fibre-reinforced concrete. In: RILEM Proceedings of the 6th RILEM Symposium (BEFIB 2004), PRO39, BEFIB 2004. RILEM Publications S.A.R.L., Bagneux (2004) 6. Rossi, P., Chanvillard, G.: Fibre reinforced concrete. In: RILEM Proceedings of the 5th RILEM Symposium (BEFIB 2000), PRO15, BEFIB 2000. RILEM Publications S.A.R.L., Bagneux, France 7. Grzybowski, M., Shah, S.P.: Shrinkage cracking of fiber reinforced concrete. ACI Mater. J. 87(2), 138–148 (1990) 8. Pasini, F., et al.: Experimental study of the properties of flowable fiber reinforced concretes. In: di Prisco, M., Felicetti, R., Plizzari, G.A. (eds.) 6th International RILEM Symposium on Fibre Reinforced Concretes, pp. 279–288. RILEM Publications SARL (2004) 9. Barros, J., et al.: Post-cracking behaviour of steel fibre reinforced concrete. Mater. Struct. 38 (1), 47–56 (2005) 10. di Prisco, M., Ferrara, L., Lamperti, M.G.L.: Double edge wedge splitting (DEWS): an indirect tension test to identify post-cracking behaviour of fibre reinforced cementitious composites. Mater. Struct. 46(11), 1893–1918 (2013) 11. Buratti, N., Mazzotti, C., Savoia, M.: Post-cracking behaviour of steel and macro-synthetic fibre-reinforced concretes. Constr. Build. Mater. 25(5), 2713–2722 (2011) 12. fédération internationale du béton (fib), Model Code 2010 First complete draft (2010) 13. Kurtz, S., Balaguru, P.: Postcrack creep of polymeric fiber-reinforced concrete in flexure. Cement Concrete Res. 30(2), 183–190 (2000) 14. MacKay, J., Trottier, J.F.: Post-crack creep behavior of steel and synthetic FRC under flexural loading. In: Shotcrete: More Engineering Developments, pp. 183–192. Taylor & Francis (2004) 15. Zerbino, R.L., Barragan, B.E.: Long-term behavior of cracked steel fiber-reinforced concrete beams under sustained loading. Mater. J. 109(2), 215–224 (2012) 16. Kiang Hwee, T., Mithun Kumar, S.: Ten-Year Study on Steel Fiber-Reinforced Concrete Beams Under Sustained Loads. ACI Struct. J. 102(3), 472 (2005)

358

R. Vrijdaghs et al.

17. Arango, S., et al.: A comprehensive study on the effect of fibers and loading on the flexural creep of SFRC. In: 8th RILEM International Symposium on Fibre Reinforced Concrete (BEFIB 2012). RILEM Publications S.A.R.L., Guimarães, Portugal (2012) 18. Babafemi, A.J., Boshoff, W.P.: Tensile creep of macro-synthetic fibre reinforced concrete (MSFRC) under uni-axial tensile loading. Cement Concrete Composites 55, 62–69 (2015) 19. European Committee for Standardization, EN 14651 Test method for metallic fibered concrete - Measuring the flexural tensile strength (limit of proportionality (LOP), residual), p. 17 (2005) 20. Vrijdaghs, R., di Prisco, M., Vandewalle, L.: A numerical model for the creep of fiber reinforced concrete. In: Hordijk, D.A., Luković, M. (eds.) FIB Symposium 2017: High Tech Concrete, pp. 366–373. Springer, Maastricht (2017) 21. di Prisco, M., Colombo, M., Dozio, D.: Fibre-reinforced concrete in fib Model Code 2010: principles, models and test validation. Struct. Concrete 14(4), 342–361 (2013)

Shrinkage of Steel-Fibre-Reinforced Lightweight Concrete Hasanain K. Al-Naimi and Ali A. Abbas(&) University of East London, London, UK [email protected]

Abstract. Long-term behaviour of steel fibre reinforced concrete remains rather unknown and to a large extent unquantified by equations and standards. This paper studies experimentally the free drying shrinkage of steel fibre reinforced lightweight concrete during the first 28 days using 100  100  500 mm beams. The coarse lightweight material tested (LYTAG) is recycled and offers an alternative to gravel and quarry resources which are subjected to depletion in the future. Also, this material can lead to reduction in the mass of the structure which results in economical designs. However, LYTAG aggregate can absorb up to 15% of its own weight in water. This makes it susceptible to drying shrinkage both at young age and long-term due to environmental diffusion. Shrinkage can have a detrimental effect on the concrete by inducing cracks, creating therefore weak zones in the concrete. It is thought that fibres can have a favourable effect on the reduction of shrinkage due to their ability to bridge cracks. This could be vital particularly in large concrete flat slabs, joints, beams and even columns. This project uses modern hooked-end DRAMIX 3D and 5D fibres with different dosages Vf and number of hooks and evaluates shrinkage for concrete with different characteristic strengths fck. KEYWORDS: Lytag
Hooked-end fibres
Early-age shrinkage shrinkage
Shrinkage beams
Environmental diffusion
Cracks


Drying

1 Introduction Shrinkage is the time-dependent change in volume of unrestrained concrete when tensile stresses due to contraction exceeds that of the concrete itself, although creep can play a factor in counter acting the latter due to stress relaxation in a restrained structure (Hossain, 2003; Havlasek 2014). Shrinkage can take place due to either internal reactions usually before concrete hardening, responsible for by autogenous, plastic and chemical shrinkage, or due to their surrounding environment responsible for by thermal and drying shrinkage that leads to water evaporation through concrete pores (Neville 2011). This leads to shrinkage cracking at the surface which can negatively affect the strength and integrity of the concrete from a young age to long-term (Mindess and Young 1981). Shrinkage can be affected by wind, humidity, temperature, cement fineness, water content and curing (Shoya 1979; Mehta 1994). The lightweight aggregate LYTAG used in this work is porous and capable of absorbing water of up to 15% of its weight. Besides, no pore size reducing materials such as silica fume are used © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 359–367, 2021. https://doi.org/10.1007/978-3-030-58482-5_33

360

H. K. Al-Naimi and A. A. Abbas

for the concrete mixed. Hence, this makes it easy for the water to evaporate out of the concrete and act as a catalyst for drying shrinkage unlike the denser high strength concrete (Mehta and Monteiro 1993; Koenders 1997). The lower modulus of elasticity of lightweight concrete can also facilitate the shrinkage of concrete (Kayali et al. 1835). Therefore, lightweight concrete is more susceptible to shrinkage than normal weight concrete (Neville 2011). This however has been proven wrong in some other studies where lightweight concrete showed less shrinkage than normal weight concrete (Zhang et al. 2005). Lightweight concrete is noted by its brittle nature in tension due to the absence of tension toughening aggregate interlock mechanism which causes micro- and macrocracking. The incorporation of steel fibres in the concrete mix has been shown and proven to increase the tensile strength of the concrete by bridging the crack and maintain tensile stress flow in the concrete (Gao et al. 1997; Campione and La Mendola 2004; Abbas et al. 2014; Di Prisco et al. 2013; Grabois et al. 2016; Mo et al. 2017). Provided that shrinkage cracking only takes place if the contraction due to water dissipation to environmental diffusion or cement hydration is higher than the concrete tensile strength, steel fibre reinforcement can offer a control mechanism to shrinkage cracking at all ages. Previous studies on lightweight concrete agree with the latter (Tan et al. 1994; Zhang et al. 2001). Since the current design codes such as the Eurocode 1992-1 use the drying shrinkage to calculate the long term shrinkage strain due to drying shrinkage, while as erroneous as it might be, autogenous shrinkage is assumed to be controlled in the choice of material mixing and curing conditions, this work focuses on the measurement of the free unrestrained drying shrinkage of SFRLC.

2 Experimental Study 2.1

Experimental Programme

The experimental program included 3 mixes with 3 different fibre dosages: Vf = 0%, 1% and 2%. Each mix constituted of 3 specimens. The mixes also aimed to study the effect of different W/C ratios and type of fibre (Table 1). The latter was either 3D fibres regarded as the most commonly used in industry (Sadoon et al. 2018) or the 5D fibres with the highest tensile strength and most comprehensive hooking system which promises of being capable of primary reinforcement substitution. Table 1. The specimens cast Mix 1 2 3

fck (MPa) Vf (%) 30 0, 1, 2 40 30

Fibre type 3D 5D 5D

Shrinkage of Steel-Fibre-Reinforced Lightweight Concrete

361

It should be noted that the specimens were kept in an environmental chamber with unrestrained supports. 2.1.1 Materials Portland-Limestone cement (CEM 11) according to the specification supplied in EN 197-1 was used. Coarse aggregate Lytag, also known as Sintered Pulverised Fuel Ash Lightweight Aggregate (LYTAG) was provided by LYTAG Ltd. The loose dry density of LYTAG was calculated in the lab to be approximately 760 kg/m3 while the water absorption was estimated to be around 15% per mass of LYTAG. For this reason the aggregates were pre-soaked for 24 h before mixing. This was proven to help reduce shrinkage. Natural river sand with a 4.75 mm maximum size was used as the fine aggregate of the concrete. The sand had a water absorption coefficient of 0.09% and specific gravity of 2.65 complying with BS EN 12620. The properties of the fibres used are summarized in the table below. It should be noted that to prevent the possibility of balling, fibres were collated from the manufacturer (Table 2). Table 2. Properties of fibres Fibre type ru (MPa) lf (mm) df (mm) 3D 65/60 1160 60 0.9 5D 65/60 2300 60 0.9

2.1.2 Mix Design The mix designs used are summarized in Table 3 below. The mix design of the Lytag concrete for the characteristic cylinder and cube compressive strengths used are summarised below. These were directly adopted from Lytag (2011) manuals. Table 3. Mix design used (fck/fck, cube) LC30/33 LC40/44

Cement (kg/m3) 370 480

Sand (kg/m3) 592 485

Loose bulk Lytag (kg/m3) 668.8 668.8

Effective water (kg/m3) 175 175

Calculating the water content of Lytag was of high importance as Lytag aggregates were found to absorb water of approximately 15% of their weight which is also confirmed by Lytag manual. For this reason and as suggested by Lytag manual 5 (2011), excess water to saturate Lytag aggregates was added 30 min before mixing (Fig. 1).

362

H. K. Al-Naimi and A. A. Abbas

Fig. 1. Mixing process for plain and fibrous lightweight concrete

2.2

Experimental Tests

To measure the free axial early age and drying shrinkage, 100  100  500 mm beam specimens designed according to the European standard UNI 6555 with steel inserts fixed into both ends of the specimens were cast and the method of measuring shrinkage was identical to UNI 6555. This relatively large specimen was preferred to other smaller specimens such as the more conventional 75  75  280 mm specimen according to BS ISO 1920-8:2009 in an attempt to avoid any possible favorable orientation of fibres in the shrinkage direction i.e. orthogonal to crack plane for the SFRLC beams. The specimens are required to be housed in a measuring apparatus with a digital displacement guage for the duration of the test with a constant relative humidity of (50 ± 5)% and temperature of (22 ± 2) °C (Fig. 2). To measure shrinkage strain, simply

Fig. 2. Concrete specimen about to be housed in the measuring apparatus

Shrinkage of Steel-Fibre-Reinforced Lightweight Concrete

363

the digital readings of the displacement are taken and divided by the total length of the specimen. Shrinkage strains from repeated specimens are averaged. The drying shrinkage is assumed to start from day 2 following early age shrinkage which is recorded 24 h after concrete hardening.

3 Results and Discussion Figure 3 below shows the recorded free drying shrinkage for the 2 different W/C ratios chosen. Throughout the duration of the test, the mix with the higher W/C ratio (fck = 30 MPa) seemed to shrink more than that with the lower W/C ratio (fck = 40 MPa). Both the readings reduced in gradient as the days of testing were increased. It should be noted that while the shrinkage of the mix with the low W/C ratio relatively plateaued at the end of the test, that of the mix with the high W/C ratio continued to increase (75% more shrinkage). Figure 4 below studies the effect of hook geometry on drying shrinkage for 2 fibrous beams with Vf = 1%. It can be seen that while the free shrinkage of the fibrous beam reinforced with the stronger bonded 5D was slightly higher during the first 14 days, the free shrinkage of the fibrous beam reinforced with 3D fibres became higher during the remaining days of the test. Figure 5 below illustrates the effect of increasing the fibre volume fraction Vf on shrinkage. It is clear that the higher the fibre dosage, the lower the shrinkage as the tensile strength of fibres prevent cracking of the concrete beams. The axial shrinkage strain of the beam reinforced with Vf = 1% was about 56% of that of the plain lightweight concrete beam, while that reinforced with Vf = 2% was only 23%.

Drying shrinkage for different W/C ratios 140

Free shrinkage (με)

120 100 80 fck=30MPa, Vf=0% 60

fck=40MPa, Vf=0%

40 20 0 0

5

10

15 Days

20

25

30

Fig. 3. Effect of W/C ratio on shrinkage

364

H. K. Al-Naimi and A. A. Abbas

Drying shrinkage for different fibre types 250

Free shrinkage (με)

200 150 fck=30MPa, Vf=1%, 5D

100

fck=30MPa, Vf=1%, 3D 50 0 0

5

10

15 Days

20

25

30

Fig. 4. Effect of fibre type on shrinkage

Drying shrinkage for differnet Vf 160

Free Shrinkage (με)

140 120 100 80

fck=30MPa, Vf=0%, 5D

60

fck=30MPa, Vf=1%, 5D fck=30MPa, Vf=2%, 5D

40 20 0 0

5

10

15

20

25

30

Days

Fig. 5. Influence of increasing fibre dosage on the drying shrinkage of lightweight concrete

The column chart below (Fig. 6) displays the early age shrinkage due to drying shrinkage and cement hydration process of the concrete. Overall, it is evident that the early age shrinkage is responsible for the majority of the shrinkage at the end of the 28 day test which makes it impactful. This agrees with findings reported by Holt (2001).

Shrinkage of Steel-Fibre-Reinforced Lightweight Concrete

365

For example, during the first day after hardening for the specimen with fck = 30 MPa and Vf = 0%, shrinkage strain was recorded to be 742 microstrains which is 542% higher than that recorded at the end of the testing at 28 day. This is due to the porous nature of the lightweight concrete and the ability of the aggregates to absorb 15% of their weight in water which makes water loss due to environmental diffusion more likely. Another observation that can be made is with regards to fibre geometry. It appears that the more complicated anchorage system of the 5D fibres created more air voids in the concrete lattice making it more easily for the water to dissipate as compared to the specimens reinforced with the less complicated hook-ended 3D fibres. Similarly, the higher fibre fraction led to higher early age shrinkage. Hence, it can be deduced that a better vibration process, smaller fibres, less extended hooks or high humidity curing process is recommended for better shrinkage control.

Early age shrinkage (με) 800 700 600 500 400 300 200 100 0

Fig. 6. Early age shrinkage for SFRLC beams

4 Conclusions • The methodology adopted in this work was successful at measuring the early age and drying shrinkage for fibrous lightweight concrete beams. • The higher the water cement ratio, the higher the drying shrinkage measured. • The higher the fibre dosage, the lower the drying shrinkage recorded. • The more extensive the hook anchorage system is the less the drying shrinkage. However, it was seen that the extensive 5D hooks lead to a more pronounced increase in early age shrinkage as compared to the more conventional 3D fibres.

366

H. K. Al-Naimi and A. A. Abbas

• For the duration of the test and the material used, the early age shrinkage during the first day after hardening appeared to be 300% to 500% higher than the drying shrinkage by the end of the test. • The porous nature of lightweight concrete and the ability of its aggregate to absorb up to 15% of their weight in water make the concrete more susceptible to shrinkage. • The addition of fibres creates more voids in the lightweight concrete, making it likely to shrinkage especially during early age. • Therefore, to control early shrinkage, a better compacting process, smaller fibres, less extended hooks or 99% humidity curing is recommended during early age.

References Hossain, A.B.: Assessing residual stress development and stress relaxation in restrained concrete ring specimens. Ph. D Disseratation. Purdue University (2003) Havlasek, P.: Czech Technical University in Prague (2014). http://mech.fsv.cvut.cz/wiki/images/ c/ca/PhD_thesis_Havlasek_2014.pdf. Accessed 8 July 2015 Neville, A.: Properties of Concrete, 5th edn. Pearson, Harlow (2011) Mindess, S., Young, J.F.: Concrete, vol. 181. Patience-Hall, Inc., New York, (1981). 671 p. Shoya, M.: II-5 Drying Shrinkage and Moisture Loss of Super Plasticizer Admixed Concrete of Low Water Cement Ratio. Transactions of the Japan Concrete Institute (1979) Mehta, P.K.: Concrete technology at the crossroads-Problems and opportunities. American Concrete Institute, SP-144, pp. 1–30 (1994) Mehta, P.K., Monteiro, J.M.: Concrete: Structure, Properties and Materials, 2nd ed. Prentice Hall, Inc., New York (1993) Koenders, E.A.B.: Simulation of Volume Changes in Hardened Cement-Based Materials. Delft University Press, Delft (1997) Kayali, O., Haque, M., Zhu, B.: Drying shrinkage of fibre-reinforced lightweight aggregate concrete containing fly ash. Cem. Concr. 121 Res. 29(11), 1835–1840. http://www. sciencedirect.com/science/article/pii/S0008884699001799. Accessed 15 Jan 2020 Zhang, M., Zakaria, M., Dang, L., Paramasivam: Shrinkage of high-strength lightweight aggregate concrete exposed to dry environment. ACI 129 Mater. J. 102(2). http://www. researchgate.net/publication/250613564_Shrinkage_of_highstrength_lightweight_aggregate_ concrete_exposed_to_dry_environment. Accessed 15 Jan 2020 Gao, J., Sun, W., Morino, K.: Mechanical properties of steel fiber-reinforced, high-strength, lightweight concrete. Cement Concr. Compos. 19(4), 307–313 (1997) Campione, G., La Mendola, L.: Behavior in compression of lightweight fiber reinforced concrete confined with transverse steel reinforcement. Cement Concr. Compos. 26(6), 645–656 (2004) Abbas, A., Syed Mohsin, S., Cotsovos, D.: Seismic response of steel fibre reinforced concrete beam–column joints. Eng. Struct. 59, 261–283 (2014) Di Prisco, M., Colombo, M., Dozio, D.: Fibre-reinforced concrete in fib Model Code 2010: principles, models and test validation. Struct. Concr. 14(4), 342–361 (2013) Grabois, T., Cordeiro, G., Filho, R.: Fresh and hardened-state properties of self-compacting lightweight concrete reinforced with steel fibers. Constr. Build. Mater. 104, 284–292 (2016)

Shrinkage of Steel-Fibre-Reinforced Lightweight Concrete

367

Mo, K., Goh, S., Alengaram, U., Visintin, P., Jumaat, M.: Mechanical, toughness, bond and durability-related properties of lightweight concrete reinforced with steel fibres. Mater. Struct. 50(1) (2017) Tan, K., Paramasivam, P., Tan, K.: Creep and shrinkage deflections of rc beams with steel fibers. J. Mater. Civ. Eng. 6(4) (1994). http://ascelibrary.org/doi/abs/10.1061/(ASCE)0899-1561 (1994)6:4(474). Accessed 15 Jan 2020 Zhang, J., Li, V.: Influences of fibers on drying shrinkage of fiber-reinforced cementitious composite. J. Eng. Mech. 127(1). http://ascelibrary.org/doi/abs/10.1061/%28ASCE%2907339399%282001%29127%3A1%2837%29?journalCode=jenmdt. Accessed 15 Jan 2020 Sadoon, A., Rees, D.W.A., Ghaffar, S.H., Fan, M.: Understanding the effects of hooked-end steel fibre geometry on the uniaxial tensile behaviour of self-compacting concrete. Constr. Build. Mater. 178, 484–494 (2018) Lyag: Technical manual. Lytag ltd., London (2011) Holt, E.: Early age autogenous shrinkage of concrete. Doctoral dissertation ed. University of Washington, Seatle (2001)

Time Dependent Deflection of FRC Members Under Sustained Axial and Flexural Loading Murray Watts1, Ali Amin1(&), R. Ian Gilbert2, and Walter Kaufmann3 1

3

School of Civil Engineering, The University of Sydney, Sydney, Australia [email protected] 2 School of Civil and Environmental Engineering, The University of New South Wales, Sydney, Australia Institute of Structural Engineering (IBK), ETH Zürich, Zurich, Switzerland

Abstract. The inclusion of steel or polypropylene fibres into a concrete matrix can considerably improve the serviceability performance of reinforced concrete members. The benefits of including fibres in structural concrete have been extensively studied, and as a result, provisions for strength, and short-term serviceability conditions are contained in national codes of practice such as the Australian Standards for Concrete Structures and Concrete Bridges. Provisions relating to the long-term serviceability behaviour of fibre reinforced concrete (FRC) are either not included or can be seen to provide limited guidance to designers. This paper describes a method of analysis that can be applied to predict the time-dependent behaviour of cracked fibre (steel or macro-synthetic) reinforced concrete. The model is versatile and can handle a wide range of geometries, material properties and loading conditions. The layered modelling approach provides a high level of flexibility which allows for the consideration of variable creep, shrinkage and fibre properties, as a function of time. Results from the model have been compared to existing experimental data available in the literature and have been shown to correlate well. In addition, a sample analysis is presented to demonstrate the effects of residual tensile stress, tensile creep and variable shrinkage gradients on a FRC flexural section. Keywords: Fibre
Concrete
Combined axial/flexural loading loading
Serviceability
Deformation
Cracking
Creep


Sustained

1 Introduction Satisfying the serviceability limit state is one of the primary objectives when designing reinforced concrete structures. This is partly due to aesthetic concerns, since cracked or excessively deformed of concrete may detract from architectural finishes and impair the functionality of a structure. It is also important from a durability perspective, as wider crack widths can allow the penetration of aggressive solutions that can cause damage to internal reinforcing steels [1]. The inclusion of fibres into a concrete matrix has been investigated and shown to significantly improve the serviceability performance of reinforced concrete members [2, 3]. To date, with respect to in-service behaviour, modelling and subsequent provisions included in National Standards have revolved around short-term behaviour. Currently the Australian Standards for Concrete © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 368–379, 2021. https://doi.org/10.1007/978-3-030-58482-5_34

Time Dependent Deflection of FRC Members

369

Structures and Concrete Bridges [4, 5] contain no provisions for predicting the longterm serviceability performance of FRC. This can be partially attributed to the lack of experimental data available due to the onerous nature of long-term testing but, also, the research focus on the strength, and short-term behaviour found in the literature. Failure to account for creep and shrinkage of the concrete and creep of the fibres, if any, can lead to a serious under-estimation of both the extent and severity of cracking and long-term deformations of FRC members [6]. The development of simple and reliable methods to accurately predict long-term cracking and deformation is essential if the full potential of FRC is to be realised in practice. The model presented herein aims to solve this problem. The model is versatile as seen in its ability to handle a range of inputs but remains simple enough for routine use. The model presented in this paper is an extension of the modelling approaches found in Gilbert and Bernard [6], Gilbert and Amin [7] and Watts et al. [8].

2 Assumed Constitutive Relationships The first step towards the development of serviceability models for efficient analysis and design of FRC members is to formulate accurate and reliable underpinning models of material behaviour. These include the material constitutive behaviour of the FRC as well as the long-term deformational characteristics of FRC in tension and compression. 2.1

Uncracked FRC

Under in-service conditions, maximum compressive stresses rarely exceed 50% of the maximum compressive strength of the concrete. For the modelling undertaken in this paper, it is assumed that the compressive concrete remains in this range and is modelled as linear elastic. Specific compressive properties can be determined through experimental testing, or through formulations given in [4, 5, 9]. If an initial stress, r0, is applied to a concrete element, an elastic strain, ece, develops in the concrete. If the stress is sustained over a time interval (i.e. between tk-1 & tk), a creep strain, Decc, develops in the concrete as a result of the initial stress. Over the same time interval, a shrinkage strain, Decs, will also develop. These conditions are summarised in Eq. (1): ece ¼ r0 =Ec ; Decc ¼ ucc ðtk ; tk1 Þ  ece ; Decs ¼ Decs ðtk Þ  Decs ðtk1 Þ

2.2

ð1Þ

FRC in Tension

The most fundamental property when considering the design of structural concrete containing fibres is its post cracking, or residual tensile strength [10–12]. This is typically characterised by its stress verses crack opening displacement (COD) (r−w) relationship. This relationship is primarily influenced by the geometry, material and dosage of the fibres as well as the strength of the matrix in which the fibres are cast. The recent revision of the Australian Standard [5] characterises the tensile response of

370

M. Watts et al.

FRC through the concrete cracking stress fct, the stress at a COD of 0.5 mm, f0.5, and the stress at a COD of 1.5 mm COD, f1.5. From these defined points the entire shortterm behaviour of the fibre can be reasonably determined by interpolating and extrapolating between these points provided the stress remains greater than 0 MPa. In this paper, wf is defined as the crack width in which the fibres carry no stress: wf ¼

1 ½mm þ 0:5 ½mm 1  f1:5 =f0:5

ð2Þ

The stress carried by the fibres, rw, at any COD may be determined as: rw ¼ f0:5

wf  w wf  0:5 ½mm

ð3Þ

The r−w relationship is the most suitable method of modelling the post cracking behaviour of FRC. However, it can be troublesome and often not compatible when implementing into structural analysis and design procedures. For this reason, it is beneficial to convert the r−w relationship to a r−e relationship. This is conveniently achieved through the calculation of a characteristic length, lcs which is accounted for in the fib Model Code 2010 [9] and facilitates the definition of longitudinal strain as a function of crack width: e ¼ w=lcs

ð4Þ

For FRC elements with conventional longitudinal reinforcing steel, lcs may be taken as: lcs ¼ minðsr ; ðD  dn ÞÞ

ð5Þ

where sr is the distance between primary cracks, D is the total depth of the section and dn is the depth to the neutral axis from the extreme compressive fibre. In this paper, the crack spacing in FRC members is predicted based on the adaptation of the Tension Chord Model proposed by Marti et al. [13] accounting for the presence of the fibres [14, 15]: sr ¼

ðfct  f0:5 Þdb ðD  dn Þb 12fct Ast

ð6Þ

where db is the diameter of the longitudinal reinforcement, D is the depth of the member, b is the width of the member, dn is the neutral axis of the section and Ast is the total area of tensile longitudinal reinforcement.

Time Dependent Deflection of FRC Members

371

σ

σ fct

fct w

1 0

Ec=E t ε cr

f0.5 f1.5

ε w

0

ε cr 0

0

(a)

0.5 mm

1.5 mm

wf

w

(b)

Fig. 1. Stress vs strain/COD(w) relationship for softening FRC: (a) Actual Behaviour; (b) Modelled Behaviour

2.3

Creep and Shrinkage of FRC Matrix

In this paper, uncracked concrete creep and shrinkage are modelled using the provisions in fib MC2010 [9]. The creep coefficients are defined as a function of the concrete strength, relative humidity and time at first loading. The tensile creep of cracked FRC is modelled through a creep amplification factor, kfcc. This is a crude approximation of the creep that occurs within the fibres, if any (i.e. which might be the case for polypropylene fibres), as well as any loss of frictional bond between the fibres and concrete matrix with time. For the purposes of modelling, we take the tensile creep as: uc;av ¼ kfcc ucc ¼ kfcc ðecc =ece Þ

ð7Þ

where kfcc = 1 for steel fibres and kfcc > 1 for polymer fibres. We note that this is an area that requires further research, testing and modelling. 2.4

Conventional Reinforcing

When serviceability levels of loading are applied to FRC members, the stress in the conventional reinforcing steels are typically well below the yield stress of the steel. In this paper it is assumed that the reinforcing steels remain in the linear elastic range through time and therefore the reinforcing strain, es, can be determined as: es ðtÞ ¼ rs =Es

ð8Þ

3 Layered Method of Analysis The layered approach presented herein is a common approach to the solution of structural problems. Implementing a discretized layered approach provides significant flexibility in analysis and allows the designer to assign variable properties to each of the layers in the analysis. The assignment of variable properties is well suited to the analysis of FRC structures where the fibres response in tension contributes in cracked

372

M. Watts et al.

layers only. As part of the analysis, a conventionally reinforced FRC section symmetrical about its vertical axis is discretised into n layers of D/n thickness. This is represented in Fig. 2 [8].

εr

ε0

σs1

1

σ0

As1

f ct

z As2

i

εi

Dn

ds2

D

ds1

y

κ

fi

σs2

n

bi

(a) Section

(b) Strains at crack

(c) Stresses at crack

Fig. 2. Layered Approach. (a) Typical Section. (b) Strain Resultant. (c) Layer Stress Resultant.

The sectional analysis is somewhat complicated by the softening behaviour of the cracked concrete region in which the fibres are engaged and resist tension. For implementation of the method described below, a simplification is made in the evaluation of the stiffness provided by cracked FRC layers. In this paper, we define the effective modulus of FRC (refer to Fig. 3), Ec,eff(i), as: ( Ec;effðiÞ ¼

fw;i ec;i

Ec  Ec ¼ efctcr

. . . for ec;i \ecr . . . for ec;i  ecr

ð9Þ

σ Ec 1

fct

1 Ec,eff(i)

fi 0

ε cr

εi

ε

Fig. 3. Effective Modulus of FRC

If the layer is uncracked, then the effective modulus is equal to the elastic modulus of the concrete, Ec. However, if the layer is cracked then the effective modulus is defined based on the fibre tensile response. The effective modulus of each cracked layer is defined at the completion of the instantaneous analysis and is dependent on the extent of cracking identified. It is assumed that the effective modulus remains constant for

Time Dependent Deflection of FRC Members

373

each layer throughout each subsequent time interval. The concrete rigidities (cracked or uncracked) for the cross-section are therefore defined as: RA;c ¼

n X

AcðiÞ Ec;effðiÞ ; RB;c ¼

i¼1

n X

AcðiÞ Ec;effðiÞ yi ; RI;c ¼

i¼1

n X

AcðiÞ Ec;effðiÞ y2i

ð10Þ

i¼1

where Ac(i) is the cross-sectional area of the i-th concrete layer, and yi is the distance between the centroid of the i-th layer and the chosen reference axis. Similarly, the steel rigidities are defined as a function of the number of steel layers, ns: RA;s ¼

ns X

AsðiÞ Es ;

i¼1

RB;s ¼

ns X

AsðiÞ Es yi ;

i¼1

RI;s ¼

ns X

AsðiÞ Es y2i

ð11Þ

i¼1

The corresponding instantaneous strain at the reference axis and the instantaneous curvature produced by the applied actions can then be directly evaluated as: 

   RI RB Next þ Mext RA RI  R2B RA RI  R2B

ð12aÞ

   RB RA Next þ Mext RA RI  R2B RA RI  R2B

ð12bÞ

er ¼  j¼

where RA, RB and RI are the overall cross-sectional rigidities and equal to: RA ¼ RA;c þ RA;s ;

3.1

RB ¼ RB;c þ RB;s ;

RI ¼ RI;c þ RI;s

ð13Þ

Instantaneous Cracked Sectional Analysis at First Loading (t0)

To determine the instantaneous response at time step t0 for a section subjected to an applied moment and/or axial force, the section is first assumed to be uncracked. In the uncracked state, Ec,eff(i) is constant for all layers and is equal to the intact concrete elastic modulus, Ec (see Eq. (8)). Using the corresponding sectional rigidity characteristics, Eq. (13), the strain at the reference axis, er and curvature of the section, j are calculated using Eq. (12). A search is then undertaken to identify whether the tensile stress of the concrete has been exceeded in one or more concrete layers (ec,i > ecr = fct/Ec). If cracking is not identified, the section is solved, and computation is terminated. If cracking is identified, the cracked concrete tensile stress in each cracked layer is set equal to the residual tension carried by the fibres based on the FRC material stress verses strain relationship (as defined in Fig. 1b and Eq. (2) and (3)). The stress in each of the remaining uncracked layers is equal to rc(i),0 = Ec,eff(i)  ec,i with Ec,eff(i) equal to the intact concrete modulus, Ec. The stress in each of the steel layers is equal to rs(i) = Es  es,i. Based on the cracked residual tensile stresses and the current strain and curvature profile, Ec,eff(i) is updated for each cracked concrete layer within the section (Eq. (9)). These yields revised cross-sectional rigidities which in turn updates the strain (Eq. (12)) and stress

374

M. Watts et al.

profile of the section. This iterative procedure is repeated until axial and rotational equilibrium is achieved on the cross-section: Nint;0 ¼

n X

Ac;i rc;i þ

i¼1

Mint;0 ¼

n X

n X

    As;i rs;i ¼ RA;c þ RA;s er;0 þ RI;c þ RI;s j0 ¼ Next

ð14aÞ

i¼1

Ac;i rc;i yi þ

i¼1

n X

    As;i rs;i yi ¼  RB;c þ RB;s er;0 þ RB;c þ RB;s j0 ¼ Mext

i¼1

ð14bÞ

3.2

Time Analysis During Each Time Interval (Tk-1 to Tk)

During each time interval, a relaxation procedure is adopted to determine the change of strains and stresses in the section induced by creep and shrinkage. The strain state is initially frozen; that is, the strain distribution at the start of a time interval is assumed equal to that determined in the previous time interval. If the total strain in a concrete layer is held constant, the creep and shrinkage components are fully restrained and stresses develop due to the restrained strain. The restrained strain is equal in magnitude to the creep and shrinkage strains but of opposite sign. The restrained creep strain in a particular layer during the time interval tk-1 to tk is the creep strain that would develop due to the initial stress in the layer at time t0 and the sum of creep strains caused by the stress increments that have developed in each of the previous time intervals. When transitioning from time interval tk-1 to tk, the layered approach presented above involves the determination of a fictitious transitional force (Ntrans;tk ) and a fictitious transitional moment (Mtrans;tk ). These transitional actions may be interpreted as the equivalent force and moment produced by the increments of creep and shrinkage in a particular time interval. The creep and shrinkage components are calculated independently and then added together to determine the total transitional strain in each individual layer through the section in that time interval: Detrans;i;tk ¼ Decreep;i;tk þ Deshrink;i;tk

ð15Þ

The creep strain that develops in the interval tk-1 to tk in a particular layer is equal to the summation of the increments of creep caused by the stress increments in each previous time interval Each creep increment is the product of the stress increment and the change in the associated creep coefficient versus time curve between tk-1 and tk divided by Ec: Decreep;i;t1 ¼

Decreep;i;tk [ 1 ¼

ri;t0 ut1 ;t0 Ec   i 1 h   kP ri;t0 utk ;t0  utk1 ;t0 þ ri;tj  ri;tj1
utk ;tj  utk1 ;tj j¼1

Ec

ð16Þ

Time Dependent Deflection of FRC Members

375

The shrinkage component of the transitional strain is determined by applying the increment of shrinkage strain to each of the individual layers. The increment of shrinkage strain is the additional shrinkage that occurs in each individual layer during the time interval tk-1 to tk only. Deshrink;i;tk ¼ esh;i;tk  esh;i;tk1

ð17Þ

Once Decreep;i;tk and Deshrink;i;tk are defined, Ntrans;tk can be evaluated as the summation of the restrained change in strain due to creep and shrinkage multiplied by the effective modulus and the area of each layer through the depth of the section: Ntrans;tk ¼

n X

Ntrans;tk ;i ¼

i¼1

n X

Detrans;i;tk AcðiÞ Ec;effðiÞ

ð18Þ

i¼1

Similarly, the transitional moment is taken as the summation over all the layers of the product of the axial force determined in Eq. (18) and the distance between the centroid of each individual layer and the reference axis location: Mtrans;tk ¼

n X

ð19Þ

Ntrans;tk ;i yi

i¼1

The cumulative effects of creep and shrinkage Incurred from t1 onwards are determined through the summation of all transitional forces and moments from all previous time step analyses undertaken combined with the externally applied loads. This represents the cumulative (total) impacts of creep, shrinkage and externally applied loads on a section: Ncumulative;tk ¼ Next þ

k X

Ntrans;tj

ð20aÞ

Mtrans;tj

ð20bÞ

j¼1

Mcumulative;tk ¼ Mext þ

k X j¼1

The corresponding strain profile at time tk is calculated by replacing Next and Mext with Ncumulative;tk and Mcumulative;tk , respectively in Eq. (12). From the resultant strain profile a revised stress profile is determined by taking the resolved strain in each layer at time step tk, ec;i;tk and subtracting the summed transitional strains for that layer obtained from the previous time interval analyses and then multiplying by Ec,eff(i): rc;i;tk ¼ Ec;effðiÞ ec;i;tk 

k X j¼1

! Detrans;i;tj

ð21Þ

376

M. Watts et al.

The iterative solution presented above is necessary because the FRC exhibits strain softening characteristics after cracking and the concrete rigidities (RA,c, RB,c, RI,c) are a function of the effective modulus of each layer (Ec,eff,(i)), which in turn are dependent on the strain (er,tk & jtk) and hence stress resultants of the FRC layers. A simplification is made after the completion of the instantaneous analysis that assumes the concrete section rigidities remain constant in time during each subsequent time interval.

4 Model Validation Results obtained using the proposed model are plotted against deformation vs time and crack width vs time data from, Vasanelli et al. [15], Gilbert and Nejadi [16] and Aslani et al. [17] for beams ST1E, ST2E, ST3E, ST4E, B1a, B1b, SSCC-a. The output curvatures at each time step were translated into member deflections through the application of Mohr’s Analogy (the Conjugated Beam Method). The correlations to existing experimental data sets are presented in Fig. 4. 15

0.20

25

Experiment

Experiment

Model

Model B1a

20

0.15

0.10

δ [mm]

B1b

δ [mm]

w [mm]

10

15 Experiment Model (kfcc = 1) Model (kfcc = 5) Model (kfcc = 10)

10 5

0.05

0.00 0

b = 250 mm; D = 350 mm; d = 300 mm; l = 3500 mm; fc = 25 MPa; fct = 2.8 MPa; Ec = 24.9 GPa; Ast = 400 mm 2; ρf = 0%; lf = N/A; df = N/A; f0.5 = 0 MPa; f1.5 = 0 MPa

b = 250 mm; D = 250 mm; d = 205 mm; l = 2800 mm; fc = 21.4 MPa; fct = 2.3 MPa; Ec = 27.7 GPa; Ast = 462 mm 2; ρf = 0.60%; lf = 30 mm; df = 0.60 mm; f0.5 = 0.70 MPa; f1.5 = 0.61 MPa

100

200

300

400

t [Days]

500

600

700

(a)

0 0

100

200

t [Days]

(b)

300

400

5

0 0

b = 400 mm; D = 161 mm; d = 130 mm; l = 3500 mm; fc = 38 MPa; fct = 3.1 MPa; Ec = 35.8 GPa; Ast = 452 mm 2; ρf = 0.56%; lf = 65 mm; df = 0.85 mm; f0.5 = 0.49 MPa; f1.5 = 0.47 MPa

50

100

150

t [Days]

200

250

(c)

Fig. 4. Model comparison: (a) Vasanelli et al. (2014): Series ST; (b) Gilbert and Nejadi (2004): B1a/b; (c) Aslani et al. (2014): SSCC-a.

5 Sample Analysis In this section, the results of a sample cross-sectional analysis are presented. The geometry of the cross-section is D = 250 mm, b = 1000 mm, ds1 = 30 mm, ds2 = 220 mm and As1 = As2 = 452 mm2. The section is discretised into 50 layers of equal thickness (i.e. D/n = 5 mm). The cross-sections are subjected to a sustained bending moment Mext = 45kNm (applying tension to the soffit of the section) and axial compression of Next = 90kN that are first applied at t = 28 days. This level of loading induces cracking of the section at first loading, and the steel and concrete stresses are typical of in-service conditions. The creep coefficient and shrinkage strains are modelled using the fib MC2010 [9], with ucc(t1000, t28) = 1.39 and esh(t1000) = −432  10−6. A variable tensile creep amplification coefficient, kfcc is also considered in the sample analysis. kfcc is varied from 1 to 10 with the results plotted in Fig. 5.

Time Dependent Deflection of FRC Members

377

Two different cases are analysed in Fig. 5, the first one being an increase in the residual tensile stress provided by the fibres. In this case the f0.5 stress is modelled as 0 MPa (plain concrete) (see Fig. 5a), 0.5 MPa (Fig. 5b) and 1 MPa (Fig. 5c) with f1.5/f0.5 equal to 0.8 in the latter two cases. In Fig. 6, we have taken f0.5 = 1 MPa but have varied the shrinkage profile through the depth of the section. A constant, increasing and decreasing shrinkage profile is applied to the section through the layered model analysis–these distributions are shown in Figs. 6a, 6b and 6c, respectively. This analysis is not specific to a particular FRC however it does illustrate the flexibility of the model and the importance of understanding the influence of shrinkage profiles in the analysis of structures.

Fig. 5. Effect of varying residual stress on curvature in time (a) f0.5 = 0 (Plain Concrete); (b) f0.5 = 0.5 MPa; (c) f0.5 = 1.0 MPa.

Fig. 6. Effect of varying shrinkage profile on curvature in time & corresponding strain profiles (a) Constant shrinkage profile. (b) Increasing shrinkage. (c) Decreasing shrinkage.

378

M. Watts et al.

6 Conclusions A method of analysis to predict the time dependent behaviour of FRC has been presented. The model can be used to predict member deflection and crack widths of members subjected to sustained axial and flexural loads. The model can predict the behaviour of either macro-synthetic or steel fibre reinforced concrete and has the versatility to implement variable tensile creep properties in order to simulate the behaviour of synthetic fibres. The model was validated against various results available in the literature and shown to correlate well. The results reveal that inclusion of fibres in the concrete reduces time-dependent deformations and can significantly reduce maximum crack widths when used in combination with conventional reinforcing bars. A sample analysis was undertaken to provide an insight to the sensitivity of the outputs as a function of the residual stress provided by the fibres along with any tensile creep of the fibres. Additionally, the impacts of a variation in the shrinkage profile on an FRC flexural section was investigated. The sample analyses demonstrate the flexibility of the model but also the importance of understanding the impacts of residual stress, tensile creep and shrinkage on the long-term serviceability performance of FRC structures. Acknowledgements. This work was supported by an Australian Research Council Discovery Grant (DP 200102114) awarded to the second and third Authors.

References 1. Chiaia, B., Fantilli, A.P., Vallini, P.: Evaluation of crack width in FRC structures and application to tunnel linings. Mater. Struct. 42, 339–351 (2009) 2. Amin, A., Foster, S., Watts, M.: Modelling the tension stiffening effect in steel fiber reinforced-reinforced concrete. Mag. Concr. Res. 68(7), 339–352 (2016) 3. Vrijdaghs, R., di Prisco, M., Vandewalle, L.: Uniaxial tensile creep of a cracked polypropylene fiber reinforced concrete. Mater. Struct. 51(1), 1–12 (2018). https://doi.org/ 10.1617/s11527-017-1132-5 4. AS5100.5.: Bridge Design Part 5: Concrete. Australian Standard, Standards Association of Australia (2017) 5. AS3600.: Concrete Structures. Australian Standard, Standards Association of Australia (2018) 6. Gilbert, R.I., Bernard, E.S.: Creep analysis of macro-synthetic fibre cross-sections in combined bending and axial force. Concr. Aust. 41(1), 33–40 (2015) 7. Gilbert, R.I., Amin, A.A.: Time-dependent behaviour of fibre reinforced concrete. Concr. Aust. 45(1), 31–38 (2019) 8. Watts, M.J., Amin, A., Gilbert, R.I., Kaufmann, W.: Behavior of fiber reinforced concrete members under sustained axial/flexural load. Struct. Concr, pp. 1–17 (2019) 9. fib Model Code 2010. Fédération Internationale du Béton, p. 402 (2013) 10. di Prisco, M., Plizzari, G., Vandewalle, L.: Fibre reinforced concrete: new design perspectives. Mater. Struct. 42, 1261–1281 (2009) 11. Pfyl, T.: Tragverhalten von Stahlfaserbeton. PhD Dissertation, IBK-Report No. 279. Swiss Federal Institute of Technology, Switzerland. (in German) (2003)

Time Dependent Deflection of FRC Members

379

12. Amin, A., Foster, S., Muttoni, A.: Derivation of the r-w relationship for SFRC from prism bending tests. Struct. Concr. 16(1), 93–105 (2015) 13. Marti, P., Alvarez, M., Kaufmann, W., Sigrist, V.: Tension chord model for structural concrete. Struct. Eng. Int. 8(4), 287–298 (1998) 14. Amin, A., Foster, S., Watts, M.: Modelling the tension stiffening effect in steel fiber reinforced-reinforced concrete. Mag. Concr. Res. 68(7), 339–352 (2016) 15. Vasanelli, E., Micelli, F., Aiello, M.A., Plizzari, G.: Crack width prediction of FRC beams in short- and long-term bending condition. Mater. Struct. 47, 39–54 (2014) 16. Gilbert, R.I., Nejadi, S.: An experimental study of flexural cracking in reinforced concrete members under sustained loads. UNICIV Report R-435. School of Civil and Environmental Engineering, The University of New South Wales, Australia (2004) 17. Aslani, F., Nejadi, S., Samali, B.: Long-term flexural cracking control of reinforced selfcompacting concrete one way slabs with and without fibres. Comput. Concr. 14(4), 419–444 (2014)

Influence of the Residual Tensile Strength on the Factor for Quasi-permanent Value of a Variable Action w2 Darko Nakov1(&), Goran Markovski1, Toni Arangjelovski1, and Peter Mark2 Faculty of Civil Engineering, University “Ss. Cyril and Methodius”, Skopje, North Macedonia [email protected] Faculty of Civil and Environmental Engineering, Ruhr-University Bochum, Bochum, Germany 1

2

Abstract. Steel fibres are known to aid in deflection control, control the process of cracking and mainly improve the toughness of structural elements and the whole structure. Large experimental program was performed at the Faculty of Civil Engineering-Skopje to find out how steel fibres and the residual tensile strength affect the time-dependent deformation properties and deflections of concrete. Specific realistic loading with permanent and repeated variable loads in loading interval of 8 h per day was applied on full scale beams that were monitored up to an age of concrete of 400 days. The beams were with cross section dimensions 15/28 cm and total length of 300 cm, manufactured from concrete class C30/37. They were reinforced with same percentage of longitudinal and shear reinforcement, but with different amount of steel fibres (0, 30 kg/m3 and 60 kg/m3). Using the experimental results, detailed analysis of the time-dependent deformation properties of concrete and their effect on the timedependent behaviour was done. A value for the factor for quasi-permanent value of variable action w2 is proposed for each type of concrete. It was concluded that the factor for quasi-permanent value of a variable action w2 depends linearly on the residual tensile strength. Keywords: Steel fibre reinforced concrete Repeated variable loads
Factor w2


Residual tensile strength


1 Introduction The effect of long-term actions is usually connected with the permanent load. However, there are certain concrete structures such as: storage areas at warehouses, traffic areas at parking garages and city bridges under severe traffic conditions, where the variable loads are acting longer and are with significant magnitude. In these structures, the variable actions could overcome serviceability limit states criteria of concrete structures. In the serviceability limit states design, effects due to creep and shrinkage of concrete caused by variable load, are taken into consideration using quasi-permanent © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 380–391, 2021. https://doi.org/10.1007/978-3-030-58482-5_35

Influence of the Residual Tensile Strength

381

combination of actions [1]. The level of quasi-permanent load is defined by the factor for quasi-permanent value of variable action w2, Eq. (1): X X G þPþ w
Qk;i ð1Þ j  1 k;j i  1 2;i Where: Gk;j P– Qk;i w2;i

– permanent actions, prestressing force, – variable actions, – factors for quasi-permanent values of variable actions

The National Annex of each country could also set the values for the factor w2. This gives opportunity for research of effects of variable load and its replacement by a quasipermanent load. Proposed values for the factor w2 for buildings in Eurocodes [1], are presented in Table 1. Table 1. Recommended values of factor w2 for buildings [5] Action w2 Imposed loads in buildings, category (see EN1991-1-1) 0.3 Category A: domestic, residential areas 0.3 Category B: office areas 0.6 Category C: congregation areas 0.6 Category D: shopping areas 0.8 Category E: storage areas 0.6 Category F: traffic area, vehicle weight  30 kN Category G: traffic area, 30 kN < vehicle weight  160 kN 0.3 0.0 Category H: roofs

The recommended value for the factor w2 for road bridges is 0 [1]. In DIN report 102 “Concrete bridges” based on Eurocodes, for city bridges, w2 = 0.2, while in the National application document of Finland for EN 1992-2, w2 = 0.3. The recommended value of the factor w2 for railway bridges is 0. But if deformation is considered for persistent and transient design situations, the factor w2 should be taken equal to 1.00 for rail traffic actions [1]. Several studies have been conducted on long-term behavior of SFRC beams under sustained loads [2–6], while studies which include the effect of variable repeated load are uncommon [7]. Up to now, there is no research dealing with the factor w2 for steel fibre reinforced concrete. Having in mind that the residual tensile strength is one of the main characteristics which differ ordinary from fibre concretes, with this research an attempt was made to find out the influence of the residual tensile strength to the factor w2.

382

D. Nakov et al.

2 Experimental Program The experiment was carried out at the “Ss. Cyril and Methodius” University, Faculty of Civil Engineering in Skopje, Republic of North Macedonia. It involved testing of 24 full scale beams constructed from reinforced concrete and steel fibre reinforced concrete with additional reinforcement. The beams were with cross section dimensions 15/28 cm and total length of l = 300 cm, Fig. 1. Together with each series of beams, control specimens were cast in order to test the compressive strength, flexural (and residual) tensile strength [8], splitting tensile strength, elastic modulus and deformations due to creep and shrinkage. In addition to the tests on mechanical [9] and timedependent properties of concrete [10, 11], the used reinforcement was also tested.

Fig. 1. Geometry, reinforcement and loading scheme of full scale beams.

All 24 beams were manufactured with concrete class C30/37. According to the used type of material, they were divided into three series: – Series A, reinforced concrete (C30/37); – Series B, SFRC with 30 kg/m3 steel fibres and additional reinforcement (C30/37 FL1.5/1.5); – Series C, SFRC with 60 kg/m3 steel fibres and additional reinforcement (C30/37 FL2.5/2.0). In each series, the plain reinforcement was kept the same. The longitudinal reinforcement was ribbed and of RA 400/500-2 quality, while the shear reinforcement was smooth, with GA 240/360 quality. Reinforcement 2Ø10, 2Ø8 and Ø6/10/20 cm was used as tension, compression and shear reinforcement, respectively. The used steel fibres were hooked-end HE1/50, produced of cold-drawn wire, manufactured by Arcelor Mittal, with a diameter of 1 mm, length of 50 mm and tensile strength of 1100 N/mm2. The mixture proportions are presented in Table 2. Regarding the loading history, the beams were divided into four groups, each with 6 beams. In this paper, only the results for group “4” are presented. On the beams from group “4”, a long term permanent load with intensity “g” was applied at the age of concrete of 40 days and was held for a year as a long term load. On the fortieth day, variable repeated load “±q” was also applied in an interval of 8 h +q and 16 h −q, for a year. This means that the beams were loaded additionally with

Influence of the Residual Tensile Strength

383

Table 2. Mixture proportions Mixture proportions Cement CEM II/A-M 42.5 N Water Water/Cement ratio, w/c Aggregate: 0–4 mm (river sand), 50% 4–8 mm (limestone), 20% 8–16 mm (limestone), 30% Fibres: C30/37 C30/37 FL 1.5/1.5 C30/37 FL 2.5/2.0

(kg/m3) 410 215 0.524 875 350 525 0 30 60

load “q” for 8 h every day, whereat the strains, deformations and crack widths were measured. After 8 h, the beams were unloaded from load “q” and all measurements were performed again. The long term loading scheme is presented in Fig. 2.

Fig. 2. Long term loading scheme.

The beams and control specimens were cured for 8 days and then they were transported to the Laboratory at the Faculty of Civil Engineering – Skopje, where they were kept under almost constant temperature with an average of 19.5 °C and constant relative ambient humidity with an average of 60.2%, which was regulated with special humidifiers and dehumidifiers. In each step, the concrete strains in middle section of the beam through the thickness as well as on the top of the beam, were measured by a mechanical deflection meter, type Hugenberger, Switzerland, with a base of 250 mm. The mechanical measurement of the deflections was done at 5 points through the length of the beam and 2 points over the supports by using deflection meters produced by Stopani, Italy. The crack widths were also measured in each load step, in the region with constant moment, by use of a crack microscope - product of Controls, Italy. The positions of the measurement points are presented in Fig. 3. The long term load, which consists of permanent sustained load “g” and repeated variable load “q”, was applied by gravitation levers, which enabled an increase of the

384

D. Nakov et al.

Fig. 3. Positions of measurement points of full scale beams.

load for 13 times. The permanent load acts all the time, while the variable load was applied and removed each day by secondary hand gravitation levers. The bending moments are as follows: from self-weight of the beam, Msw = 1 kNm, from permanent load “g”, Mg = 5.0 kNm, from variable load “q”, Mq = 3.1 kNm, from self-weight, permanent and variable load (service) Msw+g+q = 9.1 kNm. The bending crack moment was Mcr = 6.1 kNm, while the ultimate bending moment Md = 15.6 kNm. The intensity of the load was chosen so that the Mcr is bigger than Msw+g and smaller than Msw+g+q. The permanent load is 0.39 times the flexural strength, while the service load is 0.58 times the flexural strength of the beam without fibres.

3 Analytical Analyses The analytical analyses of the results from the experimental research were performed in two parts: • Analytical analysis of time – dependent deformation properties, • Analytical analysis of time – dependent deflections. Data on the time – dependent deformation properties were later used to calculate the time-dependent deflections using the Age-Adjusted Effective Modulus Method (AAEMM). The quasi-permanent load procedure and the principle of superposition were used to obtain the factor for the quasi-permanent value of the variable action w2 . 3.1

Analytical Analysis of Time – Dependent Deformation Properties

The analytical analysis of time – dependent deformation properties, drying shrinkage and creep, were performed by the B3 model [12] and Fib Model Code 2010 [13]. In the beginning, the analyses were done only for the time period considered in this research, which was 400 days.

Influence of the Residual Tensile Strength

385

The B3 model offers the possibility of improvement of the model by its users and updating of its predictions based on short-time measurements. The updating of the drying shrinkage strain was done very efficiently by using the scaling parameter p6. The experimental and analytical results for the drying shrinkage up to the age of 400 days are presented in Fig. 4. It can be noticed that the Fib Model Code 2010 underestimates the drying shrinkage strain for 29%, while the original B3 model underestimation is 11.5%. The obtained scaling parameter in the improved B3 model is p6 = 1.123. It can be noticed that there is a very good agreement between the experimental results and the improved B3 model. Having in mind the service life of designed structures, it is very important to be able to predict the time – dependent deformation properties for their serviceability period. Therefore, based on the results obtained for the age of up to 400 days, the analyses according to the previously mentioned models were extended to the serviceability period of the structures of 100 years. The results are presented in logarithmic scale in Fig. 5.

Drying shrinkage εds [10-6]μs

900 800

700 600 500

C30/37

400

C30/37 FL 1.5/1.5

300

C30/37 FL 2.5/2.0 FIB MC 2010

200

B3 model 100 B3 model IMP. 0 0

50

100

150

200

250

300

350

400

t [days]

Fig. 4. Experimental and analytical results for drying shrinkage up to 400 days.

Due to the differences in the creep strain between different types of concrete, an analytical analysis of the creep strain was performed for each concrete type taken separately, by the B3 model and Fib Model Code 2010. For the concrete type C30/37, the experimental and analytical results are presented in Fig. 6. In addition to the previously mentioned models, the improvement of the B3 model is also presented. On the basis of linear regression analysis, the following values for adjustment of the creep compliance in the B3 model were obtained: p1 = 1.143 and p2 = 1.122. These values are valid only for the concrete type C30/37. Taking into consideration the coefficients of variation of each model code, a good agreement was found in all cases. The Fib Model Code 2010 overestimates the experimentally obtained creep coefficient at 400 days for 13%, while the B3 model underestimates it for 9.5%. The experimental results for both steel fibre reinforced concretes, C30/37 FL 1.5/1.5 and C30/37 FL 2.5/2.0, show a very small difference in the final strain after 400

386

D. Nakov et al.

Drying shrinkage εds [10-6]μs

900 800 700 600 500

C30/37

400

C30/37 FL 1.5/1.5 C30/37 FL 2.5/2.0

300

FIB MC 2010 200

B3 model 100

B3 model IMP.

0 1

10

100

1000

10000

36500

100000

t [days]

Fig. 5. Experimental and analytical results for drying shrinkage up to 100 years.

800

Creep εcc [10-6]μs

700 600 500 400

300

C30/37

200

FIB MC2010 B3 model

100

B3 model IMP.

0 0

50

100

150

200

250

300

350

400

t [days]

Fig. 6. Experimental and analytical results for creep of C30/37 up to 400 days.

days when the original B3 model is used. Therefore, for these types of concrete, modification of the flow compliance q4 in the B3 model is proposed. The modification includes addition of an amount of fibers Gf to the amount of aggregate a, multiplied by the ratio between the moduli of elasticity of the fibers and the aggregate Ef/Ea, Eq. (2). 0 q4 ¼ 20:3@

a þ Gf c

 10:7 E f

Ea

A

ð2Þ

The results for the concrete C30/37 FL 2.5/2.0 with 60 kg/m3 steel fibers are presented in Fig. 7. Based on the results obtained for an age of up to 400 days, the analyses of the creep strain according to the previously mentioned models were

Influence of the Residual Tensile Strength

387

extended to the serviceability life of the structures of 100 years. The results for the concrete type C30/37 FL 2.5/2.0 are presented in logarithmic scale in Fig. 8.

800

Creep εcc [10-6]μs

700 600

500 400 300

C30/37 FL 2.5/2.0

200

B3 model

100

B3 model MOD.

0 0

50

100

150

200

250

300

350

400

t [days]

Fig. 7. Experimental and analytical results for creep of C30/37 FL 2.5/2.0 up to 400 days.

3.2

Analytical Analysis of Time – Dependent Deflections

There are many available analytical and numerical methods for obtaining the time – dependent response of concrete structures. However, the big number of necessary input data, complicated analysis and the big number of unknown parameters in the design phase make these methods impractical for most of the engineers. The influence that variable load has on the time – dependent response of concrete structures is another aspect that, in certain structures, has a big effect on the total behaviour. With the experimental program of this research, it was planned to use a simple quasi – permanent load procedure to obtain a factor by which only one part of the variable load will be included in the time – dependent analysis. This procedure is intended for calculation of the creep effects when serviceability limit state design should be performed using the quasi – permanent combination of actions. This combination of actions enables inclusion of the variable loads in the calculation of creep effects. One part of the variable load is added to the permanent load and is named as quasi – permanent load. The factor that defines the part of the variable load is called factor w2 [1]. This factor depends on the category of the building and the loading history. The loading history on different types of concrete has been the subject of research at the Faculty of Civil Engineering – Skopje for almost 12 years. In this research, the loading history has been chosen such that the variable load acts for 8 h each day in the period of one year. Four approaches to determination of the effects of variable load were considered (Fig. 9): The simplest solution of the problem is taking into consideration approach 1, because the initial and time-dependent deflection can be obtained with intensity of the load as a sum of permanent and quasi-permanent load. On the basis of the results obtained by the experimental research, an analytical solution was proposed in which

388

D. Nakov et al.

Creep εcc [10-6]μs

1000

800

600

C30/37 FL 2.5/2.0

400

FIB MC2010

200

B3 model B3 model MOD.

0

10

100

1000

10000

36500

100000

t [days]

Fig. 8. Experimental and analytical results for creep of C30/37 FL 2.5/2.0 up to 100 years.

Fig. 9. Four considered approaches to determination of w2 .

the total deflection due to the permanent load “g” and variable load “q” obtained from the experiments at,exp(g + q) was determined as a sum of the initial deflection ao(g + w2 q) and long-term deflection at(g + w2 q) due to the permanent load “g” and variable load represented as quasi permanent load “w2 q”, Eq. (3): at;exp ðg þ qÞ ¼ ao ðg þ w2 qÞ þ at ðg þ w2 qÞ

ð3Þ

Factor w2 was determined by the Age – Adjusted Effective Modulus Method (AAEMM) using the quasi – permanent load procedure and the principle of superposition. The CRACK computer program developed by Ghali A. and Elbadry M. from the University of Calgary, Canada, was used for calculation of the time – dependent deflections. Except the geometry, materials and bending moments, all other necessary input data needed for the AAEMM were obtained by the previously presented analysis of the time – dependent deformation properties, based on the experimentally obtained results. Those data include values which vary each day, like the values for shrinkage,

Influence of the Residual Tensile Strength

389

creep coefficient, relaxation function and aging coefficient. For the concrete type C30/37, the total experimentally observed deflection was obtained with intensity of load g + 0.37q, which meant that factor w2 had a value of 0.37. In the case of the steel fibre reinforced concrete types C30/37 FL 1.5/1.5 and C30/37 FL 2.5/2.0, the total experimentally observed deflection was obtained with intensity of load g + 0.22q and g + 0.18q, which meant that the factor w2 had a value of 0.22 and 0.18, respectively. The results for C30/37 FL 1.5/1.5 are presented in Fig. 10, while all results are presented in Table 3.

Time-dependent deflection Concrete type: C30/37 FL 1.5/1.5

Total deflection a(mm)

5 C30/37 FL 1.5/1.5 (EXP.)

4

3

C30/37 FL 1.5/1.5 (AAEMM) g+0.22q

ψ2 =0.22 2

1

0 0

50

100

150

200

250

300

350

400

t (days)

Fig. 10. Experimental and analytical time-dependent deflection for concrete type C30/37 FL 1.5/1.5. Table 3. Factor w2 as a function of the ratio of residual tensile strength to the concrete compressive strength, af Concrete type C30/37 C30/37 FL 1.5/1.5 C30/37 FL 2.5/2.0 0 0.043 0.063 af w2 0.37 0.22 0.18

ψ2 (αf) 1

0.8 C30/37

ψ2

0.6

C30/37

Linear (C30/37)

C30/37 FL 1.5/1.5

0.4

C30/37 FL 2.5/2.0 0.2

0 0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

0.1

αf

Fig. 11. Factor w2 as a function of the ratio between the residual tensile and compressive strength, af.

390

D. Nakov et al.

In the following figure (Fig. 11), the results are presented as a function of the ratio between the residual tensile strength and the concrete compressive strength, af, which is zero for ordinary concrete. A simple linear dependence between factor w2 and the ratio af can be noticed.

4 Conclusions • Analytical analysis of the drying shrinkage and creep, performed by the B3 model and fib Model Code 2010, demonstrated agreement with the experimental results. For concrete C30/37, an improvement of the B3 model was done on the basis on linear regression analysis. New values of the coefficients p1 and p2 for adjustment of the creep compliance were obtained, as well as scaling parameter for the drying shrinkage p6. For steel fibre reinforced concrete types, a modification of the flow compliance q4, which includes the amount of fibres is proposed. • The repeated variable load has significant influence on the time-dependent behaviour of reinforced and steel fibre reinforced concrete beams. • Using the experimental results and analytical analysis, the following factors for quasi-permanent value of a variable action were obtained: for concrete type C30/37, w2 = 0.37, while for concrete type C30/37 FL 1.5/1.5, w2 = 0.22 and for C30/37 FL 2.5/2.0, w2 = 0.18. • The factor w2 depends linearly on the residual tensile strength.

References 1. European Standard EN1990+A1. Basis of structural design, Standardization Institute of R. Macedonia, Skopje, R.Macedonia (2002) 2. Tan, K.H., Saha, M.K.: Ten-year study on Steel fiber-reinforced concrete beams under sustained loads. ACI Struct. J. (2005) 3. Vasanelli, E., Micelli, F., Aiello, M.A., Plizzari, G.: Long term behaviour of fiber reinforced concrete beams in bending. In: BEFIB2012 – Fibre Reinforced Concrete, Guimaraes (2012) 4. Casucci, D., Thiele, C., Schnell, J.: Behavior of cracked cross-section of fibre reinforced UHPFRC under sustained load. In: Serna, P., et al. (ed.) Creep Behaviour in Cracked Sections of Fibre Reinforced Concrete. RILEM Bookseries, vol. 14. Springer (2017) 5. Nishiwaki, T., Kwon, S., Otaki, H., Igarashi, G., Shaikh, F.U.A., Fantilli, P.: Experimental study on time-dependent behaviour of cracked UHP-FRCC under sustained loads. In: Serna, P., et al. (ed.) Creep Behaviour in Cracked Sections of Fibre Reinforced Concrete. RILEM Bookseries, vol. 14. Springer (2017) 6. Candido, L., Micelli, F., Vasanelli, E., Aiello, M.A., Plizzari, G.: Durability of FRC beams exposed for long-term under sustained service loading. In: Serna, P., et al. (ed.) Creep Behaviour in Cracked Sections of Fibre Reinforced Concrete. RILEM Bookseries, vol. 14. Springer (2017) 7. Nakov, D., Markovski, G., Arangjelovski, T., Mark, P.: Creeping effect of SFRC elements under specific type of long term loading. In: Serna, P., et al. (ed.) Creep Behaviour in Cracked Sections of Fibre Reinforced Concrete. RILEM Bookseries, vol. 14. Springer (2017)

Influence of the Residual Tensile Strength

391

8. RILEM TC 162-TDF: Test and design methods for steel fiber reinforced concrete, bending test, final recommendation. Mater. Struct. 35 (2002) 9. Nakov, D., Markovski, G., Arangjelovski, T.: Influence of steel fibre reinforcement on the properties of concrete. In: 1st International Conference COMS, Zadar (2017) 10. Nakov, D., Markovski, G., Arangjelovski, T., Mark, P.: Experimental and analytical analysis of creep of steel fibre reinforced concrete. Periodica Polytechnica Civil Engineering 62(1) (2017) 11. Nakov, D., Markovski, G., Mark, P., Arangjelovski, T.: Analytical analysis of drying shrinkage of SFRC based on experimental results. In: Conference Fibre Concrete 2015, Prague (2015) 12. Bazant, Z.P., Baweja, S., Creep and shrinkage prediction model for analysis and design of concrete structures: model B3. In: Al-Manaseer, A. (ed.) The Adam Neville Symposium: Creep and Shrinkage-Structural Design Effects, SP-194, American Concrete Institute, Farmington Hills, Michigan (2000) 13. Fib Model Code 2010, fib Bulletin 65, Final draft (2012)

Compressive and Tensile Creep and Shrinkage of Synthetic FRC: Experimental Results and Comparison to Codes Razan H. Al Marahla and Emilio Garcia-Taengua(&) School of Civil Engineering, University of Leeds, Leeds, UK [email protected]

Abstract. The mechanical properties of FRC mixes, namely compressive and tensile strength, are generally improved with respect to their unreinforced counterparts due to the contribution of fibres. This has implications in aspects like crack propagation or the development of time-deferred strains such as those resulting from creep and shrinkage, which in turn influence the long-term deformation of structural elements. The effect of fibres on compressive creep has been studied by various researchers, and codes and guidelines for the design of concrete structures include provisions to predict creep under compression. However, tensile creep of FRC has not attracted the same level of coverage despite its relevance to the loss of tension stiffening. This paper presents the results of an experimental study in concrete specimens reinforced with synthetic fibres were subjected to constant loading in uniaxial tension and in compression in order to evaluate the effect that increasing dosages of synthetic fibres have on the resulting time-dependent strains. The evolution of shrinkage, creep strains and the creep coefficient were analyzed in relation to the fibres dosage. The experimental strain-time curves were compared to the theoretical curves from the models adopted in the Eurocode 2, the Model Code, or by the ACI Committee 209. Keywords: Synthetic fibres


Strains
Time-dependant

1 Introduction Concrete exhibits volume changes with time. Due to the partial or total restraint of concrete deformation, shrinkage induce stresses even in the absence of externally applied loads, and the resulting tensile stresses can result in cracking [1, 2]. The longterm behaviour of concrete structures is also influenced by creep deformations under sustained stresses. Time-dependent strains due to shrinkage and creep affect the mechanical capacity of reinforced concrete sections [3]. The creep and shrinkage equations included in the Eurocode 2 [4], Model Code 2010 [5], or in ACI 209.2R [6] are tools for estimating shrinkage and creep of concrete as a function of time. The presence of fibres in concrete plays a significant role in controlling the propagation and growth of microcracks, and increase the tensile strength of concrete [7] as well as its toughness in the cracked state [8]. The behaviour of FRC under sustained © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 392–401, 2021. https://doi.org/10.1007/978-3-030-58482-5_36

Compressive and Tensile Creep and Shrinkage

393

tensile and flexural loads, and how this behaviour is influenced by the combined effect of creep and shrinkage, remain issues that attract considerable research interest. The behaviour of concrete under sustained compressive loads has been studied by different authors, who have investigated the effect of different variables such as the compressive strength of concrete, the specimen shape, the stress level applied, the duration of the sustained loading, or the temperature and relative humidity conditions [9, 10]. An empirical expression to estimate concrete creep under compression was developed by [11], based on experimental data obtained for FRC mixes with different types and dosages of steel fibres, and the effect of fibres geometry and dosage was studied in terms of bond of fibres to the cementitious matrix. Different methodologies have been proposed for evaluating tensile creep. Domone [12] evaluated the tensile creep of both sealed and immersed concrete specimens, adopting the uniaxial tensile test introduced by Elvery and Haroun [13], in which cylindrical (bobbin) specimens are tapered at both ends and a concentric load is applied. Tensile creep reduces the stresses caused by restrained shrinkage to some extent [14–16]. Swamy and Stavrides investigated the influence of fibres on restrained shrinkage and subsequent cracking and concluded that the addition of fibres to normal and lightweight concrete mixes can reduce shrinkage deformations up to 20% [17]. Various experimental and analytical investigations considered different fibre types and contents and studied the effect that higher fibre dosages have on interfacial bond and to what extent fibres contribute to crack control [18–20]. The effect of steel fibres on concrete time-dependent deformations has been investigated widely and results show that steel fibres improve the tensile and flexural capacity of FRC under sustained loads [21, 22]. This paper is concerned with the effect that synthetic fibres have on creep and shrinkage under sustained tensile and compressive loads. Different contents of synthetic fibres were added to the same reference mix design and the sensitivity to this parameter was evaluated.

2 Experimental Programme 2.1

Materials and Reference Mix Design

The reference mix design considered in this study had a water-to-cement (w/c) ratio of 0.29 and an average compressive strength of 60 MPa at 28 days. It was adjusted for a slump value of 120–150 mm so it could incorporate synthetic fibres at different dosages without further adjustments. The mix proportions are given in Table 1. The synthetic fibres were high-modulus 54 mm long polymeric fibres, with a tensile strength of 600 MPa. A picture of these fibres is shown in Fig. 1. 2.2

Mixes Production and Testing Methodology

In total, three different mix designs were considered in this study, all based on the reference mix design summarised in Table 1 and differing in the synthetic fibre content only: 0, 5, and 10 kg/m3. Two identical batches of each of these mixes were produced,

394

R. H. Al Marahla and E. Garcia-Taengua Table 1. Mixture proportions kg/m3. Specimen Cement Water Fine aggregate Coarse aggregate (20 mm) Coarse aggregate (10 mm) Superplasticiser

w/c = 0.29 510 147 950 300 580 9

Fig. 1. Synthetic fibres used in this study.

and the same number of specimens were cast from each batch, with the objective of generating a sufficient number of replicates for better accuracy of the results. The same mixing sequence was followed in all cases. Prior to the mixing of every batch, the total amount of water to be added was separated in two buckets: one containing only 80% of the water, and the other containing the mixture of the required amount of superplasticiser and the remaining 20% of the water. First, cement and all aggregates were poured into the mixer and dry-mixed for 2 min. After that, 80% of the water was added and mixed with the cement and aggregates for 3 min. During this time, the fibres were poured gradually into the mixer. Finally, the remaining 20% of the water with the superplasticiser predispersed in it was added, and the mixing continued for 4 min. In all cases, a uniform distribution of the fibres in the mix was observed. From each batch, specimens were cast and tested as follows: • Characterisation: 3 cubic specimens (100 mm side) and 3 cylindrical specimens (150  300 mm), to determine the compressive strength and splitting tensile strength at 28 days. • Creep and shrinkage under compression: 8 prismatic specimens with a 75  75 mm cross-sectional area and a length of 200 mm. Of these specimens, 4 were tested under sustained compressive load, and 4 were not loaded but instrumented to measure free shrinkage strains. • Creep and shrinkage under tension: 3 cylindrical specimens with a diameter of 75 mm and a length of 365 mm. Throughout this paper, these are referred to as ‘bobbins’, to distinguish them from the 150  300 mm cylindrical specimens. Out of these 3 bobbins, 2 were tested under sustained tensile load, and 1 was not loaded but instrumented to measure free shrinkage.

Compressive and Tensile Creep and Shrinkage

395

Characterisation specimens were tested at the age of 28 days. Prior to this, they were kept in a fog room with a controlled temperature of 20 °C and a relative humidity of 90%. Creep and shrinkage specimens were instrumented with DEMEC points, which were glued and fixed to their sides and positioned 150 mm apart from each other, in order to measure the average surface strain. A picture of the specimens and test instrumentation is shown in Fig. 2. For the specimens tested under compression, a sustained load corresponding to 25% of the average compressive strength measured at 28 days was applied. For the specimens tested in tension, the sustained tensile stress applied was 1 MPa. Measurements were taken daily for 90 days. In both cases, the tests were set up in a controlled room where the temperature and relative humidity were kept at 21 ± 2 °C and a relative humidity of 50 ± 5% throughout the duration of the tests.

Fig. 2. Experimental setup for specimens in compression (left) and in tension (right).

3 Results and Discussion 3.1

Compressive Strength and Splitting Tensile Strength

For each of the batches, compressive strength and tensile splitting strength were measured at 28 days. The average results as well as the corresponding standard deviation values are shown in Table 2. Table 2. Compressive and splitting tensile strength results. Fibre content (kg/m3) Compressive strength (MPa) Average Std. deviation 0 60.9 0.8 0 60.4 0.6 5 61.1 0.7 5 61.6 0.7 10 60.4 0.6 10 59.8 1.4

Splitting tensile strength (MPa) Average Std. deviation 3.9 0.1 3.9 0.4 4.4 0.2 4.3 0.2 4.6 0.5 4.6 0.4

396

R. H. Al Marahla and E. Garcia-Taengua Table 3. Strains measured in prismatic specimens under compression (in microstrain).

Fibre content Elastic strains Strains at 14 days Loaded Unloaded 3 −395 −807 −156 0 kg/m −390 −811 −162 −412 −789 −151 −389 −819 −149 −145 −728 −365 5 kg/m3 −134 −730 −359 −128 −724 −373 −149 −731 −376 10 kg/m3 −100 −655 −356 −92 −661 −353 −110 −647 −347 −118 −639 −354

Strains at 30 days Loaded Unloaded −968 −231 −966 −237 −228 −962 −958 −240 −201 −858 −189 −864 −193 −851 −210 −863 −168 −786 −178 −797 −179 −814 −163 −789

Strains at 90 days Loaded Unloaded −1146 −325 −1140 −326 −1150 −319 −1136 −312 −1033 −288 −1042 −286 −1038 −280 −1022 −277 −267 −881 −256 −876 −274 −890 −265 −871

Regarding the average compressive strength, the addition of synthetic fibres at 5 or 10 kg/m3 did not introduce statistically significant differences with respect to the batches without fibres. The same can be said in relation to the standard deviation values. In terms of the splitting tensile strength, an improvement was observed due to the introduction of synthetic fibres. With respect to the batches without fibres, the average splitting tensile strength increased by 11.5% when synthetic fibres were dosed at 5 kg/m3, and by 18% when the fibres content was 10 kg/m3. 3.2

Creep Under Sustained Compression

For the analysis of creep and shrinkage under compression, strains were measured over a period of 90 days on unloaded as well as loaded specimens. Measurements were taken daily on a total of eight specimens per case. Table 3 shows the measurements corresponding to 14, 30, and 90 days, as well as the elastic strains, which were measured immediately after the compressive load was applied. Individual values for the compressive creep strain at any age were determined as the difference between the loaded and unloaded specimen strains, minus the corresponding elastic strain. The average compressive creep strains for the three synthetic fibre contents considered are shown in Fig. 3, together with the theoretical strains as predicted by the equations in Model Code and Eurocode 2. The compressive creep strains corresponding to the reference mix without fibres showed good agreement with the theoretical values. In particular, strain values after 5 days were observed to be very close to the predictions calculated according to the Model Code.

Compressive and Tensile Creep and Shrinkage

397

Fig. 3. Average compressive creep strains and comparison to codes.

The addition of fibres was correlated with a reduction in compressive creep strains, consistently observed at all ages. Strain values corresponding to FRC mixes with 5 kg/m3 of synthetic fibres were between 10% and 15% lower than those observed in mixes without fibres (strains were reduced in 15%, 11% and 10% at 14, 30, and 90 days, respectively). For the FRC mixes where the fibre content was 10 kg/m3, these reductions were between 15% and 22% (strains were reduced in 22%, 16% and 15% at 14, 30, and 90 days, respectively). The analysis of the results was also made in terms of the creep coefficient values, and these are shown in Fig. 4, together with the theoretical values as predicted by the equations in the Model Code, the Eurocode, and the ACI Committee 209 report. Again, it was observed that the values corresponding the case without fibres showed very good agreement with the theoretical values as per the Model Code. At all ages, the presence of synthetic fibres was correlated with lower creep coefficients than without fibres, and the reduction associated with a fibre content of 10 kg/m3 was higher than that observed for a fibre content of 5 kg/m3. If the compressive creep coefficient at 90 days is considered, reductions of 5% and 8% were observed for fibres contents of 5 kg/m3 and 10 kg/m3, respectively.

Fig. 4. Average compressive creep coefficient values and comparison to codes.

398

3.3

R. H. Al Marahla and E. Garcia-Taengua

Shrinkage

Shrinkage strains were determined from the strains measure in unloaded conditions (values in Table 3). The average shrinkage strains for the different synthetic fibre contents considered, together with the theoretical values according to the Model Code, Eurocode 2 and the ACI Committee 209 report, are shown in Fig. 5. Shrinkage strains observed in specimens without fibres were well in agreement with the ACI 209 predictions up to approximately 50 days. For later ages, the model proposed by the Eurocode 2 yielded better predictions. In any case, the experimental results seemed conclusive in showing that the theoretical models tend to overestimate shrinkage strains.

Fig. 5. Average shrinkage strains and comparison to codes.

Similarly to what was observed in relation to creep under compression, the addition of synthetic fibres led to generalised reductions in shrinkage strains, and these reductions were directly related to the fibre content. The addition of synthetic fibres at a dosage of 10 kg/m3 decreased shrinkage strains in between 20% and 26% (the reductions observed were 23%, 26%, and 20% at 14, 30, and 90 days, respectively). In those cases where the fibre dosage was 5 kg/m3, shrinkage strains were between 8% and 16% lower than the values corresponding to the specimens without fibres. 3.4

Creep Under Sustained Tension

To characterise the tensile creep, strains were measured for 90 days on unloaded as well as loaded specimens. Table 4 summarises the measurements corresponding to 14, 30, and 90 days, as well as the elastic strains, which were measured immediately after the tensile load was applied.

Compressive and Tensile Creep and Shrinkage

399

Table 4. Strains obtained from loaded and unloaded bobbins under tensile creep Fibre content Elastic strains Strains at 14 days Loaded Unloaded 3 31.6 101 −244 0 kg/m 27.9 99 26.2 78 −191 5 kg/m3 27.4 75 10 kg/m3 24.4 65 −172 23.8 66

Strains at 30 days Loaded Unloaded 145 −322 139 85 −241.5 88 85 −218 79

Strains at 90 days Loaded Unloaded 168 −381 155 126 −311 101 98 −282 94

Individual tensile creep values were determined as the difference between the loaded and unloaded specimen strains, minus the corresponding elastic strain. The average tensile creep strains observed for the three synthetic fibre contents considered in this study are shown in Fig. 6.

Fig. 6. Tensile creep strain for samples with different fibre dosages.

Tensile creep strains were significantly reduced by the addition of synthetic fibres. The differences observed between the specimens without fibres and those with 5 kg/m3 of fibres were significantly more pronounced than the differences observed when the fibre content was increased from 5 kg/m3 to 10 kg/m3. The introduction of 5 kg/m3 of synthetic fibres led to reductions in tensile creep strains between 20% and 11% with respect to the unreinforced specimens (reductions were 20%, 16%, and 11% at 14, 30, and 90 days, respectively). With 10 kg/m3 of synthetic fibres, reductions between 15% and 24% were observed (24%, 22%, and 15%, at 14, 30, and 90 days, respectively). These observations led to the conclusion that, in terms of tensile creep reduction, the additional gains achieved by doubling the fibre content were less significant than those achieved by the incorporation of fibres at the intermediate dosage considered in this study. In consequence, dosing the fibres at the maximum dosage did not led to the most advantageous control of tensile creep from a cost-benefit point of view.

400

R. H. Al Marahla and E. Garcia-Taengua

4 Conclusions In this study, a reference mix with w/c ratio of 0.29 and an average compressive strength of 60 MPa at 28 days was considered as representative of most of the medium to high specification mixes prevalent in real scale production. The effect of synthetic fibres on compressive creep, tensile creep and shrinkage when dosed at 5 kg/m3 and 10 kg/m3 was analysed. To this end, daily measurements were taken on loaded and unloaded specimens over a period of 90 days. Multiple replicates were produced and tested to ensure accuracy of the final results and conclusions, and a comparison was made to the predictive models proposed by Eurocode 2, Model Code, and the ACI committee 209. The following conclusions were drawn: • In terms of compressive strength, no significant differences were observed between the reference mix without fibres and the FRC mixes, irrespective of the fibre content. However, the addition of 5 kg/m3 or 10 kg/m3 of synthetic fibres increased the spitting tensile strength by 11.5% or 18%, respectively. • The incorporation of synthetic fibres was associated with a general, consistent decrease in compressive creep, shrinkage, and tensile creep strains. The magnitude of these reductions was observed to increase with the fibre content. • Compressive creep strains measured on FRC specimens with 5 kg/m3 of synthetic fibres were up to 15% lower than those corresponding to the reference mix without fibres. When the fibre content was 10 kg/m3, hihger reductions (up to 22%) were observed. In terms of the creep coefficient at 90 days, it decreased by 5% and 8% for fibre contents of 5 kg/m3 and 10 kg/m3, respectively. • Tensile creep strains observed in FRC specimens with 5 kg/m3 of synthetic fibres were up to 20% lower than the values obtained for the reference mix without fibres. The additional gains achieved by doubling the fibre content were less significant: reductions of up to 24% were observed when the fibre content was 10 kg/m3. • The addition of 10 kg/m3 of synthetic fibres led to shrinkage strains being up to 26% lower than those corresponding to the reference specimens without fibres. When the fibre content was 5 kg/m3, reductions of up to 16% were obtained. • The models proposed by the Model Code, Eurocode 2 and the ACI Committee 209 for the prediction of creep and shrinkage strains showed good agreement with the compressive creep strains observed in specimens without fibres, but they consistently overestimated shrinkage strains. Acknowledgements. The authors wish to acknowledge the contribution of Oscrete Construction Products, part of Christeyns UK Ltd, and Sika Ltd (UK), which very kindly provided some of the materials used in this study, as well as the support and assistance provided by the technical staff of the School of Civil Engineering, University of Leeds. The authors are also thankful to AlZaytoonah University of Jordan for the financial support granted to Ms Al Marahla in undertaking her PhD studies at the University of Leeds.

Compressive and Tensile Creep and Shrinkage

401

References 1. Kristiawan, S.A.: Strength shrinkage and creep of concrete in tension and compression. Civ. Eng. Dim. 8(2), 73–80 (2006) 2. Gilbert, R.I., Ranzi, G.: Time-Dependent Behaviour of Concrete Structures. CRC Press, Boca Raton (2010) 3. Bazant, Z.P., Wittmann, F.H.: Creep and shrinkage in concrete structures, pp. 12–16 (1982) 4. BS EN 1992-1-1. Eurocode 2: Design of concrete structures: Part 1-1: General rules and rules for buildings. European Committee for Standardization (2004) 5. FIB: fib model code for concrete structures 2010, Wiley, Germany (2013) 6. ACI Committee, ACI 209.2 R-08: Guide for Modeling and Calculating Shrinkage and Creep in Hardened Concrete. American Concrete Institute Committee (2008) 7. Monteiro, P.: Concrete: Microstructure, Properties, and Materials. McGraw-Hill Publishing, London (2006) 8. Buratti, N., Mazzotti, C., Savoia, M.: Post-cracking behaviour of steel and macro-synthetic fibre-reinforced concretes. Constr. Build. Mater. 25(5), 2713–2722 (2011) 9. Gamble, B.R., Parrott, L.J.: Creep of concrete in compression during drying and wetting. Mag. Concr. Res. 30(104), 129–138 (1978) 10. Brooks, J., Neville, A.M.: A comparison of creep, elasticity and strength of concrete in tension and in compression. Mag. Concr. Res. 29(100), 131–141 (1977) 11. Mangat, P.S., Azari, M.M.: A theory for the creep of steel fibre reinforced cement matrices under compression. J. Mater. Sci. 20(3), 1119–1133 (1985) 12. Domone, P.L.: Uniaxial tensile creep and failure of concrete. Mag. Concr. Res. 26(88), 144– 152 (1974) 13. Elvery, R.H., Haroun, W.: A direct tensile test for concrete under long-or short-term loading. Mag. Concr. Res. 20(63), 111–116 (1968) 14. Igarashi, S.I., Bentur, A., Kovler, K.: Stresses and creep relaxation induced in restrained autogenous shrinkage of high-strength pastes and concretes. Adv. Cem. Res. 11(4), 169–177 (1999) 15. Igarashi, S.I., Bentur, A., Kovler, K.: Autogenous shrinkage and induced restraining stresses in high-strength concretes. Cem. Concr. Res. 30(11), 1701–1707 (2000) 16. Forth, J.P.: Predicting the tensile creep of concrete. Cem. Concr. Compos. 55, 70–80 (2015) 17. Swamy, R.N., Stavrides, H.: Influence of fiber reinforcement on restrained shrinkage and cracking. J. Proc. 76(3), 443–460 (1979) 18. Chern, J.C., Young, C.H.: Factors influencing the drying shrinkage of steel fiber reinforced concrete. Mater. J. 87(2), 123–139 (1990) 19. Mangat, P.S., Azari, M.M.: A theory for the free shrinkage of steel fibre reinforced cement matrices. J. Mater. Sci. 19(7), 2183–2194 (1984) 20. Shah, S.P., Rangan, B.V.: Fiber reinforced concrete properties. J. Proc. 68(2), 126–137 (1971) 21. Llano-Torre, A., Arango, S.E., García-Taengua, E., Martí-Vargas, J.R., Serna, P.: Influence of fibre reinforcement on the long-term behaviour of cracked concrete. In: Creep Behaviour in Cracked Sections of Fibre Reinforced Concrete, pp. 195–209. Springer, Dordrecht, (2017) 22. García-Taengua, E., Arango, S., Martí-Vargas, J.R., Serna, P.: Flexural creep of steel fiber reinforced concrete in the cracked state. Constr. Build. Mater. 65, 321–329 (2014)

Creep in FRC – From Material Properties to Composite Behavior Martin Hunger1(&), Jürgen Bokern1, Simon Cleven2, and Rutger Vrijdaghs3 1

3

BASF Construction Solutions GmbH, Trostberg, Germany [email protected] 2 Institute of Building Materials Research, RWTH Aachen University, Aachen, Germany Materials and Building Technology Section at KU Leuven, Leuven, Belgium

Abstract. Although being the subject of numerous studies during the last 5 decades, fiber reinforced concrete (FRC) only very recently started to penetrate the construction market to a larger extent. This is mainly due to the lack of appropriate standards in the past years and the recent availability of structural codes such as fib Model Code 2010, the German DAfStb guideline “Stahlfaserbeton”, the Italian code, or codes under development, such as the Eurocode 2. However, FRC still struggles with some issues that potentially hinder, at least at a first glance, a further penetration into structural applications. Creep under tensile stress is one of these major obstacles. The creep response of a structural member in tension results from the interaction of the creep of the fiber material itself and the creep of the bond between fiber and surrounding matrix. The latter at last leads to the fact that creep is also a subject of further investigation for steel fiber reinforced concrete which shows that this is not an exclusive problem for synthetic fibers, as publicized in common literature. Unfortunately, commonly agreed test methods to investigate creep are missing. In this regard more research is required in order to specify which load levels and crack openings should be chosen to provide a realistic scenario for creep tests on FRC in general. Testing of creep in FRC is a very complex and time-consuming matter. Therefore, the subject of this paper is an investigation of a possible correlation of the creep properties of a single filament and the bond behavior of a single fiber in mortar. The most important subject of this paper is whether it is possible to derive the overall creep response of the composite material FRC from the determined component-specific characteristics. Based on a thorough test plan, two distinctly different performing polypropylene fibers (moderate to highperforming) are tested and compared. The determination of short- and long-term behavior of the filaments in this test plan is carried out in varying temperatures in a range from −10 °C to 60 °C. Results indicate that the creep performances of a filament and the bond between a filament and a mortar correlate with the overall creep of the corresponding FRC. Moreover, a broad range of creep performance is revealed which strongly suggests that the performance of synthetic fibers and, in particular the creep performance, cannot be generalized. Keywords: Creep
Fiber reinforced concrete (FRC) Tension
Filament
Bond
Temperature


Synthetic fibers


© RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 402–413, 2021. https://doi.org/10.1007/978-3-030-58482-5_37

Creep in FRC – From Material Properties to Composite Behavior

403

1 Introduction About 10 years ago, the state-of-the-art in terms of creep of synthetic fibers could be summarized as follows: (i) neither a model to include creep in the design of FRC is available nor threshold values in regard to tolerated creep deformations are defined, (ii) a long-term load level of 50% of residual load in short-term test seems like a maximum for well performing polypropylene (PP) fibers and (iii) long-term tests are lacking as creep may occur after two years and even later [1]. Not much has changed in the general perception of synthetic fibers in literature since then although results have become available, in particular during the last two years, that would first put the above-mentioned three limitations into perspective and even suggest synthetic fibers for structural applications [2, 3]. However, one of the most critical factors hindering a wider application or more specifically a structural application of synthetic macro fibers is creep performance. This paper aims to shed some light on the entire causal chain of creep – from material creep (fiber level), over creep in the interface (bond creep) until creep performance of structural members. An evaluation seems challenging as testing is not standardized and some decisive test parameters often reported in literature (e.g. load levels, pre-cracking width) seem at least arguable. This paper presents a potential testing plan for the creep analysis of fibers for structural applications. Moreover, some indicative results are presented to show how based on fiber creep and single-fiber pullout the creep performance of FRC under sustained load can be modelled, a very valuable approach in simplifying the time-consuming and cumbersome creep testing of FRC and a potential way forward in design of FRC.

2 State-of-the-Art in FRC Creep The long-term behavior of FRC under sustained load remains one item that still requires intensive research. FRC (not necessarily with polymeric fiber reinforcement) is in the meantime part of a number of international codes for structural concrete (e.g. ACI 318, fib Model Code 2010 (MC 2010), the German DAfStb Guideline for steel fiber-reinforced concrete or the Italian guideline (CNR-DT 204/2006). More codes considering structural behavior of FRC are in development or about to include FRC, such as Eurocode 2. In general, creep even for normal RC structures is only treated with simplified models and design rules. A good overview of how few creep related inputs are considered in MC2010 is given in [3]. MC2010 highlights the need to take creep of cracked FRC into account but does not provide any design guidelines. The German DAfStb-guideline for steel fiber-reinforced concrete (November 2012) only mentions the possibility that deformations can increase due to creep in the interface of steel fiber and matrix. However, no testing or any form of proof is required. Creep is a phenomenon to be considered in Serviceability Limit State (SLS) as deformations have to be limited and crack width growth has to be controlled. In SLS, quasi permanent loads in combination will only sum up to 30–40% of fR,1 (according to EN 14651) of the FRC component which in turn is even less the 30–40% of the

404

M. Hunger et al.

ultimate tensile strength (UTS) of the fiber [3]. Hence, most of the creep testing done with 60% load level or the like is too demanding and does not represent practical application. This way a 40% load level can be accepted as a sound upper limit. When it comes to creep in FRC, MC2010 does not consider any structural issues other than creep of the fiber material itself. There is general consensus that the addition of fibers, not exceeding a critical amount, does not change the creep behavior in uncracked state in comparison to normal concrete, neither in compression nor in tension [4]. After cracking, however, time-dependent strains in FRC under tension or in bending mainly stem from three sources: (i) creep in the compression zone, (ii) creep in the interface between fiber and matrix and (iii) creep of the fiber material itself. The authors believe that for a deterministic description of the phenomenon creep (on composite level) the knowledge of single fiber creep and interfacial bond is indispensable. So far, literature is very scarce when it comes to a systematic investigation of creep on these three levels [5, 6]. However, for the technical approval of fibers for structural purpose the ‘Deutsches Institut für Bautechnik’ prepared a guidance paper (DIBt (2013)) that defines the necessary testing [7]. For structural purposes this also includes a section on creep. Here, the above suggested split in fiber creep, interfacial bond and creep on the composite level is followed. According to this test plan, next to creep testing on concrete beams by means of 4–point bending, optionally testing of fiber creep and interfacial bond and creep at temperatures of −10, 20 and 60 °C is required. In order to define the performance envelope of typically used PP-fibers this testing plan was used as guidance and two differently performing PP-fibers were chosen to undergo this extensive test plan.

3 Applied Materials and Testing Setup 3.1

Applied Materials

A summary of the materials and test methods used in this research is given in the following chapters. Two PP-fibers with different mechanical properties and shape were selected (Table 1). For single fiber and fiber pull-out testing long filaments were used whereas for the FRC bending creep tests short fibers in their nominal lengths were applied. Table 1. Properties of the applied PP fibers* Specimen

Length Equivalent Shape Strength Modulus [mm] diameter [mm] [N/mm2] [N/mm2] Type A 48 0.85 embossed 363 11,000 Type B 54 0.81 crimped 408 15,000 *properties determined according to EN 14889-2, modulus shown here is tangent modulus at first loading, i.e. the slope of the stress-strain curve at zero stress

To investigate the short- and long-term fiber-matrix interaction (pull-out) a mortar defined in DIN EN 14649 (SIC test for glass fibers) is prescribed in the DIBt guideline. It is composed, by weight, of three parts cement, one part CEN standard sand and 0.427

Creep in FRC – From Material Properties to Composite Behavior

405

parts water. The composition of the FRC for bending tests is also specified in the DIBt guideline. It contained 300 kg/m3 CEM I 32,5 R, 1.835 kg/m3 siliceous gravel, 180 kg/m3 water and 5.0 kg/m3 of either fiber Type A or B. 3.2

Single Fiber Testing

Single fiber tensile strength properties are determined according to DIN EN ISO 5079. Both ends of the filament are embedded in a metal polymer resin to prevent any slippage or edge pressure when clamping. The free length of filament is 150 mm and the test machine is operated in position-controlled mode at a rate of 15 mm/min. For the single fiber creep trials, apart from the free length of the filament, which is reduced to 75 mm, filaments are prepared same as for the single fiber tensile tests. One side of the filament cast in metal polymer resin is fixed at the top of the creep test setup whereas the embedded filament end hanging down is loaded with steel plates and allowed to freely move. The creep deformation is measured by an analogue dial gauge. This setup is chosen to ensure an unhindered filament deformation and a centric loading. Load levels of 40% and 70% of the fiber tensile strength are applied. All tests described before are performed at temperatures of −10, 20, 40 and 60 °C. 3.3

Fiber Pull-Out and Pull-Out-Creep

For the fiber pull-out and pull-out-creep tests, mortar blocks of 30 mm side length are cast, in which one side of the filament is centrally embedded in such way that the filament is perpendicularly oriented to the mortar block surface. The thickness of the mortar block is similar to the embedment depth of the fiber, which is 24 mm. The other side of the filament again is cast in a steel polymer resin. Between mortar and resin a free filament length of 150 mm is maintained to enable a stress-free vertical and centric pull-out of the fiber. After demolding at the age of one day the specimens are stored for 6 days under water and afterwards for 7 days at 65% RH until testing. The temperature is kept constant at 20 °C. For pull-out testing the mortar block is placed on a notched steel plate where the notch serves as lead-through for the fiber. Then the downwards oriented resin block is clamped and the mortar block is pulled up with a rate of 25 mm/min until complete pull-out. The test is performed at temperatures of −10, 20 and 40 °C for both fiber types and additionally at 60 °C for fiber type B. Pull-out-creep tests are conducted with the same test set-up as the fiber pull-out tests, but only with fiber type B and a free fiber length of 75 mm. As load level 40% of the respective pull-out load is defined. The creep deformation is again measured by an analogue dial gauge, that is placed against the steel plate the mortar was positioned on. The test is performed at temperatures of 20 °C and 40 °C for at least three months. 3.4

Fiber Reinforced Concrete Bending and Bending Creep

Finally, the composite creep behavior of the FRC with fiber type A and B is investigated. Therefor six fiber-reinforced concrete beams per fiber type with dimension of 150  150  700 mm3 are cast, kept in the molds protected from evaporation for one

406

M. Hunger et al.

day and wrapped in plastic foil after demolding. At an age of 21 days the beams are notched according to EN 14651 and wrapped again in plastic foil for further 7 days. In order to determine the creep load the beams are first subjected to 3-point-bending test in accordance with EN 14651 at an age of 28 days. The test is stopped when a CMOD of 0.3 mm is reached. The load applied at that point of time is documented and the beam is unloaded. Room temperature while storing and testing is always kept at 20 °C. Out of the series of six beams finally three beams are chosen which show a similar behavior and load level in the short-term tests. Those beams are subsequently installed in 3-point bending test frames where each beam could be loaded with its specific creep load, which is defined as 50% of the beam’s residual strength at 0.3 mm crack width. The midspan deflection of the beams is measured by two indicating calipers over a time of one year at a controlled temperature of 20 °C and 65% RH.

4 Results 4.1

Single Fiber Tensile Strength

Tensile strength properties of both fiber types were investigated as a first step. Stressstrain curves obtained at a temperature range from −10 °C to 60 °C are shown in Fig. 1. Comparing the two fibers, type B shows at each temperature level an about 20– 30% higher ultimate tensile strength (UTS) than type A. Moreover, the mechanical response to tensile stress is much stiffer for type B. The impact of temperature on tensile fiber properties is different as well. While for fiber type A the UTS drops at temperatures beyond 20 °C it remains in a comparable range up to 40 °C for fiber type B. A similar conclusion can be derived concerning fiber stiffness. While for fiber type A the response softens substantially with each temperature increment, the effect of temperature on fiber type B becomes especially noticeable beyond 40 °C with a reduction in strength and stiffness. Overall it can be concluded that increasing temperatures have a negative impact on tensile strength of polymer fibers, but critical temperatures at which a relevant decrease may occur seems to be fiber specific. 4.2

Single Fiber Creep

As shown in Fig. 2 where the creep deformation until 120 days of both fiber types is presented, the difference in tensile strength properties transfers to their creep behavior. For fiber type B, no unstable creep development occurs at load levels of 40% and temperatures from −10 °C to 60 °C. Up to 40 °C it even seems that creep tends to reach asymptotically a finite value. Instead, for fiber type A, an unstable creep development may occur beyond 20 °C already. A more precise determination of the critical temperature is unfortunately impossible, because a test at 40 °C is missing. At the higher load level of 70%, fibers of both types show unstable creep development and rupture. However, while for fiber type A rupture occurs at 20 °C already, it takes for type B until 40 °C. Considering that fiber type B is – because of its higher UTS – exposed to much higher absolute loads, this is in particular striking. In summary it is observed that there are PP-fibers with an asymptotic stable performance in single fiber creep even at high temperatures.

Creep in FRC – From Material Properties to Composite Behavior

407

Tensile Stress in N/mm²

Tensile Stress in N/mm²

500

500

-10 °C

-10 °C 20 °C

400

40 °C

20 °C

400

40 °C 60 °C

60 °C

300

300

200

200

100

100 0

0 0

5

10

0

15 20 Strain in %

5

(a)

10

15 20 Strain in %

(b)

Fig. 1. Stress-strain curves for fiber type A (left) and B (right)

Creep strain in %

Creep strain in %

0

0

-15

-15

-30

-30

-45

-10 °C -10 °C 20 °C 20 °C 40 °C 40 °C 60 °C

-45 -10 °C -10 °C 20 °C 20 °C 60 °C

-60

40 % 70 % 40 % 70 % 40 %

-75

-60

40 % 70 % 40 % 70 % 40% 70 % 40 %

-75 0

30

60 90 120 Creep duration in d

(a)

0

30

60 90 120 Creep duration in d

(b)

Fig. 2. Single fiber tensile creep behavior of fiber type A (left) and B (right)

Finally, it can be concluded that quality of PP-fibers differs regarding creep, but high-quality PP-fibers are capable to sustain relevant stress levels even at increased temperatures. 4.3

Fiber Pull-Out

For characterizing the bond behavior of fibers and cementitious matrix, pull-out tests at different temperatures are performed (Fig. 3). At a temperature of −10 °C both fibers show the same behavior with a pull-out-load of about 150 N following a decrease in bond with increasing pull-out-deformation. At higher temperatures of 20 °C to 60 °C fiber type B shows the same behavior. In contrast, the bond of fiber type A is greatly

408

M. Hunger et al.

influenced by temperature. At 20 °C the pull-out-load is reduced by 50% and at 40 °C by about 65%. For a temperature of 60 °C there is no result, however, considering the impact on fiber tensile strength (Fig. 1) a significant further decrease is expected. In conclusion it seems that fiber type B itself and eventually its shape (crimped) lead to a much better bond at higher temperature and therefor allow a broader application scope than for fiber type A. Tensile Load in N

Tensile Load in N

200

200 -10 °C

175

-10 °C

175

20 °C

20 °C

150

150

40 °C

125

125

60 °C

100

100

75

75

50

50

25

25

40 °C

0

0 0

5

10 15 20 25 30 Pull-Out-deformation in mm

0

(a)

5

10 15 20 25 30 Pull-Out-deformation in mm

(b)

Fig. 3. Pull-Out behavior of fiber type A (left) and fiber type B (right)

4.4

Fiber Pull-Out Creep

The pull-out creep behavior is examined only on fiber type B at 40% load level and 2 temperature levels in order to investigate whether besides fiber creep also long-term bond phenomena might influence the fiber pull-out creep (short-term bond is not influenced by temperature for fiber type B). Results obtained at 20 °C and 40 °C are shown in Fig. 4. At both temperature levels this fiber type first shows an initial steep creep deformation which turns into an almost asymptotic behavior after about 60 days. This is in line with fiber creep observations. A first estimate of the fiber creep deformation at roughly similar loads indicates that in absolute terms, the fiber creep deformation is in the same order of magnitude as the pull-out creep deformation. Therefore, the effect of long-term bond phenomena appears limited and the pull-out creep seems to be dominated by the fiber creep deformation. This confirms findings in [6]. 4.5

Fiber Reinforced Concrete Bending Creep

Finally, the creep performance of the two fiber types is tested as composite material, hence, in FRC. After pre-cracking of six beams for creep-test preparation, three of them were fully tested in accordance with EN 14651 to characterize the short-term bending behavior of the FRC (left side of Fig. 5). Fiber type A provides an average residual

Creep in FRC – From Material Properties to Composite Behavior

409

Creep deformation in mm 0 20 °C 40 %

-1

40 °C 40 %

-2 -3 -4 -5 -6 0

30

60

90 120 Creep duration in d

Fig. 4. Pull-Out creep behavior of fiber type B

strength of fR,1 = 1.5 N/mm2 and fR,3 = 1.6 N/mm2. Type B fibers exceed this performance especially at higher deformations and residual strength of fR,1 = 2.1 N/mm2 and fR,3 = 2.4 N/mm2 are achieved. These results perfectly fit to the short-term fiber bond performance, where fiber type B showed much higher pull-out loads. FRC bending deformation in mm

Residual tensile strength in N/mm² 0,00

5,0 Type A precracking

4,5

Type A postcracking

4,0 3,5

Type B precracking

-0,10

Type B postcracking

-0,15

3,0

Type A

-0,05

Type B

-0,20

2,5

-0,25

2,0

-0,30

1,5 1,0

-0,35

0,5

-0,40 -0,45

0,0 0

1

2

3

4 5 CMOD in mm

0

50

100

150 200 250 300 creep duration in d

Fig. 5. Results of 3-point-bending tests to determine the creep load (left) and creep behavior of fiber reinforced concrete beams under bending load (right)

As a consequence, the absolute creep load of beams with fiber type B is about 15% higher. Despite the higher applied load, fiber type B concrete beams show a smaller creep deformation than beams with fiber type A (compare Fig. 5 right). This, according to previous analysis, should predominately be related to the better creep performance of fiber type B (Fig. 2). But, in contrast to fiber creep and fiber bond creep performance,

410

M. Hunger et al.

which both show an almost asymptotic trend after 60 days, for the beams unfortunately an uncertainty remains in this respect and further investigations are required to create a better understanding of potential phenomena behind and to build further confidence. If the FRC creep deformation level for fiber type B should remain at this level or slightly above, it would be acceptable for construction.

5 Modeling The experimentally obtained results highlight the different behavior for both fiber types on all scales, but time and space constraints limit the number of test samples as well as the duration of the creep test to 1 or several years at most. However, most structures are designed with at least a 50 years life span in mind. Finite element modelling (FEM) can be used to verify the SLS and ULS requirements at these time scales. In this chapter, a numerical approach to determine the crack width growth of cracked FRC under uniaxial tension is presented. This approach is based on previous work [6] where cored cylinders were notched at mid-height and subjected to sustained uniaxial tensile loading. The load was applied after pre-cracking the specimen to 0.2 mm. The numerical approach does not consider the pre-cracking procedure itself, rather it only simulates the time-dependent crack widening upon and after load application. The approach is validated for 2 types of polymeric FRC investigated before and is implemented as such in this research project. The numerical approach consists of three steps: (i) pre-processing and fiber generation, (ii) construction and analysis of a finite element model with discrete fibers and (iii) post-processing and analysis. 5.1

Fiber Generation Algorithm

The fibers are modelled discretely in the numerical model, and their positions within a cylindrical concrete core (diameter 100 mm, height 300 mm) are determined in a fiber generation algorithm (FGA), written in MATLAB. The FGA requires the input of the concrete volume, fiber geometry and fiber fraction and distributes straight fibers within the volume without overlapping or intersecting, according to a user-defined distribution. Here, a random 3D distribution is assumed, and since the pre-cracking itself is not modeled, the crack faces are assumed perfectly flat. The fibers are divided into bondslip and fully embedded fibers, whether the fiber crosses the crack face or not, respectively. In Fig. 6, all fibers are shown in the core on the left and middle figure, with the contributing fibers, i.e. those crossing the flat crack faces, shown in red. On the (right), one half of the numerical model is also shown. 5.2

Finite Element Model: Material Models and Analysis Procedure

From the fiber distribution, a finite element model (FEM) with discrete fibers is built in the numerical program DIANA. The material models are calibrated based on the experimental results, discussed before. The model consists of 4 element types: (1) 3D solid elements for concrete, (2) embedded reinforcement elements for the embedded

Creep in FRC – From Material Properties to Composite Behavior

411

Fig. 6. Fiber distribution in the whole core (left and center), one half of the FEM model with fibers crossing the crack (right)

fibers, (3) bond-slip reinforcement elements for the contributing fibers and (4) 3D beam elements for the fiber part in the crack itself. For the concrete (1), a linear-elastic material model is assumed as the tested material is uniaxial strain softening and the initial pre-crack is explicitly modelled at the start of the finite element analysis (FEA). The embedded fibers (2) adapt the stiffness of the mother material in which they are embedded, based on the stiffness of the fibers, but since they do not cross the crack, they take up only very limited forces. As such, the embedded fibers are elastic with the elastic parameters determined in the short-term tests and are assumed to be perfectly bonded with the concrete. The bond-slip fibers (3) behave elastic as well, but in contrast with the embedded fibers, they can undergo deformations relative to the mother elements. The pull-out displacement upon loading is calibrated with the pull-out tests, where the material model requires the input of the pull-out stiffness parameter, i.e. the initial slope in Fig. 3. The effect of the embedded length and angle is effectively neglected by adopting an identical pull-out stiffness for all fibers, which is verified for the pull-out data set presented here. The 3D beam elements in the crack (4) can undergo creep deformations based on the single fiber creep deformations presented in Fig. 2. In the FEM, fiber creep is implemented with a time-dependent stiffness and based on analyses presented in [6] the pull-out creep can be lumped in the single fiber creep for polymeric fiber types. As such, the model can capture single fiber creep as well as (implicitly) pull-out creep. The applied load is 50% of the residual strength of the specimen at the pre-crack width of 0.2 mm as discussed before. The residual strength is numerically determined as no uniaxial tensile tests are performed. The imposed load is 950 N and 1450 N for fiber type A and B, respectively, and it is applied centrically and is kept constant for 50 years, divided into 125 logarithmically spaced time steps. A quasi-Newton iterative scheme with a simultaneous force and displacement convergence criterion of 5% is used.

412

5.3

M. Hunger et al.

Analysis and Discussion

The evolution of the crack width in mm and the corresponding creep coefficient evolution, u, is shown in Fig. 7 on the left and right respectively. It is noted that the absolute value of the crack width is greater for type B fibers, even though they outperform the type A fiber, as shown in Fig. 1. It is attributed to the nearly 50% increase in load (950 N to 1450 N), even if the load ratio remains constant at 50%. The superior performance of the type B fiber is more clearly shown in the creep coefficient evolution. 1.8

0.4

Type A

Type A

1.7

Type B

Type B

0.35

Creep coefficient [-]

1.6

COD [mm]

0.3

0.25

1.5

1.4

1.3

1.2

0.2 1.1

0.15 1 s.

1

1 min.

1 h.

1 m.

1 d.

1 y.

1 s.

10 y. 50 y.

1 min.

1 h.

1 d.

1 m.

1 y.

10 y. 50 y.

Time

Time

Fig. 7. (left) Crack width evolution (right) creep coefficient evolution for both fiber types

While the load application is fundamentally different for the experimental results shown at right in Fig. 5 (bending) and the numerical model (uniaxial tension), the ratio of the creep coefficients of type A with respect to type B can be compared (Fig. 8). While both methods yield different values for the coefficient, a similar conclusion can be drawn, as both methods indicate that the type B fiber outperforms type A (ratio is greater than 1, higher creep coefficients for type A). Additionally, for the time period considered in the bending test, an increasing ratio in time can be observed (unsteady course of experimental graph is a result of calculating the ratio of parameters only slightly changing). In that sense, the numerical methodology and results are in line with the experimental, bending results. 1.3 Uniaxial (FEM)

1.25

Bending (EXP)

ratio [-]

1.2

1.15

1.1

1.05

1 1 s.

1 min.

1 h.

1 d.

1 m.

1 y.

10 y.50 y.

Time

Fig. 8. Ratio of the creep coefficient for fiber type A with respect to type B in bending and uniaxial tests

Creep in FRC – From Material Properties to Composite Behavior

413

6 Conclusion The work presented in this paper can be summarized as follows: 1. Standards and guidelines, due to limited understanding, are not yet considering creep of FRC in a satisfying way. 2. A holistic approach considering creep from material, over interface to composite level is suggested to better understand decisive phenomena. 3. Fiber creep and pull-out creep results indicate that a high performing PP-macrofiber can sustain a load level of 40% related to maximum strength in short-term tests even at a temperature of 40 °C. This would be appropriate for a lot of applications in construction. 4. While for a temperature of 20 °C the superior performance is proven on composite level (FRC), results at 40 °C still need to be verified in beam tests. 5. The importance of fiber creep for the creep of FRC with synthetic fibers is confirmed. Other factors, such as fiber shape (crimping vs. embossing) may have an influence as well. 6. Finally, it seems feasible to simulate the crack opening due to creep on synthetic FRC by considering test results in a model recently developed at KU Leuven [6]. Further proof is needed. Investigations presented here shall be continued to understand creep of FRC up to an extent that engineers are able to consider this factor in structural design calculations.

References 1. Kusterle, W.: Viscous material behavior of solids – creep of polymer fibre reinforced concrete. In: 5th Central European Congress on Concrete Engineering, Baden, pp. 95–100 (2009) 2. Pujadas, P., Blanco, A., Cavalaro, S., de la Fuente, A., Aguado, A.: The need to consider flexural post-cracking creep behavior of macro synthetic fiber reinforced concrete. Constr. Build. Mater. 149, 790–800 (2017) 3. Plizzari, G., Serna, P.: Structural effects of FRC creep. Mater. Struct. 51(6), 1–11 (2018) 4. Balaguru, P., Ramakrishnan, V.: Properties of fibre reinforced concrete: workabilty, behaviour under long-term loading and air-void characteristics. ACI Mater. J. 85(3), 189–196 (1988) 5. Babafemi, A.J.: Tensile creep of cracked macro synthetic fibre reinforced concrete. Ph.D. thesis. Stellenbosch University, South Africa (2015) 6. Vrijdaghs, R.: Creep of synthetic fiber reinforced concrete - a multi-scale and two-phased approach. Ph.D. thesis. KU Leuven, Belgium (2019) 7. DIBt: Prüfplan für die Zulassungsprüfung von Polymerfasern zur Verwendung in Beton nach DIN EN 206-1 in Verbindung mit DIN 1045-2 mit nachgewiesener Wirksamkeit. Deutsches Institut für Bautechnik, Berlin, January 2013

Durability

Morphology of Corrosion of Metallic Fibers in Aggressive Media Carmen Andrade1 and Miguel A. Sanjuán2(&) 1

International Center of Numerical Methods in Engineering, CIMNE-Madrid, Madrid, Spain 2 Institute for Cement and Its Applications, IECA, Madrid, Spain [email protected]

Abstract. In marine and industrial environments, the steel bars embedded in concrete can corrode producing the cracking of the cover and impacting negatively in the load-bearing capacity. Metallic fibres have shown a record of good mechanical performance as reinforcing material for concrete. Those on carbon steel can corrode in the aggressive environments and although publications have reported a better impermeability in the fibre reinforced concretes, it still remains the question of how the steel fibres corrode and, in this case,, whether they can microcrack the surroundings. In present communication are presented results of long-term performance of the fibres due to chlorides and carbonation attacks during more than 20 years of exposure. The carbon fibres corrode but not cracking seems to be produced in spite of the full conversion of the fibres into oxides. The galvanized fibres corrode comparatively less, and the stainless-steel ones are in perfect condition. Keywords: Corrosion
Metallic fibres
Durability
Aggressive environment
Concrete

1 Introduction The introduction of fibres into the concrete is due to the historical disadvantage that both cement mortar and concrete have with regard to their low tensile strength and high brittleness. To counteract these characteristics, they are normally controlled by means of steel reinforcement. As a measure to control the cracking, fibre reinforced concretes were also used from the second half of the 20th century. Steel fibre reinforced concrete (SFRC) is a composite material, combining a Portland cement and steel fibres as a discontinuous reinforcement, which are randomly distributed in the cement paste. The cement content in SFRC usually is ranged between 300 kg/m3 and 450 kg/m3. Lower amount of cement leads to a loss of workability. SFRC needs from 35% to 45% by volume of cement paste while the plain concrete only needs from 25% to 35%. SFRC is increasingly being used for the production of slabs, tunnel linings, foundations, thin-shell structures and so on [1, 2]. From the durability viewpoint, corrosion of metallic fibers in aggressive environments is a risk due to the discontinuous nature of the metallic fibres in the SFRC. The first effect is aesthetic as consequence of the surface corrosion [3]. Later, the corrosion © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 417–422, 2021. https://doi.org/10.1007/978-3-030-58482-5_38

418

C. Andrade and M. A. Sanjuán

of the internal steel fibres may affect to the strength of the concrete structure [4, 5]. Given that, currently, the total replacement of steel bars reinforcement by SFRC is controversial when the long-term durability of SFRC under severe environments is addressed [6]. In present communication results are presented on the morphology of the corrosion of fibers and whether the oxides have the expansive character as in normal reinforcements.

2 Experimental The main components found in the fibre reinforced concretes are the Portland cement, both fine and coarse aggregates, water, superplasticizer and steel fibres. Therefore, the only significant variation from a normal concrete is the addition of the fibers. Different types of steel fibre reinforced concrete pipes have been studied for 30 years. 2.1

Materials

Portland cement CEM I 42.5 R, siliceous aggregates (sand and gravel), tap water and superplastifier were used in the present research program. Three types of hooked steel fibres were selected to produce the fibre reinforced concrete: carbon steel, stainless steel and galvanized steel. 2.2

Aggressive Solutions

An aggressive solution of Na2S 9H2O (2.9 g/l) and NH4Cl (29 g/l) was prepared. In addition, tap water was used as reference media. 2.3

Fibre Reinforced Concrete

Fibre reinforced concrete made of carbon steel, stainless steel and galvanized steel cylindrical specimens (ø15  30 cm) were manufactured according to the mix design shown in Table 1. Concrete slump test was performed to check the workability of freshly made concrete, and the result was 5 cm. These specimens have an internal hole of 5 cm of diameter in order to simulate a concrete pipe. Concrete was cured for 28 days under water and, later on, two series of the pipe specimens were subjected to the aggressive solution (Na2S 9H2O (2.9 g/l) and NH4Cl (29 g/l)) and other two series of the pipe specimens were kept in tap water. The concrete specimens’ layout is shown in Fig. 1 (Table 2). Table 1. Concrete mix design (kg/m3). Constituent Cement Sand Gravel Water Water with superplastifier Steel fibres Content 350 820 1110 190 154 35

Morphology of Corrosion of Metallic Fibers in Aggressive Media

419

SoluƟon level

Fig. 1. Steel fibre reinforced concrete specimens’ layout. Table 2. Steel fibre reinforced concrete codes. Water/cement ratio 0.55 0.44 (with superplastifier)

Witout fibres 1 6

Steel fibres 2 7

Steel fibres with 3%wt Ca(NO2)2 3 8

Stainless steel fibres 4 9

Galvanized steel fibres 5 10

After the first months of testing, some of the specimens were held at the atmosphere for 8–9 years (sheltered from rain) while other were kept in the aggressive solutions. Some remaining after the atmospheric conservation where introduced in NaCl solution for further years. The analysis of the results was not systematic and is out of present work to give rates for deterioration. However, having been submitted to aggressive solutions, some of the fibres fully corrode and have allowed us to study which is the kind of damage produced.

3 Results Some aspects of the specimens will be presented in order to appreciate the oxides and their impact in the concrete. The specimens immersed in tap water with and without nitrites, presented corrosion at 4 months of exposure in the superficial fibres as shown in Fig. 2 left. However, galvanized and stainless-steel fibres were found free of corrosion. In the specimens held in the aggressive solutions, those without fibres presented cracks and expansion. On the opposite the stainless-steel fibres were not corroding as can be appreciated in Fig. 3.

420

C. Andrade and M. A. Sanjuán

Fig. 2. Aspect of the specimens after 4 months in tap water (left) and in the aggressive solutions (right).

Fig. 3. Concrete with stainless steel fibres after 28 years of being in contact to aggressive solutions, air and sodium chloride solution.

The aspect of the corroding fibres is shown in Fig. 4. In the left is shown that all the internal fibres were corroding 4 months after of attack and in the right after 28 years.

Fig. 4. Aspect of the concrete with the fibres fully corroded.

Morphology of Corrosion of Metallic Fibers in Aggressive Media

421

Figure 5 shows aspects of the oxides and how they diffuse out of the fibre. In the right can be detected that the concrete surrounding the fibre is not cracks in spite that in some places the fibre has disappear completely.

Fig. 5. Aspect for the oxide’s diffusion from the fibres when corrosion is active. No cracking around the fibres was identified.

4 Discussion Due to the diverse environments and conditions where the specimens were held a correct rate of degradation was not feasible, but however the work enables to deduce well the morphology of the corrosion and that the concrete surrounding the fibres does not seem to crack. That is, the fibres dissolve without producing visual damage in the concrete, even at its surface. In the surface the fibres dissolve too until disappearing but without visual cracking. This performance can be explained in Fig. 6 where it has been tried to indicate that the controlling factor for producing cracking due to corrosion oxides depends on the ratio diameter of the bar/cover depth. This ratio expresses also the ratio of the surface where the pressure is produced due to the expansive character of the oxides with respect to the volume around the rod.

Fig. 6. Importance of the ratio cover depth/bar diameter for the cracking due to reinforcement corrosion.

422

C. Andrade and M. A. Sanjuán

The oxides occupy a higher volume than the parent metal due to oxides have oxygen and water in addition of the metallic atoms. The cracking of the cover due to the pressure they impose depends on how large the cover thickness is. If it is large enough the cracks do not reach the external surface. In addition to this effect, we attribute to the cover thickness much larger in the case of internal fibres, in the case of superficial fibres, very close to the surface but not protruding, the oxides can however produce cracking which is mixed with the diffusing oxide and the spalling is very superficial. The fibres protruding from the surface they directly are dissolved and detached form the concrete mass due to the loss of bond. The performance of the oxides detected enables to support the idea of a sacrificial thickness in fibre reinforced concretes when in contact to aggressive solutions. The corrosion is going to produce a concrete surface thickness where the fibres will disappear and with this disappearance their reinforcing effect.

5 Conclusions The specimens with stainless steel have been maintained for 28 years free of corrosion in spite of the very aggressive solutions tested. Galvanized fibres presented a not so good performance because some of them corrode and other not, Bare steel has corroded in the aggressive media and in tap water. The morphology pf the corrosion shows however that not microcracking is visually detected in the concrete surrounding the bars which is indicative that the concrete in the attacked thickness is not to weaken and become more permeable.

References 1. Serna, P., Arango, S., Ribeiro, T., Núñez, A.M., Garcia-Taengua, E.: Structural cast-in place SFRC: technology, control criteria and recent applications in Spain. Mater. Struct. 42, 1233– 1246 (2009) 2. di Prisco, M., Plizzari, G.A.: Precast SFRC elements: from material properties to structural applications. In: di Prisco, M., Felicetti, R., Plizzari, G.A. (eds.) 6th International RILEM Symposium on Fibre-Reinforced Concretes, pp. 81–100. RILEM Publications SARL, Varenna, Italy (2004) 3. Balouch, S.U., Forth, J.P., Granju, J.-L.: Surface corrosion of steel fibre reinforced concrete. Cem. Concr. Res. 40, 410–414 (2010) 4. Mangat, P.S., Gurusamy, K.: Permissible crack widths in steel fibre reinforced marine concrete. Mat. Struct. 20, 338–347 (1987) 5. Mangat, P.S., Gurusamy, K.: Corrosion resistance of steel fibres in concrete under marine exposure. Cem. Concr. Res. 18, 44–54 (1988) 6. Hwang, J.P., Jung, M.S., Kim, M., Ann, K.Y.: Corrosion risk of steel fibre in concrete. Constr. Build. Mater. 101, 239–245 (2015)

Effects of Fibres on the Flexural Behaviour of Sound and Damaged RC Beams Raúl L. Zerbino1(&), María C. Torrijos1, Graciela M. Giaccio2, and Antonio Conforti3 1

CONICET, LEMIT-CIC, Faculty of Engineering, UNLP, La Plata, Argentina [email protected] 2 LEMIT-CIC, CIC Researcher, Faculty of Engineering, UNLP, La Plata, Argentina 3 Department of Civil, Environmental, Architectural Engineering and Mathematics (DICATAM), University of Brescia, Brescia, Italy

Abstract. The incorporation of fibres in Reinforced Concrete (RC) beams controls the width and evolution of cracks leading to positive effects on the durability of the element. The study of damage processes in concrete and their effects on the residual properties represents a key point related to the service life of RC structures. The contribution of fibres on the bending behaviour of sound and damaged RC beams was investigated. In order to use alkali silica reaction as a damaging tool, RC beams with and without fibres and reactive aggregates were subjected to service loading conditions during eight months in an environment with high humidity. The evolution of deformations and the distribution and propagation of cracks were recorded. As reference, similar RC beams without reactive aggregates were also evaluated. After the treatment, all RC beams were loaded up to failure. The free expansion, the compressive strength and the bending residual capacity of plain and fibre concrete were measured on companion specimens for material characterization. The effect of both fibres and alkali silica reaction on bearing capacity and ductility of RC beams were analysed. Results showed that alkali silica reaction damage provokes a significant reduction of RC beam ductility, while the flexural strength is preserved. Keywords: Alkali silica reaction concrete
Steel fibres


Crack control
Flexure
Reinforced

1 Introduction Alkali Silica Reaction (ASR) represents one of the processes of degradation of concrete that generates most interest. Much progress has been made in terms of minerals and causes that produce ASR as well as in the criteria and methods of evaluation and prevention. Nevertheless, there is not much work on the development of ASR under load [1–6] and even less under tensile loads. However, considering that the presence of cracks substantially affects concrete permeability and water is essential for the ASR, it is possible to infer that the kinetics of the reaction should be related to the evolution of tensile cracking. © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 423–432, 2021. https://doi.org/10.1007/978-3-030-58482-5_39

424

R. L. Zerbino et al.

Fibre Reinforced Concrete (FRC) is a high-performance material with great possibilities for structural design as seen in the fib Model Code 2010 [7]. The incorporation of fibres into concrete controls the propagation of fissures and increases residual capacity and toughness. In addition to FRC traditional applications such as floor slabs or tunnel lining, the combined use of FRC y Reinforced Concrete (RC) structural elements improves shear resistance allowing the reduction of conventional steel bars and improves steel-concrete bond, then, reductions in anchorage length are possible. In addition, as fibers control crack width they give important benefits in terms of the extension of service life. Many papers confirm that the use of steel fibers reduces cracks width and spacing [8–10]. In a study on steel FRC elements loaded under uniaxial tension, increases in toughness and reductions in crack spacing were found [11]. Although when combined with conventional reinforcements, fibers do not significantly increase flexural strength and ductility in Ultimate Limit State [12] there are benefits in Service Limit State referred to the control of cracks and deflections, as they enhance the transfer of tensile stresses from the reinforcements to the concrete. The synergy between this mechanism and the FRC residual tensile capacity reduces the width and spacing of cracks. Finally, although the presence of fibres does not prevent the ASR, it reduces the expansions; mainly in the case of steel ones [13]. As a part of a research project on the advantages of using FRC for the extension of the service life and durability of the structures, this paper shows the effects of fibres on the development of ASR in RC beams subjected to long-term loads.

2 Experimental Program Four different types of concrete were used: two incorporating non-reactive aggregates and two with potentially reactive coarse aggregates, and, in each case, one without fibres and the other with steel fibres. Four RC beams, one for each concrete, were subjected to service loads. Twin beams were also cast and they remain unloaded and exposed to the same environmental conditions. 2.1

RC Beams

Beams of 150  150  900 mm including two 8 mm diameter steel bars as main reinforcement (q = 0.53%) and 6 mm diameter stirrups (spaced 50 mm) were cast. They were loaded using 4 PB configuration on a span of 840 mm; Fig. 1 shows the geometry and load configuration adopted. Tensile and compression deformations were continuously recorded by mechanical extensometers. The study was carried out in a temperature controlled chamber (23 ± 2 °C) and, to promote ASR all samples were covered with wet cotton cloths and then were isolated by a plastic film and remain in that condition during all the creep test; in addition, water was periodically injected (see Fig. 2). The beams were demoulded at 24 h and moist cured (as described) for 7 days. Then, they were removed from the bags, instrumented, wrapped again with humid cloths and plastic film and loaded by a lever system. Unloaded reference RC beams also remain in the same conditions. The RC beams

Effects of Fibres on the Flexural Behaviour 5 stirrups Ø6@5cm

5 stirrups Ø6@5cm 150

60

25

150

150

2Ø8 400

35 80

425

280

280

280

900

Fig. 1. Geometry and reinforcement details of the RC beams (in mm).

remain during 8 months under loading and then they were tested in bending up to failure using the same loading configuration.

Fig. 2. Load RC beams (left) and reference unloaded ones (right).

2.2

Concretes Mixture Design and Properties

Two plain concrete (R, P) with similar mixture proportions, w/c ratio 0.42; 380 kg/m3 of ordinary portland cement (Na2Oeq 0.73%), natural siliceous sand (fineness modulus 2.07) and granitic crushed stone of 19 mm maximum size were done. To promote ASR, 40% of the coarse aggregate was replaced by a very reactive quartzitic sandstone in the first of them (R) and NaOH was added in the water to achieve a total alkali content in concrete equal to 4 kg/m3. Two FRC, one reactive and the other non-reactive (FR, FP), were prepared adding 40 kg/m3 of low carbon hooked ended steel fibres to the same matrices (R, P). The slump was 170 ± 20 mm on concretes R and P and 90 ± 20 mm when fibres were incorporated. In addition to RC beams, three 70  70  300 mm prisms to measure the evolution of the length variations, six 100  200 mm cylinders to evaluate the compressive strength and the modulus of elasticity (only in R and P concretes) and six 75  105  430 mm prisms to characterize the flexural response were cast. All the specimens were compacted by external vibration and protected to prevent water evaporation. Both RC beams and the rest of the specimens were demoulded after 24 h, covered with wet cotton cloths and stored in plastic bags. Throughout the study the ambient temperature was maintained at 23 ± 2 °C.

426

R. L. Zerbino et al.

At 7 days RC beams were loaded in the frames as already described. Simultaneously, the compressive strength and the modulus of elasticity and the flexural behaviour of each concrete were evaluated on three cylinders and three notched prisms respectively. Bending tests were performed following the general guidelines of EN 14651 [14, 15]. The rest of the cylinders and prisms were tested at the end of the experience (near 9 months) in order to evaluate the impact of the ASR on the mechanical properties of the concrete.

3 Results 3.1

Plain and FRC Properties

Figure 3 shows the linear expansions measured on 70  70  300 mm prisms of each concrete. It can be seen that after near 3 months R and FR show expansions greater than 0.04%, while in the reference mixtures (P and FP) the dimensional changes did not exceed 0.01%. It can be seen that ASR expansions are smaller in FRC; although the incorporation of fibres cannot inhibit ASR, it leads to some benefits such as a decreases in expansions and, even more important, reductions in the width and length of cracks [16]. 0.15 FR 0.12

R

Expansion (%)

FP 0.09

P

0.06 0.03 0.00 -0.03 0

60

120

180

240

300

360

Time (days)

Fig. 3. Evolution of expansions (FR: reactive FRC; R; reactive concrete; FP: non-reactive FRC; P: non-reactive concrete).

Figure 4 shows typical stress-CMOD curves of each concrete at 7 days and 9 months. The effect of ASR damage on the mechanical behaviour of plain and FRC is clearly appreciated when compared the post-peak responses. The average results of compression and bending tests at both ages are given in Table 1. As it can be seen, no great differences in the mechanical properties at 7 days are present, indicating that ASR damage has not yet occurred at that age. On the contrary, at later ages there is a significant decrease in compressive strength and especially in the elastic modulus of concrete R.

Effects of Fibres on the Flexural Behaviour

427

Fig. 4. Stress – CMOD curves on notched prisms at 7 days (left) and 9 months (right) (FR: reactive FRC; R; reactive concrete; FP: non-reactive FRC; P: non-reactive concrete). Table 1. Mechanical properties of concretes (CV between brackets). Concrete Fibre content (kg/m3) Age f’c (MPa) 7 days 9 months E (GPa) 7 days 9 months fL [MPa] 7 days 9 months fR,1 [MPa] 7 days 9 months fR,3 [MPa] 7 days 9 months

R 0 42.5 (4) 32.2 (18) 38.5 (4) 19.3 (20) 5.5 (3) 4.4 (12)

FR 40

6.9 5.1 5.8 4.1 5.6 3.2

(3) (1) (6) (31) (4) (28)

P 0 44.5 (4) 47.8 (6) 40.2 (4) 38.5 (3) 5.8 (6) 6.2 (9)

FP 40

6.3 7.8 5.1 7.6 5.0 6.1

(4) (11) (17) (12) (12) (8)

Regarding the bending behaviour, reactive concretes (R, FR) evidence the internal damage in a lower initial slope (due to internal cracking) when compared to the one observed in concretes P and FP, a reduction in the first peak load and, in the case of plain concrete, an increase in the softening branch. In FR the first peak, the maximum and the residual loading capacity decrease, however, it conserves a residual stress near 4 MPa. The comparison against P and FP is difficult since the crack opening before testing in FR and R prisms is different than cero (i.e. samples are pre-cracked by ASR). 3.2

Instantaneous and Creep Behaviour of RC Beams Under Loading

Once placed in the frames, RC beams were loaded measuring tensile and compression deformations in the middle third. Considering that the ultimate estimated load in bending would be between 50 and 65 kN, 24 kN were applied enhancing the development of flexural cracks as in SLS condition. In every case, when the load exceeded 18 kN (near tensile stress 4 MPa) a clear deviation from the linear behaviour was observed during initial loading, indicating the presence of cracks (see Fig. 5).

428

R. L. Zerbino et al.

Stress (MPa)

6 4 Compression face

2 0 -500

-250

Tensile face

250

0

500

750

Strain (10-6)

Fig. 5. Example of instantaneous deformations of a beam during the loading process.

Figure 6 shows the evolution of deferred deformations measured on the tensile face of loaded and unloaded RC beams. It can be seen that the specimens incorporating reactive aggregates (R, FR) show a clear increase in the deformations; no great differences between specimens with and without fibres can be detected. Nevertheless, in the case of beams which are not subject to long-term loading (and therefore are not precracked), slight expansions (150 MPa). The structural performance was assessed on pre-cracked elements exposed to accelerated corrosion and tested to failure in comparison with control, non-exposed elements. Here are the most important findings:

Fiber Reinforced Concrete Elements Exposed

443

• Half-cell potential measurements showed more than 50% chances of corrosion, but the extracted tensile reinforcement showed neither significant signs of corrosion products, nor a significant mass loss. The steel fibres might have acted as anodes in the surrounding concrete taking up all the corrosion deterioration from the tensile bar. The exception might have been in the element containing a sacrificial zinc piece, which was found half-way destroyed, thus being the element taking up the deterioration. • A crack width of 0.1 mm was reached at a value of (0.64–0.84) of the maximum bending moment, which indicates the ductility limitations of low scale elements. However, there were no significant quantitative differentiation between the exposed elements and the control elements. • The values of the maximum experimental bending moment of all the elements were in the same range, with 0.62% to 13.97% less than the greatest value experienced by control element 13. • Furthermore, the design resistant bending moment was MRd = 3.66 kNm, which represents approximately 50% of the maximum bending moment for all elements, exposed and control. Although these findings are proof that the corrosion process did not affect the tested elements, more specimens are in order for viable conclusions and the research will be continued in this direction. Acknowledgements. This research was partially supported by the project 21 PFE in the frame of the programme PDI-PFE-CDI 2018.

References 1. Dong, Y.: Performance assessment and design of ultra-high performance concrete (UHPC) structures incorporating life-cycle cost and environmental impacts. Constr. Build. Mater. 167, 414–425 (2018) 2. Zhou, M., Lu, W., Song, J., Lee, G.C.: Application of Ultra-High Performance Concrete in bridge engineering. Constr. Build. Mat. 186, 1256–1267 (2018) 3. Wang, D., et al.: A review on ultra-high performance concrete: part II. hydration, microstructure and properties. Constr. Build. Mat. 96, 368–377 (2015) 4. Zhou, Z., Qiao, P.: Durability of ultra-high performance concrete in tension under cold weather conditions. Cement and Concr. Composites 94, 94–106 (2018) 5. An, M., Wang, Y., Yu, Z.: Damage mechanisms of ultra-high-performance concrete under freeze–thaw cycling in salt solution considering the effect of rehydration. Constr. Buil. Mater. 198, 546–552 (2019) 6. Fan, L., Meng, W., Teng, L., Khayat, K.H.: Effects of lightweight sand and steel fiber contents on the corrosion performance of steel rebar embedded in UHPC. Constr. Buil. Mater. 238, 117709 (2020)

444

C. Negrutiu et al.

7. Berrocal, C.G., Lundgren, K., Löfgren, I.: Corrosion of steel bars embedded in fibre reinforced concrete under chloride attack: state of the art. Cem. Concr. Res. 80, 69–85 (2016) 8. Hou, L., Liu, H., Xu, S., Zhuang, N., Chen, D.: Effect of corrosion on bond behaviors of rebar embedded in ultra-high toughness cementitious composite. Constr. Build. Mater. 138, 141–150 (2017) 9. Hou, L., Guo, S., Zhou, B., Chen, D., Aslani, F.: Constr. Build. Mater. 196, 185–194 (2019) 10. Nguyen, W., Duncan, J.F., Devine, T.M., Ostertag, C.P.: Electrochemical polarization and impedance of reinforced concrete and hybrid fiber-reinforced concrete under cracked matrix conditions. Electrochim. Acta 271, 319e336 (2018) 11. Azad, A.K., Ahmad, S., Azher, S.A.: Residual strength of corrosion – damaged reinforced concrete beams. ACI Mater. J. 104(1), 40–47 (2007). Title no. 104 – M05 12. Mokhtar, K.A., Loche, J.M., et al.: Report no. 2–2. Concrete in marine environment. MEDACHS, Interreg IIIB Atlantic Space – Project no. 197 (2006)

Effect of Corroded Steel Fibers on Mechanical Behavior of Steel Fiber Reinforced Concrete Minoru Kunieda1(&), Masaki Tsutsui2, and Le V. Tri2 1

Department of Civil Engineering, Gifu University, Gifu, Japan [email protected] 2 Graduate School of Natural Science and Engineering, Gifu University, Gifu, Japan

Abstract. This paper presents the exposure test results of Steel Fiber Reinforced Concrete (SFRC) beams having crack width of 0.2 mm for 3 months, 1 year and 2 years. By flexural loading test, initial stiffness of the exposed SFRC beams was decreased with increasing of exposure period. And corrosion of steel fibers themselves was observed at crack surface. The corrosion depth from the specimen surface was about 15–35 mm. There was no significant effect of fiber content on corrosion depth. Corrosion of the re-bar across the exposed crack was observed, and it may affect the reduction of stiffness of the beams due to loss of bond properties between re-bar and SFRC. Keywords: SFRC


Cracking
Corrosion of fiber
Durability

1 Introduction Fiber Reinforced Concrete (FRC) is used to improve mechanical properties of concrete structures such as shear capacity, ductility in addition to flexural capacity. In ordinary design concept of reinforced concrete structures, cracking in service period is allowed (e.g. allowable crack width less than 0.2 mm). Durability of FRC after cracking is not, however, well known. Various kinds of short fibers are proposed for FRC, steel fibers are familiar with practical applications. When Steel Fiber Reinforced Concrete (SFRC) is used after cracking, corrosion of steel fibers across a crack may be concerned [1], even if the crack width is less than allowable crack width in design. J. L. Granju [2] reported the test results on cracked SFRC samples with 0.5 mm crack mouth openings exposed to marine-like environment for 1 year. The results confirm the small sensitivity of SFRC to corrosion. The factors affecting the corrosion of the fibers and the reasons for the increase in flexural strength after corrosion are discussed. The researches on stainless steel and/or other fibers having higher resistance against corrosion have been conducted [3]. This paper presents the results of exposure test of cracked SFRC beams with re-bar, and flexural failure behavior of SFRC beams with corroded steel fibers was investigated.

© RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 445–452, 2021. https://doi.org/10.1007/978-3-030-58482-5_41

446

M. Kunieda et al.

2 Outline of Experiments 2.1

Specimen

Mix proportions of used SFRC are shown in Table 1. Water to cement ratio was 50%, and high early strength Portland cement was used. Volume fractions of fibers were 0.5, 1.0 and 1.5. The length and diameter of used steel fibers, which had hoock in both ends, were 30 mm and 0.62 mm, respectively. Compressive strength was measured by using cylindrical specimens having the size of d100 x h200 mm (JIS A 1108 Method of test for compressive strength of concrete), and flexural strength was measured by using rectangular specimens having the size of h100 x w100 x l400 mm (JIS A 1106 Method of test for flexural strength of concrete). Compressive strength and flexural strength at the age of 28 days are summarized in Table 2. The compressive strength of the specimen in Vf = 0.5% series was slightly lower than those of other cases. Figure 1 shows the tension softening curves of each series. The curves were obtained from the flexural tests was converted to tension softening curves by means of modified J integral method [4]. The fiber bridging stress was observed corresponding to fiber content. The difference of fiber bridging stress in Vf = 1.0% and Vf = 1.5% series was not, however, significant. As shown in Table 3, 4 specimens were prepared for each series having different fiber content (only Vf = 1.5% series was 3 specimens). Two of them were tested after 3 months after the initial loading. One specimen was tested after 1 year, and rest one specimen was also tested after 2 years. Table 1. Mixproportions of SFRC Vf (%) W/C (%) Unit content (kg/m3) Water Cement Sand 0.5 50 168 335 851 1.0 50 168 335 851 1.5 50 168 335 851

Gravel Fiber 913 39 898 79 888 118

Table 2. Compressive strength of SFRC (age: 28 days) Vf (%) 0.5 1.0 1.5

2.2

Compressive strength (N/mm2) 28.5 37.8 34.1

Flexural strength (N/mm2) 4.8 6.2 6.6

Initial Loading Test

At the age of 28 days, four-point bending tests were carried out to induce initial cracks in the SFRC beam. Figure 2 shows the image of loading test. Size of specimen was

Effect of Corroded Steel Fibers on Mechanical Behavior of SFRC

447

h200 x w100 x l1600 mm, and effective depth was 170 mm. Two deformed rebar having a diameter of 13 mm (SD345, fsy = 395 N/mm2) was used. Loading span was 1400 mm, load and displacement at loading points were measured. 2.3

Treatment of Occurred Cracks

5

4

4

Tensile stress (N/mm2

5

3 2 1 1

2 3 Crack width (mm)

4

5

3 2 1 0

1

(a) Vf =0.5%

2 3 Crack width (mm)

(b) Vf =1.0% 5

2

0

Tensile stress (N/mm

Tensile stress (N/mm2

Target crack (major crack) in the specimen after initial loading tests was selected for the exposure test. Figure 3 shows schematic image of the specimen with initial cracks, and photo of the treated specimen. Epoxy resin was painted to the specimen surface except for 50 mm around the target crack. The maximum crack width (residual crack width after unloading) at tension side was 0.2 mm approximately.

4 3 2 1 0

1

2 3 Crack width (mm)

4

(c) Vf=1.5% Fig. 1. Tension softening curves

5

4

5

448

M. Kunieda et al. Table 3. Number of the specimens in this test program Initial loading 3 months 1 year 2 years 2 1 1 Vf = 0.5% 4 Vf = 1.0% 4 2 1 1 Vf = 1.5% 3 2 1 0

Fig. 2. Test setup

2.4

Exposure Test and Re-Loading

The specimens with cracks were exposed at outside for 3 months, 1 year and 2 years, and loading tests were conducted. The exposure test was conducted in the campus of Gifu University (started from Nov. 2015). Note that, the specimens were affected by the climate condition includes rain and temperature. The specimen was arranged so that the side of the specimen is on the upper side. The NaCl solution (3.0%) was sprayed to the specimen with interval of once a week during 3 months.

Epoxy coating

Fig. 3. Schematic image of exposed specimen (cracks were coated except for a target crack)

Effect of Corroded Steel Fibers on Mechanical Behavior of SFRC

3 months (No.1) 3 months (No.2) 1 year 2 years 0

2

4 6 8 Disp. (mm)

70 60 50 40 30 20 10 0

Load (kN)

Load (kN)

70 60 50 40 30 20 10 0

3 months (No.1) 3 months (No.2) 1 year 2 years 0

10

449

2

4 6 8 Disp. (mm)

(a) Vf=0.5%

10

(b) Vf=1.0%

Load (kN)

70 60 50 40 30 20 10 0

3 months (No.1) 3 months (No.2) 1 year 0

2

4 6 8 Disp. (mm)

10

(c) Vf=1.5%

Fig. 4. Load-displacement curves

3 Experimental Results 3.1

Load-Displacement Relations

Figure 4 shows the load-displacement curves obtained from the re-loading. Note that the load displacement curves were displayed with the origin set to zero. The decrease of initial stiffness was observed at the age of 1 year and 2 years, comparing to that of 3 months. In addition, no significant difference was observed in initial stiffness between 1 year and 2 years. In the re-loading test, the target crack, which was selected as an exposed crack, was re-opened in all series. 3.2

Corrosion of Steel Fibers Across a Crack

Figure 5 shows exposed crack surface of FRC having fiber content of 1.0% at the age of 3 months. In addition, exposed crack surfaces at the age of 1 year and 2 years are also shown in Figs. 6, 7, 8 and 9. As shown in Fig. 5, there was no corrosion of steel fibers in the crack surface at the age of 3 months. In the case of exposure for 1 and 2 years, corrosion of steel fibers near specimen surface (depth about 15–35 mm) was observed. Note that the depth was measured by ruler directly. There was no significant effect of fiber content on the

450

M. Kunieda et al.

observed depth of corrosion. The corrosion of fibers was not, however, so severe as to cause the steel fibers to lose its cross section. It seems that this corrosion affects the decrease of initial stiffness in load-displacement curves.

Fig. 5. Fibers at crack surface (3 months, Vf = 1.0%)

Fig. 6. Observed corrosion of fibers at crack surface (1 year, Vf = 0.5%)

Fig. 7. Observed corrosion of fibers at crack surface (2 years, Vf = 0.5%)

Effect of Corroded Steel Fibers on Mechanical Behavior of SFRC

451

Fig. 8. Observed corrosion of fibers at crack surface (1 year, Vf = 1.0%)

Fig. 9. Observed corrosion of fibers at crack surface (2 years, Vf = 1.0%)

Figure 10 shows the photo of the re-bar across a crack. Corrosion on the surface of re-bar was slightly observed. Although loss of cross-sectional area of re-bar was not observed, and the corrosion may affect the decrease of initial stiffness in loaddisplacement curves too. Further researches on the contribution of re-bar corrosion to the decrease of the stiffness in load-displacement curves are needed.

Fig. 10. Observed corrosion of re-bar (2 years, Vf = 0.5%)

452

M. Kunieda et al.

4 Conclusions Experimental study of corroded Steel Fiber Reinforced Concrete (SFRC) beams subjected to flexural was carried out, and following results were obtained. • The exposure tests for SFRC beams having crack width of 0.2 mm was carried out for 3 months, 1 year and 2 years. By flexural loading test, initial stiffness of the exposed SFRC was decreased with increasing of exposure period. • Corrosion of steel fibers themselves was observed at crack surface. The corrosion depth from the surface was about 15–35 mm. There was no significant effect of fiber content on corrosion depth. • Corrosion of the re-bar across the exposed crack was observed. Although loss of cross-sectional area of re-bar was not observed and it may affect the reduction of stiffness of the beams due to loss of bond properties between re-bar and SFRC.

References 1. Kosa, K., Naaman, A.E.: Corrosion of steel fiber reinforced concrete. Mater. J. 87(1), 27–37 (1990) 2. Granju, J.L., Balouch, S.U.: Corrosion of steel fibre reinforced concrete from the cracks. Cem. Concr. Res. 35(3), 572–577 (2005) 3. Fu, X., Chung, D.D.L.: Bond strength and contact electrical resistivity between cement and stainless steel fiber: their correlation and dependence on fiber surface treatment and curing age. Mater. J. 94(3), 203–208 (1997) 4. Uchida, Y., Rokugo, K., Koyanagi, W.: Determination of tension softening diagrams of concrete by means of bending tests. J. JSCE 426, 203–212 (1991). (in Japanese)

Self-healing of Fibre Reinforced Concrete Containing an Expansive Agent in Different Exposure Conditions K.-S. Lauch, C. Desmettre, and J.-P. Charron(&) Department of Civil, Geological and Mining Engineering, Polytechnique Montreal, Montreal, Canada [email protected]

Abstract. While most studies about self-healing concrete investigated several healing conditions such as water immersion, humid chamber and wet/dry cycles, very few assessed the self-healing capacity of concrete in realistic outdoor condition. Furthermore, self-healing capacity is usually determined with a single testing procedure (mechanical or permeability measurements), sometimes with visual observations. Hence, this project aimed to evaluate, through mechanical and water permeability tests, as well as optical and microscopic observations on the same specimens, the self-healing capacity of high performances fibre reinforced concretes (HPFRC) containing different admixtures. This paper focuses on the water permeability evolution of two HPFRC (water to binder ratio of 0.43 and 0.75%-vol of steel macrofibres), one control and one containing calcium sulfoaluminate-based expansive agent (CSA), under different laboratory and outdoor conditions. Prisms were pre-cracked at 28 days by means of a 3-point bending test and then exposed to laboratory conditions (air, water immersion and wet/dry cycles) for 3 months and outside for one year. The results showed a better self-healing performance of the CSA HPFRC in wet/dry cycles compared to water condition. For the control HPFRC, the opposite is observed. Selfhealing of outdoor prisms is slower than in laboratory (6 months outside corresponds to around 2 weeks in water immersion for CSA). Self-healing performance of the CSA HPFRC is better than the control mix in outdoor condition. Keywords: Self-healing concrete Outdoor condition


Expansive agent
Water permeability


1 Introduction Most reinforced concrete structures prematurely reach the end of their service life because of gradual degradations caused by the environment. Twenty-five percent of Quebec Ministry of Transportation (QMT) bridges suffer from various disorders such as rebar corrosion and concrete cracking and require major repairs [1], bridge upgrade being of the order of $2634 M/year [2]. Concrete cracks, which are inevitable, accelerate the penetration of aggressive agents such as CO2 or chlorides and hence reduce the durability of the structures. As the durability and safety of structures are rising concerns, concretes with self-healing capabilities are a promising solution in line © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 453–465, 2021. https://doi.org/10.1007/978-3-030-58482-5_42

454

K.-S. Lauch et al.

with sustainable development. Such construction materials could not only increase the service life of bridges, but also cut down the huge costs of maintenance and repair, and reduce the traffic and disturbances around construction sites for the road users. Cementitious materials have a natural self-healing capacity, called autogenous healing. Hydration of anhydrous cement and carbonation are the main mechanisms involved in this autogenous capacity, being efficient to heal very small cracks (10 to 100 µm) [3]. In practice, cracks are often wider. The cracks width limits stand at 0.3 mm in the Canadian standard CSA A23.3 for reinforced concrete in outdoor environment and at 0.25 mm in the Canadian standard CSA S6 for highway bridges. To improve and better control this self-healing capacity, different techniques have been developed, such as the use of bacteria, encapsulated chemicals or mineral admixtures [4]. The latter has the advantages of being available products, less complex, more economical and having a long shelf life. Among mineral admixtures, crystalline admixtures [5] and calcium sulfoaluminate (CSA)-based expansive agents [6] or a combination of these two [7] have been investigated for their self-healing potential. Sisomphon et al. [8] found that the addition of CSA allow crack closing, even for crack widths of 0.25–0.40 mm. As numerous testing methods are used to evaluate the self-healing capacity, it is difficult yet to establish the optimal measurement techniques. Some studies consider optical measurements [7], while others run water permeability [9], bending tests, or non-destructive tests [5]. In most cases, each test is carried out on different specimens. Besides, self-healing depends on many parameters such as concrete composition, age of pre-cracking, crack width and the exposure condition. While most studied only the water immersion as a healing condition [10], Sisomphon et al. found that wet/dry cycles (12 h dry and 12 h wet) was the best condition for a mortar mix with CSA [7]. Furthermore, as pointed out by Li and Herbert [4], there is a lack in research of selfhealing under realistic outdoor conditions. While most studies showed promising results in controlled laboratory conditions, very few investigated the self-healing capacity under long-term outdoor conditions involving a wide and random range of temperatures and precipitations. Self-healing can occur in real condition and be efficient under multiple damage events, as demonstrated by Herbert and Li [11], Sherir et al. [12] and Cuenca et al. [13]. Although the number of papers about self-healing concrete is increasing exponentially, the research area is still relatively recent and not mature yet. This project has been launched to respond to some issues mentioned. The project goal is to evaluate the self-healing capacity of high performance fibre reinforced concretes (HPFRC) containing different admixtures (control, CSA, crystalline admixture, superabsorbent polymer), subjected to different conditions (laboratory and long-term outdoor exposure). The presence of fibres limits the crack widths, while the admixture promotes the healing of the cracks. The self-healing performance is assessed by mechanical tests, water permeability measurements, optical observations and microscopic analyses on the same specimens. This paper focuses only on the self-healing capacity of control and CSA HPFRC, exposed to different laboratory conditions (air, water immersion, wet/dry cycles) and outdoor exposure via water permeability tests.

Self-healing of Fibre Reinforced Concrete

455

2 Experimental Program 2.1

Materials

A high performance fibre reinforced concrete (HPFRC) with a water to binder ratio of 0.43 and containing a blended General Use Portland cement with 8% by mass of silica fume (type GUb-SF) was used. 0.75% volume of hooked-end steel macrofibres (lf = 35 mm and /f = 0.55 mm) was included to limit crack width of testing specimens. Two HPFRC mixes, one control without any admixture and one with an expansive agent (Denka Power CSA) at a dosage of 3.33% by mass of cement as recommended by the manufacturer, were produced. The compositions of the control and the CSA mixtures are summarized in Table 1. A first batch was made to produce specimens intended for the long-term outdoor exposure. For each mix, 12 cylindrical specimens (/100 mm  200 mm) for compressive strength characterization at different ages, 4 prisms (75 mm  125 mm  450 mm, Fig. 1) to be pre-cracked under flexure before healing and 2 reference uncracked prisms, were cast. Then a second batch was made for the specimens subjected to laboratory exposure conditions. 12 cylinders and 14 prisms were produced per mix. The prisms included 2 uncracked prisms per exposure condition. The specimens were demoulded at 24 h and then immersed in lime-saturated water for 28 days of curing. The compressive strength (fc) and the Young’s modulus (Ec) were determined at 28 days, 6 months and 12 months for outdoor specimens and at 28 days and 3 months for indoor specimens, in accordance with ASTM C39 and ASTM C469 respectively. Table 1. Compositions of the HPFRC mixtures Material Cement (kg/m3) CSA (kg/m3) Water (kg/m3) Superplasticizer (l/m3) Viscosity agent (l/m3) Sand (kg/m3) Coarse aggregates (kg/m3) Steel fibre (kg/m3)

2.2

Control 550 – 237 10 0.7 779 631 58.5

CSA 550 18.3 237 10 0.85 771 624 58.5

Methodology to Assess Self-healing

The prisms (Fig. 1) were first notched at mid-span (20 mm deep) and pre-cracked at the age of 28 days by means of a 3-point bending test according to EN 14651. The specimens were loaded up to a crack mouth opening displacement (CMOD) of 0.9 mm

456

K.-S. Lauch et al.

measured by an extensometer. After unloading, the residual CMOD was around 0.5 mm at the bottom of the notch, which corresponded to crack widths of around 0.1– 0.3 mm at the notch root (beginning of the actual crack), as measured by a digital microscope. These values fall into the Canadian standards crack width limits for exposed structures. After pre-cracking, the initial water permeability of prisms was measured. Because of the inherent variability of HPFRC properties, the characteristics of the flexural cracks varied widely in width, length, and tortuosity, which led to a large scatter of initial water permeability coefficients (Kwi). Therefore, prisms showing a fair dispersion of initial water permeability (i.e. prisms with high and low initial permeability) were attributed to each exposure conditions to compare objectively the impact of the exposure on the healing process. Prisms intended for outdoor conditions were kept in laboratory air condition another 28 days after pre-cracking to allow a partial drying before they were placed outside (winter period at that time). They were kept for one year in Montreal climate. Prisms intended for laboratory conditions were submitted 3 months either to water immersion, wet/dry cycles (3.5 days wet and 3.5 days dry) and air condition (only for the control mix). The air and drying conditions took place in a climatic chamber at 21 ± 3 °C and 45 ± 10% relative humidity. Water permeability Kw was then measured after 3, 6, 9 and 12 months for outdoor prisms and at 11 days, 25 days, 2 months and 3 months for indoor prisms. Before each water permeability measure, pictures of the crack were taken with a digital microscope at 4 equidistant points along the crack, on both prism faces (face 1 “front ! inlet” and face 2 “back ! outlet”, Fig. 1). These visual observations were only used to assess the self-healing qualitatively. Once the healing period terminated (3 months for indoor specimens and 1 year for outdoor ones), prisms were reloaded until failure with the same 3-point bending test as in the pre-cracking phase to assess the mechanical recovery provided by the healing products. The two uncracked prisms kept in the same conditions as the others were consecutively pre-cracked and loaded to failure as references. Following the reloading, some samples were sawn at healed cracks for scanning electron microscopy (SEM) and energy dispersive X-ray (EDX) analyses to identify the healing products. This paper

Aluminium box

Notch filled with elastomer

Face 1, inlet

Face 2, outlet Aluminium box

Flexural crack

Fig. 1. 3D diagram of the notched and cracked concrete specimen

Self-healing of Fibre Reinforced Concrete

457

focuses only on the water permeability measured on the control and CSA concrete mixtures, subjected to the laboratory and outdoor conditions. 2.3

Water Permeability Device

A new water permeability set-up was developed to measure the permeability through the pre-cracked prisms. This new set-up was inspired by a testing device previously developed at Polytechnique Montreal [14]. A significant benefit from this new water permeability set up is that it allows to measure transport properties on a prism than can then be reloaded, enabling to investigate correlations between a self-healing state and mechanical recovery. The set-up consists of two water-filled aluminium boxes clamped on the two opposite cracked lateral faces of the prism (Fig. 2). The boxes are linked to inlet and outlet tanks. Water in the inlet tank is put under pressure, while the outlet tanks remains at the atmospheric pressure to subject the specimen to a pressure gradient of 30 kPa (corresponding to a 3-m water depth). This pressure gradient initiates a water flow through the cracked specimen from the inlet to the outlet tank. The incoming and outgoing water flows through the prism are recorded. The permeability coefficient Kw (m/s) is calculated after 5 min, insuring that the inlet and outlet flows are equal, using Darcy’s law (Eq. (1)), with Q (m3/s) the water flow, L (m) the flow path length, A (m2) the specimen cross-section and Dh (m) the drop in the hydraulic head across the specimen. Kw ¼

QL ADh

ð1Þ

Patm

Pressure regulator Water flow

Pressure transmitter

Clamping system

Aluminium boxes

Notched and cracked concrete specimen Notch filled with elastomer

Outlet tank

Inlet tank

Fig. 2. New water permeability device

458

K.-S. Lauch et al.

3 Results 3.1

Characterization Tests

All concrete batches (for outdoor and indoor exposures) and mixes (control and CSA) reached similar slump flows of 550 ± 40 mm and Young’s modulus of 33 000 ± 1000 MPa. The evolution of the compressive strength fc (average and standard deviation) is shown in Fig. 3. At a same time, fc was quite similar for the control and CSA mixes. Furthermore, there were no significant differences regardless of the exposure condition. Thus, the fresh state properties and compressive strength of the HPFRC mixtures only had a limited impact on the self-healing capacity measured. 90 Compressive strength (Mpa)

Control

CSA

80 70

65.2 66.1

69.5

72.8

71.1 73.1

74.50 66.9 57.3

60

82.0

63.7

50 40 30 20 10 0 28 days (indoor)

119 days (water)

119 days (wet/dry)

119 days (air)

28 days (outdoor)

247 days (outdoor)

Fig. 3. Results of compressive strength at various times.

3.2

Water Permeability Measurements

3.2.1 Prisms in Controlled Laboratory Conditions The average results and standard deviations of water permeability coefficients Kw measured at different exposure times are shown in Fig. 4 and Fig. 5 for the control and CSA HPFRC respectively. These values came from 4 to 5 specimens for the water and wet/dry exposures and from 2 specimens for the air condition. In these figures, the values obtained for the water and wet/dry conditions have been slightly shifted by 0.5 and 1 day respectively for clearer display purpose. The healing ratios (HR), calculated for each specimen by Eq. (2) with Kwt the permeability coefficient at a time t and Kwi the initial permeability coefficient, are averaged and illustrated in Fig. 6. HR ¼

  Kwt 1 :100 Kwi

ð2Þ

Self-healing of Fibre Reinforced Concrete

459

Permeability coefficient Kw (m/s)

8.E-06 Control water

7.E-06

Control wet/dry

6.E-06

Control air

5.E-06 4.E-06 3.E-06 2.E-06 1.E-06 0.E+00 0

10

20

30

40 50 Exposure time (days)

60

70

80

90

Fig. 4. Evolution of permeability coefficient for the control HPFRC.

Permeability coefficient Kw (m/s)

8.E-06 CSA water

7.E-06

CSA wet/dry

6.E-06 5.E-06 4.E-06 3.E-06 2.E-06 1.E-06 0.E+00 0

10

20

30

40 50 60 Exposure time (days)

70

80

90

Fig. 5. Evolution of permeability coefficient for the CSA HPFRC.

The control specimens presented average initial permeabilities of 2  10−6 to 3  10−6 m/s. At 11 days, Kw decreased rapidly to 1.5  10−7, 3.6  10−7 and 1.8  10−6 m/s for water, wet/dry cycles and air exposures respectively (Fig. 4), which represents average self-healing ratios (HR) of 95%, 87% and 12% (Fig. 6). Then the decreasing rate slowed down and Kw stabilized after 2 months to reach average values of 5.3  10−8, 1.1  10−7 and 1.2  10−6 m/s after 3 months in water, wet/dry cycles and air respectively, which represents HR of 98%, 95% and 46%. Self-healing of the control HPFRC is greater in water than in wet/dry condition and, as expected, lower in air exposure. For CSA HPFRC, the self-healing kinetics was similar (fast self-healing rate during the first 11 days and then the rate slowed down). Their average initial permeabilities was around 2.8  10−6 m/s. At 11 days, Kw decreased to 1  10−6 and 1.6  10−7 m/s

460

K.-S. Lauch et al. 100% 11 days

90%

25 days 2 months

80%

3 months

Healing ratio

70% 60% 50% 40% 30% 20% 10% 0% Control water

CSA water

Control wet/dry

CSA wet/dry

Control air

Fig. 6. Evolution of healing ratios for control and CSA HPFRC for different exposure conditions.

for the water and wet/dry cycles exposures respectively (Fig. 5), which represents average HR of 62% and 87% (Fig. 6). Then Kw dropped to 5.4  10−7 and 9.1  10−9 m/s respectively for the water and wet/dry cycles exposures, which represents HR of 86% and 99%. Unlike the control HPFRC, self-healing of the CSA HPFRC is superior in wet/dry than in water condition. The standard deviations for the initial permeability measured from one specimen to another for the control and CSA mixes are addressed in the Discussion section below. Figure 6 also compares the average healing ratios and standard deviations of the HPFRC mixes when submitted to a same exposure condition. When kept in water, the control mix showed higher HR at all times than the CSA mix. After 3 months, the control HPFRC had an average HR of 98% against 86% for the CSA HPFRC. In wet/dry condition, both mixes showed an average HR of 87% at 11 days, then the CSA mix healed better than the control mix (HR of 99% against 95% after 3 months). In air condition, the HR of the control specimens were much lower than the other conditions, but they increased with time (HR from 12% at 11 days to 46% after 3 months). 3.2.2 Prisms in Realistic Outdoor Conditions Figure 7 shows the average permeability coefficients and their standard deviation at different times (0, 3, 6 and 9 months) of the prisms exposed to outdoor conditions. The results of the prisms kept in indoor laboratory conditions (water immersion), described in the previous section, are also shown in Fig. 7 for comparison purpose. A different time scale was used to compare the self-healing kinetics between the indoor and outdoor prisms, in months for outdoor prisms and in days for indoor prisms. As for the results of the indoor prisms, the Kw values of the control and CSA outdoor prisms have been slightly shifted in the x-axis for clearer display purpose. It should be noted that the initial water permeability coefficients Kwi of the outdoor specimens varied from the indoor specimens. The difference in Kwi for the control and CSA outdoor concretes was also large (Kwi for the control mix was almost 4 times

Self-healing of Fibre Reinforced Concrete

0

Time exposure for outdoor specimens (months) 4 6 8

2

10

461

12

7.E-06

Permeability coefficient Kw (m/s)

Control water 6.E-06

CSA water Control outdoor

5.E-06

CSA outdoor

4.E-06 3.E-06 2.E-06 1.E-06 0.E+00 0

10

20

30 40 50 60 Time exposure for indoor specimens (days)

70

80

90

Fig. 7. Evolution of permeability coefficient for prisms exposed to outdoor and indoor conditions.

higher than the CSA mix). This is because a second batch of the control and CSA concretes had to be prepared for indoor specimens and some discrepancies were obtained in the permeabilities between the two batches. The average water permeability coefficients of the 4 prisms exposed to outdoor conditions decreased from initial values of 5.2  10−6 and 1.3  10−6 m/s to 2.2  10−6 and 4.1  10−7 m/s at 9 months for the control and CSA HPFRC respectively. These correspond to average healing ratios of 59% (control) and 81% (CSA) after 9 months. As observed for the indoor specimens, the self-healing kinetics (decrease of Kw) of the outdoor specimens was faster at the beginning and then slowed down with time until stabilization. Table 2 shows the evolution of healing ratios for indoor and outdoor prisms. The self-healing of the outdoor control prisms after 3 months was equal to 46%. The same HR was obtained in less than 11 days for the indoor specimens exposed to water or wet/dry cycles and after 3 months for the ones subjected to air condition (with a relative humidity of 45%). The self-healing of the outdoor CSA specimens after 6 months was 66%, which is close to the HR of 62% obtained after 11 days of water immersion. After 9 months, HR for the CSA outdoor prisms reached 81%, which is close to the HR of 85% obtained after 25 days of water immersion or to the HR of 87% obtained after 11 days of wet/dry cycles. In conclusion, the self-healing obtained in 2 to 4 weeks in laboratory conditions (water or wet/dry cycles) was accelerated compared to the effective self-healing process in outdoor prisms for which 6 to 9 months were necessary to reach the same self-healing level.

462

K.-S. Lauch et al.

Table 2. Evolution of healing ratios for prisms exposed to outdoor and indoor conditions. Indoor Control water Control wet/dry Control air CSA water CSA wet/dry Outdoor Control outdoor CSA outdoor

11 days 95% 87% 12% 62% 87% 3 months 46% 51%

25 days 98% 85% 10% 85% 98% 6 months 50% 66%

2 months 98% 93% 35% 86% 99% 9 months 59% 81%

3 months 98% 95% 46% 86% 99 12 months

4 Discussion As mentioned in the methodology section, initial permeability coefficient (Kwi) of prisms varied due to the inherent variability of the crack pattern (width, length, tortuosity) within HPFRC specimens. The impact of the initial permeability coefficient was reduced as much as possible in the research program by distributing fairly prisms with high and low permeability to each exposure conditions, thus the average Kwi of prisms for different exposure condition was similar. Then, the healing ratio was evaluated on each specimen and averaged for prisms of the same condition. This methodology should limit the impact of the Kwi variability of prims in the research program and does not modify the general trends observed from one exposure condition to another. Whether for the indoor or outdoor prisms, the self-healing kinetics is globally similar with a first phase of fast decrease of the water permeability (during around 11 days for the indoor and around 3 months for the outdoor specimens) followed by a second phase where permeability slowed down and stabilized over time. For example, in the first 11 days of exposure of prisms in laboratory conditions, water permeability coefficients (Kw) dropped by 95% and 87% respectively for water and wet/dry conditions for the control HPFRC mix. Afterwards and until 3 months of exposure, permeability decreased by 23% (water) and 43% (wet/dry). Such self-healing kinetics was already reported in literature [15, 16] and is explained by the calcium carbonate precipitation mechanism. First, precipitation is rapid as calcium ions Ca++ are readily available at the concrete surface. Then it is controlled by the diffusion of the calcium ions through the concrete matrix, which is a slower process. Another trend observed is that the variability of the initial permeability coefficients (Kwi) of prisms (due to the variability of the cracks pattern) decreased sharply with the evolution of the healing process, as shown with the errors bars in Fig. 4 and Fig. 5. This trend is logical as all the specimens converged over time to the same completely healed state. This can also be seen in Fig. 6 where the magnitude of error bars of the healing ratios decrease slightly with time, except for the control specimens in air condition. The control specimens immersed in water presented low variability of the healing ratios, in contrast to the large variability of the specimens in wet/dry and air

Self-healing of Fibre Reinforced Concrete

463

conditions, which can be related to a more efficient healing in water. For the CSA prisms, the opposite trend is observed as self-healing proceeds: high magnitude of error bars is noted in water immersion and low magnitude in wet/dry condition, which can be related to a more efficient healing in wet/dry condition. These trends are coherent with the decreasing variability of permeability observed for the outdoor CSA prisms but not for the control prisms (Fig. 7). This high variability for the outdoor control prisms may also be explained by the still high permeability of these specimens after 9 months. According to these previous observations, the control HPFRC prisms performed better than the CSA HPFRC prisms in water condition, while it is the opposite when exposed to wet/dry cycles. However, Sisomphon et al. [8] found a greater permeability decrease of mortars specimens immersed in water made with CSA in comparison with a control mortar (water to cement ratio of 0.25, pre-cracked at 28 days). When different exposure conditions were tested, they found however that wet/dry cycles was the best exposure for mechanical regain, for CSA and for control specimens [7]. The better performance of the control HPFRC in water in this paper, in comparison to the study of Sisomphon et al., could be explained by the silica fume (SF) content of HPFRC. It has been shown that a mortar with SF showed outstanding self-crack closing ability [10]. Another reason would be the high alkalinity of the water. As lime is added in containers used for water and wet/dry conditions in laboratory, the water has a high pH fostering healing [6]. The results of the SEM and EDS analyses planned in the project will provide a better understanding of the different self-healing behaviours of the control and CSA HPFRC mixes subjected to different exposure conditions. Concerning the outdoor prisms, this paper showed a high healing ratio (81%) for the CSA HPFRC mix at 9 months compared to the control HPFRC mix (59%). This would suggest that the CSA specimens show a better self-healing capacity in real environment than the control ones. This observation is in line with the results found earlier for indoor prisms as CSA HPFRC showed the best performance in wet/dry cycles, which better replicate real climate exposure. Herbert and Li [11] and Sherir and al. [12] also confirmed with Engineered Cementitious Composites (ECC) the selfhealing capability of specimens in realistic environment. They assessed the mechanical recovery and Resonant Frequency, thus no correlation can be established with permeability measurements shown in this paper. Cuenca et al. [13] investigated the selfhealing capacity of FRC with crystalline admixture (CA). They calculated a sealing index (SI) based on image analysis of crack surfaces and found low SI for open air exposure for both CA and control specimens (SI around 5 to 30% at 6 months). Sealing of crack widths at specimen’s surfaces may not provide the same trend for self-healing as permeability measurement represents the healing of the interior of the crack. Thus results comparison with this study is not obvious.

5 Conclusions The self-healing capacities of two high performance fibre reinforced concretes (HPFRC), one control and one containing a calcium sulfoaluminate-based (CSA) expansive agent, subjected to different indoor laboratory conditions (ambient air, water immersion and wet/dry cycles), and to a long-term outdoor condition (Montreal climate

464

K.-S. Lauch et al.

for 1 year) have been investigated and compared. Self-healing was assessed by means of water permeability measurements. Based on the results of this experimental program, the following conclusions can be drawn: • Self-healing capacity depends greatly on the mixture composition and the exposure condition. While control HPFRC healed better in water immersion, CSA HPFRC achieved greater healing in wet/dry cycles. • The high healing ratios of the CSA HPFRC were also observed in realistic environment. Thus, HPFRC with a CSA-based expansive agent seems to have a strong self-healing capacity in outdoor conditions. • The self-healing kinetics was slower for the outdoor exposure condition than for the water and wet/dry laboratory exposures. The self-healing state of the CSA HPFRC exposed 9 months in outside conditions corresponds approximatively to 25 days in water or less than 11 days in wet/dry condition. This paper only focused on the water permeability measurements carried out on prisms. At the end of their healing conditions, all specimens will be reloaded to assess their mechanical recovery and healing products will be identified with SEM/EDX analysis. These results will be presented in future papers. Acknowledgements. The authors would like to thank the Québec Research Fund on Nature and Technologies (FRQNT) for the financial support and Denka for providing the admixture.

References 1. Quebec Ministry of Transport: General condition index. In: QMT statistics (2019) 2. Quebec Ministry of Transport: Quebec infrastructure plan PQI 2011–2016 (2013) 3. De Belie, N., et al.: A review of self-healing concrete for damage management of structures. Adv. Mater. Interfaces, 28 (2018) 4. Li, V., Herbert, E.: Robust self-healing concrete for sustainable infrastructure. J. Adv. Concr. Technol. 10, 207–218 (2012) 5. Ferrara, L., Krelani, V., Carsana, M.: A “fracture testing” based approach to assess crack healing of concrete with and without crystalline admixtures. Constr. Build. Mater. 68, 535– 551 (2014) 6. Jiang, Z., Li, W., Yuan, Z.: Influence of mineral additives and environmental conditions on the self-healing capabilities of cementitious materials. Cem. Concr. Res. 57, 116–127 (2015) 7. Sisomphon, K., Copuroglu, O., Koenders, E.A.B.: Effect of exposure conditions on self healing behavior of strain hardening cementitious composites incorporating various cementitious materials. Constr. Build. Mater. 42, 217–224 (2013) 8. Sisomphon, K., Copuroglu, O., Koenders, E.A.B.: Self-healing of surface cracks in mortars with expansive additive and crystalline additive. Cem. Concr. Compos. 34, 566–574 (2012) 9. Roig-Flores, M., Moscato, S., Serna, P., Ferrara, L.: Self-healing capability of concrete with crystalline admixtures in different environments. Constr. Build. Mater. 86, 1–11 (2015) 10. Jaroenratanapirom, D., Sahamitmongkol, R.: Self-crack closing ability of mortar with different additives. J. Met. Mater. Miner. 21(1), 9–17 (2011) 11. Herbert, E., Li, V.: Self-healing of microcracks in Engineered Cementitious Composites (ECC) under a natural environment. Materials 6, 2831–2845 (2013)

Self-healing of Fibre Reinforced Concrete

465

12. Sherir, M.A.A., Hossain, K.M.A., Lachemi, M.: Development and recovery of mechanical properties of self-healing cementitious composites with MgO expansive agent. Constr. Build. Mater. 148, 789–810 (2017) 13. Cuenca, E.T., Tejedor, A.; Ferrara, L.: A methodology to assess crack-sealing effectiveness of crystalline admixtures under repeated cracking-healing cycles. Constr. Build. Mater. 179, 619–632 (2018) 14. Desmettre, C., Charron, J.-P.: Novel water permeability device for reinforced concrete under load. Mater. Struct. 44, 1713–1723 (2011) 15. Escoffres, P., Desmettre, C., Charron, J.-P.: Effect of a crystalline admixture on the selfhealing capability of high-performance fiber reinforced concretes in service conditions. Constr. Build. Mater. 173, 763–774 (2018) 16. Homma, D., Mihashi, H., Nishiwaki, T.: Self-healing capability of fibre reinforced cementitious composites. J. Adv. Concr. Technol. 7(2), 217–228 (2009)

Characterisation of Strain-Hardening Cementitious Composite (SHCC) Under Cyclic Loading Conditions for Self-healing Applications Zixuan Tang(&), Chrysoula Litina, and Abir Al-Tabbaa Department of Engineering, University of Cambridge, Cambridge, UK [email protected]

Abstract. The ground shaking in an earthquake often imposes cyclic loadings on infrastructures placing them in danger. Concrete is quasi-brittle in tension and easy to crack under cyclic loadings. Fibre reinforced strain-hardening cementitious composites (SHCC) featuring high ductility, high energy absorbing capacity and controlled multiple cracking have been proposed for seismicresistant applications. The fine cracks have been proved to not only improve the durability but also enhance the autogenous self-healing ability. This study focuses on investigating the material behaviour and self-healing potential of SHCC under cyclic flexural loading conditions. Four-point flexural tests (including monotonic and cyclic tests) were performed to study its mechanical properties and cracking behaviour. The surface crack widths were measured by optical microscopy. Results showed that 28-day air cured specimens exhibited a deflection capacity of up to 9.6 mm and an average crack width of 28 lm under monotonic flexural loading. Regarding the flexural stress-deflection curves, the envelops of cyclic testing results were close to monotonic results with both elastic and hardening phases. SHCC could still maintain fine crack widths (below 60 lm) under cyclic loading conditions. SEM/EDX test was conducted to investigate the fibre-matrix interface. Keywords: Strain-hardening loading


Self-healing
Multiple cracking
Cyclic

1 Introduction During an earthquake, the ground shaking initiated by seismic waves often applies cyclic forces to infrastructures and threatens the integrity and stability of constructions. Concrete is inherently brittle and is extremely weak under tensile or cyclic shear loadings. Although shear reinforcement is often applied to avoid brittle failure and enhance shear strength, large cracks will still occur affecting the durability of the reinforced concrete section. Traditional fibre reinforced concrete (FRC) is still quasibrittle with low ductility under tensile stress, and the crack widths are not well controlled. Therefore, a more durable material featuring high ductility and energy

© RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 466–476, 2021. https://doi.org/10.1007/978-3-030-58482-5_43

Characterisation of SHCC Under Cyclic Loading Conditions

467

absorption capacity to bear cyclic loadings without brittle failure is required for seismic resistance. Moreover, substantial and timely repair work is usually required after a large earthquake due to unavoidable damage, which is often costly and difficult. Cementitious materials inherently have the ability to heal small cracks and regain transport and mechanical properties mainly due to continuing hydration, carbonation and pozzolanic reactions; called autogenous healing [1]. However, this mechanism can only heal very small cracks (mostly < 100 lm) depending on 1) intrinsic features such as age and mix design of the cement-based material, crack widths and depths; 2) external healing conditions such as environmental exposure (e.g. availability of water and CO2 surrounding cracks, thermal conditions), stress state and steadiness of the cracked state (mechanical loadings) and healing duration [2, 3]. Hence, it is beneficial to develop a ductile material with a more consistent selfhealing ability against seismic damage or cyclic loadings. Fibre-reinforced strainhardening cementitious composites (SHCC), represented by Engineered Cementitious Composites (ECC) can be a good candidate [4]. Designed based on micromechanics models and using special tailoring techniques, SHCC incorporates a low fibre content (typically  2%) and relies on a synergistic interaction between the fibres, matrix and the interface to realize a tensile strain-hardening effect. SHCC exhibits a multiple cracking behaviour under tension due to effective fibre bridging and load transfer mechanisms. Accordingly, SHCC could achieve much higher strain capacity or deformation ability than traditional FRC under tension or flexural stress. The high ductility and fracture energy have made SHCC a promising seismic-resistant material [5]. Moreover, the cracking control capacity of SHCC has also been proved to effectively enhance the durability and promote autogenous self-healing of the cementitious material in different environments [6–9]. This research aims at characterising and studying the performance of SHCC under cyclic loading conditions and evaluating its potential for self-healing. Based on existing studies on SHCC, M45-ECC [10] is one of the most investigated mixtures and has exhibited overall good mechanical performance, featuring a high tensile strain capacity (>3%), moderate tensile strength and fine crack opening (crack widths < 100 lm) [6]. In this study, this standard mix was selected to assess the mechanical behaviour and crack patterns of SHCC under monotonic and cyclic flexural loading. Results of this preliminary investigation will inform further self-healing studies.

2 Materials and Mix Proportions 2.1

Materials

CEM-I 52.5 N High strength Portland cement (C) supplied by Hanson UK was used as the clinker source. Silica sand (S) with a maximum grain size of 250 lm was adopted as fine aggregates. Fly ash (FA) with a fineness category of N according to BS EN 4501 was supplied by CEMEX UK as a supplementary cementitious material. No coarse aggregates were adopted in order to maintain the micromechanical properties between the matrix, fibres and interfaces. A polycarboxylate-based superplasticiser (SP) with a

468

Z. Tang et al.

feature of enhancing workability retention and a viscosity modifying admixture (VMA) were incorporated to control the fresh properties of SHCC and to achieve uniform fibre distribution, both of which were supplied by Sika UK. Polyvinyl alcohol (PVA) fibres produced by Kuraray, Japan were coated with oil (1.2% by weight) to reduce the excessively strong interfacial bonding with the matrix. The properties of the PVA fibres are provided in Table 1. Table 1. Properties of PVA fibres. Density (g/cm3) 1.3

2.2

Tensile strength (GPa) 1.56

Elastic modulus (GPa) 41

Elongation (%) 6.5

Diameter (lm) 40

Length (mm) 12

Mixture Proportions

The mixture proportions were similar to [11]. The ratio of water to binder (W/B, or W/(C + FA)), was set as 0.26, the sand to binder ratio (S/B) was fixed at 0.36, and FA/C was 1.2. The fibres were added as 2% of the total mixture volume. The mix proportions are shown in Table 2.

Table 2. The mix proportions of SHCC investigated. Ingredients

Mass ratio Unit weight (kg/m3)

2.3

Cementitious materials/binders C FA 1 1.2 542 650

Aggregates

Chemical admixtures

S 0.8 434

SP 1.69% 9.18

W 0.58 314

VMA 0.49% 2.67

Fibre 4.80% 26

Specimen Preparation and Curing

Cement, sand and fly ash were first dry mixed in a Hobart mixer, after which a premixed water solution containing SP and VMA was added and uniformly mixed for 3.5 min. Then a flow table test according to BS EN 12350-5 was performed to quantify and check the workability. The flow spread was 290 ± 10 mm. At last, fibres were added and mixed for 3 min. The fresh SHCC mixture was then cast into moulds. The specimens included 40  40  40 mm3 cubes and Ø75  150 mm3 cylinders for compressive tests, and 300 (length)  50 (width)  26 (height) mm3 prisms for fourpoint flexural tests. Two curing environments were selected; high moisture and ambient. For the former the specimens were stored in a high moisture container at 98 ± 2% relative humidity (RH) and 20 ± 2 °C until the required age was reached. Conversely for the latter, the specimens were placed in ambient conditions with 50 ± 5% RH and a temperature of 20 ± 2 °C.

Characterisation of SHCC Under Cyclic Loading Conditions

469

3 Methods 3.1

Physical and Mechanical Properties

3.1.1 Unconfined Compressive Strength and Density Unconfined compressive strength tests were conducted on cubic specimens (40  40  40 mm3) of different ages. Triplicate specimens were tested in a 250kN testing cell with a vertical loading speed of 2400 N/s according to BS EN 196-1. The density of SHCC was calculated by measuring and averaging the weights and volumes of 24 cubic specimens. 3.1.2 Secant Modulus of Elasticity Secant modulus of elasticity in compression was also tested based on BS EN 12390-13. Three cylindrical specimens with a dimension of Ø75  150 mm3 were cured in a high moisture condition for 7 days before testing. One day before testing, the top surfaces of specimens were capped to make the surface smooth and flat. Using a 2000 kN testing frame (Controls), two specimens were loaded till failure to achieve an average compressive strength fc. Based on fc, a loading profile containing three preloading cycles and three loading cycles within the elastic range were designed, and then another specimen from the same batch was loaded to obtain the secant modulus of elasticity. 3.1.3 Four-Point Flexural Tests (Monotonic and Cyclic) Prisms with a dimension of 300 (length)  50 (width)  26 (height) mm3 were loaded using a 30 kN Instron machine under displacement control. The loading span was 80 mm and the supporting span was 240 mm. Three types of tests, i.e. monotonic loading tests, one-side loading/unloading cyclic tests and reversed cyclic tests were conducted. For monotonic loading tests, the loading rate regarding the crosshead displacement (u) was 0.2 mm/min and all the specimens were loaded until failure. For cyclic loading tests, the loading speed was a combination of 0.2 and 0.5 mm/min considering the testing efficiency. The midspan deflection (v) of each specimen was monitored by using a laser extensometer. 3.2

Microstructural Observation

3.2.1 Optical Microscopy A Leica DM 2700 M stereoscope was used for measuring all the crack widths generated during the four-point flexural tests. Since flexural cracks are often in a wedge shape, the maximum width of each crack was measured from the bottom surface of the specimens. For specimens with a total crack number of no more than ten, each individual crack was identified and nine evenly-spaced measuring points were positioned along each crack. For those with a crack number greater than ten, since individual cracks were difficult to classify and one crack often connected or branched to another, three parallel lines along the specimen span were drawn (Fig. 1) and all the crack widths crossed by the lines were measured and averaged.

470

Z. Tang et al.

80mm Line 1 Line 2 Line 3

Fig. 1. Multiple cracking pattern of a 28-day air-cured SHCC specimen and the crack width measuring method.

3.2.2 SEM-EDX Analysis Scanning electron microscopy with an energy-dispersive X-ray spectroscopy detector (SEM-EDX) was applied to observe the fibre-matrix interfaces, specifically the fibre bridging behaviour and the substance on the fibre surface. Phenom ProSuite software was used for EDX analysis. 0.5  0.5  0.5 mm3 specimens were cut from large cracked SHCC specimens and completely dried by storing in an oven with a constant temperature of 40 °C for one day before testing.

4 Results and Discussion 4.1

Physical and Mechanical Properties

4.1.1

Density, Unconfined Compressive Strength and Secant Modulus of Elasticity Results indicated that the development of compressive strength was not significantly affected by the curing conditions after seven days of high moisture curing. In particular, under consistent high moisture condition, the compressive strength was 37.52 (±0.83) MPa at 3 days, 50.33 (±2.71) MPa at 7 days and 75.42 (±2.00) MPa at 28 days, whereas for specimens cured in high moisture for 7 days and then cured in ambient conditions (RH = 50 ± 5%) until 28 days the corresponding strength was 71.55 (±0.71) MPa at 28 days. The average density of SHCC specimens was 2050 kg/m3. Moreover, the 7-day modulus of elasticity was obtained. Two specimens were first tested to obtain the compressive strength, and then by loading the third specimen, the initial modulus of elasticity (13.3 (±2.1) GPa) and the stabilised modulus of elasticity (14.0 (±0.7) GPa) were calculated. 4.1.2 Behaviour Under Monotonic Flexural Loading 7-day old prisms were tested under monotonic flexural loading until failure, which all exhibited deflection-hardening behaviour. The flexural stress-deflection curves are presented in Fig. 2(a). After the initial crack, the stress continued to rise with some fluctuations while the midspan deflection of SHCC increased significantly and multiple fine cracks were generated. According to visual observation, as the deformation

Characterisation of SHCC Under Cyclic Loading Conditions

471

increased, the crack number increased until a localised crack occurred. Then the width of the localised crack expanded until the fibres could not sustain the tensile stress, and finally the specimen softened with a continuous decreasing stress. Representative salient points are summarised in Table 3. The modulus of rupture (MOR) is defined as the flexural stress just before softening, and the deflection capacity is the corresponding midspan deflection. The flexural stiffness is defined as the slope of flexural stressdeflection curve when the flexural stress is between 1 and 3 MPa, where the slope is almost linear [12]. Table 3. Characteristics of 7-day SHCC under monotonic flexural loading. First cracking strength (MPa) 4.46 (±0.40)

First cracking deflection (mm) 0.26 (±0.04)

MOR (MPa) 7.90 (±0.91)

Deflection capacity (mm) 3.79 (±0.44)

Flexural stiffness (MPa/mm) 19.13 (±1.41)

28-day cured specimens were also tested under monotonic flexural loading, and the curing conditions included both high moisture and ambient curing. Under the former curing condition, the 28-day specimens exhibited higher flexural strength than the 7day cured specimens due to the longer hydration duration. For the 28-day cured specimens, the air-cured specimens showed significantly higher deflection capacity (up to 9.6 mm), slightly lower flexural strength (11.7 MPa) and lower stiffness than the moisture-cured specimens (5.5 mm and 13.2 MPa respectively). The explanation could be that the matrix tended to develop a higher fracture toughness and was more difficult to crack after high moisture curing; accordingly, based on micromechanical analysis, the crack tip toughness would increase while deflection-hardening behaviour and flexural ductility would decrease. Typical flexural stress-deflection curves under different curing conditions are given in Fig. 2(b).

14 12

8

10

6 4 Specimen No.1 Specimen No.2 Specimen No.3

2 0

Flexural stress (MPa)

Flexural stress (MPa)

10

0

1

2

3

Midspan deflection (mm)

(a)

4

5

8 6 4

28d high moisture 28d air 7d high moisture

2 0

0

2

4

6

8

Midspan deflection (mm)

10

12

(b)

Fig. 2. Flexural stress-deflection curves of (a) 7-day SHCC (high moisture curing); (b) SHCC under different curing conditions or durations.

472

Z. Tang et al.

4.1.3 Behaviour Under Cyclic Loading-Unloading 7-day SHCC prisms were tested under loading-unloading cycles until failure. During each cycle, the specimen was loaded until the crosshead displacement reached a certain value um and then unloaded. When um < 1 mm, the displacement increment was Du = 0.2 mm, and the loading-unloading rate was 0.2 mm/min; when um > 1 mm, Du = 0.5 mm and the loading-unloading speed was 0.5 mm/min. The resulting stressdeflection curve is given in Fig. 3. The envelop curve of the cyclic testing result was close to the monotonic stress-deflection profile, showing that the specimen still experienced an elastic phase, a hardening phase and a softening phase. The resulting MOR was slightly higher probably because a higher loading rate typically results in higher fibre strength and composite strength [13]. The cracking pattern of multiple parallel fine cracks was also similar to that for monotonic tests, as shown in Fig. 3.

80mm

12

Flexural stress (MPa)

10 8 6 4 2 0

0

1

2

3

4

5

Midspan deflection (mm)

6

7

Fig. 3. Flexural stress-deflection curve and cracking pattern of a 7-day SHCC specimen under cyclic loading-unloading.

4.1.4 Performance Under Reversed Flexural Loading Cycles 7-day SHCC specimens were tested under symmetrically reversed cyclic loading up to three cycles. The increase in the midspan deflection Dv was 1 mm between cycles. The loading and unloading speed in both directions was 0.5 mm/min. Typical flexural stress-deflection curves are shown in Fig. 4. Based on visual observation, in every cycle, the reversed force could help close previous cracks in one half part of the specimen while re-opening old cracks or generating new cracks in the other half part. On the other hand, when the specimen was unloaded to zero stress after each cycle, a residual deformation was generated. This shows that when the SHCC has been

Characterisation of SHCC Under Cyclic Loading Conditions

473

deformed to crack and beyond elastic phase, plastic deformation would occur and the material could not retrieve its original shape, and the cracks would not completely close without external forces. This is because the process of fibre debonding, pull-out or rupture could not be reversed. The indication is that if the cracked SHCC went through autogenous self-healing, the cracks could be filled with healing products and the bonding strength of fibres might recover to some extent, although the ruptured fibres could no longer bridge the crack.

Flexural stress (MPa)

10

5

-3

-2

-1

0

0

1

2

3

-5

-10

Cycle 1 Cycle 2 Cycle 3

Midspan deflection (mm)

Fig. 4. Hysteresis loop under reversed cyclic loading.

4.2

Microstructural Observation

4.2.1 Crack Width Measurement Under flexural loading, most cracks generated were approximately perpendicular to the neutral axis. An example of micro-cracks under stereoscope is shown in Fig. 5. The total crack numbers and non-localised crack widths of specimens subjected to monotonic flexural loading are summarised in Table 4. Overall, the air cured specimens exhibited a higher number of cracks and smaller crack widths than the high moisture cured specimens. These were also in accordance with the flexural performance and geometric properties of SHCC. A larger product of the average crack width and total crack number tended to contribute to a higher deflection capacity. The crack widths were 58 (±15) lm for cyclic loading-unloading tests and 38 (±14) lm for reversed cyclic tests. The average crack widths for either monotonic or cyclic loading condition were below 60 lm thanks to the bridging effect of fibres, indicating that these cracks could potentially be completely self-healed in favourable environments [8].

474

Z. Tang et al.

Fig. 5. Micro-cracks under stereoscope. Table 4. Crack number and widths of SHCC. Curing condition 7-day high moisture 28-day high moisture 28-day air Crack number 12 32 61 Crack width (lm) 51 (±19) 35 (±17) 28 (±17)

4.2.2 SEM/EDX Analysis From SEM images, it is shown that PVA fibres in cracks either bridged the crack, ruptured or were pulled out due to the crack opening. When the crack widths were small (generally below 100 lm), fibres worked together to bridge the crack, and being in the debonding stage were able to generate steady-state cracks. As a crack localised and its width increased, fibres across this crack started to lose function. They tended to rupture when the pull-out force exceeded their strength (Fig. 6(a)), or they were pulled out, as shown in Fig. 6(b). This confirms previous findings [14]. The fibres were covered with cementitious materials according to both SEM images and EDX analysis (Fig. 6(c)). Fibres are composed of carbon and oxygen. The materials attached on the fibre surfaces exhibited an element combination of either Ca, Si, O or Ca, Si, O, Al, showing that they were probably calcium silicate hydrate (C-S-H), unhydrated cement particles or fly ash (in circular shapes). The latter two are essential for continuing hydration and pozzolanic reaction, which promote autogenous self-healing. This is in agreement with [15], i.e. the fibres could not only bridge cracks mechanically, but also act as a scaffolding for autogenous healing products to precipitate on due to the high polarity of PVA fibres, thus the cracks could be filled more efficiently. Larger cracks (up to 0.3 mm) in PVA-SHCC could also be more effectively healed compared with those in normal concrete under favourable conditions. Further work is underway to assess the self-healing performance of the investigated mixes.

Characterisation of SHCC Under Cyclic Loading Conditions

(a)

(b)

475

(c)

Fig. 6. SEM images of (a) fibres in a crack and ruptured fibres; (b) bridging fibres being partially pulled out; (c) hydration products and unhydrated particles on fibre surface.

5 Conclusions In the work herein we investigated the behaviour of SHCC under monotonic loading, cyclic loading-unloading and reversed cyclic flexural loading conditions respectively and studied the cracking behaviour and microstructural features. The main findings are summarised below: • The compressive strength of SHCC was 50.33 MPa at 7 days and 75.42 MPa at 28 days under high moisture curing. • The SHCC specimens exhibited deflection-hardening and multiple cracking behaviour under either monotonic or cyclic flexural loading conditions. The envelop curves of the cyclic testing results were close to the monotonic stress-deflection profile, including elastic, hardening and softening phases. • 28-day air-cured specimens showed higher deflection capacity and smaller crack width than the 7-day and 28-day high-moisture-cured specimens, exhibiting the highest potential self-healing. Further mechanical testing is required for valid experimental proof. • In the reversed cyclic tests, the reversed force could help close previous cracks at one half part of the specimen while generating new cracks or re-opening old cracks in the other half of the specimen. • The specimens could not return to their original shape when unloading after plastic deformation, showing that the process of cracking, fibre debonding, pull-out or rupture was irreversible. However, if the cracked SHCC went through autogenous self-healing, the cracks would be filled with healing products and the bonding strength of fibres might recover to some extent, which also requires further validation and is the object of ongoing work. • The fibres were covered with cementitious materials and hydration products according to both SEM images and EDX analysis, confirming that PVA fibres with a high polarity could not only bridge the cracks mechanically, but also act as a medium for autogenous healing products to participate on, thus larger cracks could potentially be healed.

476

Z. Tang et al.

Acknowledgements. The financial support from the EPSRC for the Resilient Materials for Life (RM4L) Programme Grant (EP/P02081X/1) is gratefully acknowledged. The financial support from the Cambridge Commonwealth, European and International Trust (CCEIT) for the first author’s PhD research, is highly appreciated.

References 1. de Rooij, M., Van Tittelboom, K., De Belie, N., Schlangen, E.: Self-healing phenomena in cement-based materials. Springer (2013) 2. Huang, H., Ye, G., Qian, C., Schlangen, E.: Self-healing in cementitious materials: Materials, methods and service conditions. Mater. Des. 92, 499–511 (2016) 3. De Belie, N., Gruyaert, E., Al-Tabbaa, A., Antonaci, P., Baera, C., Bajare, D., Darquennes, A., Davies, R., Ferrara, L., Jefferson, T., Litina, C., Miljevic, B., Otlewska, A., Ranogajec, J., Roig-Flores, M., Paine, K., Lukowski, P., Serna, P., Tulliani, J.M., Vucetic, S., Wang, J., Jonkers, H.M.: A review of self-Healing concrete for damage management of structures. Adv. Mater. Interfaces. 15(7), 28 (2018) 4. Li, V.C.: From micromechanics to structural engineering - the design of cementitious composites for civil engineering applications. J. Struct. Mech. Earthq. Eng. 10(2), 37–48 (1993) 5. Li, V.C.: On Engineered Cementitious Composites (ECC) - a review of the material and its applications. J. Adv. Concr. Technol. 1(3), 215–230 (2003) 6. Yildirim, G., Keskin, Ö.K., Keskin, S.B.I., Şahmaran, M., Lachemi, M.: A review of intrinsic self-healing capability of engineered cementitious composites: recovery of transport and mechanical properties. Constr. Build. Mater. 101, 10–21 (2015) 7. Li, V.C., Herbert, E.: Robust self-healing concrete for sustainable infrastructure. J. Adv. Concr. Technol. 10, 207–218 (2012) 8. Wu, M., Johannesson, B., Geiker, M.: A review: self-healing in cementitious materials and engineered cementitious composite as a self-healing material. Constr. Build. Mater. 28, 571– 583 (2012) 9. Van Zijl, G.P.A.G., Wittmann, F.H.: Durability of strain-hardening fibre-reinforced cementbased composites (SHCC). Springer (2011) 10. Wang, S., Li, V.C.: Engineered Cementitious Composites with high-volume fly ash. ACI Mater. J. 104(3), 233–268 (2007) 11. Yang, Y., Lepech, M.D., Yang, E.-H., Li, V.C.: Autogenous healing of engineered cementitious composites under wet-dry cycles. Cem. Concr. Res. 39, 382–390 (2009) 12. Qian, S.Z., Zhou, J., Schlangen, E.: Influence of curing condition and precracking time on the self-healing behavior of Engineered Cementitious Composites. Cem. Concr. Compos. 32, 686–693 (2010) 13. Yu, K., Li, L., Yu, J., Wang, Y., Ye, J., Xu, Q.: Direct tensile properties of engineered cementitious composites: a review. Constr. Build. Mater. 165, 346–362 (2018) 14. Redon, C., Li, V.C., Wu, C., Hoshiro, H., Saito, T., Ogawa, A.: Measuring and modifying interface properties of PVA fibers in ECC matrix. J. Mater. Civ. Eng. 13(6), 399–406 (2001) 15. Nishiwaki, T., Koda, M., Yamada, M., Mihashi, H., Kikuta, T.: Experimental study on selfhealing capability of FRCC using different types of synthetic fibers. J. Adv. Concr. Technol. 10, 195–206 (2012)

Corrosion Pattern and Mechanical Behaviour of Corroded Rebars in Cracked Plain and Fibre Reinforced Concrete E. Chen1(&), Carlos G. Berrocal1,2, Ingemar Löfgren1,2, and Karin Lundgren1 1

Division of Structural Engineering, Chalmers University of Technology, Gothenburg, Sweden [email protected] 2 Thomas Concrete Group AB, Gothenburg, Sweden

Abstract. This paper presents experimental results of corrosion pattern and tensile behaviour of corroded rebars extracted from 4 uncracked and 18 precracked plain concrete and fibre reinforced concrete (FRC) beams. The beams were pre-cracked through three-point bending to a target maximum crack width of 0.1 and 0.4 mm, and then subjected to natural corrosion through cyclic exposure to a 16.5% NaCl solution for more than three years. 3D-scanning was used to characterise the pit morphology and evaluate the maximum local corrosion level of extracted rebars. Under the same loading condition and crack width, most rebars in FRC had smaller maximum local corrosion level than those in plain concrete. Subsequently, tensile tests were carried out on the extracted rebars, with Digital Image Correlation (DIC) technique adopted to investigate the influence of pit morphology on the local strain development. Finally, the time-dependent influence of transverse and longitudinal cracks on the pit morphology which governs the ultimate strain of corroded rebars was discussed. The time-varying nature of corrosion morphology should be considered when predicting the durability and long-term safety of conventional reinforced concrete and FRC structures with reinforcing bars under chloride environments. Keywords: Fibre reinforced concrete crack
Tensile behaviour
Durability


Pitting corrosion
Corrosion-induced

1 Introduction Reinforced Concrete (RC) structures inevitably possess cracks, originating from shrinkage, thermal gradients and/or mechanical loading. These cracks accelerate the ingress of various adverse agents, leading to earlier corrosion initiation. However, regarding to the effect of cracks on the corrosion propagation, contradictory results have been obtained, as shown in the state-of-the-art report [1]. Some current codes dictate the maximum allowable crack width in addition to the minimum cover depth to limit the risk of corrosion. The restrictive requirement of controlling crack width often results in congested reinforcement layouts for structures exposed to marine or de-icing © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 477–488, 2021. https://doi.org/10.1007/978-3-030-58482-5_44

478

E. Chen et al.

salt environment. Fibre reinforcement combined with conventional reinforcing bars is an attractive alternative to avoid deploying heavy steel bars. Fibres are an effective means of crack control and can also improve the mechanical performance of concrete structures [2]. Nevertheless, the long-term performance of fibre reinforced concrete (FRC) combined with conventional steel bars under natural corrosion condition has not be adequately studied so far. A long-term experimental program on conventional RC beams and concrete beams reinforced with steel bars and various types of fibres (with a volume fraction of 0.3 mm [7]) in prestressed concrete sleeper is predominantly aimed at the partial or complete replacement of steel to further improve the design efficiency in terms of weight, adaptability, structural performances, corrosion resistance and sustainability [8, 9]. Yet, most studies which investigated the behaviour of fibre reinforced concrete concluded that fibre content within the moderate range of 0.5% to 5.0% has limited effect on the mechanical properties of uncracked concrete [10, 11]. In fact, the significant benefits of macro-synthetic fibres are observed across the failure plane providing bridging and confinement of cracks towards considerably improving the energy absorption capacity (i.e. toughness), ductility and post-cracking behaviour [5, 12]. Furthermore, the orientation and distribution of the synthetic fibres in the cementitious material are critical towards inducing a more isotropic behaviour as well as ductile fracture mechanisms of FRC as highlighted in Fig. 2. Even though limited research on

Post-cracking Strength Classification of Macro-synthetic Fibre

719

FRC sleeper has been conducted, comparable reduction in cracks propagation are expected when implementing macro-synthetic fibre reinforcement in sleepers.

Fig. 2. Schematic illustration of the fibre/matrix bridging mechanism.

Accordingly, in an effort to understand and categorise the benefits of fibre reinforced concrete for structural applications, the post-cracking tensile behaviour (i.e. r-e relationship) must be known [13]. Currently, there are limited accepted standards for the design of macro-synthetic fibre reinforced concrete (MSFRC) which further highlights the lack of research towards this innovative material [6, 13]. These standards include predominantly the Italian National Code (CNR-DT 204/2006), American Concrete Institute (ACI 544.4R-18) and Fib Model Code 2010 that introduces a constitutive law model in uniaxial tension for the post-cracking behaviour of MSFRC as further discussed in this paper. Therefore, the study herein presented will focus on the reinforcing capabilities (i.e. post-cracking) of macro-synthetic fibre reinforced concrete towards efficiently reducing the cost while optimising the load-bearing capacity, failure mechanisms and sustainability for sleeper applications.

2 Methodology 2.1

Materials and Specimen Preparation

A total of fifteen batches (i.e. 1 plain & 14 MSFRC mixes) were cast with a high strength concrete mix which was designed and further adjusted to achieve a characteristic compressive strength of 50 MPa as specified for railway sleeper applications [14]. The mix proportion herein implemented for the evaluation of macro-synthetic fibre reinforced concrete is provided in Table 1. The nominal maximum size of coarse aggregates used was 10 mm to further increase the fibre-matrix interface towards ensuring adequate packing density as well as homogenous distribution of the fibres. In addition, a high-range water reducer (superplasticizer) was incorporated to further improve the workability, dispersion and prevent fibre clumping throughout the mixes.

720

C. Camille et al. Table 1. Mix proportions used for MSFRC. Component Dosage (kg/m3) Cement (30% fly ash) 446 Coarse aggregate 870 Nepean paving sand (coarse) 510 Newcastle sand (fine) 350 Water 175 High-range water reducer (HRWR) 4.4 – 5.3 Water-cement ratio ~ 0.39

The macro-synthetic fibres herein investigated are predominantly made from highperformance polypropylene base materials whose physical and mechanical characteristics are provided in Table 2. Accordingly, both BC48 and BC58 are characterised by quite similar mechanical properties including a continuously embossed surface as presented in Fig. 3. In this study, the macro-synthetic fibres were introduced at various fibre volumes ranging from 0.2%–2.0% (i.e. 1.82–18.2 kg/m3) to enable the evaluation of MSFRC post-cracking behaviour for sleeper applications. Table 2. Properties of the macro-synthetic fibres [15].

Fibre type BC48 BC58

Base material Virgin polypropylene Bi-component polymer

Length (mm)

Diameter (mm)

Tensile strength (MPa)

Elastic modulus (GPa)

Specific gravity (kg/m3)

48

> 0.3

640

12

890 – 910

58

> 0.3

640

10

910 – 920

For each fibre-dosage combination, three cylindrical and four notch beams specimens of size Ø100  200 mm and 150  150  575 mm were cast respectively to further assess the compressive and residual flexural strengths at 28 days. 2.2

Experimental Tests

The testing procedures undertaken for both the compression and flexural specimens comply with AS1012.9-2014 [16] and EN14651-2005 [17] respectively, thus ensuring the accuracy and reliability of the experimental results herein presented. Accordingly, the specimen’s geometry and experimental setup are displayed in Fig. 4.

Post-cracking Strength Classification of Macro-synthetic Fibre

721

Fig. 3. Macro-synthetic fibre reinforcement: BC48 (left), BC58 (right).

Fig. 4. Specimens geometry and testing configuration (mm).

Prior to testing, notches were cut at mid-span across the tensile side of the beam specimens to be further assessed in a closed-loop displacement control three-point bending test arrangement (i.e. 1 constrained DOF at each roller). As a result, crack initiation and propagation occur within a predefined section of the beam for which the load versus crack mouth opening displacement (CMOD) was continuously recorded (i.e. 5 Hz). Correspondingly, the flexural strengths (i.e. pre & post-cracking) of the macro-synthetic fibre reinforced concrete specimens were evaluated from the loadCMOD curves as expressed through Eq. (1): fR;j ¼

3:FJ :L 2 2:B:Hsp

ð1Þ

where FJ is the load corresponding with CMOD; L = span length; B = average beam width and Hsp = distance between tip of notch to the top of specimen. Similarly, the

722

C. Camille et al.

effective fracture energy (Gf) of the notched specimens in a three-point bending configuration was evaluated through Eq. (2) as recommended by RILEM Technical Committee 50 [18]. The fracture energy of synthetic fibre reinforced concrete is characterised by the area under the load-deflection curve which is also typically defined as the energy required to open a unit crack surface area [19]. " Gf ¼

dmax Z

# F ðdÞdd þ m:g:dmax =½ðd  ao Þ:B

ð2Þ

0

where dmax corresponds to the maximum deflection; m = beam mass; g = 9.81 ms−2; d = beam depth and ao = notch depth. In comparison to plain concrete, it is predominantly the tensile bridging mechanism of the macro-synthetic fibres which enables a substantial increase in the energy absorption capacity and post-cracking behaviour of the MSFRC as further demonstrated in the herein presented study. 2.3

Post-cracking Flexural Strength Behaviour

In the study herein presented, the classification of FRC’s post-cracking strength comply with the Fib Model Code 2010 which assumed a linear-elastic relationship [9]. Indeed, this approach defined two main characteristic flexural strength values that are crucial towards the structural design of MSFRC for serviceability (fR1k) and ultimate (fR3k) conditions. For instance, these particular strength parameters correspond with CMOD of 0.5 and 2.5 mm respectively. Table 3 presents the classification of fibre reinforced concrete in terms of strength interval and residual strength ratios. Table 3. Residual strength class & ratio in accordance with Fib Model Code 2010 [9]. Strength interval (MPa)

1.0, 1.5, 2.0, 2.5, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, …

Residual strength ratio (fR3k / fR1k)

a if 0.5 ≤ fR3k / fR1k < 0.7 b if 0.7 ≤ fR3k / fR1k < 0.9 c if 0.9 ≤ fR3k / fR1k < 1.1 d if 1.1 ≤ fR3k / fR1k < 1.3 e if 1.3 ≤ fR3k / fR1k

In addition, a stress-crack opening constitutive law can also be defined towards characterising the post-cracking strength of macro-synthetic fibre reinforced concrete as a linear softening/hardening behaviour. The linear model identifies two reference values defined as the serviceability (fFts) and ultimate residual strengths (fFtu) evaluated using Eq. (3) and Eq. (4) respectively [9].

Post-cracking Strength Classification of Macro-synthetic Fibre

fFts ¼ 0:45:fR1 fFtu ¼ fFts 

wu :ðfFts  0:5:fR3 þ 0:2:fR1 Þ  0 CMOD3

723

ð3Þ ð4Þ

where wu = maximum acceptable crack opening (taken as 2.5 mm), fR1 and fR3 corresponding to the loads at CMOD of 0.5 and 2.5 mm respectively. As a result, the stress-strain relationship of FRC (shown in Fig. 5) in uniaxial tension can be derived wherein which the various cases (i.e. I, II & III) may describe the post-cracking behaviour through the softening or hardening of the material. In fact, the first branch represents the pre-peak behaviour of plain concrete followed by the second branch which characterises the post-peak crack propagation until intersecting with the residual post-cracking behaviour (third & fourth branch). Hence, resuming the contribution of the fibres throughout the material towards a characteristic softening or hardening [9].

Fig. 5. Stress-strain relationship of fibre reinforced concrete [9].

3 Results and Discussion 3.1

Compressive Strength

At 28 days of curing, the macro-synthetic fibre concrete specimens were tested for their ultimate compressive strength (fc) for which most of the specimens incorporating BC48 and BC58 at different fibre content were observed to exceed the minimum 50 MPa sleeper’s requirement. Figure 6 summarises the average compressive strength behaviour of MSFRC specimens. BC48 specimens were observed to peak at 62.7 MPa (i.e. 0.2% fibres) characterising an increase of approximately 18% over plain concrete which only achieved an ultimate capacity of 53.4 MPa. This particular behaviour of BC48 specimens at 0.2% fibre volume ratio can be justified by the slightly lower spread-flow obtained. On the other hand, specimens reinforced with BC58 (up to 1%) experienced on average a 6.6% increase in compressive strength as compared to BC48. The ultimate strengths of the MSFRC encompassing BC58 fibres were observed to be closely consistent up to 1.0% of fibres, reaching a peak of 61.8 MPa at 0.8%. However, as the fibre dosage exceeds

Compressive Strength, fc (MPa)

724

C. Camille et al. 70 60 50 40 30 20 10 0

Fibre Volume Ratio, % (kg/m3) BC48

BC58

Control

Fig. 6. Compressive strength behaviour of MSFRC at 28 curing days.

1%, the reported strength decreases due to the balling effect of longer fibres which also involved undesirable air voids, poorer workability and segregation. Therefore, it is recommended that the design mix is improved towards a more workable concrete especially at higher fibre dosage, although the addition of macro-synthetic fibres has no consensual benefits on the ultimate compressive performances. Yet, the reinforcing fibres significantly improved the failure mechanisms towards more ductile concrete specimens with the ability to absorb energy without shattering into pieces. 3.2

Flexural Behaviour

In general, concrete is considered to undergo quasi-brittle failure wherein which the specimen experiences a complete loss of loading capacity once the failure has been initiated. Therefore, the flexural behaviour presented in this study fundamentally identifies the important benefits of incorporating macro-synthetic fibres into concrete towards improved post-cracking characteristics as presented in Fig. 7. Although the addition of macro-synthetic fibres had negligible impact on the precracking strength, both types of fibre are observed to significantly increase the residual flexural (post-cracking) capacity as the fibre volume ratio increases up to 2.0%. The main difference across the two types of fibre used in this study is the residual strength which peaked at around a CMOD of 2.5 mm for BC48 as compared to 3.5 mm for BC58. Correspondingly, BC58 specimens tend to possess a greater ductility due to the longer and more flexible fibres. Overall, the specific post-cracking behaviour of the MSFRC is justified by the mechanics of crack formation and propagation where both fibre types interact with the concrete matrix. As a result, BC 48 and BC58 adequately stabilised and suppressed crack growth towards more ductile failure modes, a desirable material behaviour for sleeper’s application. Table 4 classifies the post-cracking flexural strength of macro-synthetic fibre reinforced concrete with various fibre dosages and types. Characteristically, these residual strength classes could be used for design purposes wherein which the fibre

Post-cracking Strength Classification of Macro-synthetic Fibre

725

Flexural strength (MPa)

9 8 7 6 5 4 3 2 1 0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 CMOD (mm)

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 CMOD (mm)

Fig. 7. Residual flexural-CMOD behaviour for BC48 (left) & BC58 (right). Table 4. Residual flexural strength classification of MSFRC.

Fibre dosage % kg/m3 0.2 1.82 0.4 3.64 0.6 5.46 0.8 7.28 1.0 9.10 1.5 13.7 2.0 18.2

Residual flexural strength (MPa) BC48 BC58 fR1 fR3 fR1 fR3 1.26 1.18 1.02 0.62 1.32 1.87 1.36 1.71 1.96 2.98 1.88 2.87 2.54 3.70 1.92 3.07 3.69 5.92 2.70 4.76 4.53 7.12 3.86 6.87 5.09 7.70 3.75 6.65

Residual strength class BC48 BC58 1e 1.5e 2e 3e 3e 4e

1d 1.5e 1.5e 2e 3e 3e

reinforcement partially or completely replace conventional reinforcement in sleepers at ultimate limit state. 3.3

Fracture Energy

The effective fracture energy (Gf) of MSFRC derived in this study characterises the energy dissipated across the notched specimens up to a corresponding mid-span deflection of approximately 3.4 mm (i.e. CMOD = 4 mm). Indeed, this value was designated to consistently evaluate and compare the Gf of the macro-synthetic fibre reinforced concrete specimens within its design residual strength classification as

726

C. Camille et al.

Fracture energy, Gf (N/mm)

previously discussed. Figure 8 illustrates the changes in Gf with an increase in the fibre volume ratio of the macro-synthetic fibres. 5.0 4.0 3.0 2.0 1.0 0.0

Fibre Volume Ratio, % (kg/m3) Control

BC48

BC58

Linear (BC48)

Linear (BC58)

Fig. 8. Effect of fibre type and volume fraction on the effective fracture energy (vertical bars represent standard deviation).

In general, a linear increase in the effective fracture energy was observed for both BC48 and BC58 specimens when tested up to a fibre volume fraction of 2.0%. This linear behaviour of polymer fibres has already been observed previously, thus supporting the accuracy of the results. As a result, the macro-synthetic fibre reinforcement herein investigated has the ability to improve the toughness of concrete due to its relatively low elastic modulus and flexible fibres which can be uniformly distributed throughout the cement matrix [19]. Undeniably, this atypical behaviour of concrete is desired in the railway industry wherein which the sleeper does not need to be immediately replaced once cracked. 3.4

Stress-Strain Relationship

The design stress-strain (r-e) relationships were plotted for both BC48 and BC58 at various fibre volume ratios as represented in Fig. 9. The r-e results obtained in this study characterises the post-cracking stages (i.e. softening or hardening) of macrosynthetic fibre reinforced concrete when subjected to uniaxial tension as predefined through the crack-opening constitutive law. Indeed, both BC48 and BC58 are observed to enhance the post-cracking performances as compared to plain concrete, predominantly characterising a Case I (Fig. 5) constitutive relationship. For instance, up to a fibre volume fraction of 0.6%, both types of fibres exhibited a fairly improved softening behaviour as opposed to plain concrete. Nevertheless, when reinforced with a fibre volume ratio of 0.8% up to 2.0%, an uncharacteristic hardening branch (i.e. 4th branch) is developed due to the high fibre

Post-cracking Strength Classification of Macro-synthetic Fibre

727

Fig. 9. Design stress-strain relationship for BC48 (top) and BC48 (bottom) at various fibre dosages.

dosage bridging the crack surfaces. Correspondingly, the stress-strain relationship of macro-synthetic fibre reinforced concrete in the post-cracking region is characterised through an improved Case I exhibiting hardening of the material as presented in Fig. 9.

4 Conclusions In this paper various impacts of macro-synthetic fibre reinforcement on the postcracking performances of concrete are discussed from which the following conclusions were achieved: • Compressive strength – The experimental results demonstrated that the macrosynthetic fibres provide evident bridging effect characteristically reducing the crack propagation towards more ductile failure modes. However, the addition of fibres did not noticeably influence the ultimate compressive strengths. • Flexural behaviour & fracture energy – The addition of fibres was observed to insignificantly impact the ultimate flexural strength (i.e. pre-cracking) of the specimens. Nevertheless, the fibre reinforcement improved the fracture mechanisms towards a more ductile behaviour, naturally minimising the loss in capacity sustained after the initial cracks. In general, higher fibre dosages exhibited better performance in terms of residual flexural strength (i.e. post-cracking), ductility and

728

C. Camille et al.

toughness towards structural design implementation involving the partial or complete substitution of traditional steel reinforcement. • Stress-strain behaviour – The incorporation of the macro-synthetic fibres into the concrete matrix especially at high fibre content generate an atypical yet desired hardening behaviour once the specimens endured a loss in capacity after the first crack. Acknowledgements. The authors gratefully acknowledge the Australian Research Council’s Industrial Transformation Training Centres Scheme (ARC Training Centre for Advanced Technologies in Rail Track Infrastructure; IC170100006) which provided the catalyst for undertaking this research as well as Western Sydney University through the School of Computing, Engineering and Mathematics for their support to the authors work described herein.

References 1. Kumar, D.K., Sambasivarao, K.: Static and dynamic analysis of railway track sleeper. Int. J. Eng. Res. Gener. Sci. 2(6), 1–10 (2014) 2. Tzanakakis, K.: The Railway Track and Its Long Term Behaviour, 1st edn. Springer Tracts on Transportation and Traffic. Springer-Verlag, Berlin Heidelberg (2013) 3. Sharma, R.C., Palli, S., Sharma, S.K., Roy, M.: Modernization of railway track with composite sleepers. Int. J. Vehicle Struct. Syst. 9(5), 321–329 (2017) 4. Ferdous, W., Manalo, A., Van Erp, G., Aravinthan, T., Kaewunruen, S., Remennikov, A.: Composite railway sleepers–recent developments, challenges and future prospects. Compos. Struct. 134, 158–168 (2015) 5. Buratti, N., Mazzotti, C., Savoia, M.: Post-cracking behaviour of steel and macro-synthetic fibre-reinforced concretes. Constr. Build. Mater. 25(5), 2713–2722 (2011) 6. Juhász, K.P.: The effect of the synthetic fibre reinforcement on the fracture energy of the concrete. In: IOP Conference Series: Materials Science and Engineering, vol. 613, p. 012037, 4 Nov 2019 7. Fibres for concrete-Part 2: Polymer fibres-Definitions, specifications and conformity, EN14889–2:2006 (2006) 8. Kohoutková, A., Broukalová, I.: Optimization of fibre reinforced concrete structural members. Procedia Eng. 65, 100–106 (2013) 9. Fib Model Code for Concrete Structures 2010, 9783433604083 (2013) 10. Yehia, S., Douba, A., Abdullahi, O., Farrag, S.: Mechanical and durability evaluation of fiber-reinforced self-compacting concrete. Constr. Build. Mater. 121, 120–133 (2016) 11. Enfedaque, A., Alberti, M.G., Gálvez, J.C., Domingo, J.: Numerical simulation of the fracture behaviour of glass fibre reinforced cement. Constr. Build. Mater. 136, 108–117 (2017) 12. Merta, I., Tschegg, E.K.: Fracture energy of natural fibre reinforced concrete, (in English). Constr. Build. Mater. 40, 991 (2013) 13. Jafarifar, N., Pilakoutas, K., Angelakopoulos, H., Bennett, T.: Post-cracking tensile behaviour of steel-fibre-reinforced roller-compacted-concrete for FE modelling and design purposes. Materiales de Construcción 67(326), e122 (2017) 14. Railway Track Materials-Part 14: Prestressed concrete sleepers, AS1085:14:2012 (2012) 15. BarChip, 20 October 2018. https://barchip.com/product/

Post-cracking Strength Classification of Macro-synthetic Fibre

729

16. Methods of testing concrete-Methods 9: Compressive strength tests-Concrete, mortar and grout specimens, AS 1012.9:2014 (2014) 17. Test method for metallic fibered concrete-Measuring the flexural tensile strength (limit of proportionality (LOP), residual), EN14651:2005 (2005) 18. RILEM, Recommendation TC 50-FMT-Determination of the Fracture Energy of Mortars and Concretes by Means of Three-Point Bend Tests on Notched Beams. Mater. Struct. vol. 18 (1985) 19. Kosior-Kazberuk, M.: Post-cracking Behaviour and Fracture Energy of Synthetic Fibre Reinforced Concrete. Mater. Sci./Medziagotyra, Article 22(4), 542–547 (2016)

Structural Behaviour of Steel-Fibre-Reinforced Lightweight Concrete Hasanain K. Al-Naimi and Ali A. Abbas(&) University of East London, London, UK [email protected]

Abstract. This paper evaluates the influence of adding hooked-end 3D fibres to recycled lightweight concrete beams with and without shear reinforcement. Lightweight concrete used is from recycled fly ash and can lead to reductions in the mass of the structure. However, it is typically more brittle than normal weight concrete. To enhance ductility, steel fibres were used and the complete and partial substitution of stirrups with fibres was investigated. The flexural and shear behaviour of SFRLC beams under 3-point loading with adequate, substandard and no shear reinforcement are studied. The experimental study includes Vf = 0%, 1% and 2% as well as stirrups spacing of 120 mm, 240 mm and ∞. The results in this paper are given in terms of strength, ductility, crack patterns and nature of failure. It was found out that fibre reinforcement is capable of partially replacing stirrups, however, when stirrups were completely removed the failure became brittle. Nonlinear finite element analysis (NLFEA) using ABAQUS is also used to validate a SFRLC constitutive model whose properties were derived from compressive and pullout-tests. Keywords: Recycled lightweight concrete beams
3D hooked-end fibres
3-point-bending test
Shear design
Stirrups replacement
Ductility
Load bearing capacity
Validation
NLFEA
ABAQUS

1 Introduction The use of structural lightweight aggregate concrete brings a number of advantages as compared to the conventional normal weight concrete such as thermal insulation and fire resistance. Besides, the lightweight aggregate used in this work is recycled and offers reduction in CO2 emissions as well as an alternative to the depleting gravel and quarry resources (Gerritse 1981). More importantly, the improved strength-to-weight ratio results in smaller cross sections which in turn leads to a decrease in reinforcement and savings in material transport costs due to lower inertial and gravity loads. The usage of lightweight concrete can therefore be ideal and competitive in industry for the growing need for taller and longer span structures, especially in seismic and dynamic zones (Libre et al., 2011; Campione, 2014; Dias-Da-Costa 2014; Mo et al., 2017). A study on Lytag (Lytag, 2014) showed that lightweight concrete can bring about 34% savings in CO2 as well as a reduction of up to 48% of concrete and reinforcement when compared to conventional gravel concrete. These advantages however come as a tradeoff for the increased brittleness of lightweight concrete. The latter is the case due to the © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 730–744, 2021. https://doi.org/10.1007/978-3-030-58482-5_65

Structural Behaviour of Steel-Fibre-Reinforced Lightweight Concrete

731

lightweight material having poor aggregate interlock which translates into lacking toughening mechanisms in tension post-crack. This causes lower tensile strength which ultimately lowers shear resistance in structures such as beams and slabs. Also, the porous nature of lightweight concrete leads to lower modulus of elasticity (Chen et al., 2010; Badogiannis and Kotsovos 2013) which leads to excessive deflections and cracking (Lim et al., 2006; Wu et al., 2011). Nonetheless, the brittle nature of lightweight concrete can be addressed by incorporating traditional reinforcement such as rebars and stirrups. However, the latter solution can become inherently counterproductive and impractical when reduction in structural elements is sought by employing lightweight concrete especially at critical zones such as joints. Therefore, fibre reinforcement which has long proven its effectiveness in controlling and bridging tensile and shear cracks in the past for both lightweight and normal weight concretes can become an adequate solution (Gao et al., 1997; Campione and La Mendola, 2004; Abbas et al., 2014a; Di Prisco et al., 2013; Grabois et al., 2016; Mo et al., 2017). For over 40 years the usage of steel fibres in concrete mixes has been experimented with and used (Ritchie and Kayali 1975), however, comprehensive studies on fibrous lightweight concrete is still scarce with most work being merely theoretical (Swamy et al., 1993; Kang and Kim 2010; Di Prisco et al., 2013; Iqbal et al., 2015; Grabois et al., 2016). It should be noted that, at present, there is no international standards specific for steel fibre reinforced lightweight concrete (SFRLC) with current guidelines being usually adapted from fibrous normal weight concrete. This work is a continuity to an accompanying paper that evaluates the behaviour of plain and fibrous lightweight concrete on the material level in both tension and compression. The latter paper’s constitutive model is used to validate the reinforced concrete beams on the structural level by studying the shear and flexural behaviour via NLFEA using ABAQUS (Habbit et al., 2000). The parameters used in this study include fibre dosage (Vf), and stirrups spacing (s). The first parameter will be either 0%, 1% or 2% while the second will be 120 mm, 240 mm or ∞. The choice of the beam spans and transversal reinforcement was as such so that a/d = 3 with d as the effective depth and a as the shear span. This anticipated Type II shear failure behaviour according to Kani’s valley (Kotsovos and Pavlovic 2013) and hence a brittle failure is predicted for the beams unless adequately shear reinforced. It should be noted that a stirrups spacing of 120 mm is adequate to Eurocode 2 shear design and is therefore expected to fail in a ductile flexural manner while the spacing of 240 mm is inadequate to Eurocode 2 beam design and is consequently expected to fail in brittle shear manner. No stirrups spacing i.e. s = ∞, is an even more severe case of substandard reinforced concrete beam design. The fibre type used in this study will be the conventional hooked-end 3D fibres while the target compressive strength chosen was 30 MPa for all the reinforced concrete beams tested. This fibre type was chosen for experimental tests because hooked-end 3D fibres are currently the most commonly used in industry (Abdallah et al., 2018a), and they provide the least tensile upgrade to plain concrete as compared to 4D and 5D. Also, a compressive strength of 30 MPa which consequently provides the lowest tensile strength possible for structural lightweight concrete.

732

H. K. Al-Naimi and A. A. Abbas

2 Methodology The methodology of the work presented in this paper consists of two main parts. The first will involve experimental testing of the structural reinforced concrete beams under 3-point loading. The main focus of this part will be on the shear and flexural strength, ductility and cracking pattern. The second part will cover nonlinear numerical finite element analyses of the structural beams using ABAQUS’ concrete damaged plasticity by adopting the proposed model suggested in a previous paper as well as a comparison with 3 available SFRC models mentioned previously.

3 Experimental Study 3.1

Experimental Programme

As previously noted, an accompanying paper with more work detailing the material properties of plain and fibrous lightweight concrete was conducted. Figure 1 below summarises the complete proposed model adopted for fcm = 30 MPa with Vf = 0%, 1% and 2% and used for the structural reinforced beams on ABAQUS. Also, for over 100 compression tests, Poisson’s ratio was recorded between 0.15 and 0.20 with no particular pattern in relation to strength, Vf, or fibre type. It should be noted that this model was input into ABAQUS concrete material relationships.

The proposed constitutive model for fcm=30MPa and 3D fibres 5

-5

-3

-1

0

1

3

5

7

9

Stress (MPa)

-5 -10

Vf=0% Vf=1%

-15

Vf=2% -20 -25 -30 Strain ε (‰)

Crack width ω (mm)

Fig. 1. The proposed constitutive model for SFRLC

The study in this paper focuses on reinforced concrete beams. As mentioned before, 2 parameters are studied, namely; Vf and stirrups spacing s. 7 beams are tested, 3 for

Structural Behaviour of Steel-Fibre-Reinforced Lightweight Concrete

733

s = 120 mm, 3 for s = 240 mm with Vf = 0%, 1% and 2%, and one beam for s = ∞ with Vf = 1%. The latter aimed at studying the possibility of fibres to completely replace stirrups. The choice to skip lower dosages such as Vf = 0.5% is so, since most common structures such as tunnels and pavements require at least a dosage of Vf = 1% for efficient crack control and thickness reduction (The Concrete Society, 2007). 3.1.1 Materials Portland-Limestone cement (CEM 11) according to the specification supplied in EN 197-1 was used. Coarse aggregate Lytag, also known as Sintered Pulverised Fuel Ash Lightweight Aggregate (LYTAG) was provided by LYTAG Ltd. The loose dry density of LYTAG was calculated in the lab to be approximately 760 kg/m3 while the water absorption was estimated to be around 15% per mass of LYTAG. Natural river sand with a 4.75 mm maximum size was used as the fine aggregate of the concrete. The sand had a water absorption coefficient of 0.09% and specific gravity of 2.65 complying with BS EN 12620. The properties of the fibres used are summarized in Table 1 below. It should be noted that to prevent the possibility of balling, fibres were collated from the manufacturer. The reinforcement steel bars and stirrups modulus of elasticity was 210 GPa while the yield tensile stress for 6 mm bars was 512 MPa and that for 12 mm bars was 548 MPa. Table 1. Properties of fibres Fibre type ru (MPa) lf (mm) df (mm) 3D 65/60 1160 60 0.9

3.1.2 Mix Design The mix design used is summarized in Table 2 below. The mix design of the Lytag concrete for the characteristic cylinder and cube compressive strengths used are summarised below. These were directly adopted from Lytag (2011) manuals.

Table 2. Mix design used (fck/fck, cube) LC30/33

Cement (kg/m3) 370

Sand (kg/m3) 592

Loose bulk Lytag (kg/m3) 668.8

Effective water (kg/m3) 175

Calculating the water content of Lytag was of high importance as Lytag aggregates were found to absorb water of approximately 15% of their weight which is also confirmed by Lytag manual. For this reason and as suggested by Lytag manual 5 (2011), excess water to saturate Lytag aggregates was added 30 min before mixing (Fig. 2).

734

H. K. Al-Naimi and A. A. Abbas

Fig. 2. Mixing process for plain and fibrous lightweight concrete

3.2

Experimental Tests

For the 3 point- flexural beam tests, the LVDTs were glued using high strength epoxy after the concrete surface in touch with the LVDT was ground in the mid-section at the front of the beam to enable the LVDTs to fully adhere onto the concrete. The LVDT’s are connected to a computer software. For the purpose of estimating the vertical deflection accurately, a steel bar inspired by a technique similar to JSCE-SF4 recommendations was made. The beams are placed onto two frictionless steel supports. This was deemed adequate as the loading was symmetrical (Fig. 3). A displacement controlled constant loading of 0.2 mm/min was adopted using the hydraulic machine which has a load capacity of 500 kN. It should be noted that the machine is also supplied with a calibrated displacement transducer.

Fig. 3. On the left, vertical LVDT touching the Japanese Yoke and on the right, structural beam under 3-point loading

A relatively low longitudinal reinforcement ratio in tension of q = 1.8% with 2  T12 rebars was chosen to avoid possible over-reinforcement of section in tension as fibres were added. The longitudinal reinforcement ratio was similar to that adopted by Kodur et al. (2018) for SFRC beams. For compression reinforcement, 2  T6 rebars were used to support the compression flange of the lightweight concrete beams to avoid the possibility of concrete crushing as Vf was increased to 2%. Figures 4, 5 and 6 below show the 3 sections used for the structural beams.

Structural Behaviour of Steel-Fibre-Reinforced Lightweight Concrete

Fig. 4. Cross section for beam with S = 240 mm

Fig. 5. Cross section for beam with S = 120 mm

735

736

H. K. Al-Naimi and A. A. Abbas

Fig. 6. Cross section for beam with S = ∞

4 Numerical Study Three dimensional nonlinear finite element analyses will be carried out using ABAQUS (Zienkiewicz and Taylor, 2005). This software has shown to be successful at predicting the behaviour of SFRC in tension, compression, flexure and shear as well as the cracking pattern and mode of failure of plain and reinforced elements to a good level of accuracy (Tlemat et al., 2006; Syed Mohsin 2012; Abbas et al., 2014b; Behinaein et al., 2018). The approach adopted in this work will involve modelling both plain and fibrous concrete. The fibres will not be modelled discretely, instead they will be introduced directly into the constitutive tensile and compressive models of the concrete in ABAQUS. This methodology was opted for since modelling fibres discretely can be time consuming and will produce a nonflexible FE-based model difficult to be adopted by designer engineers. Moreover, although modelling fibres discretely using probabilistic techniques such as Monte Carlo is aimed to account for the random distribution of fibres (Cunha et al., 2011), its usage can itself be unrealistic as the prediction might as well be likely to completely miss the actual distribution and location of fibres in a particular structural element. Hence, modelling fibres as part of the concrete matrix as shown in a number of design guidelines can offer an easier and perhaps safer prediction of the behaviour of composite material (Lok and Xiao 1999; RILEM TC 162-TDF 2003; Barros et al., 2005; Di Prisco et al., 2013). However, unlike the discrete 3D modelling, it should be noted that the homogenous fibrous concrete modelling is not aimed to detect local failures on the mesoscale level explicitly such as fibre rupture and concrete fracturing at the IFZ (Zhang et al., 2018).

Structural Behaviour of Steel-Fibre-Reinforced Lightweight Concrete

737

Generally, there are two approaches that can be used in FEM to predict the tension stiffening behaviour of fibrous concrete: the discrete crack approach (r-x) (Ngo and Scordelis, 1967) or smeared crack approach (r-e) (Rashid, 1968). Although more accurate at post-crack, the discrete crack approach can be impractical and numerically intensive to use (Tlemat et al., 2006), while the more accepted smeared crack approach that assumes the crack is smeared over an element can be more useful for design. In this work, a smeared crack approach with (r-x) is adopted using ABAQUS option to define cracking displacement rather than cracking strain. ABAQUS derives the strains based on the characteristic length which is defined as the mesh size for hex elements using e = x/lc. The values of crack width is input into ABAQUS directly from the pullout test. As compared to the strain, the r-x approach offers a number of advantages since it represents the actual behaviour of the fibrous material, is member size-effect independent and can be directly applied to FEM (ABAQUS) from the pullout tests used (De Montaignac et al., 2011). On ABAQUS, mesh sensitivity is not an issue for r-x relation as compared to r-e. Also, r-x relationship can provide the necessary information needed to design for service limit state including fatigue and shrinkage. Concrete damaged plasticity (CDP) was calibrated and chosen to model the notched beam specimens to check the validity of the r-x model. The calibration of CDP on the material level for both cylinder compression test and pullout test can be found with more detail elsewhere (Al-Naimi and Abbas 2019). Table 3 below summarises the CDP parameters used. Table 3. Parameters usage for CDP Dilation angle Eccentricity fb0/fc0 K Viscosity 25 0.1 1.16 0.666 0

The explicit dynamic solver is adopted. The explicit dynamic analysis was found to be a more computationally efficient tool at solving the problems used in this project as compared with the implicit solver which had a tendency to terminate. The following properties of the finite element model were based on a comprehensive convergence study. To ensure a quasi-static solution the ratio between kinetic energy and internal energy was kept below 1% throughout the analysis and any spikes in energy even below 1% were monitored as they could represent an indication of model failure. The analysis was ran using a displacement induced loading rate of 0.1 mm/step in a smoothened step with a mass scaling of 50. The finite element model investigated to validate the structural beams is shown below. Due to the symmetrical setting only a quarter of the beam was modelled with symmetry boundary conditions along the Z and X directions. The beam was restricted from moving in the Y direction by applying a displacement rotation boundary condition (in the initial step) along the middle line of the support. The displacementinduced load was applied (in the explicit dynamic step) on the surface of the loading steel plate in a similar manner to the experiment. To estimate the load, the reaction force was calculated by summing up the load along the boundary condition line on the bottom of the support then multiplied by 4. The displacement, however, was calculated

738

H. K. Al-Naimi and A. A. Abbas

by taking the average Y displacement of the surface or line of the beam in middle point of the total span of the full beam. With regards to meshing the concrete, hexagonal brick element C3D8R of 20 mm element size with reduced integration and hourglass control was used. For the reinforcement stirrups and rebars, wire elements T3D2 were embedded in concrete with true stress-strain material properties based on tensile tests of the rebars. The supports and load blocks were made to be rigid and a tie coupling to the concrete was applied Fig. 7.

Fig. 7. FE model used for the reinforced concrete beams with boundary conditions for beam of S = 120 mm (ABAQUS)

5 Results and Discussion 5.1

Experimental Results

Figure 8 below shows the cracking pattern for the beams tested. The beam with no stirrups fails in shear in an identical patter to a-. Aside from a- which fails in shear and f- which fails in flexure all the beams show a shear-flexure failure.

Fig. 8. The tested structural beams

Structural Behaviour of Steel-Fibre-Reinforced Lightweight Concrete

739

Figure 9 below shows the load-deflection curves for the 7 reinforced concrete beams. For all the beams illustrated in the figure, it can be seen that the addition of fibres increases the load bearing capacity and ductility. The beam which failed in shear (solid brown line), failed in a ductile manner once fibres are added (dashed brown line). The interval o-t indicates can be seen between 0 kN to 10−12 kN for all the beams tested confirming that the effect of fibres on stiffness in lightweight concrete can only be activated after cracking takes place. This was detailed in an accompanying paper. At point t tension cracking takes place followed by tension steel rabar yielding at y. The maximum load is denoted by the letter m while the ultimate load is denoted by the letter u characterized by either complete failure (for beams failing in shear) or 85% of maximum load at post peak (for beams failing in flexure).

Fig. 9. Load-deflection for the tested beams

Table 4 below details the load and deflection at yield y, peak p, and ultimate u. It is evident that all the 3 types of loads increase with the increase of either fibre dosage or reduction of stirrups spacing. It is interesting to see that the maximum load bearing capacity of beam S = 240 mm, Vf = 1% is not very far from that of beam S = 120 mm, Vf = 0% while similarly the maximum load bearing capacity of beam S = 240 mm, Vf = 2% is close to that of beam of S = 120 mm, Vf = 1%. For all these beams, the higher the Vf the higher the ductility regardless of stirrup spacing with the beam with S = 120 mm, Vf = 2% recording the highest while the beam with S = 240 mm, Vf = 0% recording the lowest. It can be deduced from the latter that the maximum load bearing capacity can be maintained while the ductility improved by doubling the stirrups spacing and adding fibre volume fraction of approximately 1%.

740

H. K. Al-Naimi and A. A. Abbas

The beam with S = ∞, Vf = 1% bears a maximum loading slightly higher than that of S = 240 mm, Vf = 1%. This slight difference can be attributed to variances in the compressive strength. A high fcm consequently leads to stronger fibre-matrix interfacial bond and thus a higher residual tensile strength of the beam. This is seen with the beam of S = ∞, Vf = 1% which records an average fcm of 36 MPa as compared to that of S = 240 mm, Vf = 1% which has an fcm of 31 MPa. However, it is vital to notice that the beam with S = ∞, Vf = 1% still failed in a brittle shear manner and recorded a ductility as low as that of S = 240 mm, Vf = 0%. The latter beam also failed in shear which proves the adequacy of the adopted methodology. Figure 10 below shows both the strength and ductility ratios. Both control ductility and load bearing capacity were chosen as those of the adequately designed beam with S = 120 mm, Vf = 0%. Table 4. Load bearing capacity and ductility details. dy S Vf (mm) (%) (mm) 240 0 0.67 1 0.76 2 0.62 120 0 0.77 1 0.69 2 0.50 ∞ 1 0.77 *S=Shear F=flexure

Py dp Pm (kN) (mm) (kN) 27.7 1.37 32.1 30.1 3.90 59.3 32.2 5.17 78.4 28.3 3.95 51.8 31.1 4.32 78.1 33.8 3.69 92.7 30.2 3.11 62.5 SF=shear flexure

du (mm) 1.86 4.96 9.59 5.31 5.79 24.3 3.56

Pu (kN) 27.3 50.4 66.8 44.1 66.4 78.8 53.4

Max crack spacing (mm) 79.4 60.4 53.1 79.0 60.1 53.3 60.2

Nature of failure S SF SF SF SF F S

lu 2.77 6.61 15.41 6.91 8.41 48.24 4.64

Ductility Ratio μu (Solid line) and Strength Ratio P0 (dashed line) 8 S=120mm

Ductility Ratio (μu) Strength Ratio (P0)

7

S=240mm

6

S=∞

5

S=120mm S=240mm

4

S=∞

3 2 1 0 0

1 Vf (%)

Fig. 10. Strength and ductility ratios

2

Structural Behaviour of Steel-Fibre-Reinforced Lightweight Concrete

5.2

741

Numerical Results

Figure 11 below illustrates the cracking patterns of the beam with S = 120 mm, Vf = 2% which failed in flexure and that of S = ∞, Vf = 1% which failed in shear on ABAQUS.

Fig. 11. At ultimate deflection: on the left: beam failing in flexure, on the right: beam failing in shear

Validation of the tested SFRLC beams is illustrated in Fig. 12 below. It can be seen that the model adopted closely resembles the load-deflection behaviour of the SFRLC beams tested. Once high spikes were noticed in the KE/IE graph on ABAQUS, the simulation was stopped since this was an indication of the instability of the solution beyond that point as it represents a non-quasi-static behaviour. Both load bearing capacity and ductility were accurately and safely predicted by ABAQUS.

Behaviour of fibrous beams using the proposed constitutive material model on ABAQUS

100

EXP (S=120mm. V=1%) ABAQUS (S=120mm, V=1%) EXP (S=120mm. V=2%) ABAQUS (S=120mm. V=2%) EXP (S=240mm. V=1%) ABAQUS (S=240mm. V=1%) EXP (S=240mm, V=2%) ABAQUS (S=240mm, V=2%) EXP (S=∞, V=1%) ABQUS (S=∞, V=1%) EXP (S=∞, V=2%) ABAQUS (S=∞, V=2%)

90 80

Load (kN)

70 60 50 40 30 20 10 0 0

2

4

6

8

Displacement (mm)

Fig. 12. Validation of the SFRLC beams using the suggested constitutive model

742

H. K. Al-Naimi and A. A. Abbas

6 Conclusions • The methodology used was successful at predicting the behaviour of fibrous lightweight concrete beams with and without web reinforcement. • The addition of fibres can alter the behaviour of beams from brittle shear to ductile flexural such as the case with the deficiently reinforced beam of S = 240 mm and Vf = 0%. However, the latter statement only remains valid provided that a sort of minimum reinforcement is provided. • It is safe to deduce that adding fibres of Vf = 1% permits the stirrups spacing to be doubled without compensation of strength or ductility. In fact the behaviour of beam with S = 240 mm, Vf = 2% was more favourable in terms of ductility than the beam with S = 120 mm, Vf = 1%, however the maximum load bearing capacity Pm remained nearly identical for both beams. • The numerical NLFE results show excellent correlation between the proposed constitutive model predictions and experimental results.

References Gerritse, A.: Design considerations for reinforced lightweight concrete. Int. J. Cem. Compos. Lightweight Concr. 3(1), 57–69 (1981) Libre, N., Shekarchi, M., Mahoutian, M., Soroushian, P.: Mechanical properties of hybrid fiber reinforced lightweight aggregate concrete made with natural pumice. Constr. Build. Mater. 25 (5), 2458–2464 (2011) Campione, G.: Flexural and shear resistance of steel fiber-reinforced lightweight concrete beams. J. Struct. Eng. 140(4), 04013103 (2014) Dias-da-Costa, D., Carmo, R., Graça-e-Costa, R., Valença, J., Alfaiate, J.: Longitudinal reinforcement ratio in lightweight aggregate concrete beams. Eng. Struct. 81, 219–229 (2014) Mo, K., Goh, S., Alengaram, U., Visintin, P., Jumaat, M.: Mechanical, toughness, bond and durability-related properties of lightweight concrete reinforced with steel fibres. Mater. Struct. 50(1), 46 (2017) Lytag: Ramboll Frame Comparison Study (2014). https://www.aggregate.com/our-businesses/ lytag, Accessed 31 Dec 2019 Chen, H., Huang, C., Tang, C.: Dynamic Properties of Lightweight Concrete Beams Made by Sedimentary Lightweight Aggregate. J. Mater. Civ. Eng. 22(6), 599–606 (2010) Badogiannis, E., Kotsovos, M.: Monotonic and cyclic flexural tests on lightweight aggregate concrete beams. Earthquakes Struct. 6(3), 317–334 (2014) Lim, H.S., Wee, T.H., Mansur, M.A., Kong, K.H.: Flexural behavior of reinforced lightweight aggregate concrete beams. In: Asia-Pacific Structural Engineering and Construction Conference. Kuala Lumpur: APSEC, pp. 68–82 (2006) Wu, C., Kan, Y., Huang, C., Yen, T., Chen, L.: Flexural behavior and size effect of full scale reinforced lightweight concrete beam. J. Mar. Sci. Technol. 19(2), 132–140 (2011) Gao, J., Sun, W., Morino, K.: Mechanical properties of steel fiber-reinforced, high-strength, lightweight concrete. Cem. Concr. Compos. 19(4), 307–313 (1997)

Structural Behaviour of Steel-Fibre-Reinforced Lightweight Concrete

743

Campione, G., La Mendola, L.: Behavior in compression of lightweight fiber reinforced concrete confined with transverse steel reinforcement. Cem. Concr. Compos. 26(6), 645–656 (2004) Abbas, A., Syed Mohsin, S., Cotsovos, D.: Seismic response of steel fibre reinforced concrete beam–column joints. Eng. Struct. 59, 261–283 (2014a) Di Prisco, M., Colombo, M., Dozio, D.: Fibre-reinforced concrete in fib Model Code 2010: principles, models and test validation. Struct. Concr. 14(4), 342–361 (2013) Grabois, T., Cordeiro, G., Filho, R.: Fresh and hardened-state properties of self-compacting lightweight concrete reinforced with steel fibers. Constr. Build. Mater. 104, 284–292 (2016) Ritchie, A., Kayali, O.: The effects of fiber reinforcement on lightweight aggregate concrete. In: Neville, A. (ed.) Proceedings of RILEM Symposium on Fiber Reinforced Cement and Concrete, The Construction Press Ltd, pp. 247–256 (1975) Swamy, N., Jones, R., Chiam, A.: Influence of steel fibers on the shear resistance of lightweight concrete i-beams. ACI Struct. J. 90(1), 103–114 (1993) Kang, T., Kim, W.: Shear Strength of Steel Fiber-Reinforced Lightweight Concrete Beams, pp. 1386–1392. Korea Concrete Institute, Oklahoma (2010) Iqbal, S., Ali, A., Holschemacher, K., Bier, T.: Mechanical properties of steel fiber reinforced high strength lightweight self-compacting concrete (SHLSCC). Constr. Build. Mater. 98, 325–333 (2015) Habbitt, Karlsson and Sorensen Inc. Abaqus User’s Manual, vol. II, version 6.1, pp. 11.5.1.1– 11.5.1.14, USA (2000) Lok, T.-S., Xiao, J.R.: Flexural strength assessment of steel fiber reinforced concrete. J. Mater. Civ. Eng. 11(3), 188–196 (1999) Kotsovos, M.D., Pavlovic, M.N.: Ultimate Limit-State Design of Concrete Structures: A New Approach, pp. 31–164. Thomas Telford, London (1999) Sadoon, A., Rees, D.W.A., Ghaffar, S.H., Fan, M.: Understanding the effects of hooked-end steel fibre geometry on the uniaxial tensile behaviour of self-compacting concrete. Constr. Build. Mater. 178, 484–494 (2018) The Concrete Society. Guidance for the Design of Steel-Fibre-Reinforced Concrete. Technical Report No. 63. Cement and Concrete Industry (2007) Lytag, : Technical manual. Lytag Ltd, London (2011) Zienkiewicz, O.C., Taylor, R.L.: The Finite Element Method for Solid and Structural Mechanics, 6th edn. Butterworth-Heinemann, Oxford, UK (2005) Tlemat, H., Pilakoutas, K., Neocleous, K.: Stress-strain characteristic of SFRC using recycled fibres. Mater. Struct. 39, 365–377 (2006) Syed Mohsin, S.M.: Behaviour of Fibre-reinforced Concrete Structures under Seismic Loading. PhD thesis, Imperial College London, London, UK (2012) Abbas, A., Syed Mohsin, S., Cotsovos, D., Ruiz-Teran, A.: Shear behaviour of steel-fibrereinforced concrete simply supported beams. Proc. Inst. Civ. Eng.-Struct. Build. 167(9), 544– 558 (2014b) Behinaein, P., Cotsovos, D.M., Ali, A.A.: Behaviour of steel-fibre-reinforced concrete beams under high-rate loading. Comput. Concr. 22(3), 337–353 (2018) Cunha, V., Barros, J., Sena-Cruz, J.: Tensile behavior of steel fiber-reinforced self-compacting concrete. In: Fiber-Reinforced Self Consolidating Concrete: Research and Applications (ACI SP-274), Detroit, USA: American Concrete Institute, pp. 51–68 (2010) RILEM TC 162-TDF.: r-e design method: final recommendation. Mater. Struct. vol. 36, pp. 560–567 (2003)

744

H. K. Al-Naimi and A. A. Abbas

Barros, J., Cunha, V., Ribeiro, A., Antunes, J.: Post-cracking behaviour of steel fibre reinforced (2005) Zhang, Y.J., Huang, Z.J., Yanga, S.L., Xua, X.W.C.: A discrete-continuum coupled finite element modelling approach for fibre reinforced concrete. Cem. Concr. Res. 106(2018), 130– 143 (2018) Ngo, D., Scordelis, A.C.: Finite element analysis of reinforced concrete beams. J. ACI 64(3), 152–163 (1967) Rashid, Y.R.: Ultimate strength analysis of prestressed concrete pressure vessels. Nucl. Eng. Des. 7(4), 334–344 (1968) De Montaignac, R., Massicotte, B., Charron, J.-P., Nour, A.: Design of SFRC structural elements: post-cracking tensile strength measurement. Mater. Struct. 45(4), 609–622 (2012) Al-Naimi, H., Abbas, A.: Ductility of steel-fibre-reinforced lightweight concrete. In: Eccomas Proceedia COMPDYN (2019), Crete, Greece, pp. 4009–4023, 24–26 June 2019. www. eccomasproceedia.org, Accessed 31 Dec 2019 Kodur, V., Solhmirzaei, R., Agrawal, A., Aziz, E.M., Soroushian, P.: Analysis of flexural and shear resistance of ultra-high performance fiber reinforced concrete beams without stirrups. Eng. Struct. 174, 873–884 (2018)

Experimental Analysis of Beams Produced in Self-compacting Concrete Reinforced with Different Contents of Steel Fibres Ana R. L. Pires(&), Rafael R. Polvere, Sidiclei Formagini, and Andrés B. Cheung FAENG/UFMS, Federal University of Mato Grosso do Sul, Campo Grande, Brazil [email protected]

Abstract. The insertion of steel fibres into the concrete matrix can provide improvements in the control of crack propagation and increase on stiffness of structural elements. This research aims to evaluate the behavior of reinforced concrete beams in different percentages of carbon steel fibres subjected to bending. So, three reinforced beams were made with two 12.5 mm bars on the lower edge, in the dimensions of (12.5  25.5  180 cm), which were tested by four-point flexural test. One beam was used as reference, without the presence of steel fibres, and another’s were added carbon steel fibres in different volumetric fractions, 0.5% and 1.0%. During the flexural tests, measurements were made of the vertical displacements in the middle of the span, also the deformations in the longitudinal reinforcement and in the compressed region of the concrete. Results showed that the insertion of the metal fibres gave a better performance on crack control, lower deflections and deformations on the compressed region and longitudinal armour of the concrete as well as promoted slight gain in the load capacity of the beam. Keywords: Fibrous self-compacting concrete Four-point flexural


Reinforced concrete beams


1 Introduction The use of fibrous self-compacting concrete seeks to improve the properties of conventional concrete (CC). In its fresh state, self-compacting concrete (SCC) provides greater functionality to the material due to its fluidity and moderate viscosity. The fresh mixture has the ability to fill all the spaces of a mold in a uniform manner, under its own weight without the need for vibration [1]. Thus, there is a more uniform dispersion of the fibers in the structural elements, in addition to a better alignment of the fibers along the flow of fresh concrete [2]. The SCC has a structure denser than a CC, due to the large amount of fines material. Thus, there is a reduction in the presence of voids to the point of providing a better adhesion between concrete and steel [3, 4]. The introduction of steel fibres into the SCC provides more benefits and application possibilities, which makes the material more efficient both in the fresh and hardened state. The enhanced properties with the incorporation of fibers are tensile mechanical © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 745–756, 2021. https://doi.org/10.1007/978-3-030-58482-5_66

746

A. R. L. Pires et al.

parameters, ductility and energy absorption capacity under impact [5, 6]. In addition, this incorporation can delay the propagation of cracks, which provides a longer life and durability to the material [7]. Steel fiber reinforced concrete (SFRC) is commonly used in structural engineering, in road pavements and lining. The fibers exert their main effect in the state of post-crack, intercepting the progressions of the microcracks, in order to avoid a sudden rupture [8]. Connecting bridges are formed that serve to transfer the tensile forces through the crack. The ability of SFRC to transfer tensile forces across cracked sections decreases the crack spacing and increases the tension stiffening, both of which contribute to an improved crack control [9]. According to Chenkui and Guofan [10], the fibers tend to settle at the interface between the large aggregate and the mortar matrix, where the crack is more likely to develop. When steel fibers are added to reinforced concrete, the steel bars are more protected against external agents since the fibres control the cracking of the concrete matrix when the traction is requested [11]. There are several studies that evaluate the influence of steel fibres in the cracking of structural elements. In SFRC, Abdalkader et al. [12] showed that the addition of 1.0% resulted in about 81% reduction in maximum crack width compared to concrete beam without fiber. In steel fiber reinforced expanded-shale lightweight concrete (SFRELC) beams [13], the fibers effectively restricted the extension and opening of cracks. And also, the flexural capacity of reinforced SFRELC beams increased 23.3% with steel fiber volume fraction varied from 0% to 1.6%. Yoo and Moon [14] have reported to the implication of steel fibers on the flexural behavior of reinforced concrete beams with reinforcement ratios below the minimum reinforcement ratio and first evaluates feasibility of replacing some of conventional steel rebars to discontinuous steel fibers. They observed that by including steel fibers, the crack propagation into the compressive zone was effectively limited and more flexural cracks, indicating better stress redistribution, were formed by the steel fibers. In SFRC beams subjected to bending, Ashour et al. [15] observed their mechanical behavior and verified that with the increase in the volume of steel fibres, they became larger: the moment of cracking of the beam; the moment of beginning of flow of the longitudinal reinforcement; and, the last moment. OH [16] studied the behavior of reinforced concrete beams with addition of fibrous reinforcement, varying the volume of steel fibres from 0 to 2%, achieving better results of ultimate resistance to bending, ductility and the ability to absorb energy with the increase of the amount of fibres. In real-scale beams, Meda et al. [17] have evidenced that the fibres significantly enhance the behavior at service conditions by increasing the stiffness in the cracked-stage and, therefore, by limiting the crack openings and deformations. The aim of this article is to evaluate the influence of the different content addition of carbon steel fibres in the self-compacting concrete, in relation to the mechanical performance of beams submitted to simple bending by means of an experimental analysis. Special attention is given for the general ductility and distribution of the cracks of beams. In addition, the deflections and deformations in the longitudinal bars and compressed region of the concrete are analyzed for each beam according to the applied load.

Experimental Analysis of Beams

747

2 Experimental Program In this study, three steel fibers reinforced self-compacting concrete (SFRSCC) beams with dimensions of 12.5  25.5  180 cm3 (base, height and length) were produced. The first beam was used as reference (RB), with SCC without the incorporation of steel fibres and the others (B1 and B2) were made with the presence of fibres in different volumetric fractions (Vf ) of 0.5% e 1.0%, respectively. 2.1

Materials and Concrete Mixture

The dosage and characterization were obtained in a previous study [18], using normal Portland cement type CP II 32, silica fume, natural quartz sand, gravel sand, gravel of basaltic origin (characteristic maximum size of 12.5 mm) and a polycarboxylate based superplasticizer (FORT FLOW 2558). A water/cement ratio of 0.5 was adopted to achieve a concrete compressive strength of around 40 MPa. Steel fibres chosen have hook anchorage at the ends, having a length of 30 mm and a diameter of 0.60 mm (aspect ratio L/D = 50). The reinforcement steel of the beams was of type CA-50 and CA-60 for transverse reinforcement, with resistance of 50 kN/cm2 and 60 kN/cm2, respectively. Bars with nominal diameters of 6.3 mm, 5.0 mm and 12.5 mm were used. 2.2

Reinforced Concrete Beams

First, the preparation of the beams involved cutting and installing the steel reinforcement (Fig. 1). The longitudinal reinforcement of the beams was composed of two steel bars with a diameter of 12.5 mm (N1) and the transverse reinforcement consists of 14 stirrups made of steel with a diameter of 5 mm (N2). The stirrups were spaced every 10 cm between the supports and the load application, where the shear action is present. In the central section, between the two loading application points were placed two stirrups with spacing of 20 cm. In addition, a constructive reinforcement with a diameter of 6.3 mm (N3) was used. For the measurement of deformations (steel and concrete), at each point of interest four electric strain gauges were used as shown in the diagram in Fig. 2. Strain gauges were placed on the longitudinal bars (E1) and compressed region of concrete (E2). The strain gauges were placed in the longitudinal and transverse direction (½ Wheatstone Bridge circuit configuration). After bonding, the terminals were connected to the waiting conductors for later connection to the data acquisition system. Strain gauges and their terminals have been protected with silicone to prevent direct contact of the terminals with the metal surface of the armature. In the end, self-fusing tape was used to harden and increase the insulation resistance, and tape was used to isolate the strain gauges and avoid contact with water during concreting. Prior to concreting, the beam molds were lubricated with mineral oil to facilitate the demoulding process of the parts. The cover of the reinforced beams was ensured by the use of spacers on the sides and bottom of the beam form. The beams were removed after 5 days of concreting (during this period only an upper surface remained exposed

748

A. R. L. Pires et al.

Fig. 1. Geometry of the beam (measurements in mm).

Fig. 2. Strain gauges Placement.

to the environment). The concrete beam remained exposed to relative laboratory conditions. Four days before the test, the joint compound was applied to the faces of each beam to obtain a better response to application requests. In addition to the beams, for each concrete produced were moulded: cylindrical specimens (/10  20 cm) for simple compression tests, tensile compression diametral and modulus of elasticity; and, prismatic specimens with dimensions of 15  15 50 cm3 for performing four-point flexural tests.

Experimental Analysis of Beams

2.3

749

Four-Point Flexural Test

To study the mechanical behavior of beams (RB, B1 and B2) with different percentages of steel fibers, a four-point bending test was performed. The static scheme used simulates a bi-supported beam configuration in which the four-point bending test beams had a total length of 180 cm and a clearance of 5 cm between the support shaft and the end, with a clearance of 170 cm. Concentrated loads were applied equidistant at 56.5 cm from the supports (Fig. 3).

Fig. 3. Static scheme.

Fig. 4. Illustration of flexural test (measurements in cm).

750

A. R. L. Pires et al.

In each beam, a 5.0  5.1 cm checkered mesh was demarcated on one of the side faces. The strain gauges of the compressed concrete region were glued at 5,1 cm from the upper edge, 72 h before the flexural test. The beams were tested by a four-point bending test, with load applied to the thirds of the span, checking the evolution of the cracks, reading the deflection by dial gauges and reading the deformations using strain gauges. Deformations, deflections and loads were recorded on HBM’s QuantumX data acquisition system. The method of the test was divided into the following steps: installation of support devices, concrete beam positioning, transmission beam support devices installation, transmission beam positioning, hydraulic piston positioning, load cell positioning, plumb check, marking and positioning of measuring devices. The position of the instrumentation used in the assay can be identified in the illustration of Fig. 4. Vertical loads were incrementally applied by two hydraulic pistons installed in a reaction frame. The marking of the cracks on the side faces of the beam was performed using load increments of approximately 3 kN.

3 Results and Discussion 3.1

Mechanical Properties of Concrete Mixtures

Table 1 shows the mechanical properties of concrete used at 28 days from a previous study [18], referring to the simple compression tests - fcm, splitting test tensile - fct,sp, elastic modulus - Em, and tensile strength tests at flexion - fct,f. The concrete nomenclature are SCC (standard concrete), SFRSCC 0.5% (0.5% of steel fibres in volume addition) and SFRSCC 1.0% (1.0% of steel fibres in volume addition). Table 1. Mechanical properties of the concrete used at 28 days. Properties fcm [MPa] fct,sp [MPa] Em [MPa] fct,f [MPa]

SCC 41.8 3.69 26.7 3.79

SFRSCC0.5% 45.6 5.71 30.5 3.87

SFRSCC 1.0% 41.5 5.06 29.4 5.15

In concretes with fibre incorporation there was an increase of the modulus of elasticity and tensile strength by diametral compression because the fibers addition can limite the transverse strain of concrete. The results of the compression tests indicated a decrease in the simple compressive strength for SFRSCC 1.0%. It is also observed that the tensile strength in prism bending is greater the higher the content of steel fibres. This can be explained by the fact that with a higher content there is a greater number of steel fibres acting as a stress transfer bridges along the crack, which increases the postcracking reinforcement of the concrete.

Experimental Analysis of Beams

3.2

751

Cracks Patterns and Ultimate Strength

The incorporation of steel fibres in the self-compacting concrete increased the ultimate loads (Pu .), as well as the loading of the first crack (Pcrack .) of beams B1 and B2. Figure 5 illustrate the evolution of cracks in each beam tested. In beams B1 and B2, the flexural test was terminated shortly before the rupture. As a result, the load registered at the end of the test was considered as the last load of the beam, since it no longer presented growth.

(a) RB

(b) B1

(c) B2

Fig. 5. Evolution of cracks in the beams.

In all three samples the first crack occurred in the central region, between the applied loads where the bending is maximum. The rupture of the RB beam occurred in a ductile form by flexion, with the crushing of the concrete in the compressed side and another side with yield strain of the reinforcement flow in the tensioned region. At the moment of rupture, there was a greater presence of multiple cracks in relation to the

752

A. R. L. Pires et al.

beams with the addition of metallic fibres. In addition, significant typical shear cracks appeared in the region between the concentrated loads and the supports. The inclined cracks almost reached the upper edge of the beam. It can be noticed that cracks result smaller with increasing fiber content; they become more diffused, also with few shear inclined cracks. It’s noted that in steel fibre reinforced beams, the propagation velocity of the cracks decreases in the concrete, which makes the material, according to [19] exhibit a pseudo-ductile behavior, that is, it presents a certain bearing capacity post-cracking. The reduction of the inclined cracks in the fibrous concrete beams may indicate an action of these fibres in the absorption of part of the shear, sewing these cracks that appear in the beam regions between the load and the support. It’s also visually observed that the crack spacing is also influenced, becoming slightly larger as the fiber volume is increased. 3.3

Load-Deflection Relationship

The experimental results, in terms of load (P) versus deflection for mid-span measurements of self-compacting without fibres concrete beams are shown in Fig. 6. And also, Table 2 shows the deflections at the critical points - cracking (dcrack ), yield (dy ) and ultimate (du ). 140 120

Load P(kN)

100 80 60

RB (SCC)

40

B1 (SFRSCC 0.5%)

20 B2 (SFRSCC 1.0%) 0 0

5

10

15

20

Deflection (mm) Fig. 6. Load-Deflection Curves of the beams.

The B1 beam had a last load 4.3% higher than the RB beam made without the addition of steel fibres. And B2 beam indicated a 10.3% increase in the last load. This shows that the increase in ultimate load was proportional to the increase in the amount of steel fibres used. This shows that the increase in ultimate load was proportional to the increase in the amount of steel fibres used. In relation to the appearance of the first crack, the B2 beam showed an approximately 63% increase in the first crack load compared to the crack initiation load applied to the RB beam. The B1 beam presented a cracking load 5.3% lower than the RB. In addition, there are increases of 8.3% and 23% of the yield load for RB beam compared to beam B1 and B2, respectively.

Experimental Analysis of Beams

753

Table 2. Deflections and loads at critical points. Beams Pcrack (kN) dcrack (mm) Py (kN) dy (mm) Pu (kN) du (mm) RB B1 B2

19 18 31

1.0 0.85 1.20

97 105 119

8.7 8.0 7.85

116 121 128

12 9.7 18

The concrete beam with higher volumetric fraction of steel fibres indicated greater contributions to reduce tensile stresses in reinforcement bars. It is noted that B2 presents its first crack with a deformation 41.2% greater than that of RB beam. And, the ultimate load of the reference beam, occurred when the beam had deflection of approximately 12 mm, while in B2 for the last load the deflection was of approximately 18 mm, which represents an increase of 50%. B2 beam presented more ductile behavior, corresponding to a greater ability to absorb energy. 3.4

Load-Deformation Relationship

Figures 7 and 8 show the curves obtained with the measurements of the compressed concrete deformation in the beams and deformation in the longitudinal bars, respectively. 140 120

Load P(kN)

100 80 60 RB (SCC)

40

B1 (SFRSCC 0.5%) 20 B2 (SFRSCC 1.0%) 0 -1.4

-1.2

-1.0

-0.8

-0.6

-0.4

-0.2

0.0

Deformation [‰] Fig. 7. Load-Deformation Curves of the compressed concrete region.

Through the analysis of the curves obtained with the measurements of the compressed concrete deformation in the beams (Fig. 7), it was found that the incorporation of steel fibres in the self-compacting concrete gave the beams B1 and B2 smaller deformations in the compression region for the beams with the same load level when compared to RB. This can be explained by the fact that the modulus of elasticity presented a lower value for the concrete of the RB beam, as presented in Table 1. It’s observed that the higher the fiber percentage, the better the concrete behavior in the

754

A. R. L. Pires et al. 140 120

Load P(kN)

100 80 60

RB (SCC)

40

B1 (SFRSCC 0.5%)

20 B2 (SFRSCC 1.0%) 0 0

1

2

3

Deformation [‰]

4

5

6

Fig. 8. Load-Deformation Curves of longitudinal bars.

compressed region. The effect of fiber addition was significant as beam B2 withstood greater deformation with higher loading than other beams. It’s noteworthy that the strain gauges fixed in the concrete detached at the end of the bending test on the beams RB and B1. The results of the measurements made in the longitudinal reinforcement (Fig. 8) differed from some indications in the literature and other experimental studies, since the deformations in steel are greater for the fiber reinforced beams in the initial phase given the same loading level. However, in the final phase of the carrying capacity, beams B1 and B2 achieved higher load capacities for the same longitudinal bar deformation. This may indicate that part of the tensile stresses were absorbed by the fibres immersed in the concrete, which absorb the tensile stresses during bending, thus relieving the action of the main reinforcement.

4 Conclusions In this research, the beams made with the incorporation of fibres presented a better structural behaviour. The following conclusions can be drawn, based on the analysis of the experimental results: • A greater amount of load was achieved in the fibrous concrete beams, with less deformation in the compressed concrete region. Smaller deflections mid-span to the same load level on beams B1 and B2. In addition to lower deflections mid-span to the same loading level. • The post-cracking behaviour in the beams with the presence of fibres showed greater ductility, absorbing more energy. Factors that may be interesting for the design of structures subjected to earthquake and dynamic load.

Experimental Analysis of Beams

755

• The self-compacting fibrous concrete beams presented a better crack control, with higher first crack load for the B2 beam. Visually, a better distribution of these cracks is also observed, with smaller openings. And the beam of 1.0% of fibres showed a significant increase of 10.3% higher than the reference beam in the final load. The costs of steel fibre self-compacting concrete structures and flexural reinforcement may be higher, but overall, with the use of steel fibres you can get great advantages such as: lower maintenance cost and longer durability the structure.

References 1. Okamura, H., Ouchi, M.: Self-compacting concrete. J. Adv. Concr. Technol. 1, 5–15 (2003) 2. Vandewalle, L., Heirman, G., Van-Rickstal, F.: Fiber orientation in self compacting fibre reinforced concrete. In: Proceedings of the 7th International RILEM Symposium on Fibre Reinforced Concrete: Design and Applications (2008) 3. Almeida Filho, F.M.: Contribuição ao estudo da aderência entre barras de aço e concretos autoadensáveis. In: São Carlos, Tese (doutorado)–Escola de Engenharia de São Carlos, Universidade de São Paulo, p. 310 (2006) 4. Hossain, K.M.A., Lachemi, M.: Bond behavior of self-consolidating concrete with mineral and chemical admixtures. J. Mater. Civ. Eng. ASCE 20(9), 608–616 (2008) 5. Katzer, J., Domski, J.: Quality and mechanical properties of engineered steel fibres used as reinforcement for concrete. Constr. Build. Mater. 34, 243–248 (2012) 6. Paja˛K, M., Kühn, T.: The influence of steel fibers on the dynamic response of self compacting concrete. Key Eng. Mater. 711, 179–186 (2016) 7. Falkner, H., Huang, Z., Teutsch, M.: Comparative study of plain and steel fibre reinforced concrete ground slabs. Concr. Int. 17(1), 45 (1995) 8. Sorelli, L., Meda, A., Plizzari, G.A.: Steel fibre concrete slabs on grade: a structural matter. ACI Struct. J. 103(4), 551 (2006) 9. Bischoff, P.: Tension stiffening and cracking of steel fiber-reinforced concrete. J. Mater. Civ. Eng. 15(2), 174–182 (2003) 10. Chenkui, H., Guofan, Z.: Properties of steel fibre reinforced concrete containing larger coarse aggregate. Cem. Concr. Compos. 17, 199–206 (1995) 11. Noghabai, K., Noghabai, K.: Effect of various types of fibers on bond capacity– experimental, analytical, and numerical investigations. Struct. Appl. Fiber Reinforced 182, 109 (1999) 12. Abdalkader, A., Elzaroug, O., Abubaker, F.: Flexural cracking behavior of steel fiber reinforced concrete beams. J. Sci. Technol. Res. 06(08), 6 (2017) 13. Zhao, M., Li, C., Su, J., Shang, P., Zhao, S.: Experimental study and theoretical prediction of flexural behaviors of reinforced SFRELC beams. Constr. Build. Mater. 208, 454–463 (2019) 14. Yoo, D., Moon, D.: Effect of steel fibers on the flexural behavior of RC beams with very low reinforcement ratios. Constr. Build. Mater. 188, 237–254 (2018) 15. Ashour, S.A., Wafa, F.F., Kamal, M.I.: Effect of concrete compressive strength and tensile reinforced ratio on the flexural behavior of fibrous concrete beams. Eng. Struct. 22, 1133– 1146 (2000)

756

A. R. L. Pires et al.

16. Oh, H.B.: Flexural analysis of reinforced concrete beams containing steel fibers. J. Struct. Eng. ASCE 118(10), 18 (1992) 17. Meda, A., Minelli, F., Plizzari, G.: Flexural behaviour of RC beams in fibre reinforced concrete. Compos. Part B: Eng. 43(8), 2930–2937 (2012) 18. Polvere, R.R., Pires, A.R.L., Formagini, S., Cheung, A.B.: Mix Design and Properties of Self-Compacting Fibrous Concrete. RILEM-fib X International Symposium on Fibre Reinforced Concrete, Valencia (2020). (in press) 19. Figueiredo, A.D.: Concreto com fibras de aço. Boletim Técnico da Escola Politécnica da USP. Departamento de Engenharia de Construção Civil e Urbana. BT/PCC/260. São Paulo (2000)

Innovation in Durable Segments for CSO Tunnels Ralf Winterberg1(&), Michael R. Garbeth2, and Brian Glynn3 1

BarChip Inc., Tokyo, Japan [email protected] 2 Super Excavators Inc., Menomonee Falls, USA 3 Black & Veatch Corporation, Kansas City, USA

Abstract. The Blacksnake Creek and storm water runoff in St. Joseph, MO, was piped along with sewage in a 100-year old pipe not large enough to carry all the storm water and sewage to the wastewater treatment plant and it overflowed to the Missouri River after most rainstorms. The Blacksnake Creek Storm Water Separation Project will convey storm water directly to the Missouri River. This will reduce water quantity in the existing sewer during storms and the quantity of combined storm water and wastewater overflowing to the river. A 2.74 m ID and 2.0 km long segmental tunnel was constructed as part of the Separation Project. This project is an America’s First, using segments solely reinforced with macro synthetic fibre (MSF). The project further features a challenging umbilical TBM launch in a small diameter secant pile shaft. This paper addresses the solutions to the technical challenges of the project, the design of the segments and the benefits associated with the use of MSF. Keywords: Macro synthetic fibre
Fibre reinforced concrete lining
TBM
Utility tunnel
Sewer tunnel


Segmental

1 Introduction The Blacksnake Creek Storm Water Separation Improvement Project was required as part of the City of St. Joseph, Missouri’s Combined Sewer Overflow (CSO) Long Term Control Plan in order to improve water quality as mandated by the Federal Clean Water Act. The Blacksnake Creek used to be directed into the City’s combined sewer system through a double box culvert, and the creek flow was conveyed through the sewer system to the Water Protection Facility and unnecessarily treated 365 days of the year. During wet weather events, storm water runoff exceeded the capacity of the combined sewer system and caused combined sewer overflows to the Missouri River. The overflows were a mix of storm water and sanitary sewage and resulted in adverse water quality problems. To combat the discharge, the project intercepts and redirects Blacksnake Creek stream flows away from the City’s combined sewer system to a new and dedicated storm water conveyance system that flows directly to the Missouri River, thus reducing the frequency, volume, and impacts of combined sewer overflows to the river. The project was designed by Black & Veatch, Kansas City, MO.

© RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 757–769, 2021. https://doi.org/10.1007/978-3-030-58482-5_67

758

R. Winterberg et al.

The tunnel alignment and profile, shown in Fig. 1 and Fig. 2 respectively, heads west from the Drop Shaft located near the Second Harvest Food Bank, following underneath the Highland Avenue right of way, and curves slightly to the south as it nears Interstate 229. The alignment crosses below both Interstate 229 and the Burlington Northern Santa Fe (BNSF) Railroad Tracks before terminating at an Energy Dissipation Structure. From the Energy Dissipation Structure, flows are conveyed by Roy’s Branch, which is a tributary to the Missouri River.

Fig. 1. Tunnel alignment - plan view

The tunnel alignment is approximately two kilometres in length and terminates at a junction structure located between MacArthur Drive and the BNSF Railroad Tracks. The TBM launch shaft is located at the western end of the alignment, near the BNSF Railroad Tracks.

Fig. 2. Tunnel alignment - profile view

Innovation in Durable Segments for CSO Tunnels

759

The USD 27 million Tunnel Bid Package (TBP), awarded to the contractor Super Excavators, Inc., was a critical portion of the project, as the topography of the area did not allow for construction of a new and dedicated storm water canal via traditional trench excavation without significant impacts to the public. The TBP not only comprised the segmental tunnel, but also an 11 m ID baffle drop structure, a 15 m reinforced concrete box culvert and an energy dissipation structure at the Missouri end of the alignment.

2 Geological Conditions The tunnel was to be constructed in soft ground and mixed ground conditions, including soils and shale, and combinations of both at the interfaces. The soils to be encountered varied from silty clay, silty sand, clayey sand, and sandy clay. The shale encountered is low in strength, has low to medium-low durability, is easy to excavate, and has low hydraulic conductivity. Groundwater along the tunnel varied along with the topography from 12 m to 45 m of groundwater head. The general geography and topography of the project site is located in northwestern St. Joseph, Missouri between St. Joseph Avenue and the Missouri River. The overburden along the tunnel alignment consists primarily of fill material and alluvial deposits with alternating beds of silt, clay, silty-clay, and clayey silt. The minimum overburden on the project occurs at the launch shaft with approximately 9.0 m of cover. Progressing away from the launch shaft at the peak of the topography on Highland Avenue between Main Street and 2nd Street there is 55 m of cover and at the receiving shaft near St. Joseph Ave. this reduces to 18 m. The primary rock conditions predominantly consisted of shale and claystone along the tunnel horizon. A limestone bed was present above the tunnel crown for approximately 1,280 m. The percentage of the alignment for the various different ground conditions consisted of soft-ground (23%), rock (63%), and mixed-ground for 14% of the alignment. The tunnel transitioned between soft ground and rock along each end of the tunnel alignment and mixed-faced conditions were anticipated for reaches along these subsurface transitions. The soil to rock transitions on each end of the alignment were gradual. It was anticipated that the tunnel groundwater inflow will not exceed a steady state of flow greater than 50 GPM (gallons per minute, approx. 190 litres per minute). Figure 3 shows the main parts of TBM Carrie. Considering the ground conditions along the tunnel alignment and the contract specifications, the tunnel had to be excavated with an EPB tunnel boring machine in order to mitigate risks during tunnelling in soft or mixed ground below the water table.

760

R. Winterberg et al.

Fig. 3. EPBM ‘Carrie’ assembly and testing on site

3 TBM Launch A cradle was placed inside the secant pile launch shaft prior to the arrival of the TBM. This cradle consisted of two I-Beams cast in a concrete floor to fit the curvature of the machine. All the shaft and tunnel utilities were established inside the shaft prior to lowering any machine components. A special electrical and hydraulic umbilical assembly was designed and used to help launch the TBM. Note the secant pile launch shaft diameter is 15 m where the entire TBM assembly is 78 m in length. The entire machine was methodically aligned in two rows consisting of Row #1, Main TBM Parts: cutter head, telescopic shield, gripper shield, and tail skin; and Row #2, Backup System: all the gantries and backup equipment including the electrical substation (Fig. 4 left). The TBM cutter head, gripper shield and telescopic shields

Fig. 4. TBM Carrie’s main parts (left) and launch cradle in the secant shaft (right)

Innovation in Durable Segments for CSO Tunnels

761

were nearly completely assembled at the surface due to the ease of assembly at the surface versus in the shaft. A 500 ton mobile crane was mobilized to the project site and used to lower the heavy components into the shaft. Once the major components were lowered down into the shaft, one crew began hooking up the electrical and hydraulic lines while another crew continued to work on assembly of the gantries. The gantries included such items as: the operator’s cabin, electrical substation, air compressor, main drive motors, grease pumps, ground conditioning system, ventilation cassette and cable festoons. The special hydraulic and electrical umbilicals were used to link the gantries on the surface to the main TBM body in the launching cradle on the bottom of the shaft. At this point of time, the TBM could begin the functional testing and troubleshooting process. After finishing assembly of the initial configuration for the machine, the excavation has commenced. During this process, the primary propulsion cylinders pushed off of beams welded to the I-Beam in the cradle. This allowed for the cutter head, telescopic shield and gripper shield to advance forward and penetrate the secant shaft. A short screw conveyor was utilized to control the material in front of the TBM (Fig. 4 right). After the TBM was launched, and most of the components were installed and in the ground behind the TBM, the next step was to disconnect the complete electrical and hydraulic cables/hoses, which allowed the following tasks to be completed: • • • • •

Remove the short launch screw conveyor Install the longer main screw conveyor Install the tail shield Install the segment transporter belly pan Re-hook up the electrical and hydraulic cables/hoses.

During this stage, a precast concrete segmental tunnel liner “half-ring” structure was used to push and advance the TBM further into the excavation (Fig. 5 right). Once the tailskin reached the secant shaft wall the full ring installation was initialized. A support bracket was braced off the secant piles to hold the first segment ring in place while the TBM advanced forward. The gantries were lowered down into the shaft one by one until all nine were advanced directly behind the TBM. Then, full production tunnelling began. The initial tunnel drive installation has proven that the fibre reinforced segments are robust enough to withstand the temporary load cases, i.e. transportation, hoisting, handling and installation, without experiencing damage to the segments. In addition, the half-ring segment sections used in the shaft for the launching process (Fig. 5 right) were inspected after the initial push and no damage occurred to the segments.

762

R. Winterberg et al.

Fig. 5. Launching the TBM (left) and half-ring installation for TBM propulsion (right)

4 Segmental Lining 4.1

Introduction

In the past 20 years Fibre Reinforced Concrete (FRC) has become widely utilized in segmental tunnel linings due to the improved mechanical performance, robustness and durability of the segments. Further, significant cost savings can be achieved in segment production and by reduced repair or reject rates during temporary loading conditions [1]. The replacement of traditional rebar cages with fibres further allows changing a crack control governed design to a purely structural design with more freedom in detailing. Macro Synthetic Fibres (MSF) are non-corrosive and thus ideal for segmental linings in critical environments. The market share of FRC tunnel segments compared to traditionally reinforced segments continuously grows. Recent publications such as the ITA WG2 Report [2], the ITAtech guideline [3], the British PAS 8810 [4], and the FIB state-of-the-art report [5] have given more credibility to the use of this reinforcement type as well as its basis for design to support its application in tunnel lining segments. The use of MSF in tunnel segments has been increasing on a global scale in the tunnel industry, mainly for the durability benefits that the synthetic fibres provide, as compared to steel fibres or rebar cages. A major reference for the use of MSF as sole structural reinforcement for precast tunnel segments is the Santoña–Laredo General Interceptor Collector in Northern Spain [6]. This project demonstrated very robust and satisfactory performance of the MSF reinforced segments even under difficult conditions [1].

Innovation in Durable Segments for CSO Tunnels

763

The tunnel industry in the US has also started to embrace this technology. Based on the successful completion of the final lining of the starter and end tunnels, and successful trials with tunnel segments for the Euclid Creek tunnel in the NEORSD Clean Lake Project [7], the precast concrete segment manufacturer proposed to use MSF for the Blacksnake tunnel project. 4.2

Ring Geometry and Segmentation

The lining for the Blacksnake Creek Storm Water Separation Tunnel consists of a precast concrete segmental tunnel lining. The segmental lining is composed of six trapezoidal fibre reinforced concrete segments with rubber gasket frames inserted between the segments and adjacent rings to create a watertight lining. When assembled, the six trapezoidal segments form a ring with an internal diameter of 2743 mm (Fig. 6).

Fig. 6. Ring segmentation (left) and tunnel ring assembly (right)

The universal rings of 1219 mm nominal length have a taper of 12.7 mm in order to accommodate a turning radius of 300 m on the alignment. Kinematic control during installation is provided by means of dowels and alignment indicators on the segments. The radial joints are connected with galvanized steel bolts. The precast concrete segments were produced by CSI Tunnel Systems Inc.’s plant in Macedonia, Ohio, and were shipped by truck to the project site in St. Joseph, Missouri (Fig. 7). Segments were installed within the tunnel during the excavation process by completing a segment ring with each advancement. Each individual segment was placed using a mechanical segment erector located in the trailing shield of the TBM, and the segments were manually bolted together. Annular grouting was performed through grout ports built into each segment after the rings were erected. BarChip 54 Macro Synthetic Fibre was used to reinforce the precast concrete segments as an alternative to steel fibre or traditional deformed steel bar reinforcement. This alternative was chosen to reduce material costs while ensuring compliance with American Iron and Steel provisions of the Contract Documents. The use of synthetic

764

R. Winterberg et al.

Fig. 7. MSFRC segments in the Macedonia factory (left) and storage on site (right)

fibre reinforcement further eliminates the risk of corrosion and increases the service life of the tunnel lining. This, in turn, reduces future maintenance cost of the tunnel. The Blacksnake tunnel is the first tunnel in North America to be lined with precast concrete segments that are solely reinforced with synthetic fibre reinforcement. To better accommodate the fibre reinforced concrete solution, smaller segments were adopted in order to limit the segments’ aspect ratio. The tunnel has an internal diameter of 2.743 m with a segment thickness of 190.5 mm. The segmentation was selected to be six trapezoidal segments (Fig. 6), each having a developed centre arc length of 1.536 m. This yields a segment aspect ratio of 8.1, which is well below the acknowledged limit of 10 to ensure segment robustness for temporary load cases [3]. 4.3

Lining Structural Design

The MSFRC segments are made using concrete class f’c 7.0 ksi (48 MPa) at 28 days of age, with a specified stripping strength of f’c 2.0 ksi (14 MPa). The required residual strength was specified to be minimum 3.2 MPa at any deflection at or beyond span/600 (0.75 mm through 3.0 mm) according to ASTM C1609 [8]. That means the same high residual strength value was specified for both ULS and SLS in order to guarantee crack width control in service as well as ultimate residual strength. BarChip 54 is the chosen fibre, based on the successful experience from the full scale segment trials carried out in the Euclid Creek tunnel project [7]. A dose rate of 7.0 kg/m3 of this fibre proved sufficient in the pre-production trials to exceed the 3.2 MPa residual strength specification as detailed above. The design approach adopted for the FRC segmental lining in ultimate limit state (ULS) is the use of Normal Force-Bending Moment interaction diagrams or MomentThrust Capacity Limit Curves (Fig. 8). The factored design load couples acting on the section must remain within the N–M envelope [9]. The FRC material properties are herein derived from the ASTM C1609 beam tests, which are eventually used as the basis to determine the stress-strain relationship of the concrete on the tension side. The idealized stress-strain diagram enables setting up the capacity limit curves, which are obtained by equilibrium iterations on the given cross-section.

Innovation in Durable Segments for CSO Tunnels

765

Fig. 8. N-M capacity envelope for ULS

All checks for temporary load cases as well as the checks for serviceability limit states (SLS) have been performed with finite element analysis [10]. These included numerical analysis of the jacking forces on the segmental lining during TBM forward thrust (Fig. 9).

Fig. 9. SLS verification using FEA (Atena)

The material model adopted in the FEA employs the increased fracture energy provided by the fibres [3, 11]. This advanced material model has proven to yield accurate numerical simulations, e.g. as shown in the numerical analysis of the full scale segmental ring testing for the Shanghai Metro extension at the Tongji University in Shanghai [11].

766

R. Winterberg et al.

5 Tunnel Excavation The initial tunnel drive was launched in soft ground, and then transitioned into mixed ground conditions fairly quickly. To control alignment and grade during the launching process, and throughout the tunnel drive, a TACS system was implemented. The TACS system provides continuous information about how the machine axis is aligned with respect to its designed alignment, and in what direction it is moving. The system also determines how the six-piece segmental lining will be installed, dependent on the position of the rings with the taper, which determines the straightness or curvature of the tunnel drive. The TBM used a bulkhead type mechanical segment erector inside the tail skin to erect the six piece segmental ring. It utilized a ball type pick up system, where the ball head bolt is screwed into a threaded socket in the centre of the segments (Fig. 10 left). The erector has a safe working load of 130% of the segment weight. The rotational speed is fully variable even when loaded between 0−2 RPM. The erector is capable of reaching segments over the two inner most rows of the tail seal brushes.

Fig. 10. Mechanical erector inside the tail skin (left) and completed rings in the tunnel (right)

The excavation of the Blacksnake Creek Tunnel project progressed through three distinct geologies. These were identified in the Geotechnical Baseline Report as being soft ground, mixed-ground, and rock conditions. The project tunnelling started out in soft ground conditions, transitioning into mixed-ground and then rock conditions. As the tunnel progressed towards the receiving shaft, the ground transitioned back from rock to mix-ground and then soft ground conditions.

Innovation in Durable Segments for CSO Tunnels

767

The project team did not experience any significantly challenging tunnelling conditions during the transitions between the soil/rock interfaces. However, the production rate increased considerably when excavating in rock conditions. The rock conditions were generally consistent for a longer period of time, thus tunnelling became repetitive for the crew. The biggest challenge for the TBM was controlling steering throughout the alignment in all three distinct geologies due to machine difficulties. The TBM had to be used in exceptionally high pressure mode to overcome steering complications at various times. During these times the fibre reinforced segments remained robust and generally did not step in, crack, or become damaged due to the high pressures. During recent tunnel inspections though, a few segments with fine longitudinal cracks were observed, which were likely due to the machine pushing off the segments in that period. However, the fibres were able to control developing crack widths to remain within the specified serviceability limits of the segmental lining design (wmax = 0.2 mm). No segment had to be rejected. The primary tunnel backfill grout was modified by adding a non-chloride accelerator. This accelerated the set time for the tunnel grout and prevented the segments from floating in the rock section of the alignment. This was important because the survey brackets were attached to the segment inside the tunnel and prevented them from moving.

Fig. 11. TBM breakthrough in October 2019

768

R. Winterberg et al.

The TBM exited from the rock geology back to mixed and soft ground conditions without any issues. However, abnormally high ground water pressure was encountered for a brief period of time in the soft ground. Tunnelling slowed down throughout this brief section while the team overcame the water pressure. The segments were able to withstand the additional water pressure and no leakage occurred. It was anticipated based on the performance of the TBM, robustness of the segments, and ground conditions that were expected to be encountered, that the advancement should be between 14.6 to 18.2 meters per 10 hour shift in full production excavation. The tunnelling advancement was sporadic throughout the drive with the best day of production at 19.5 meters per 10 hour shift in the rock conditions. This equates to installing a bit more than 1.5 rings per hour. Breakthrough was in October 2019 (see Fig. 11) and scheduled project completion is spring 2020.

6 Conclusions • Launching the EPB TBM with an extensive amount of umbilical cords, in a relatively small size shaft compared to the overall length of the TBM, and a unique launching process has been challenging but a lot of knowledge has been gained for future launching of this TBM on other projects. • This extraordinary project confirms that a strong partnership between all parties involved in this project yields an innovative, technically and economically leading solution that will deliver a quality project to the owner. • The use of MSF reinforced concrete segments on the Blacksnake Creek project has demonstrated robust, durable and dependable performance in tunnel projects even under very difficult conditions during the TBM launch phase. The durability and performance of the MSF reinforced segmental lining has outperformed expectations. • Ongoing research and continuous developments on macro synthetic fibre and macro synthetic fiber reinforced concrete have made it today being a modern and costefficient construction material. Eliminating durability issues with regard to corrosion of the primary reinforcement yields significant advantages for the design, since it is no longer governed by serviceability limits. • The successful completion of this America’s First project is expected to build further confidence in MSF reinforced segmental linings. These types of utility tunnelling projects (e.g. sewage and irrigation, power, or gas transfer tunnels) are widely existing in the world market and they present a huge opportunity for MSF reinforced concrete linings, benefiting from the proven advantages.

Innovation in Durable Segments for CSO Tunnels

769

References 1. Winterberg, R., Justa Cámara, R., Sualdea Abad, D.: Santoña–Laredo General Interceptor Collector – Challenges and Solutions. In: Proceedings FRC 2018, Fibre Reinforced Concrete: from Design to Structural Applications, Desenzano, Italy, 28–30 June 2018 (2018) 2. ITA WG2 2016. Twenty years of FRC tunnel segments practice: lessons learnt and proposed design procedure. ITA Report No. 16, ITA Working Group 2 Research, April 2016. www. ita-aites.org 3. ITAtech 2016. Guidance for Precast Fibre Reinforced Concrete Segments – vol. 1 Design Aspects. ITAtech Report No. 7, ITAtech Activity Group Support, April 2016. www.ita-aites. org 4. PAS 8810 2016. Tunnel design – Design of concrete segmental tunnel linings – Code of practice. The British Standards Institution, BSI Standards Ltd (2016) 5. FIB 2017. Precast Tunnel Segments in Fibre-Reinforced Concrete, State-of-the-art report fib WP 1.4.1. fib Bulletin 83, Fédération internationale du béton (fib), October 2017 6. Winterberg, R., Mey Rodríguez, L., Justa Cámara, R., Sualdea Abad, D.: Segmental lining design using macro synthetic fibre reinforcement. In: Proceedings FRC 2018, Fibre Reinforced Concrete: from Design to Structural Applications, Desenzano, Italy, 28–30 June 2018 (2018) 7. Wotring, D.A., Vitale, M.G., Gabriel, D.A.: Synthetic-fiber-reinforced concrete segmental lining - laboratory and field testing program and results. In: Proceedings World Tunnel Congress 2016, San Francisco, USA, 22–28 April 2016 (2016) 8. ASTM C1609/C1609M-12. Standard Test Method for Flexural Toughness of FiberReinforced Concrete (Using Beam with Third-Point Loading), ASTM International, West Conshohocken, PA (2012) 9. Nitschke, A., Winterberg, R.: Performance of macro synthetic fiber reinforced tunnel linings. In: Proceedings World Tunnel Congress 2016, San Francisco, USA, 22–28 April 2016 (2016) 10. JKP 2017. Design report on MSF reinforced concrete segments for the Blacksnake Creek tunnel. JKP Static, Budapest, Hungary, November 2017 11. Juhasz, K.P., Nagy, L., Winterberg, R.: Full-round numerical analysis of traditional steel bar and macro synthetic fibre reinforced concrete segments for the shanghai metro extension. In: Proceedings World Tunnel Congress 2015, Dubrovnik, Croatia, 22–28 May 2015 (2015)

Incorporation of Rate-Dependent Fracture Properties in the Design of Precast Concrete Tunnel Segment with Hybrid Reinforcement Stefie J. Stephen(&) and Ravindra Gettu Department of Civil Engineering, Indian Institute of Technology Madras, Chennai, India [email protected]

Abstract. The design of precast tunnel segment under service load condition is mostly based on the axial force and bending moment (P-M) interaction diagram. This diagram is usually derived using flexural or tensile parameters obtained from short-term testing procedure. Considering the fact that the tunnel segments are subjected to prolonged loading, the effect of the long-term loading rate on these structures needs to be incorporated. In this paper, the assumption and model used for deriving the P-M diagram is discussed and then the procedure to include the rate-dependent tensile constitutive model in the design is presented. The rate-dependence of the P-M interaction diagram of the tunnel segment with conventional reinforcement and different type and dosage of fibres is discussed. Keywords: Fibre reinforced concrete
Hybrid tunnel segment dependent tensile constitutive model
P-M interaction diagram


Rate-

1 Introduction Tunnels excavated using the tunnel boring machine (TBM) are most commonly supported by precast concrete segments, where conventional steel bars (rebars) are provided as reinforcement (ITA Report, 2016). A combination (hybrid) of rebars and fibre reinforcement has been used in many cases to withstand the bursting and spalling stresses developed during the manufacture and erection of segments (Caratelli et al., 2011; Briffaut et al., 2016). These tunnel linings are expected to last for at least 100 years, taking into consideration the huge initial cost. However, the cracking and spalling of concrete due to improper grouting, vibrational movement of vehicles, prolonged high overburden pressure, etc. puts the segments in risk in the long run (Asakura and Kojima, 2003). Even though the fibres in concrete help in mitigating such damage (Buratti et al., 2013; Tiberti et al., 2014), consideration of the rate-dependence of the FRC response could improve the design and performance, as addressed in this paper. The incorporation of rate effect in the derivation of P-M interaction diagram is demonstrated for a hybrid reinforced concrete (HRC) tunnel segment with conventional steel reinforcement and different types and dosages of fibres.

© RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 770–775, 2021. https://doi.org/10.1007/978-3-030-58482-5_68

Incorporation of Rate-Dependent Fracture Properties

771

2 P-M Interaction Diagram The P-M interaction diagram can be developed for an HRC segment with a particular cross section, if the tensile property of the FRC and the steel rebar and compressive property of the FRC is known. In this section, the models used, and the assumption made by Yao et al. (2018) are briefly explained. The behaviour of FRC is simulated by the stress and strain diagrams for tension and compression using Fig. 1. The tension model consists of a linear stress-strain diagram up to tensile strength ft, followed by a post-cracking constant residual tensile stress, rp = µft = µEecr, with parameter µ (0  µ < 1) representing the residual tensile strength parameter as a fraction of the tensile strength (for plain concrete, µ will be taken as 0) and ecr representing tensile cracking strain (ft/E). The parameters of the tensile model are obtained from the flexural test response of beams, as per RILEM TC 162-TDF (2003). The compressive behaviour is simulated by an elastic perfectly plastic model, with a linear stress-strain diagram up to the yield compressive strength rcy == x ft with parameter x (x > 1). When the compressive strain is above the yield strain, the stress is maintained constant up to concrete crushing (ec = ecu). The ultimate compressive strain, ecu, is assumed to be 0.35% (RILEM TC 162-TDF, 2003). The normalized yield compressive strength x is the compressive-tensile strength ratio. The compression and tension models terminate at the normalized ultimate compressive strain kcu and tensile strain btu, respectively. For steel rebars, the tensile model is assumed as elastic perfectly-plastic, with a linear stress-strain diagram up to yield tensile strength of rebar, fsy = jft, where, j is the normalized yield tensile strength, which is the ratio between the yield tensile strength of steel rebar and the tensile strength of concrete. The tensile yield strain of steel is taken as, esy = fsyEs, where Es is Young’s modulus of concrete. When the tensile strain of rebar is above the yield tensile strain, the stress maintained constant.

a)

0.002

0.0035

b)

0.025

c

c) Fig. 1. Constitutive models for a) FRC in compression, b) FRC in tension and c) steel bars in tension (Yao et al., 2018)

772

S. J. Stephen and R. Gettu

Based on the above-mentioned model and assumptions, the moment and axial force capacity can be calculated for different strain profiles. If the forces and moments due to the groundwater and soil loads are within the envelope of the developed interaction diagram, then the design is safe (Fig. 2).

Fig. 2. Procedure to design a concrete tunnel segment for service loads

3 Rate Dependence of the Tensile Constitutive Model The tunnel segments under service must withstand slow soil settlement and groundwater movement. Even though these structural elements are designed for such magnitude of loads, the variation of moment carrying capacity with loading rate is not generally considered in analysis and design. The rate-effect model described in Stephen and Gettu (2019) accounts for the tensile strength and residual bridging strength at different loading rates. The effect of loading rate on the tensile stress-crack opening (r-w) curve was determined by performing tests on notched beams made of fibre reinforced concrete under CMOD rates covering 5 orders of magnitude. The inverse analysis was performed to obtain r-w curves from the experimental P-CMOD response, and the trends of tensile strength of the material was modelled as:  :  ft;x ¼ ft;std þ m log10 CTODx

ð1Þ :

where, ft;x = tensile strength of concrete at the CTOD rate of CTODx ; ft;std = tensile strength of concrete at the standard loading rate and m is a parameter that depends on the type and dosage of fibre in concrete. The rate-dependence of the bridging stress was modelled as:

Incorporation of Rate-Dependent Fracture Properties

773

 : 

ri;x  ri;std ¼ ft;conc;x  ft;conc;std þ c log10 CTODx ; where i ¼ 1; 2; 3

ð2Þ

where c is a constant considered to be independent of the fibre type and dosage, and ft;conc;x and ri:x are the tensile strength of the plain concrete and the bridging stress at the :

CTOD rate of CTODx , respectively. The proposed Eqs. (1) and (2) consequently comprise the model for rate-dependence of the fracture of FRC that can be used in analysis and design. This rate-effect model is utilized to determine the fracture properties required to derive the interaction diagram for the Terrassa tunnel segment, details for which can be found in de la Fuente et al. (2012). The axial force-bending moment interaction diagram is derived here based on the closed-formed solution proposed by Yao

a)

b) Fig. 3. Load rate effects on moment-force interaction diagram for tunnel segment under service load condition

774

S. J. Stephen and R. Gettu

c)

d) Fig. 3. (continued)

et al. (2018) for concretes provided in Stephen and Gettu (2019). Note that the notation used for the different cases indicates the strength of the concrete, type of fibre incorporated (polymer fibre – PF; steel fibre - SF), its dosage in kg/m3 and the CMOD rate of the test in lm/s; e.g., M40SF30_0.01 denotes the 40 MPa grade concrete with steel fibres at 30 kg/m3 dosage, tested at 0.01 lm/s. It is found that the incorporation of PF and lesser dosage of SF have a greater impact on the rate-sensitivity of the segment moment carrying capacity (Fig. 3). For instance, at zero axial force, the moment carrying capacity reduces by 40% for M40PF3.75 and M40SF10, and less than 10% for higher dosage of steel fibres, when the loading rate is changed from 100 µm/s to 0.01 µm/s. So, for a hybrid tunnel segment reinforced with polymer fibres and low dosage of steel fibres, it appears mandatory to include rate effects in the derivation of the interaction diagram.

Incorporation of Rate-Dependent Fracture Properties

775

4 Conclusions In this paper, the design methodology to incorporate rate-dependent fracture properties has been proposed. The significance of the methodology is discussed in terms of variation of section capacity with loading rate. The modification factors are obtained from the rate effect model to alter the design inputs and determine the section capacity. Further, the load carrying capacity of tunnel segment was investigated at different loading rates and was found to be decreasing with decrease in loading rate. The segments reinforced with polypropylene fibres and low dosage of steel fibres are significantly affected by long-term loading. With an increase in steel fibre dosage, the long-term effects decreases. Therefore, it is crucial to incorporate rate effect model in tunnel segments reinforced with low dosage of steel fibres or polymer fibres.

References Asakura, T., Kojima, Y.: Tunnel maintenance in Japan. Tunn. Undergr. Sp. Technol. 18, 161– 169 (2003). https://doi.org/10.1016/S0886-7798(03)00024-5 Briffaut, M., Benboudjema, F., D’Aloia, L.: Effect of fibres on early age cracking of concrete tunnel lining. Part II: Numerical simulations. Tunn. Undergr. Sp. Technol. 59, 221–229 (2016). https://doi.org/10.1016/j.tust.2016.08.001 Buratti, N., Ferracuti, B., Savoia, M.: Concrete crack reduction in tunnel linings by steel fibrereinforced concretes. Constr. Build. Mater. 44, 249–259 (2013). https://doi.org/10.1016/j. conbuildmat.2013.02.063 Caratelli, A., Meda, A., Rinaldi, Z., Romualdi, P.: Structural behaviour of precast tunnel segments in fiber reinforced concrete. Tunn. Undergr. Sp. Technol. 26, 284–291 (2011). https://doi.org/10.1016/j.tust.2010.10.003 de la Fuente, A., Pujadas, P., Blanco, A., Aguado, A.: Experiences in Barcelona with the use of fibres in segmental linings. Tunn. Undergr. Sp. Technol. 27, 60–71 (2012). https://doi.org/10. 1016/j.tust.2011.07.001 ITA Report, 2016. Twenty years of FRC tunnel segments practice : Lessons learnt and proposed design principles ITA Working Group 2. International Tunnelling and Underground Space Association, Avignon. France (2016) RILEM TC 162-TDF, 2003. Final recommendation of RILEM TC 162-TDF: test and design methods for steel fibre reinforced concrete sigma-epsilon-design method. Mater. Struct, vol. 36, pp. 560–567 (2003). https://doi.org/10.1617/14007 Stephen, S.J., Gettu, R.: Rate-dependence of the tensile behaviour of fibre reinforced concrete in the quasi-static regime. Mater. Struct. 52, 107 (2019). https://doi.org/10.1617/s11527-0191405-2 Tiberti, G., Minelli, F., Plizzari, G.: Reinforcement optimization of fiber reinforced concrete linings for conventional tunnels. Compos. Part B Eng. 58, 199–207 (2014). https://doi.org/10. 1016/j.compositesb.2013.10.012 Yao, Y., Bakhshi, M., Nasri, V., Mobasher, B.: Interaction diagrams for design of hybrid fiberreinforced tunnel segments. Mater. Struct. Constr. 51, 1–17 (2018). https://doi.org/10.1617/ s11527-018-1159-2

Codes and Standards

Developments and Standardisation of Flowable Concrete Reinforced with Fibres for Structural Design, Update of fib TG 4.3 Steffen Grünewald1,2(&), Liberato Ferrara3, and Frank Dehn4 1

4

Delft University of Technology, Delft, The Netherlands [email protected] 2 Ghent University, Ghent, Belgium 3 Politecnico di Milano, Milan, Italy Karlsruhe Institute of Technology (KIT), Karlsruhe, Germany

Abstract. The fib Model Codes aim at integrating in a single document the relevant knowledge for the structural design with concrete. Fibre reinforced concrete is already integrated in fib Model Code 2010 (fib MC2010) as a general category of materials. The group of flowable concrete consists of clusters of different types of concrete among others Self-Compacting Concrete, Ultra High Performance Concrete and Strain-Hardening Cementitious Composites. Being highly flowable is the distinguishing characteristic, flowable concrete might contain or not contain fibres. Although the fibre contribution on the structural level can be assessed on short-term, the structural behaviour also depends on the behaviour of the fibres and the matrix in which they are embedded. fib Task Group 4.3 worked on identifying and characterising different types of flowable concrete and discusses in a fib bulletin the most relevant aspects with regard to mix design, manufacturing, material performance and structural behaviour of flowable concrete which can allow innovative applications to be developed and realised. This paper discusses recent developments with regard to flowable concrete in a broader perspective and addresses the progress with regard to standardisation. Keywords: Concrete Structural design


Fib Model Code
Flowable concrete
Fibres


1 Flowable Concrete Flowable concrete (FC) distinguishes itself first of all from conventional vibrated concrete (CVC) by its rheological characteristics, which are obtained by selective constituent choice and mix design. Main types of FC developed during the past years are self-compacting concrete (SCC, with or without fibres), fibre reinforced concrete (FRC) with high flowability, strain-hardening cementitious composites (SHCC), high performance fibre reinforced cementitious composites (HPFRCC), high performance concrete (HPC) and ultra high performance concrete (UHPC). The trend of increase of the workability of CVC facilitates the casting operation but also makes it difficult to distinguish between FC and CVC. The different types of FC are often treated separately © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 779–790, 2021. https://doi.org/10.1007/978-3-030-58482-5_69

780

S. Grünewald et al.

by standards and recommendations; the fib Model Code aims at integrating in a single document the relevant knowledge with regard to performance of and structural design with concrete, which also includes all types of FC. FC is an innovative concrete type and often it is a material tailored for a specific application. Mainly, FC can be distinguished from CVC through three dimensions (strength, fibre dosage and flowability), see Fig. 1. In the different dimensions all before-mentioned FC types can be located and compared with CVC. fib TG 4.3 [1] considers FC to be flowable with a slump of at least 200 mm; depending on the material density and equation applied this translates into a yield stress of roughly 600–900 Pa. Although, this flowability criterion with regard to the yield stress seems to be high, FC containing a high fibre dosage with a slump of 200 mm has a very low yield strength when tested without fibres. Such concretes require special mix design considerations and are also included in the group of FC. Categories of fibre dosages are: 1) non-structural applications (Vf  0.3 vol.%), 2) low to moderate fibre dosage (> 0.3 &  1.0 vol.%), 3) significant fibre contribution (> 1.0 &  2.5 vol.%) and 4) very high fibre dosage (>2.5 vol.%). Currently, the highest strength class of normal concrete is C50/60 [2]; under discussion is the definition of high, very high and ultra-high-strength concrete types.

Fig. 1. Three main differentiators of different concrete types.

Table 1 lists potential benefits per differentiator when adjusting the mix design of FC compared to CVC hereby enabling the design and production of special applications. The achievement of potential benefits depends on component selection, mix design, execution and structural design. Table 1. Potential benefits when applying FC. +Strength -

+Fibres

- Maximum tensile strength * Compressive strength * - Post-cracking strength * Tensile strength * - Other mechanical properties * Modulus of elasticity * Other mechanical properties * - Crack width + - Replacement reinforcement Durability * - Combine with reinforcement Service life * - Durability * Material volume + Increased early age strength - Self-healing - Material volume + Weight structure + - Weight structure + - Slender elements - Thin elements

+Flowability -

Remote casting Casting of complex shape Casting with dense reinforcement Production efficiency * Aesthetic appearance improved Fibre orientation can be beneficial/tailored Mix design SCC promotes higher fibre dosage

Developments and Standardisation of Flowable Concrete

781

2 Standardisation of Fibre Reinforced Concrete Significant progress has been made during the past years for FRC and FC with regard to material understanding, optimization of material and casting operation, the development of test methods and structural design approaches. Important knowledge and experience already have been implemented in new recommendations and standards. For example, fib MC2010’s FRC provisions [2] have been partially implemented in the Swedish Standard on SS812310 [3]. The AFGC guideline on ultra high performance concrete [4] has been transferred to EN-complementary documents [5–7] addressing concrete characteristics, execution aspects and structural design aspects. In the US, the American Concrete Institute TC 544-Fibre Reinforced Concrete has published in 2018 the ‘Guide to design with fiber-reinforced concrete’ in which fib provisions are also included [8]. Until now, most standards have a limit in scope e.g. with regard maximum strength class, level of flowability or applications covered. For the new fib Model Code, a broader approach has to be considered in order to present FRC as a unique group of concrete, flowable fibre concrete being a subgroup of it. Within fib COM 4 Concrete & Concrete Technology three task groups are working on topics related to FRC and/or FC, which are: TG 4.1 (Fibre-reinforced concrete), TG 4.2 (Ultra highperformance fibre-reinforced concrete) and TG 4.3 (Structural design with flowable concrete). At the moment of writing this paper, TG 4.1 and TG 4.3 are finalising drafts of bulletins with regard to their work. Both documents support the preparation of the next Model Code. fib TG 4.3 is finalising a state-of-the-art report and already published 9 papers during the past years addressing structural design with flowable concrete [9–17].

3 Flowable Concrete in the Fresh State 3.1

Mix Design

FC is obtained by the reduction of friction between solid particles and the optimization of the paste (content and characteristics) complements the composition of the aggregate skeleton with regard to yield strength and viscosity. Paste in FC has three functions in the fresh state: • Filling the interstices of the granular skeleton; • Reducing the friction between the solid particles through an excess of the fluid which separates the solids by means of an adequately “thick” layer of either water or cement paste (dependent on the model considered); • Keeping the aggregates suspended and preventing segregation. The addition of fibres in concrete affects its characteristics in the fresh state, due to the larger surface area of fibres, which requires more fluid phase to properly envelope and lubricate the fibres. Stiff fibres also decrease the packing density of the granular skeleton. Fibres can cause significant interparticle friction and interlocking between fibres and between fibres and aggregates [18]. Fibres have an elongated shape; the slenderness ratio (L/D) is calculated by dividing the length L by the diameter D. An

782

S. Grünewald et al.

equivalent diameter can be calculated for fibres with cross-sections that are not round or which are deformed (i.e. with hooked-ends or wave-shaped) by weighting fibres and calculating the diameter with the assumption that the fibres have a round circular crosssection. The effect of fibres on the workability of concrete depends among others on the dosage, their flexibility, surface characteristics and how they interact with other components of concrete during the flow. Fibres often are bundled with water-soluble glue (effectively reducing the slenderness ratio of fibres before mixing) in order to prevent the formation of intertwining fibres (balling) during the mixing process. The size of the fibres relative to the aggregates determines their distribution. The explained mechanisms and principles result in the fact that the mixture composition of CVC and FC can differ considerably; Fig. 2 shows examples of mix designs of different types of concrete.

100

% by volume

90 80 Cement Filler Water Air Sand Coarse

70 60 50 40 30 20 10 0 Normal concrete

SCC

HPC

UHPC

Fig. 2. Examples of compositions for normal concrete/CVC, SCC, HPC and UHPC [1].

3.2

Flow Conditions

In order to take into account material behaviour for the structural design FC needs to remain homogenous during the mixing and casting process. Two flow phenomena are discussed by fib TG 4.3, which are 1) flow induced particle migration and 2) flow induced anisotropy. A more detailed discussion with regard to flow conditions can be found in [1]. 3.2.1 Flow Induced Particle Migration Concrete components can segregate or rise to the top of the concrete surface through dynamic or static segregation. The first is associated with flow induced particle migration originating from various phenomena whereas the second one is only associated with the settling process due to the density difference between the components when the material is at rest before or after casting. Yield stress and thixotropy seem to dictate static segregation for a given granular skeleton. Depending on the considered

Developments and Standardisation of Flowable Concrete

783

concrete flow, different phenomena can dominate and dictate particle migration within the material. Different types of flow induced particle migration were identified: • • • • •

Shear induced particle migration; Gravity induced particle migration; Granular blocking; Heterogeneity in particle distribution induced by the wall effect; Induced heterogeneities during concrete casting or testing.

3.2.2 Flow Induced Anisotropy Fibres can improve mechanical properties of cementitious materials in the hardened state. This influence depends primarily on the fibres (shape, constitutive material, volume fraction) but also on the casting process. Flow can induce a preferred orientation of the fibres which influences the casting flow and, after setting, affects the mechanical properties of the resulting fibre reinforced composite. An accurate prediction of a population of fibre orientation should consider several parameters which determine the macroscopic orientation in the hardened state material, such as the geometry of the mould, the presence of free surfaces, the method of casting, or interactions between fibres. The two categories of flow relevant in industrial practice are: 1) free surface flow referred to as ‘slab casting’ and 2) confined flow referred to as ‘wall casting’.

4 Flowable Concrete in the Hardened State 4.1

Strength and Stiffness

As the fibre dosage of FC typically is up to 1.0 vol.% in ordinary FRC and 1.5−2.5 vol. % in UHPC, the influence of fibres on engineering properties such as compressive strength, modulus of elasticity, Poisson ratio and tensile cracking strength is very limited or relatively small. This does not mean that the fibres do not influence such properties when cracking is initiated. Since fibres also affect the workability, differences between mixtures with and without fibres can be caused by entrapped air or less than adequate compaction after the addition of the fibres. However, as mentioned in Sect. 3.1, in order to obtain FC within defined boundary conditions (e.g. bar spacing to pass through, fibre content, required flowability) an adjustment of the mix design is required. This can include: increase in paste content, decrease of aggregate content, size and content, increase or decrease of water-powder ratio. Such changes affect the stiffness and time-dependent deformation behaviour of concrete (see also Sect. 4.3). 4.2

Fibre Contribution

The behaviour of fibres during pull-out is determined by the characteristics of the fibres, the matrix and the interface. Single and bundled (filament) fibres behave differently, since the bond with the outer fibres in a bundle is better compared to fibres inside a bundle. A brittle material/structure reinforced with brittle fibres (i.e. glass textiles)

784

S. Grünewald et al.

loaded in tension can perform ductile, since the pull-out of inner fibres takes place by friction with other fibres. According to [19], mainly three types of interactions determine the pull-out behaviour of a single fibre: 1) physical and chemical adhesion, 2) friction and 3) mechanical anchorage induced by deformation of fibre or anchorage. Dependent on the pull-out behaviour, the dosage, distribution and orientation of the fibres, softening or hardening in direct tension is obtained (Fig. 3).

Fig. 3. Softening (a) and hardening (b) behaviour in direct tension [2].

A material with direct tension-softening behaviour still can be hardening in bending or when applied in a statically indeterminate structure. In many cases, the production with FRC is a trade-off between processing requirements, desired mechanical performance and economical aspects. Since fibres increase the material costs, an optimum fibre dosage has to be determined. The difference in contribution of softening or hardening FRCs to the structural performance can be taken into account. Figures 4a and 4b show the influence of the type and dosage of steel fibres on the compressive and tensile strengths of HPFRCC; the abscissa value (fibre factor) is the product of fibre volume and aspect ratio (ratio length to diameter) of the steel fibres. 4.3

Time-Dependent Deformations

The risk of cracking is also a consequence of an interrelation of the time-dependent viscoelastic properties, the tensile strength and the modulus of elasticity. The time- and load-dependent behaviour of FC can be rather different compared to CVC. As Fig. 2 shows, the paste contents and mixture compositions of FC can vary throughout a wide range. The different contributions to shrinkage occur at different moments after casting (chemical, autogenous and drying shrinkage). Higher cement contents can cause larger deformations due to temperature-effects in an early stage; a high heat release potential might not be obtained in practice because of a very low water-binder ratio and low degree of hydration. SHCC has a high paste content at the highest w/c-ratio of all types of FC. Accordingly, drying shrinkage and creep are relatively high. Strain caused by shrinkage and creep can add up, but creep also results in relaxation, which reduces the shrinkage strain. A heat treatment of UHPC (e.g. at 90 °C for 96 h) significantly reduces the creep and shrinkage beyond the day of heat curing. Predictive model

Developments and Standardisation of Flowable Concrete

785

Fig. 4. Influence of the fibre factor of steel fibres on a) compressive strength (cylinders of d/h = 100/170 mm) and b) maximum tensile strength (dog-bone specimens) [20]

formulae provided in standards and recommendations may not be applicable for FC [14]. A detailed discussion of the time- and load-dependent behaviour for FC can be found in [14] and the fib bulletin (to be published) on Structural Design with Flowable Concrete [1]. Main conclusions of [14] are summarised hereafter: Drying shrinkage: The total shrinkage increases at increasing paste volume [21]. When the water-powder ratio is kept constant, the relation between shrinkage and paste volume is approximately linear and can be regarded as the dominating parameter in drying shrinkage. Drying shrinkage in UHPC is very low (in the range of 0.1‰ concrete deformation) compared to autogenous shrinkage [22]. Although the shrinkage of SHCC is high (range of 1‰ at a relative humidity of 60%), its tensile strain capacity seems to be higher than the drying shrinkage deformation [23]. Autogenous shrinkage: Influence of paste volume: At a constant water/powder-ratio, the autogenous shrinkage of SCC increases at increasing paste volume [24]. As a result, CVC shrinks less than SCC when the binder composition and strength class are the same. Autogenous shrinkage of UHPC is considerably larger compared to CVC [25]. In SHCC a decrease in paste volume by the addition of fine-grained aggregates (grain size of 0.15–0.30 mm) can significantly reduce autogenous shrinkage [26]. Creep: Even when results of different studies are not consistent, there seems to be a general agreement that the creep coefficient and the specific creep are normally slightly higher for SCC compared to CVC [21]. Specific creep of UHPC is in the range of 0.01– 0.035‰/MPa, while the creep coefficient is in the range of 0.5–1.2 [27]. Because of a higher paste volume SHCC (HPFRCC) shows large creep deformations, but due to a low modulus of elasticity, creep coefficients can be even smaller than the ones of CVC [28].

786

S. Grünewald et al.

5 Model Code Provisions Essential information for the structural design with FRC is the magnitude of postcracking (residual) strength at different values of crack mouth opening displacement (CMOD) of 0.5, 1.5, 2.5 and 3.5 mm. The CMOD is generally measured from flexural tests performed on notched beams. fR1 (at CMOD = 0.5 mm) and fR3 (at CMOD = 2.5 mm) are main parameters considered for design (Fig. 5). Fibre reinforcement can substitute (also partially) conventional reinforcement in the ultimate limit state, if the following relationships are fulfilled: fR1k/fLk > 0.4 and fR3k/fR1k > 0.5. Combining both criteria means that fR3 has to be at least 20% of the flexural strength fLk, which assures a minimum ductile behaviour in bending. Dependent on the ratio fR3k/fR1k categories a-e (a 0.5 < fR3k/fR1k < 0.7, b 0.7  fR3k/fR1k < 0.9, c 0.9  fR3k/fR1k < 1.1, d 1.1  R3k/fR1k < 1.3 and e 1.3  R3k/fR1k) have been defined which allows specifying concrete performance e.g. for ready-mix concrete producers. Accordingly, the design engineer can derive structural performance without being involved in the design and testing of the concrete. The orientation and distribution of the fibres but also the variation of results within a test series also determine the flexural performance that can be taken into account for design. With flow induced anisotropy FRC can be tensile hardening in one direction and tensile softening in the other which has to be properly considered when performing a structural analysis, mainly of statically redundant structures, including slabs where biaxial stress states can be common.

a)

b)

Fig. 5. Deriving design residual strengths a) typical load F-CMOD curve for plain concrete and FRC and b) translation to simplified post-cracking constitutive laws [2].

fib MC2010 [2] includes provisions with regard to the following aspects related to the design with FRC (Table 2), provisions that take into account e.g. the contributions of fibres in bending, shear, punching shear and for crack width calculations. Where relevant the contribution of fibres can be included in the calculation of minimum reinforcement (bending, shear, punching shear, skin reinforcement) as well as on anchorage/introduction length of rebars.

Developments and Standardisation of Flowable Concrete

787

Table 2. Provisions in fib MC2010 with regard to the design of FRC. Section MC 5.6 5.6.2/5.6.3 5.6.4 5.6.5 5.6.6 7.7 7.7.2 7.7.3 7.7.3.1 7.7.3.2 7.7.3.2 7.7.3.4 7.7.3.5 7.7.4 7.7.4.1 7.7.4.2 7.7.4.3

Design aspect Fibres/fibre reinforced concrete Material behaviour and classification Constitutive laws Safety factors Fibre orientation Verification of safety and serviceability of FRC structures Design principles Verification of safety Bending and/or axial compression in linear members Shear in beams Torsion in beams Walls Slabs, members with/without reinforcement, punching Verification of serviceability Stress limitation Crack width in members with conventional reinforcement Minimum reinforcement for crack control

Bending and/or axial compression in linear members: fib Model Code 2010 includes simplified stress/strain relationships with which the cross-sectional analysis can be executed. The difference between a strain-hardening and a strain-softening material is shown by Fig. 6.

Fig. 6. Ultimate limit state for bending moment and axial force, use of simplified stress/strain relationships [2].

788

S. Grünewald et al.

Shear resistance: Due to the increase in fracture energy of FRC compared to normal concrete and the opening of more and smaller cracks, the aggregate interlock contribution of concrete also increases. Accordingly, fibres are able to increase the shear resistance of concrete. The part of Eq. (1) multiplied by 7:5
fFtuk =fck is the additional shear resistance provided by the fibres. VRd;F ¼

! 
 13 0:18 fFtuk
k
100
q1
1 þ 7:5

fck þ 0:15
rcp
bw
d c fctk

ð1Þ

Punching shear resistance: The punching shear resistance is the sum of the contributions of punching shear reinforcements of steel and fibres, which includes the contribution VRd,c, see Eqs. (2) and (3). VRd ¼ VRd;F þ VRd;s ) VRd;F ¼ VRd;c þ

fFtuk b0 dv cF

ð2 þ 3Þ

Crack width verification: The design crack width wd in FRC elements can be determined with Eq. (4). The term with fFtsm (negative contribution) reduces the value of wd; it is the tensile contribution of the fibres which reduces the crack opening. By virtue of the action of the fibres, which generate a residual tensile strength fFtsm, the force to be reintroduced by bond is reduced. (

1 /s fctm  fFtsm wd ¼ 2 k
c þ
4 qs;ef sbm

)


1
ðrs  b
rsr þ gr
2sh
Es Þ Es

ð4Þ

fib TG 4.3 identified and discussed differences between CVC and FC with regard to mixture composition, execution aspects and structural performance. Fibre orientation has to be considered; with an appropriate safety approach the potentially different characteristics can be taken into account.

6 Conclusions fib TG 4.3 aims at integrating different types of flowable concrete in a state-of-thereport; the result will be published in an oncoming fib bulletin. The following conclusions were drawn: • Significant steps have been made with regard to the development of performance tests, non-destructive test methods and standardisation; • Innovative and/or economical concrete solutions and structures can be developed with tailor-made flowable concrete; • Mix design, fibre orientation and distribution are essential aspects to consider in order to take into account the structural performance;

Developments and Standardisation of Flowable Concrete

789

Acknowledgements. The authors want to thank the members of fib TG 4.3 for their contributions; with regard to this publication T.A. Hammer, A. Leemann, N. Roussel and W. Schmidt are gratefully acknowledged.

References 1. fib bulletin: Structural design with flowable concrete. State-of-the-art report (to be published) 2. fib Model Code 2010: fib Model Code for Concrete Structures 2010. Wilhelm Ernst and Sohn, ePDF ISBN: 978-3-433-60408-3 (2013) 3. SS812310:2014: Fibre Concrete-Design of Fibre Concrete Structures, Swedish Institute for Standards (2014) 4. AFGC: Ultra High Performance Fibre-Reinforced Concretes–Recommendations, Association Française de Génie Civil (2013) 5. NF P 18-470: Bétons fibrés a ultra-hautes performances - Spécification, performance, production et conformité, AFNOR, Paris (2016) 6. NF P 18-710: Calcul des structures en béton-Regies spécifiques pour les bétons fibrés a ultrahautes performances (BFUP), AFNOR, Paris (2016) 7. NF P 18-451: Execution des structures en béton-Régies spécifiques pour les BFUP, AFNOR, Paris (2018) 8. ACI TC 544: Guide to design with Fiber-Reinforced Concrete’, 544.4R-18, ACI (2018) 9. Grünewald, S., Ferrara, L., Dehn, F.: fib task group structural design with flowable concrete. In: Hirt, M.A., Radic, J., Mandic, A. (eds.) Codes in Structural Engineering; Developments and Needs for International Practice, IABSE Dubrovnik, pp. 1333–1340 (2010) 10. Grünewald, S., Ferrara, L., Dehn, F.: fib-recommendation structural design with flowable concrete. In: Proceedings of the 3rd International Congress Incorporating the PCI Annual Convention and Bridge Conference, Chicago, Precast/Prestressed Institute PCI, pp. 1–9 (2010) 11. Ferrara, L., Grünewald, S., Dehn, F.: Design with higly flowable fiber-reinforced concrete: overview of the activity of fib TG 8.8. In: Khayat, K., Feys, D. (eds.) Proceedings SCC 2010 Design, Production and Placement of SCC, Springer, ISBN 978–90-481-9663-0, pp. 395– 406 (2010) 12. Grünewald, S., Ferrara, L., Dehn, F.: Structural design with flowable concrete–a fibrecommendation for tailor-made concrete. In Khayat, K., Feys, D. (eds.) Proceedings SCC 2010 Design, Production and Placement of SCC, Springer, ISBN 978-90-481-9663-0, pp. 13–24 (2010) 13. Grünewald, S., Bartoli, L., Ferrara, L., Kanstad, T., Dehn, F.: Translation of test results of small specimens of flowable fibre concrete to structural behaviour: a discussion paper of fib task group 4.3, fib Bulletin No. 79. In: Proceedings FRC 2014 ACI-fib Int. Workshop FibreReinforced Concrete: From Design to Structural Applications, ISBN: 978-2-88394-119-9, pp. 81–90 (2016) 14. Leemann, A., Hammer, T.A., Grünewald, S., Ferrara, L., Dehn, F.: Time- and loaddependent behaviour of flowable concrete: progress report of fib task group 4.3. In: Braestrup, M., Stang, H. (eds.) fib Symposium Concrete-Innovation and Design, Lausanne: fib, pp. 1–11 (2015) 15. Schmidt, W., Grünewald, S., Ferrara, L., Dehn, F.: Design of concrete for high flowability: Progress report of fib task group 4.3. In: Braestrup, M., Stang H. (eds.) fib Symposium 2015 Concrete-Innovation and design, Lausanne, fib, pp. 1–10 (2015)

790

S. Grünewald et al.

16. Grünewald, S., Ferrara, L., Dehn, F.: Structural design with flowable concrete. In: Nunes, S., Sousa Coutinho, J., Faria, R. (eds.) Proceedings 4th Congresso Ibero-Americano sobre Betao Auto-compactavel, BAC2015, Faculdade de Engenharia, Universidade do Porto, pp. 19–32 (2015) 17. Grünewald, S., Liberato, F., Dehn, F.: Flowable fibre-reinforced concrete: Progress in understanding and development of design standards. In: Khayat, K.H. (ed.) Proceedings 8th International RILEM Symposium on SCC, RILEM Publications, ISBN: 978-2-35158156-8, pp. 467–478(2016) 18. Bayasi, M.Z., Soroushian, S.P.: Effect of steel fiber reinforcement on fresh mix proportions of concrete. ACI Mater. J. 89(4), 369–374 (1992) 19. Bentur, A., Mindess, S.: Fibre Reinforced Cementitious Composites, 2nd edn. Taylor & Francis, New York (2007). ISBN 978-0-415-25048-X 20. Sato, Y., Mier, J.G.M. van, Walraven, J.C.: Mechanical characteristics of multi-modal fiber reinforced cement based composites. In: Lyon, Rossi, P., Chanvillard, G. (eds.) Befib 2000, Cachan Cedex, pp. 791–800, ISBN: 2-912143-18-7 (2000) 21. Leemann, A., Lura, P., Loser, R.: Shrinkage and creep of SCC-the influence of paste volume and binder composition. Constr. Build. Mater, vol. 25, pp. 2283–2289 22. Koh, K., et al.: Shrinkage properties of ultra-high performance concrete. Adv. Sci. Lett. 4, 948–952 (2011) 23. Weimann, M.B., Li, V.C.: Hygral behavior of engineered cementitious composites (ECC). Int. J. Restor. Build. Monuments 9, 513–534 (2003) 24. Rozière, E., et al.: Influence of paste volume on shrinkage cracking and fracture properties of self-compacting concrete. Cem. Concr. Compos. 29, 626–636 (2007) 25. Loukili, A., Khelidj, A., Richard, P.: Hydration kinetics, change of relative humidity, and autogenous shrinkage of ultra-high-strength concrete. Cem. Concr. Res. 29, 577–584 (1999) 26. Billington, S.L., Rouse, J.M.: Time-dependent response of highly ductile fiber-reinforced cement-based composites. In: International Symposium Brittle Matrix Composites, Warsaw, pp. 47–54 (2003) 27. Flietstra, J.C., Ahlborn, T.M., Harris, D.K., De Melo e Silva, H.: Creep behaviour of UHPC under compressive loading with varying curing regimes. In: 3rd International Symposium on UHPC, Kassel, Germany, pp. 333–40, ISBN: 978-3-86219-264-9 (2012) 28. JSCE HPFRCC: Recommendations for design and construction of High Performance Fiber Reinforced Cement Composites with multiple fine cracks (HPFRCC). Concrete Engineering Series 82, Japan Society of Civil Engineers (2008)

Assesment of Codal Provision for SFRC Beam in Minimum Shear Kranti Jain and Bichitra S. Negi(&) Department of Civil Engineering, National Institute of Technology, Srinagar, Uttarakhand, India [email protected]

Abstract. Natural fibres (straw, chip, horse tail, goat hair and plume, etc.), was being used from ancient time for construction purposes. Inspired from ancient time, artificial fibres (vitreous, synthetic, carbon and steel fibre, etc.) are commonly used nowadays in order to improve the mechanical properties of concrete. Literature has suggested that concrete matrix with steel fibre commonly known as Steel fibre reinforced concrete (SFRC) have enhanced flexural as well as shear behaviour as compared to conventional reinforced concrete (RC). Also it is recommended by ACI building code that steel fibre can use as minimum shear reinforcement in RC beams. In the present study, the experimental results for assessment of ACI building code provisions allowing the use of deformed steel fibres as minimum shear reinforcement in RC beams are presented. Hooked end steel fibres of length 35 mm and 60 mm have been used in concrete matrices at volume fraction of fibres ranges from 0.75 to 1.5% and 0.5 to 1% respectively. The experimental performance of the fibrous concrete beams has been compared with the Indian Standard and ACI codal provisions. It is found that in all the cases, even though at a volume fraction of 0.5%, lower than the ACI code-specified lower limit of 0.75%, the measured shear strengths were higher than the predicted values as per the ACI building code, Indian standard code, as well as from the lower bound value of 0.3√fc′MPa (3.5√fc′ psi). Also confirmed multiple diagonal cracking having crack width lower than the specified permissible values. On the basis, it is suggested that there is a need of relook into codal provisions of ACI. KEYWORDS: Steel fibre reinforced concrete (SFRC)
Reinforced concrete (RC)
Minimum shear reinforcement
Deformed steel fibres

1 Introduction Concrete and mortar are well known for its brittle nature therefore less able to resist tensile stresses and propagation of cracks. Various researchers have been carried out to overcome this problem. The inclusion of deformed fibres in concrete matrix mainly improves its residual strength, toughness and ductility. The improved material properties of fibre-reinforced concrete (FRC) further improves the flexural and shear behaviour of FRC structures [1–4]. It has been already recognized that the fibre reinforcement is an effective way to enhance toughness of concrete for every type of failure © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 791–800, 2021. https://doi.org/10.1007/978-3-030-58482-5_70

792

K. Jain and B. S. Negi

modes. Beam without web reinforcement fails suddenly in shear because of its brittle nature. Research over last few decades has clearly established the potential use of fibre for enhancing the shear capacity of reinforced concrete (RC) beams [5–7]. The effect of fibre reinforcement on shear strength is contributed of two main factors: 1) a direct factor imposed by post cracking strength at inclined shear crack; 2) an indirect factor which cause increase in concrete contribution to shear strength by improving aggregate interlock and dowel action of reinforcement [2]. Many researchers has been carried out in past few decade to focus on shear behaviour of SFRC beam. The result showed that the inclusion of steel fibre boost the shear capacity of concrete and enhance the shear crack distribution, therefore they are capable of replacing vertical stirrups in RC structural members. It is well known that shear failure loads vary widely about the value given by the design equations. Most of the current design codes [8–10] suggested the requirement of minimum amount of transverse reinforcement if applied shear force exceeds a definite fraction of inclined cracking shear. Such reinforcement sometime may also have such value that it may able to resist unexpected tensile force and overload. When a flexural member is subjected to any load like fatigue loading, there exists the possibility that inclined diagonal tension cracks may form under statis loading even at lower load. And therefore codes suggested to provide minimum shear reinforcement even calculation does not shows requirement of shear reinforcement. According to Dinh et al. [11] fibre in concrete matrix enhance shear resistance by transferring tensile stresses across inclined diagonal cracks and overcome further increase in crack width and spacing and also enhance contribution due to aggregate interlock. Parra-Montesinos [12] shows that the steel fibre can be used as shear reinforcement in reinforced concrete (RC) beam on the basis of past literature. A lower bound for shear strength of FRC of value 0.3 √fc was considered adequate with fibre in concrete at volume fraction greater than or equal to 0.75% After this research and data from past literature, a new provision was included in 2008 ACI building code [13] which allows use of deformed steel fibres as minimum shear reinforcement in normal strength concrete when fibre volume fraction is greater than or equal to 0.75%. This investigation explores the use of deformed steel fibres as minimum shear reinforcement in RC beams. The experimental program comprises of test on large SFRC beam which are designed to fail in shear under monotonically increasing loads in three point loading configuration. The behaviour of SFRC beams are compared with traditional beams with minimum shear reinforcement as per the recommendations of ACI building code [8, 13] and IS 456:2000 [9].

2 Experimental Programme Hooked end steel fibre of length 60 mm (aspect ratio = 80) and 35 mm (aspect ratio = 65) having ultimate tensile strength of 1050 MPa and 1100 MPa respectively were used in concrete mixture. 35 mm long steel fibre were added at volume fraction (Vf) of 0.75%, 1% and 1.5% whereas 60 mm long steel fibre were added at volume fraction of 0.5%, 0.75%, and 1%. 1.5% volume fraction of 60 mm long steel fibre were not used because of possibility of balling effect of fibre at volume fraction more than

Assesment of Codal Provision for SFRC Beam in Minimum Shear

793

1%. ACI 2008 building code [13] recommended a minimum Vf of 0.75% whereas the Vf of 60 mm long steel fibre were kept below 0.75% to expand the scope of investigation. The detail of plain and SFRC mixture is documented in Table 1. For evaluation of shear behaviour, control specimens of plain concrete mixture were used. Table 1. SFRC mixture proportions Weight (kg) per m3 V*f = 0% Vf = 0.5% Cement 396 394 Fine aggregate, FA 870 862 Course aggregate, CA 656 648 (4.75 mm–10 mm) Course aggregate, CA 353 349 (10 mm–12.5 mm) Superplasticiser – – Water 225 233 W/C ratio 0.57 0.59 Steel fibres – 39 CA/(FA + CA) 0.54 0.54 Ingredient

Vf = 0.75% Vf = 1% Vf = 1.5% 394 394 394 860 857 852 646 643 638 348

346

344

– 233 0.59 59 0.54

– 233 0.59 79 0.54

0.79 232 0.59 118 0.54

Shear behaviour of SFRC beams was investigated by testing singly reinforced beam having length 1770 mm and cross-Sect. 150 mm  300 mm. The beam have simply supported span of 1470 mm and failed under monotonically increasing threepoint loads. In order to ensure the correctness of results, nominally identical companion beams were cast for every specimen. The geometry of beam specimens and test set up configuration is presented in Fig. 1. Each beam was designed in such a way that it will fail in shear in the tested span whose shear span to effective depth ratio was 3.5. The shorter span was provided with sufficient stirrups so that the failure always occurs in tested span. In order to obtain shear failure prior to flexural failure, all beam specimen were intentionally over reinforced with 2.67% (100Ast/bd) of tension reinforcement at an effective depth of 251 mm. Based on the distribution of transverse reinforcement in the tested span, the beams were classified as follows: • To observe typical brittle failure, no transverse reinforcement was provided in the tested span. • Minimum shear reinforcement in the tested span (6 nos of 8 mm diameter twolegged closed rectangular stirrups) as per ACI building code [8]. • Minimum shear reinforcement in the tested span (4 nos of 8 mm diameter twolegged closed rectangular stirrups) as per IS 456:2000 [9]. • Minimum shear reinforcement in the form of steel fibre only.

794

K. Jain and B. S. Negi Tested span, a/d = 3.5

150

Steel bearing plate at the load point, 100 x 150 x 50

150

300

150

Load-point LVDT

Steel flats LVDT arrangement for monitoring inclined cracking Steel bearing plate at support (typical), 100 x 150 x 50

Mid-span LVDT

Test beam, 150 x 300 x 1770

875 (Tested span)

595 1470 1770 (a) Front elevation of a typical beam

300

Effective depth = 251

150

Ast, typical (All dimensions in mm)

(b) Cross-section

Fig. 1. Beam specimen geometry and test set-up configuration

The assembled reinforcement cages are shown in Fig. 2, and it should be noted that the longitudinal reinforcement bars were provided with sufficient hooked extensions at their ends as per ACI building code [8] in order to minimize chance of anchorage failure. The mechanical properties of the reinforcement bar used in the specimens are tabulated on Table 2. Before casting of beam, the prefabricated steel reinforcement cages were placed in the formwork keeping clear cover of 25 mm from bottom. The target cylindrical compressive strength of normal strength concrete was 26 MPa and initial slump was in the range of 150 mm to 175 mm for plain concrete mixture whereas 40 mm to 100 mm for fibrous concrete mixture. The beams and the control specimens are demoulded after 24 h of casting and moist curing was done for 10 days. Subsequently, the specimens were air-cured in the laboratory until testing, which was carried out after a nominal interval of 28 days from the day of casting. The summary of beam specimens along with the experimental results obtained from shear test was tabulated in Table 3. The diagonal crack would be accompanied by web deformations, which would be indicated by a sudden and significant change in the displacements measured by the

Assesment of Codal Provision for SFRC Beam in Minimum Shear

(a) Control beam (no shear reinforcement in the tested span)

(b) Minimum shear reinforcement per the ACI Building Code [8]

(c) Minimum shear reinforcement per the IS 456:2000 [9]

(d) Steel fibres as minimum shear reinforcement Fig. 2. Details of assembled reinforcement cages

Table 2. Mechanical properties of reinforcing bar Bar diameter (mm) Yield strength, MPa Ultimate strength, MPa % Elongation 8 558 723 8 10 553 646 25 16 566 692 26

795

796

K. Jain and B. S. Negi Table 3. Beam specimens and their experimental results

Beam ID

Concrete mixture and detailing of transverse reinforcement Plain concrete, no transverse reinforcement

f′c

Pu

vu

MPa

MPa

MPa

vu/(f ′c)0.5

Failure mode+

A-I 24.50 125.41 1.35 0.27 DT + ST A-II 25.50 185.01 1.99 0.39 DT + ST 28.1 394.76 4.24 0.80 SC B-I Plain concrete, transverse reinforcement per ACI B-II 25.43 374.33 4.02 0.80 SC 318-08 (ACI 2008) B-III 26.51 356.78 3.84 0.74 SC 28.15 402.20 4.32 0.81 SC C-I Plain concrete, transverse reinforcement per IS C-II 26.10 378.52 4.07 0.80 SC 456:2000 (BIS 2000) C-III 25.64 355.88 3.83 0.76 SC D-I N -35-0.75* 28.10 278.77 3.00 0.57 DT + ST + SC D-II 25.28 195.63 2.10 0.42 DT + ST + SC E-I N -35-1.00* 27.90 270.03 2.90 0.55 DT + ST + SC E-II 26.20 305.20 3.28 0.64 DT + ST + SC F-I N -35-1.50* 28.10 274.47 2.95 0.56 DT + ST + SC F-II 27.33 323.25 3.48 0.66 DT + ST + SC G-I N -60-0.50* 27.5 159.94 1.72 0.33 DT + ST + SC G-II 24.94 190.59 2.05 0.41 DT + ST + SC H-I N -60-0.75* 27.75 224.69 2.42 0.46 DT + ST + SC H-II 27.33 251.56 2.70 0.52 DT + ST + SC I-I N -60-1.00* 26.25 286.52 3.08 0.60 DT + ST + SC I-II 27.12 258.80 2.78 0.53 DT + ST + SC * N- in first place stands for normal strength, numeral value at second place stands for length of fibre, numeral value at third place stands for volume fraction of fibre. DT- Diagonal tension failure; ST- Shear tension failure; SC- Shear compression failure.

LVDT arrangement. For testing the beams specimen, hydraulic ram was used to apply monotonically increasing load in 10–15 increments until failure and beam deflections at different point was monitored by LVDTs. The data of testing was recorded by computer aided data acquisition system. The crack pattern, number of cracks, crack widths and failure modes were carefully monitored and noted for each specimen.

3 Result and Discussion Various failure modes are observed in tested span of different beams specimen and the peak load crack patterns were shown in Fig. 3 for the purpose of comparision. During testing, the initial inclined cracking first observed in shorter span but it was sufficiently reinforced in shear. Therefore the diagonal crack induced reduction in shear stiffness of the shorter span and inclined crack appeared in the longer span, where the shear failure in the beam specimen has to occur.

Assesment of Codal Provision for SFRC Beam in Minimum Shear

(a) Control beam (no shear reinforcement in the tested span)

(b) Minimum shear reinforcement per the ACI Building Code [8]

(c) Minimum shear reinforcement per the IS 456:2000 [9]

(d) 1% Vf of 35 mm long hooked-end steel fibres

(e) 1% Vf of 60 mm long hooked-end steel fibres Fig. 3. Peak load crack patterns of the beam specimens

797

798

K. Jain and B. S. Negi

The crack pattern on the tested span was different based on detailing of transverse reinforcement. Beam with no transverse reinforcement shows failure by diagonal tension (DT) in combination with shear tension (ST), Fig. 3(a). Diagonal tension was initiated by propagation of a single major crack was toward the load point and along the longitudinal reinforcement. At failure the crack is penetrated to concrete compression zone without crushing the concrete. In Shear tension, the inclined crack propagated along the longitudinal reinforcement and towards the support, which tends to reduce the anchorage capacity of the reinforcement. In opposite to Fig. 3(a), multiple diagonal cracks were observed in other tested span, Fig. 3(b), (c), (d), and (e). Figure 3(b), (c) shows shear compression (SC) failure with crushing of concrete in the compression zone near the load point. Additional multiple cracking with one single major crack were observed in fibrous beams which provide earlier warning before collapse, Fig. 3(d), (e). The measured peak loads and normalized peak average shear stresses are tabulated on Table 3 and the peak average shear stress values are plotted in Fig. 4, wherein they are compared with codal predictions and the lower bound value suggested by ParraMontesinos [12]. There was large variation in normalized shear stress value for the plain and fibrous concrete beams as compared to others, although none of the beam had value lower than the lower bound value of 0.3 of Parra-Montesinos [12]. Among all the fibrous concrete 50% Vf of 60 mm long steel fibre has relatively less normalized shear strength of value 0.33.

Shear stress (MPa)

4.50 4.00

Min.-IS 456 (Experimental)

3.50

Min.-ACI 318 (Experimental)

3.00

Hooked fibres 35 mm long

2.50

Hooked fibres 60 mm long

2.00 SFRC Predicted (Parra Montesinos)

1.50

Min. IS 456 (Predicted)

1.00

Min. ACI 318 (Predicted)

0.50 0.00 0

0.25

0.5 0.75 1 1.25 Fibre volume fraction (%)

1.5

Fig. 4. Shear stress v/s fibre volume fraction

The minimum shear reinforcement requirement in ACI building code [8] and IS 456:2000 [9] correspond to the nominal shear strength of 0.34 MPa and 0.4 MPa, respectively and values are plotted in Fig. 4, which shows that every beam specimens

Assesment of Codal Provision for SFRC Beam in Minimum Shear

799

had shear strength more than the lower bound value of 1.53 MPa. Lower bound value pffiffiffiffi is obtained by the putting fc’ = 36 MPa in the equation 0:3 fc0 suggested by ParraMontesinos [12]. Figure 4 also shows that the experimental shear strength of the specimens with the code recommended minimum shear reinforcement were much more than the predicted value, which shows that the conservative nature of the codal recommendations related to shear design.

4 Conclusions • The shear strength of 60 mm long steel fibrous concrete beam with 0.5% of Vf which was lower than the lower bound limit recommended by ACI 318 building pffiffiffiffi code shows multiple cracking and shear strength more than the 0:3 fc0 . • The failure mode of steel fibre reinforced beam was comparable to that of those beam designed for minimum shear reinforcement in accordance to ACI 318 building code and IS 456:2000. • The ACI 318 building code and IS 456:2000 specified predicted value of shear strength i.e. 0.344 MPa and 0.4 MPa, respectively was much less than the experimental results obtain by testing codal designed beams specimen. • The service load crack width of steel fibrous concrete beam was very less compared to allowable value of 0.3 mm for normal exposure condition as well as from the beam specimen designed in accordance to codal provision.

References 1. Narayanan, R., Darwish, I.Y.S.: Use of steel fibers as shear reinforcement. ACI Struct. J. 84 (3), 216–227 (1987) 2. Altoubat, S., Yazdanbakhsh, A., Rieder, K.A.: Shear behaviour of macro-synthetic fiberreinforced concrete beams without stirrups. ACI Struct. J. 106(4), 381–389 (2009) 3. Dinh, H.H., Parra-Montesinos, G.J., Wight, J.K.: Shear behaviour of steel fibre-reinforced concrete beams without stirrup reinforcement. ACI Struct. J. 107(5), 597–606 (2010) 4. Kwak, Y.-K., Eberhard, M.O., Kim, W.-S., Kim, J.: Shear strength of steel fiber-reinforced concrete beams without stirrups. ACI Struct. J. 99(4), 530–538 (2002) 5. Batson, G.B., Jenkins, E., Spatney, R.: Steel fibers as shear reinforcement in beams. ACI J. Proc. 69(10), 640–644 (1972) 6. Dupont, D., Vandewalle, L.: Shear capacity of concrete beams containing longitudinal reinforcement and steel fibers. In: Banthia, N., Criswell, M., Tatnall, P., Folliard, K. (eds.) Innovations in Fiber Reinforced Concrete for Value, SP-216, pp. 79–94. American Concrete Institute, Farmington Hills (2003) 7. Li, V., Ward, R., Hamza, A.M.: Steel and synthetic fibers as shear reinforcement. ACI Mater. J. 89(5), 499–508 (1992) 8. ACI Committee 318: Building Code Requirements for Structural Concrete (ACI 318-14) and Commentary (318R-14). American Concrete Institute, p. 503. Farmington Hills, Michigan (2014)

800

K. Jain and B. S. Negi

9. IS 456: Indian Standard Plain and Reinforced Concrete – Code of Practice,‖ Fourth Revision. Bureau of Indian Standards, BIS, New Delhi, 100 p. (2000) 10. BSI: EN 1992-1-1:2004. Eurocode 2: Design of concrete structures. Part 1-1: General rules and rules for buildings. BSI, London (2004) 11. Dinh, H.H., Parra-Montesinos, G.J., Wight, J.K.: Shear strength model for steel fiber reinforced concrete beams without stirrup reinforcement. J. Struct. Eng. ASCE 137(10), 1039–1051 (2011) 12. Parra-Montesinos, G.J.: Shear strength of beams with deformed steel fibers. Concr. Int. 28 (11), 57–66 (2006) 13. ACI: ACI 318R-08: Building code requirements for structural concrete and commentary. ACI, Farmington Hills (2008)

Laboratory Investigations on the Installation of Fasteners in Fiber Reinforced Concrete Panagiotis Spyridis1(&), Lars Walter1, Julia Dreier2, and Dirk Biermann2 1

2

Chair of Fastening Technology, TU Dortmund, Dortmund, Germany [email protected] Institute of Machining Technology, TU Dortmund, Dortmund, Germany

Abstract. A significant aspect regarding the use of post-installed anchors in concrete is related to constructability, and mainly the characteristics of possible geometrical configurations. Specifically for fibre reinforced concrete (FRC), as fibre strengths or dosages increase, these configurations may change. One of the critical parameters is the minimum allowable distance from the edge, which can significantly influence the flexibility of using post-installed anchors, or even the possibility to install them in the first place. This geometrical parameter strongly depends on the drilling and anchoring procedure, as various technologies are available. This paper will present and compare the various available technologies for fastenings, and it will focus on setting experiments in fibre reinforced concrete by use of challenging means for drilling and setting, i.e. hammer drilling, mortar injection and expansion anchors. The influence of fibres in the installation and tightening of post-installed anchors is particularly analysed. The results of the study reveal the possibilities for extended applications of fastenings in fibre reinforced concrete, as compared to plain concrete. Keywords: Concrete


Fastenings
Installation
Steel fibres

1 Introduction As a consequence of the increasing trend for the use of fibre reinforced concrete (FRC) in several types of engineering structures, advanced knowledge and standardisation in the field become more and more essential. In the last years, FRC became a favourable solution in a variety of significant structural components, including tunnel structures, high-performance floor slabs (e.g. pile-supported and plan foundations, seamless slabs, slabs on grade), immersed structures, watertight and containment infrastructure components, and various sprayed concrete applications such as structural strengthening of buildings and bridges, or geotechnical structures. The increasing interest in design guidance for FRC is now reflected, besides numerous scientific publications, in respective guidelines and standards, application specific guidance documents and most importantly in the inclusion of provisions for the design with fibre reinforced concrete in the upcoming edition of Eurocode 2. At the same time, the relevance of fastenings in the last years is also made evident by the fact that new specifically dedicated design guides and norms have been recently © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 801–812, 2021. https://doi.org/10.1007/978-3-030-58482-5_71

802

P. Spyridis et al.

published. The new Eurocode 2 – Part 4 “Design of fastenings for use in concrete” is the most prominent standardisation effort, while new design and specification guides and standards are also emerging globally for fastenings to concrete. 1.1

Overview of Fastenings Engineering and Technology

Depending on the construction approach, two main distinctions are made between castin-place and post-installed anchors, while for post-installed anchors further distinction can be made between the installation type as undercut, expansion, plastic, bonded, power actuated anchors and concrete screws. These technologies are also strongly associated to three main load bearing mechanisms in tension: – Mechanical interlock, when the anchor is engaged in the surrounding material through contact – Friction, which is developed through expansion of the anchor towards the walls of the borehole – Bond, by means of an appropriate chemical reaction that connects the anchor rod to the borehole, which is accomplished through a combination of adhesion and microkeying. Figure 1 graphically presents some of these anchor types, while details of the anchoring principles and assembly are discussed below. Figure 2 graphically represents the main anchoring mechanisms. Cast-in-place anchors: they can be inserts of several different shapes (ribbed bars, threaded sleeves, headed studs, etc.) that are installed in advance during the assembly of the formwork. Undercut anchors: in most cases for the installation of undercut anchors it is required that either the borehole is formed appropriately by means of special drilling for an expanding part of the anchor to nest, or that the anchor can itself form an appropriate tooth in concrete. Expansion anchors: the functional concept of such systems is to expand a section (bolt type) or a longer part of the body (sleeve type) so that they can develop compression against the walls of the boreholes and activate frictional forces. These expansion systems can be either torque-controlled, in which expansion forces are mechanically driven through screwing, or displacement controlled, where expansion is achieved by hammering a steel cone-shaped part through the outer steel sleeve of the anchor or vice versa. Plastic anchors employ the same installation and functional principle, by means of a plastic expansion sleeve. Bonded anchors: they are usually divided to capsule systems (the components of the bonding mortar are enclosed in separate compartments of a capsule and are mixed in the borehole by driving in the anchor) and injection systems (the mortar is filled in the borehole in liquid phase before setting the anchor). Another classification of bonded anchors can be made on the bonding material, which is usually a two-component mortar consisting of organic (polyester, vinylester, epoxy) substances, inorganic (cement-based) ones or even a combination. Special bonded anchor systems (usually intended for use in cracked concrete) may employ functional principles of expansion anchors because of their particular cone-shaped body, or of undercut anchors by means of specially shaped boreholes.

Laboratory Investigations on the Installation of Fasteners

803

Concrete screws: the installation of concrete screws demands initially drilling a borehole in the concrete surface and then driving in the screw by use of the drilling machine. These elements have properly prepared threads in order cut indentations in the walls of the borehole and provide a firm interlock throughout their length. Power actuated anchors: they are nails of high-strength steel directly installed in concrete by use of a setting tool applying pneumatic, electrical or explosive powder energy. 1.2

Status in Design and Product Harmonisation for Concrete Fastenings

The design of fastenings, particularly for safety critical applications, is primarily based on the “European Technical Product Specification” (ETPS) for each fastening product. Essentially, the ETPSs deliver information about the fabrication, installation, and associated performance for each product. For concrete fastening products, the ETPS is the “European Technical Assessment” (ETA) which is prepared on the basis of a “European Assessment Document” (EAD). The EAD (also known as ETAG until recently) typically prescribes a testing campaign for the product assessment. The ETPS also confirms the “Declaration of Performance” by the product manufacturer and allows for the product’s CE marking. A similar procedure is followed in the U.S. with the International Code Council (ICC) and the Evaluation Service Reports (ESR). To that end, fastening products are primarily characterised for their structural safety, resistance to fire, and durability. Another important aspect in the technological development of fastening products is the facilitation of safe installation, because installation defects can substantially decrease the fastening’s structural performance. This is discussed on the example of bonded anchors by [2], while it is also evident by the fact that [1] recommends an increased resistance partial safety factor for anchorages with potentially low installation quality. Typically, the installation characteristics in order to guarantee the performance declaration are dictated in a product’s ETA. These characteristics include among other things, the borehole preparation, geometry, and cleaning, the applicable drilling tools and the drill bit characteristics, the allowable substrate and fixture properties, the tightening torque, and the minimum allowable anchor spacing and the anchor’s distance from the concrete edge. The above provisions and the associated approval procedures typically refer to unreinforced or rebar reinforced concrete, while the industry often comes confronted with the design of fastenings in FRC.

Fig. 1. Various fastening systems in concrete: (a) concrete screw, (b) undercut anchor, (c) expansion anchor - sleeve type, (d) expansion anchor – bolt type, (e) cast-in-place ribbed and deformed bar, (f) cast-in-place headed stud, (g) bonded anchor with threaded rod, (h) special bonded anchor.

804

P. Spyridis et al.

Fig. 2. Distinctive load bearing mechanisms of axially loaded anchors in concrete.

1.3

Overview of the Paper

The focus of this paper lies in the influence of fibre in concrete with respect to the installation safety and defects of post-installed anchors. An overview of the fastenings engineering discipline is provided and issues related to the topic of this contribution are highlighted. Furthermore, outstanding previous studies on the performance of fastenings in fibre reinforced concrete are summarised. The designated tests and results carried out at the TU Dortmund are presented, including tests on bonded anchors, setting tests of expansion anchors close to an edge, and hammer drilling surveys. Finally the findings are summarised with the intention to contribute to practical engineering design and constructability issues. Table 1 presents the geometrical and mechanical properties of the used concrete specimens.

Table 1. Geometrical and mechanical properties of the specimens Concrete type Plain (A1) Fiber reinforced (A2) Concrete strength class C35/45 C35/45 Geometry w  h  l [mm] 900  350  1100 900  350  1100 50 Fibre content [kg/m3] – Cube compressive strength (150 mm, 6 no. per batch) Average compressive strength [N/mm2] 49.55 51.32 Coefficient of variation [%] 5.9 2.8 4-point flexural strength (150  150  700 mm, 3 no. per batch) Average flexural strength [N/mm2] 4.71 6.05 Coefficient of variation [%] 3.8 1.6

Laboratory Investigations on the Installation of Fasteners

805

2 Previous Studies for Fastenings in Concrete Extensive research has been performed on fibre reinforced concrete worldwide, ranging from material and fibre analysis to large scale applications, design methods, and codified standards. A substantial number of studies addresses the rebar bond performance in FRC. [4] provides a borad literature review on bond strength and splitting failure for rebar, in relation to various fibre properties and dosages. Rebar pull-out tests and large scale beam bending tests are addressed, which show a capacity increase of up to 55% depending on the fibre content and geometry and up to 150% for special mixes. Furthermore, significant studies have been carried out with respect to the influence of fibre reinforcement on the load bearing of fastenings in fibre reinforced concrete. [5] and [6] present investigations of headed bolts with diameters of 8 to 12 mm in various embedment depths, in shelled steel fibres reinforced concrete with normal strength. The presented tests show a rise in tensile breakout capacity of up to 35% depending on the test parameters. Steel fibre reinforced concrete and bonded anchors of typical sizes used in building construction have been tested per ASTM-E488 and presented in [7], stating a minor, “non-significant” influence of the fibres in the load bearing resistance. 95 axial tests on different types of anchors (expansion, undercut, bonded) with a 10 mm diameter and short embedment (50, 60 mm) are presented in [8]. In these tests waved sheet cut and hook-ended steel fibres with a dosage: 50 kg/m3 had been mixed in C20/25. The study concludes that there is negligible difference in resistance mean values, whereas anchorages in FRC have lower characteristic values (higher dispersion) than in normal concrete. [9] presents investigations on transverse loads toward a free concrete edge in normal strength concrete with hooked ended steel fibres at 25– 35 kg/m3. The tests have been carried out according to European standards and showed that the concrete edge breakout strength can be highly increased by adding steel fibers to the concrete, in particular for small edge distances (up to 50% increase). Normal strength concrete with steel hooked ended fibres at a dosage of 25 and 60 kg/m3 has been used for tests (also per European standards) in various types of M10 and M12 anchors with embedments of 65 and 70–75 mm respectively as reported in [10]. The tests did not show that the failure load is substantially increased, but that the failure mode changes with a higher fibre content, from pure concrete failure to pull-out or mixed failure modes. Normal and High-strength concrete reinforced with 80 kg/m3 hooked-ended fibres has been used in tensile tests reported in [11], for large diameter headed anchors. The failure was concrete cone mode and most of the tests also exhibited radial cracking/splitting, for a load increase of 25% for normal and 40% for high strength concrete, and significantly higher ductility, as compared to plain concrete. The effect of fibres on anchor group and load redistribution within the individual anchors is included in [12], presenting tensile tests in concrete with a strength of 65 MPa reinforced with 50 kg/m3 hooked-end fibres. All tests led to concrete cone failure, indicating an 18% increase in failure load for single anchors and 25–40% for anchor groups, as compared to tests in concrete without fibres. Further investigations presented in [13], additionally conclude that concrete cone, concrete edge, and presumably side blow out capacity can be increased by up to 25%, subject the dosage of fibres. However, this approach is deemed reliable, only if the anchorage load is initiated

806

P. Spyridis et al.

at an embedment depth (axial) or edge distance (shear load) at a distance of 1.7 times the fibre length (Lfiber) from the concrete surface. The study also proposes that, based on the findings, the critical edge distance for tension loads (nominal distance where the concrete boundary effects cease) may be reduced yet not below 1.7 Lfiber. Tests on castin-place anchor channels under either tensile or shear loads [14] have also indicated higher ultimate loads and larger corresponding displacements for the latter mix. As seen in the studies above, the improved ductility of fastenings in FRC is becoming evident. When it comes to the load capacity of fastenings, conclusions from the previous studies are varying, witnessing beneficial, neutral, or even adverse influence of the fibre reinforcement. In all cases, the item of installation reliability and possibly the mitigation of consequences of defects in installation is not specifically elaborated with respect to fibre reinforcement. The investigations presented below, attempt to address two significant aspects of installation, and common causes of defects in installation, namely the borehole cleaning for bonded anchors, and the torqueing level of expansion anchors close to a free edge.

3 Experimental Investigations on Borehole Cleaning To assess the influence of the steel fibres on the bond strength, nine confined pull-out tests were carried out in unreinforced and concrete reinforced with 60 mm long steel hooked end fibres at a dosage of 50 kg/m3. The concrete cube strength was 49.5 MPa and 51.3 MPa for the plain and fibre reinforced concrete respectively. For the tests M12 threaded rods and an injectable hybrid mortar with approval for concrete applications were used. The tests were calculated to induce a pull-out type of failure, in which the bond and shear strength of the injection mortar are the determining factors. The tight confinement by means of a 40 mm thick steel plate, on which the hydraulic cylinder is located, exerts pressure on the concrete surface in the immediate vicinity of the anchor during the test to prevent concrete breakout. To pursue a constant stress distribution over the length of the anchor, without stress peaks toward the surface, the upper 20 mm of the rods was wrapped with PTFE tape in several layers, which prevented the mortar from locking into the thread of the anchor rod and also prevents the formation of a load-bearing mortar layer. The boreholes were clustered into two groups. Series 1 were cleaned following a procedure of blowing out the dust twice with compressed air, brushing twice with a steel brush, and finally blowing out twice again. Series 2 did not undergo particular removal of the drill dust from the borehole surface, and it was only ensured that the sediment did not reduce the intended anchoring length; this procedure understandably leads to a substandard and potentially defective installation. The drilling was carried out by a hammer drill with a four-edge drill bit in all tests. The tests were displacement-controlled with a loading rate between 0.01 and 0.02 mm/s. The load bearing curves achieved are shown in Fig. 3. The graph developments allow to qualitatively characterise the failure situation for each bonded anchor [15]. In adequately cleaned boreholes, pull-out tests have failed at the interface between the anchor rod and the mortar, due to shearing of the mortar material. A somewhat different pattern is exhibited for tests in defect installation with uncleaned boreholes. Tests

Laboratory Investigations on the Installation of Fasteners

807

A1-3x and A2-1x show a failure between mortar and borehole wall, whereby the bond strength at the concrete interface is higher than the sliding friction forces during pullout. A mixed failure can be implied by tests A1-1x, A1-4x, A2-2x and A2-3x, where the load bearing throughout the test relies on both friction and concrete bond strength. Failure in tests A1-2x, and A2-4x is likely involving material shearing failure, while the latter appears to be entirely unaffected by the lack of borehole preparation. One average though, a significant loss of load bearing capacity is shown in tests with defective boreholes. In normal concrete, the reduction is in the range of 33%, and in FRC the reduction rises to 39% (excl. test A2-4x). Considering these differences, as well as the dispersion of load levels and modes of failure, it is statistically arduous to attribute a correlation between the presence of fibres in the mix and the borehole interface characteristics. Regarding adequately prepared boreholes, the pull-outs in unreinforced concrete exhibit a slightly higher mean value but also variation coefficient (49.4 kN and 9%) as compared to the ones in FRC (48.4 kN and 2%).

Fig. 3. Load displacement curves of confined pull-out tests.

4 Experimental Investigations on Torque Fasteners were set close to a free edge, and they exerted transverse stresses on their substrate by torqueing the head of the anchor. As a consequence, a lateral concrete breakout occurs as the expansion force increases. Understandably, this depends on the material capacity of the concrete, the level of torque, and the distance from the edge. As per each product’s specification, a minimum spacing and edge distance, and the torque to be applied are defined, beyond which these effects become critical. A steel torquecontrolled expansion anchor was used for this test, but an edge distance smaller than the minimum allowable was implemented in order to mimic the situation of a deviant

808

P. Spyridis et al.

installation. In particular, the minimum edge distance for the product used was 70 mm for an anchor spacing of 160 mm [3], while the anchors were positioned at a spacing of 200 mm, but an edge distance of only 30 mm from the edge. This distance is assumed to be the boundary of the wall-effect for the fibre distributions where fibres are expected to be isotopically distributed, acc to. The recommended torque for this application is 50 Nm, which can be observed on the dial of the torque wrench. As a fixture below the washer a 5 mm thick metal plate on top of a PTFE sheet was used. The concrete mixes compared are identical to the ones discussed in the previous section. The failure mode mostly observed in plain concrete tests resembled a side blow-out failure rather than a concrete edge failure, while all tests in FRC exhibited cracking with a small angle to the concrete surface at the fixture level (see Fig. 4). Only the applied torque was measured through an electronic torque sensor installed between the key and the bolt head. The rotation was measured in 45° steps and a rotation pace as constant as possible. Thus, a maximum value was recorded for each 45° rotation. A curve enveloping all measurement results connects the maxima of the applied torque and allows comparisons between the individual tests and the two test specimens. Assuming that the sleeve expansion responds symmetrically and uniformly to the tightening of the anchor, a consideration of the torque over the rotation course is representative of the horizontal deformation. A direct comparison of the tests in non-reinforced concrete with those in fibrereinforced concrete shows that with the same edge distance, approximately twice as much torque can be applied until the concrete develops visible cracking. A further crack propagation leads to a reduction in the torque required to expand the anchor sleeve further. As a result, the measurement curve of each test shows strong differences in gradient. The orientation of the wings of the expansion sleeve in relation to the concrete edge was found to have a major influence on the test results, regardless of the fibre inclusion. Expansion of a single wing orthogonally to the edge leads to a lower failure torque compared to expansion e.g. in parallel to the edge direction. Measurement curves in non-reinforced concrete are smoother than those in fibrereinforced concrete, as the fibres in concrete provide an increased crack resistance and influence the crack expansion, while consecutive crack bridging of the fibres is also discerned in the graphs. One torque-time development for each type of concrete is illustrated in Fig. 5. The contribution of fibres can simultaneously increase the transverse stresses required for visible cracks to develop, but also to an enhanced postcracking load bearing. For the reduced edge distance, anchor torqueing reached a mean of 16.5 Nm with a variation coefficient of 27%. The applied torque in fibre reinforced concrete reached a mean of 51.7 Nm with a coefficient of variation of 38%. Conclusively, the torque to failure, applied on anchors in both types of concrete for the given reduced edge distance, cannot safely reach the minimum value of 50 Nm required in order to comply with the product’s approval document and therefore fulfil the anchor’s designated performance. Nonetheless, the inclusion of fibres led to a multiple capacity of the concrete against a failure formation caused by the expansion transverse forces, already at a very small distance to the specimen’s surface. An interesting observation in tests with FRC, is that where an increased number of steel fibres lies in the immediate vicinity of the sleeve, the expansion cone can be pulled through the sleeve without

Laboratory Investigations on the Installation of Fasteners

809

completely expanding it, and the sleeve bends back toward it’s original shape (Fig. 4 right). Except for this, crack bridging between the anchors tested in FRC was observed, which suggests that the spacing recommended for plain concrete may not directly apply for anchors in plain concrete.

5 Analysis of Steel Fibre Separation by Hammer Drilling According to [16], a distinction is made between four different drilling methods in rock drilling technology: “rotary drilling”, “percussive drilling”, “combined drilling” and “thermal drilling”. Hammer drilling of concrete is assigned to percussive drilling, since the material separation is achieved by the notch effect of the drilling tool [16]. The drilling tool is provided with an impact impulse. An impact impulse is applied on the drill bit. This energy must be sufficient to exceed the compressive strength of the concrete and thus lead to material disintegration [17]. Due to the notching mechanism, tensile and compressive stresses occur in the effective area. In the primary cutting zone, the concrete is crushed and many micro-cracks occur, while in the secondary cutting zone, cracks occur in the concrete due to the axial cutting force, which is responsible for the formation of chips [18]. When machining reinforced concrete by hammer drilling, contact with steel rebar reinforcement should generally be avoided. In the case of ductile, carbon-steel fibre reinforcement, drilling progresses but no chip formation occurs. As a result of the percussive principle of action, plastic deformations occur with the drillings action. The material is deformed in front of the cutting edges of the tool until it is sheared off. Hammer drilling of ductile carbon-steel reinforcement is therefore more similar to forging than to cutting [19]. Plastic deformations of the steel fibres can be observed with hammer drilling such a concrete material. As a result of the plastic deformation, enlargements of the cut surface of the steel fibres occur, see Fig. 6 (left). In addition to cross-sectional expansion, the steel fibres are partially detached from the concrete and local breakouts occur in concrete, as seen in Fig. 6 (middle). Some fibres are not separated at all or they experience only a slight material separation. These fibres are plastically deformed in the direction of the drilling tool rotation (Fig. 6 - right).

Fig. 4. Aspects of testing torque application for the installation of expansion anchors: circular side breakout due to torque on expansion anchor – unreinforced concrete (left); edge breakout due to torque – fibre reinforced concrete (middle); cone pull-through with bend-in of the sleeve due to restraint from fibres.

810

P. Spyridis et al.

Fig. 5. Indicative results of torque measurement of expansion anchors being set in unreinforced (left) and reinforced concrete (right) close to a free edge.

Fig. 6. Steel fibre separation by hammer drilling: (left) plastic deformation of steel fibres; (middle) chipping-off of the concrete surrounding the fibre; (right) plastic deformation in the direction of rotation.

6 Conclusions The investigations presented herein focus on the aspect of fastenings installation in reinforced concrete. Furthermore, the inclusion of fibres in concrete as regards the performance of anchors with substandard installation configurations and defects also comes to question. It is noted, that the investigated parameters are limited, and a quantification and generalisation of the findings can only be facilitated by further research. Based on the described set of tests, the following conclusions can be summarised. • The bond performance of bonded anchors in normal strength concrete, installed by injection of a hybrid mortar according per the recommended procedures is not strongly influenced by steel fibres in the mix. • Substandard preparation of the boreholes for injection anchors leads to a significant reduction of the confined pull-out resistance. A somewhat different performance of bonded anchors under pull-out is observed between plain and steel fibre reinforced concrete. • The positive influence of steel fibres on the resistance to lateral concrete failure due to installation of torque-controlled expansion anchors became evident. This effect was observed for anchors positioned at already very small distances to the free edge (30 mm).

Laboratory Investigations on the Installation of Fasteners

811

• In consequence to the above, a reduction of the minimum edge distance may be considered for anchors in fibre concrete. The minimum spacing in such a case needs to be accordingly increased. • The installation performance of the expansion anchor shows a strong interaction with localised random occurrences such as the orientation of the sleeve wings to the free edge, or a concentration of fibres adjacent to the anchors. • An analysis of the steel fibres after hammer drilling shows that the fibres are plastically deformed and undergo enlargements of the cut surface. In addition, the concrete surrounding the fibres breaks out in many cases. Some fibres are plastically deformed in the direction of the hammer drill rotation and exhibit no or only slight material separation. Acknowledgements. The authors express their appreciation for the support of Hilti AG and Bekaert GmbH with the provision of material used in this study. The views expressed in this paper are those of the authors and do not necessarily coincide with those of the organisations mentioned above.

References 1. Technical Committee 250/European Committee for Standardisation (CEN/TC 250). EN 1992-4:2018 Eurocode 2. Design of concrete structures. Design of fastenings for use in concrete (2018) 2. Grosser, P., Fuchs, W., Eligehausen, R.: A field study of adhesive anchor installations. Concr. Int. 33(1), 57–63 (2011) 3. European Technical Assessment, ETA-02/0042 of 07/09/2015: Torque-controlled expansion anchor, made of galvanised steel, for use in concrete: sizes M8, M10, M12, M16, M20 and M24 4. ACI Committee 408: ACI 408R-03 Bond and Development of Straight Reinforcing Bars in Tension. American Concrete Institute, (2003) 5. Al-Taan, S.A., Mohammed, A.A.: Tensile strength of short headed anchors embedded in steel fibrous concrete. Al-Rafidain Eng. J 18, 35–49 (2010) 6. Al-Taan, S.A., Mohammed, A.A., Al-Jaffal, A.A.: Breakout capacity of headed anchors in steel fibre normal and high strength concrete. Asian J. Appl. Sci. 5, 485–496 (2012) 7. Gesoglu, M., Ozturan, T., Ozel, M., Guneyisi, E.: Tensile behavior of post-installed anchors in plain and steel fibre-reinforced normal-and high-strength concretes. ACI Struct. J. 102(2), 224 (2005) 8. Holschemacher, K., Wittmann, F., Klug, Y.: Structural behaviour of fastenings in steel fibre reinforced concrete. In: Yuan, Y.S., Shah, S.P., Lü, H.L. (eds.) International Conference on Advances in Concrete and Structures, RILEM Publications, pp. 939–946 (2003) 9. Grosser, P.R.: Load-bearing behavior and design of anchorages subjected to shear and torsion loading in uncracked concrete (2012) 10. Kurz, C., Thiele, C., Schnell, J., Reuter, M., Vitt, G.: Tragverhalten von Dübeln in Stahlfaserbeton. Bautechnik 89, 545–552 (2012) 11. Nilforoush, R., Nilsson, M., Elfgren, L.: Experimental evaluation of tensile behaviour of single cast-in-place anchor bolts in plain and steel fibre-reinforced normal- and high strength concrete. Eng. Struct. 147, 195–206 (2007)

812

P. Spyridis et al.

12. Bokor, B., Tóth, M., Sharma, A.: Influence of steel fiber content on the load-bearing capacity of anchorages in concrete’. In: Proceedings, 3rd International Symposium on Connections between Steel and Concrete (ConSC2017), Stuttgart (2017) 13. Toth, M., Bokor, B., Sharma, A.: Anchorage in steel fiber reinforced concrete–concept, experimental evidence and design recommendations for concrete cone and concrete edge breakout failure modes. Eng. Struct. 181, 60–75 (2019) 14. Mahrenholtz, C., Ayoubi, M., Müller, S., Bachschmid, S.: Tension and shear performance of anchor channels with channel bolts cast in Fibre Reinforced Concrete (FRC). In: IOP Conference Series: Materials Science and Engineering, vol. 615, no. 1, p. 012089. IOP Publishing (2019) 15. Meszaros, J.: Das Tragverhalten von Einzelverbunddübeln unter zentrischer Kurzzeitbelastung. Universität Stuttgart (2001) 16. DIN 20301: Gesteinsbohrtechnik Begriffe, Einheiten, Formelzeichen. Beuth, Berlin (1999) 17. Müller, C.: Einfluss der Stahlfaserzugabe auf die Spannungs-Dehnungs-Linie von Beton unter Druckbeanspruchung. Influence of steel fibers on the stress strain relationship of concrete under compression. Beton-und Stahlbetonbau 112(4), 228–237 (2017) 18. Weinert, K., Michel, O., Gillmeister, F.: Schneidengestalt und Leistung von Hammerbohrwerkzeugen. BMT, Baumaschine + Bautechnik (2), 74–78 (1994) 19. Moseley, S.G., Peters, C., Momeni, S.: Near-surface phenomena occurring in cemented carbides with different binders during rotary-percussive drilling of reinforced concrete: FE simulation and microstructural investigations. Int. J. Refract Metal Hard Mater. 86, 105–115 (2020)

Quality Control

Applications of Statistical Process Control in the Evaluation of QC Test Data for Residual Strength of FRC Samples of Tunnel Lining Segments Chidchanok Pleesudjai1, Devansh Patel1, Mehdi Bakhshi2, Verya Nasri2, and Barzin Mobasher1(&) 1

School of Sustainable Engineering and Built Environment, Arizona State University, Tempe, AZ, USA [email protected] 2 Lead Tunnel Engineer, AECOM, New York, NY, USA

Abstract. Statistical process control procedures are widely applied to improve the production efficiency of industrial products. Application of quality control procedures in monitoring the production, delivery, and construction process are essential, especially when the historical data collected on various projects can be used to gain better insight to the operational procedures. Such information will promote interaction; reduce the liabilities and benefits the owners and suppliers if any of the design parameters that are below the minimum requirement can be identified as soon as possible. The longer time it takes to detect discrepancies in the data, the more the penalty, project delays, and the higher the associated costs to the owners and suppliers. A detailed analysis of the test results of flexural closed-loop control test data conducted in accordance to ASTMC1609 test is conducted for QC requirements of precast segment in tunnel lining project. More than 378 production samples are recorded. The data set contained 1 day (demolding age) and 28 days of molding. The statistical process control and the range of the data are studied in the context of material properties as well as backcalculation of the tensile parameters. The number of parameters evaluated includes flexural strength and deflection at ultimate flexural strength, the residual strength results at L/600, L/300 and L/150. A series of statistical analysis procedures to analyze the correlation of the parameters and addressed the combination of control charts for early detection of spurious shifts in the mean of the test data (out-of-control signal). Keywords: Statistical process control
Fiber reinforced concrete lining
Quality control
Time series control process


Tunnel

1 Introduction Statistical process control identifies the sources in the variations in manufacturing and thus improves the process efficiency. QC methods that actively monitor the variations in the mean of a manufacturing process can be extremely beneficial in monitoring the © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 815–826, 2021. https://doi.org/10.1007/978-3-030-58482-5_72

816

C. Pleesudjai et al.

delivery and construction of concrete precast sections. Control charts process the data based on the historical records and provide insight into operational procedures. They may be instrumental in monitoring strength or durability data to optimize the operating process by identifying potential problems that may exist. A process typically operates under normal conditions (i.e., in control) and produces a product that satisfies the minimum specified strength. However, changes in normal operating conditions may cause the process to go out of control. An out-of-control process will eventually fail to meet the minimum specified metrics due to low-quality material, operator errors, or production changes. Early detection when this change occurs and identification of root causes that push a process to go out of control can lead to early corrective actions. Failure to implement and correct a cause will lead to operations that are continuously out of control, resulting in significant and unacceptable losses. There are several approaches for assessing the quality of concrete used in precast applications. Sykora recommended a standard deviation and coefficient of variation correction for testing error to determine the uniformity of the concrete strength [1]. Control charts are widely used to monitor and improve process performance. The CUSUM chart or the EWMA control charts are far more efficient as a moving average control tool in detecting small process shifts (1.5r or less) in comparison to Shewhart control charts [2]. Hybrid control charts for active control and monitoring of concrete strength were developed by Laungrungrong et al. [3] to support a tertiary role of improving efficiency, product formulation, and mitigating conflict through a rational acceptance/rejection criteria. The minimum strength and sample size (n) cannot be used as the only parameters to achieve the acceptable producer’s risk (rejecting “good” concrete) and consumer’s risks (accepting “poor” concrete). Models can be extended to consider multiple metrics for evaluation that are not directly tied into a single performance indicator such as a single strength criterion. This approach is useful for comparing multiple acceptance metrics, or variations within a plant, but is not recommended for comparing plant to plant variations. Leshchinsky suggested a metric that is a combination of two or more methods to improve the reliability and accuracy of concrete quality [4]. The proposed methodology presents an approach to evaluate the flexural or residual flexural strength using the combination of either the CUSUM or EWMA control chart with a standard run chart. The CUSUM and EWMA control charts often behave similarly in practice, although different weight functions can apply to current and recent past data values [5]. The CUSUM method applies a constant weight factor to the entire historical set of data. In the EWMA method, uses an exponential weight factor applied to the data; giving current or recent past observations more weight than older data values. The combination of a CUSUM (or EWMA) chart and a run chart are being used to determine the best conditions for monitoring concrete residual strength as they pertain to the full-scale test results. The performance of the CUSUM-run chart and the EWMA-run chart are compared to determine which chart is more useful for the concrete industries [6].

Applications of Statistical Process Control

817

2 Exploratory Data Background All 378 tests of ASTM C1609 were conducted as a part of the QC requirements mandated by AECOM for Hartford CSO tunnel project in 2018. This specific Fiber Reinforcement Concrete (FRC) was studied for serviceability requirement by limiting time-dependent to guarantee that the effect of creep on crack opening will satisfy residual nominal flexural strengths. As a part of processing data, the 378 tests were digitized and replotted in terms of load vs. deflection and flexural strength vs. deflection with consistent units by ASU graduate student. The test results were individually plotted one by one as a series of test data of demolding age and 28-day aged from June to November of 2018. Those data have been identified and tagged accordingly. At the same time, some of the samples were 8–10 h on the day of casting. Those samples are identified as 1-day test, accounting for 107 test results. The other 271 data results were tested on ages 28 days. Individual samples were analyzed for the following parameters, Flexural Strength and deflection at 1st crack, ultimate flexural strength and associated deflection, residual strength and toughness at deflections associated with L/600, and L/150. The effective dimensions of the specimen were 450  150  150 as per ASTM C1609 standard. A typical load-deflection graph from one of the data pool is shown below in Fig. 1a. The time history of the compression strength test results is shown in Fig. 1b with an average strength of 57 MPa.

(a)

(b)

Fig. 1. a. A typical load vs deflection plot from ASTM1609, b. Compressive Strengths test result at 28 days

The demolding age test is generally conducted to find the early strength of the specimen, which is around 8 h. An average plot of the Flexural stress vs. deflection data of all the specimens with ±1.64 standard deviations is shown in Fig. 2b. The procedure of finding average value and standard deviation is by first interpolating all of the time history data as discretized deflection, and then averaging the points at deflection steps. The discretized deflection size in this paper is 0.01 mm.

818

C. Pleesudjai et al.

Similarly, for 28 days test, an average plot of the load vs. deflection data of all the specimens with ±1.64 standard deviations is shown in Fig. 2b shows the scattered plot and the average curve. The average load at a first crack is around 5 MPa, which is almost twice the average load of the demolding age specimens, at 2.5 MPa. After the first crack, it can be observed an increase in residual stress indicating higher bond strength between concrete and fibers after 28 days.

3 Priliminary Statistic Analysis Statistical analyses were conducted to evaluate nominal peak strength f1 and nominal D D residual strengths, fD 600, f300 and f150 of the response of the studied FRC mixture in ASTM1609 as it is used in specifications for acceptance and/or rejection of samples. The proposed approaches are based on using control charts in the evaluation of variations in the test results. The control chart is a graph used to study how a process changes over time. A control chart always has a central line since an average, an upper line and lower line for the control limit. These lines are determined from historical data [7]. The residual stress from L/600 to L/150 are the primary concern in designing fiber reinforced concrete. In the studied data, the QC specifications report determined residual stress at 2.5 MPa as the levels of Lower Control Limit (LCL) for the demolding age and 4.0 MPa for 28 day age. It is true that variation exists in every manufacturing environment and damage the quality characteristic product [10]. In Fig. 2a–2b present some of individual sample fail from LCL. Normal distribution is made as an assumption in both concrete age. However, the actual data distribution was plotted with normal distribution in Fig. 3a–3b. Hence, it can say that Avg.1.64SD correspond to 90% of confidential interval. In Fig. 2a, the green line is lower than LCL by mean that confidential interval to pass QC specification in demolding age sample is lower than 90%.

(a)

(b)

Fig. 2. a. Average ASTM1609 testing plot with 107 individual samples, at demolding age, b. Average ASTM1609 testing plot 271 individual samples, at 28 days

Applications of Statistical Process Control

(a)

819

(b)

Fig. 3. a. Normal Probability distribution, residual stress L/300, b. Normal Probability distribution, residual stress L/600

Otherwise, the run chart is an alternative way to present the dynamic movement of a process changes over a period of time. Data are plotted in time order as in Fig. 4a–4b. In Fig. 4a, 15% of flexural strength at L/600 (0.75 mm), 3–4% of flexural strength at L/300 (1.5 mm) and L/150 (3 mm) were lower from QC specification. Whereas 28 days data, lower than 1% falling from QC criteria strength. However, only the run chart cannot indicate a small out of control process. Combining the run chart with a monitoring scheme that provides information about the stability process is required. The next section will explain two types of control process which depending on memory data. It calls cumulative sum (CUSUM) and exponential weighted moving average (EWMA).

(a)

(b)

Fig. 4. a. Demolding age flexural strength Run chart, b. 28 days flexural strength Run chart

820

C. Pleesudjai et al.

CUSUM and EWMA A cumulative sum (CUSUM) control is the chart for individual observations monitors sample measurement by distinguishing between a significant change (real change in slope) and non-significant change (error of aberration). The general form of the cumulative sum is given by Ci ¼

n X

ðxj  l0 Þ

ð1Þ

j¼1

Where Ci presents sum of the derivation from target values(l0) over the domain. Day recommended using the mean strength as the target value in the CUSUM chart [8– 10]. As a result of CUSUM, the sensitive value from the target can be detected. A onesided upper l0, is C+ and other sided lower l0, is C−. They were plotted on CUSUM chart will indicate the signal of process. When C+ and C− exceed Upper control level (UCL) or Lower control level (LCL) that means that the process signals the out-ofcontrol. UCL and LCL are be chosen by commonly value of H which define by hr. þ Ciþ ¼ max½0; xi  ðl0 þ KÞ þ Ci1 

ð2Þ

 Ci ¼ max½0; ðl0 þ KÞ  xi þ Ci1 

ð3Þ

The constant K is the reference value, where calculated by K = kr. Given r is a standard deviation. In the paper using k = 0.5. The value k can be adjusted where present how sensitive we expect to observe the signal out of control process. The out of control signal can show the inaccurate result when we have not reset upper and lower − + CUSUMs to zero (Ci−1 = 0, Ci−1 = 0) which out of control signal was detected. The comparison between CUSUM chart with and without resetting the signal is given in Fig. 5a and 5b. At the observation 10 was observed being exceeded LCL in Fig. 4a which did not reset C−i−1 = 0. The cumulative sum in observation 11–31 present the biased resulting above upper control level (UCL). Whereas in Fig. 5b, the signal was reset to zero by mean the historical data before out of control signal are independent on recent CUSUM. For more detail discussion on using CUSUM concrete production see in ACI 214 which provides examples of applying the CUSUM chart and discusses some of the difficulties with the CUSUM analysis (American Concrete Institute Committee 214, 2002) [9].

Applications of Statistical Process Control

Nominal Peak Strength,(MPa)

Nominal Peak Strength,(MPa)

UCL=2.62

5.0 2.5 0.0 -2.5 -5.0

LCL=-2.62 1

11 21 31 41 51 61 71 81 91

Cumulative Sum

Cumulative Sum

10.0 7.5

821

4 3 2 1 0 -1 -2 -3 -4

UCL=2.617

LCL=-2.617 1

11 21 31 41 51 61 71 81 91

Observation

Observation

(a)

(b)

Fig. 5. a. CUSUM of Peak strength at 1 day without reset signal, b. CUSUM of Peak strength at 1 day with reset signal

The exponentially weighted moving average (EWMA) control chart. The normal equation is given by: zi ¼ k

i1 X

ð1  kÞ j xij þ ð1  kÞi z0

ð4Þ

j¼0

Where k

iP 1

ð1  kÞ j xij is the weight assigned to the neighbor observations. The

j¼0

weight factor against observation has been shown in Fig. 6. It can be observed when k = 0.3, high contribution assigned to the close observations but less contribution to the far observations comparing to k = 0.1 and 0.2. UCL and LCL in EWMA chart are identified in Eq. 5–7. In this paper, h was defined as 3. The process is out of control chart is detected when the EWMA exceeds the UCL and LCL.

4 Results and Discussion The test data were analyzed based on the assumption that the data were collected on a constant time interval within the five months production period of June 2018–October 2018. While all of the testing on 1 day and 28 days were conducted in accordance with the fabrication schedule, it is possible the climatic changes during this period could have resulted in various curing and conditioning of the samples. The Fig. 7 shows frequency of the data collection based on the production schedule which can be interpreted as a uniform distribution of sampling during the period of study. It is observed that the CUSUM and EWMA charts perform in a similar manner in order to capture the variations in the test data. When CUSUM is used, a constant weighted factor is applied to the data during the observation period. The EWMA approach applies an exponential weight factor to the data that enhances recently obtained data compared to the older data sets. This indicated that EWMA gives a higher priority and

822

C. Pleesudjai et al.

a more significant concern in the current observation domain (neighbor observation) than the older data. UCL ¼ l0 þ hrb

ð5Þ

LCL ¼ l0  hrb

ð6Þ

rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi iffi k h 2i 1  ð1  kÞ b¼ 2k

ð7Þ

Fig. 6. Exponentially weighted factor to the neighbor observation, k

i1 P

ð1  kÞ j

j¼0

Fig. 7. Sample collection data points based on the production schedule of 5 months

Applications of Statistical Process Control

823

4 3 2 1 0 -1 -2 -3 -4

4.75

UCL=2.617

UCL=4.444 __ X=3.994

4.50 EWMA

Cumulative Sum

In the current evaluation, the control chart was applied to 270 specimens within subgroup = 1 (all of the available data) for 28 days and 99 specimens within the subgroup = 1 for 1-day test data. The peak strength chart is shown in Figs. 8a and 8b and indicates that using the 1-day CUSUM chart three observations were out of lower control limits (LCL), while EWMA control chart did not indicate any out of lower control results, possibly because the EWMA relies on neighbor observation than the older data. Note that the upper control limits can be ignored since exceeding the strength results is not typically a penalizing factor. The same observation can be made with the 1-day chart of the Nominal strength at L/600 and at L/150 as shown in Figs. 9a–9b and 10a–10b which indicate that the process is out of control. Similar plots for the peak strength of the 28 day data are shown in Figs. 11a and 11b. At peak strength of 28 days, CUSUM indicates observation points 55–101 showing a signal control process significantly falls below the LCL. The same event is also captured by the EWMA control chart. As a result, one can further evaluate the sample response in this period to determine the causes such as fiber distribution, mix properties, or curing conditions may have affected the residual strength of specimens. Similarly, in Fig. 12a and 12b nominal residual strength L/600 values also indicate a similar reduction trend in the data range of 55–101 observation points. Whereas, the EWMA chart of nominal strength L/150 in Figs. 13a and 13b, did not show out of control process while both peak strength and L/600 nominal strength fell lower than LCL. This may imply that at the high deflection levels, fiber reinforcement plays a more dominant role than the concrete matrix in affecting the residual strength. In other words, fiber properties have less variation than concrete matrix.

4.25 4.00 3.75

LCL=-2.617 1 11 21 31 41 51 61 71 81 91

Observation (a)

3.50

LCL=3.543 1 11 21 31 41 51 61 71 81 91

Observation

(b)

Fig. 8. a. CUSUM of Peak strength at 1 day, h = 4, k = 0.5, b. EWMA of Peak strength at 1 day, h = 3 and k = 0.1

C. Pleesudjai et al. 4 3 2 1 0 -1 -2 -3

3.8

UCL=2.360

UCL=3.662 __ X=3.256

3.6

EWMA

Cumulative Sum

824

3.4 3.2 3.0

LCL=-2.360

2.8

1 11 21 31 41 51 61 71 81 91

Observation

LCL=2.850 1 11 21 31 41 51 61 71 81 91

Observation

(a)

(b)

4 3 2 1 0 -1 -2 -3 -4 -5

UCL=3.064

EWMA

Cumulative Sum

Fig. 9. a. CUSUM of L/600 stress at 1 day, h = 4 and k = 0.5, b. EWMA of L/600 stress at 1 day, h = 3 and k = 0.1

6

UCL=4.391

4.25

__ X=3.864

4.00 3.75 3.50

LCL=-3.064 1

4.50

11 16 21 26 31 36 41 46 Observation

LCL=3.337 1 6 11 16 21 26 31 36 41 46

Observation

(a)

(b)

4 3 2 1 0 -1 -2 -3 -4

UCL=2.739

LCL=-2.739 1

2 8 55 8 2 10 9 136 16 3 190 217 24 4

Observation

(a)

EWMA

Cumulative Sum

Fig. 10. a. CUSUM of L/150 stress at 1 day, h = 4, k = 0.5, b. EWMA of L/150 stress at 1 day, h = 3 and k = 0.1

7.8 7.6 7.4 7.2 7.0 6.8 6.6 6.4 6.2 6.0

UCL=7.517 __ X=7.045 LCL=6.574 1 28

55

82 109 136 163 190 217 244

Observation (b)

Fig. 11. a. CUSUM of Peak strength at 28 day, h = 4, k = 0.5, b. EWMA of Peak strength at 28 day, h = 3 and k = 0.1

Applications of Statistical Process Control

UCL=3.42

2.5 EWMA

Cumulative Sum

5.0

0.0 -2.5 LCL=-3.42

-5.0

6.8 6.6 6.4 6.2 6.0 5.8 5.6 5.4 5.2 5.0

1 8 5 2 9 6 3 0 17 4 2 5 8 10 13 16 19 2 24

825

UCL=6.558 __ X=5.969 LCL=5.380 1

28 5 5 8 2 109 136 163 19 0 2 17 2 44

Observation

Observation

(a)

(b)

5 4 3 2 1 0 -1 -2 -3 -4

UCL=3.382

EWMA

Cumulative Sum

Fig. 12. a. CUSUM of L/600 stress at 28 day, h = 4, k = 0.5, b. EWMA of L/600 stress at 28 day, h = 3, k = 0.1

LCL=-3.382 1 8 5 2 9 6 3 0 17 4 2 5 8 10 13 16 19 2 24 Observation

(a)

6.8 6.6 6.4 6.2 6.0 5.8 5.6 5.4 5.2

UCL=6.446 __ X=5.844

LCL=5.241 1

22 4 3 6 4 8 5 106 12 7 148 169 19 0

Observation

(b)

Fig. 13. a. CUSUM of L/150 stress at 28 day, h = 4, k = 0.5, EWMA of L/150 stress at 28 day, h = 3 and k = 0.1

5 Conclusions By combining the control chart and preliminary statistic analysis section, using CUSUM or an EWMA approach, early detection of shifts in the process mean become possible. Otherwise, the run chart and average Load-Deflection plot can be identified as the quality to reject/accept a concrete batch but they can not measure stability in manufacturing process. The CUSUM and EWMA can address the stability of the production process by analysis variation in peak strength, residual strength results at L/600 and L/150 over a period of time domain. Analysis of the combination of the control charts can identify where the process is in or out of control or capable in segment fabrication. The consumer can decide on accepting or rejecting a sample set. In addition, the tunnel lining producers can use the chart to determine if the monitored process is out of control, and it provides an opportunity to identify the possible causes for the situation.

826

C. Pleesudjai et al.

References 1. Sykora, V.: Amendment of appendix X2 in ASTM C 917 evaluation of cement uniformity from a Single Source. ASTM-Cem. Concr. Aggregates 17(2), 190–192 (1995) 2. Montgomery, D.C.: Introduction to Statistical Quality Control. John Wiley & Sons, NY (2008) 3. Laungrungrong, B., Mobasher, B., Montgomery, D.: Development of rational pay factors based on concrete compressive strength Data. SPR 608 Arizona Department of Transportation, AZ (2008) 4. Leshchinsky, A.M.: Combined methods of determined control measures of concrete quality. Mater. Struct. 24(3), 177–184 (1991) 5. Bakhshi, M., Laungrungrong, B., Bonakdar, A., Mobasher, B., Borror, C.M., Montgomery, D.C.: Economical Concrete Mix Design Utilizing Blended Cements, Performance-Based Specifications, and Pay Factors, FHWA-AZ-13-633, Final Report 633, May 2013 6. Laungrungrong, B., Mobasher, B., Montgomery, D.C., Borror, C.M.: Hybrid control charts for active control and monitoring of concrete strength. ASCE J. Mater. Eng. 22(1), 77–87 (2010) 7. https://asq.org/quality-resources/control-chart 8. Day, K.W.: Concrete Mix Design, Quality Control and Specification, pp. 82–85. Taylor & Francis, New York, NY (2006) 9. ACI 214R-11 Guide to Evaluation of Strength Test Results of Concrete (2012) 10. Zaman, B., Abbas, N., Riaz, M.: Mixed CUSUM-EWMA chart for monitoring process dispersion. Int. J. Adv. Manuf. Technol. 86, 3025–3039 (2016)

Using Decades of Data to Rethink Proportioning and Optimisation of FRC Mixes: The OptiFRC Project Emilio Garcia-Taengua(&) School of Civil Engineering, University of Leeds, Leeds, UK [email protected]

Abstract. Fibres enhance the mechanical properties of concrete, but residual flexural strength parameters present significant variability. The proportioning and optimisation of FRC should integrate fresh and hardened state properties as well as their variability, and this is the urgent challenge addressed by this project funded by the ACI Foundation. It follows a meta-analytical, multivariate approach, based on the creation of an exhaustive database with information on hundreds of FRC mixes compiled from papers published over two decades and adopting a data analytics perspective. First, the relationships between the relative amounts of the mix constituents, fibre geometry and dosage are modelled. Then, the strong correlations between residual flexural strength parameters, limit of proportionality, and compressive strength are exploited to develop efficient predictive tools. The final outcome will be a software package called “OptiFRC”. This will integrate capabilities to access the database compiled, to visualise and utilise the models for the optimisation of FRC mix proportionings, and to calibrate and use the derived quality control charts. This paper presents an overview of the project, reports on its progress to date and summarises the ongoing developments and anticipated impact. Keywords: Characterisation
Optimisation
Proportioning
Residual flexural strength
Variability

1 Introduction Fibre reinforced concrete (FRC) is, by definition, any concrete made primarily of hydraulic cements, aggregates, and discrete reinforcing fibres [1]. Fibres enhance numerous mechanical properties of concrete, particularly tensile, flexural strength and toughness in the cracked state. The use of FRC has evolved from the small scale to the larger scale of routine production and field applications, with tens of millions of cubic yards produced every year [2]. However, the perception that fibres can be treated as an add-on to conventional concrete mixes is still prevalent amongst practitioners. For example, ACI 544.3R-08 [3] stated that, when fibres are incorporated to the mix, “some mixture adjustments may be required”, or “more paste may be needed to provide better workability”. This last statement also puts the focus on fresh state performance and its importance when proportioning FRC mixes. These aspects can no longer be addressed relying only on simple adjustments, considering the family of special concretes that has © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 827–838, 2021. https://doi.org/10.1007/978-3-030-58482-5_73

828

E. Garcia-Taengua

developed around FRC. Fresh state performance is crucial to fibre-reinforced selfcompacting concrete (FRSCC) [4] or ultra-high performance fibre-reinforced concrete (UHPFRC) [5]. Their relevance and increasing prevalence in industry requires that the proportioning of FRC mixes is regarded as a multi-objective optimisation problem, integrating all dimensions shown in Fig. 1.

Fig. 1. FRC proportioning redefined as a multi-objective optimisation problem.

In terms of hardened state performance, the consideration of fibres as reinforcement is now embedded in the ACI 318 code, and their structural contribution is accounted for in design equations [6]. However, the mechanical properties of FRC, particularly the residual flexural strength parameters, present considerable variability [7]. This is highly relevant, as these parameters are the basis of FRC characterisation and specification. Figure 2 shows the flexural test set-up configurations according to standards EN 14651:2005 [8] and ASTM C1609/1609M [9], and an example of stress-strain curve, where the limit of proportionality and the residual flexural strength parameters are defined.

Fig. 2. Characterisation of the flexural behaviour of FRC.

Using Decades of Data to Rethink Proportioning and Optimisation

829

The residual flexural strength parameters (fR1, fR2, fR3, fR4), limit of proportionality (fL) and compressive strength (fc) present strong correlations between them and are mutually interdependent. Treating these parameters as if they were independent variables means that part of the information obtained from characterisation tests is confounded with noise or experimental error. However, their variability and sensitivity to changes in the mix design have been analysed very rarely, and only for specific fibres in specific mixes, in separate studies. Bearing in mind that the prediction and systematic control of variability is key to the effective quality control of FRC production, a meta-analysis of the variability of FRC mechanical properties was urgently needed.

2 Overview of the Project 2.1

Main Objectives

With the aim of addressing the aforementioned challenges, the project “Optimization of Fiber-Reinforced Concrete using Data Mining”, funded by the Concrete Research Council through the ACI Foundation, has the following objectives: • To compile, pre-process, and publish an exhaustive database with mix proportionings and experimental results of FRC mixes available in literature. • To analyse the properties of FRC as a multivariate phenomenon, using data mining techniques. • To develop predictive models, based on the database compiled, that provide accurate estimates of the residual flexural strength parameters of FRC and their variability. • To implement the outcomes from the above in a software (OptiFRC) that assists in the visualisation and interpretation of the database and the predictive models developed. 2.2

Methodology and Work Programme

An overall view of the methodology followed in this project is summarised in the flowchart shown in Fig. 3. It is based on the methodological framework previously applied by the author to the meta-analysis of self-compacting concrete mixes [10, 11]. 2.2.1 Construction of the Database Detailed information on different FRC mix proportionings and their characterisation tests results has been extracted from previous literature and compiled in a database. The sources of information considered for this study are those papers published in ACI Structures, ACI Materials, and journals indexed in ScienceDirect® since 1999, resulting from the search with the terms “fiber-reinforced concrete” or “fibre-reinforced concrete”. The database is prepared in mineable format and consists of two datasets in row-to-row correspondence: one with the FRC mix proportionings, and another with the characterisation tests results. The first dataset comprises the relative amounts of the mix constituents (in kg/m3), maximum aggregate size (in mm), fibres aspect ratio and length (in mm), and fibre material. The second dataset includes values for the slump

830

E. Garcia-Taengua

Fig. 3. Overview of the methodology and correspondence with OptiFRC modules.

and average compressive strength at 28 days, in addition to fL, fR1, fR2, fR3, and fR4. To ensure sufficient statistical power for the subsequent analysis, the database size was defined to include between 500 and 3,000 complete cases [12]. 2.2.2 Meta-analysis of FRC Mix Designs The objective of this meta-analysis is to model the relationships between the relative amounts of the constituents of FRC mixes and to quantify their variation with respect to the type of fibres used, their geometry and dosage. Data mining techniques have been used to carry out a thorough analysis of the latent structure of correlations, namely: multiple linear regression, surface response methodology, and the fitting of conditional probability distributions [13]. Initially the analysis has been segmented by fibre type, and this paper presents the most salient aspects of the preliminary analysis of the database of steel FRC (SFRC). This will be followed by a global analysis of the entire dataset without disaggregation by fibre types, which is work in progress at the moment of writing this paper.

Using Decades of Data to Rethink Proportioning and Optimisation

831

2.2.3 Meta-analysis of FRC Residual Flexural Capacity The structure of correlations and association patterns between the residual flexural strength parameters (fR1, fR2, fR3, and fR4), limit of proportionality (fL) and compressive strength (fc) will be analysed by means of dimension reduction techniques [14] to compress the information represented by these six interdependent variables into a reduced set of independent factors. Multiple linear regression and analysis of variance will then be applied to find equations to relate them to the FRC mix proportioning, not only in terms of average values but also to estimate their variability and sensitivity to changes. 2.2.4 Development of Multivariate Quality Control Charts and the OptiFRC Software The models obtained as outlined in the previous section and the corresponding probability distributions will be used to define quality control charts that can be used in the monitoring of continuous production of FRC mixes. In addition to univariate tools based on Shewhart charts, CUSUM and EWMA, new multivariate charts based on Hotelling’s T2 statistic [15] will be derived. All these outcomes will be implemented in a standalone software package that can be used to access the database compiled, to visualise and utilise the models for the optimisation of FRC mix proportionings, and to use the corresponding quality control charts. It will consist of three modules: Mixes, Trends, and Control. The correspondence of these modules with the outcomes of the different stages of the work programme is also shown in Fig. 3.

3 Meta-analysis of SFRC Mixes At the time of preparation of this paper, the SFRC database consisted of 770 different cases, each case corresponding to a different SFRC mix with detailed information of their proportioning and the characterisation tests results. They have been extracted from more than 100 papers published between the years 2000 and 2019. The following sections present an overview of their analysis. 3.1

Binder Type and Content

The piechart, histogram and box-and-whisker plot shown in Fig. 4 present an at-aglance description of the binder type and contents in SFRC mixes. In almost 42% of the cases, the original sources did not report the type of cement. However, in 94% of the cases for which this information was available, the cement type was either CEM I or CEM II, and CEM I in particular was the most prevalent. In Fig. 4 (right), it can be observed that there were two distinct frequency distributions of binder contents, and therefore two sub-populations could be defined in terms of the total binder content. In 90% of the cases, the binder content was not higher than 710 kg/m3, and the median was 450 kg/m3. The first value can be regarded as the maximum, whilst the median can be considered as a typical, representative value, as it corresponds to the value below which 50% of the cases are found.

832

E. Garcia-Taengua

Fig. 4. Overview of the binder composition and distribution of total binder content values.

In terms of the binder composition, 47.3% of the cases in the database incorporated supplementary cementitious materials (SCMs) in addition to cement. Figure 5 shows the histograms and box-and-whisker plots for the cement content (left) and the SCMs content (right). For the cement content, the median was 400 kg/m3, and in 90% of the cases it was not higher than 660 kg/m3. Regarding the SCMs content, in 90% of the cases it was not higher than 230 kg/m3. However, SCMs contents higher than 100 kg/m3 were associated with cement contents higher than 660 kg/m3.

Fig. 5. Distribution of cement and SCMs contents.

3.2

Fibres and Aggregates

The histogram corresponding to the fibres volume fraction, in percentage, is shown in Fig. 6 (above). In 75% of the cases, the fibres volume fraction was not higher than 1%, and in 90% of the cases it was lower than 1.6%. When the fibre content was higher than this, in the majority of cases it ranged between 1.8% and 2%, very rarely exceeding this last value. In terms of the fibre dimensions, the aspect ratio ranged between 45 and 80 in 90% of the cases in the database (data not shown). An interesting correlation between the grading of the aggregates and the fibres content was observed. The contour plot in Fig. 6 (below) shows how the relationship

Using Decades of Data to Rethink Proportioning and Optimisation

833

between the gravel-to-sand ratio and the total aggregates content changes for increasing fibre contents. Cases with moderate fibre contents, that is, below 0.75% in volume, were associated with gravel-to-sand ratios between 1.5 and 2 but were not limited to any particular range of aggregate contents. However, increasing fibre contents were linked with a decrease in both the gravel-to-sand ratio and the total aggregates content. This can be attributed to the fact that the introduction of higher fibre contents requires the mix design to be substantially more cohesive and to have a higher relative volume of paste.

Fig. 6. Distribution of fibre contents and their relationship to aggregates content and grading.

3.3

Compressive Strength and Limit of Proportionality

The SFRC database represents a wide spectrum of mixes in terms of their average compressive strength at 28 days, as the histogram in Fig. 7 (left) shows. Similarly to what was observed in relation to the total binder content, there are two distinct clusters in terms of compressive strength. The main cluster represented 90% of the cases in the database and comprised mixes with compressive strength not higher than 110 MPa, well distributed around a median of 55 MPa. In terms of the limit of proportionality, all the cases in the database conformed to the same frequency distribution, as the histogram in Fig. 7 (right) shows. The values of this parameter were below 10 MPa in most of the cases.

834

E. Garcia-Taengua

Fig. 7. Distribution of compressive strength and limit of proportionality.

3.4

Residual Flexural Strength Parameters

The residual flexural strength parameters were observed to follow very similar distributions, and only the histogram for fR1 values is shown in Fig. 8. It can be seen that this histogram shows a distribution which is very similar to that observed for the limit of proportionality (Fig. 7), the main difference being its range or width of the distribution, with 90% of the values below 15 MPa.

Fig. 8. Distribution of fR1 values and bivariate scatterplots for fR1, fR2, fR3, and fR4.

Using Decades of Data to Rethink Proportioning and Optimisation

835

More interesting are the bivariate linear correlations between fR1, fR2, fR3, and fR4 and how clear they are. The scatterplots for fR2 versus fR1, fR3 versus fR2, and fR4 versus fR3 are shown in Fig. 8. These plots confirm that, from a multivariate perspective, any of these parameters can be expressed as a direct function of the others. As a preliminary analysis concerned with the sensitivity of these parameters with respect to the mix proportioning, a multivariate regression analysis was carried out for fR1. The following variables were considered: cement, SCMs, sand and gravel contents in kg/m3; cement-to-water and SCMs-to-water ratios; fibres volume fraction in percentage, fibre length in mm, and fibre aspect ratio. It is worth noting that not all the variables related to the mix proportioning were used in this regression analysis, as this was intended to be a simple, exploratory model. Despite that and the fact that the data come from many different sources, this analysis yielded a model with a high goodness of fit, the R-squared being 0.83.

Fig. 9. Contour plots for fR1 in SFRC mixes, based on a preliminary model.

Figure 9 shows some of the contour plots corresponding to the model obtained for fR1, and in all cases the vertical axis represents the fibre volume fraction, in percentage. Figure 9(a) shows that residual flexural strength can be improved not only by increasing the dosage of steel fibres, but also by increasing the cement content. This model predicts that, on average, the effect of increasing the fibre volume fraction in 0.1% is equivalent to that of increasing the cement content in 50 kg/m3, from the point of view of fR1 values.

836

E. Garcia-Taengua

The relationship between these parameters and water content is completed with the contour plot in Fig. 9(b), which shows that the effect of any amount of steel fibres on fR1 is affected by the cement-to-water ratio of the mix. In fact, this model predicts that, on average, the cement-to-water ratio which can be considered optimal in enhancing the effect of fibres is 2.5, which is equivalent to a water-to-cement ratio of 0.40. A similar observation can be made with respect to the sand content. According to the contour plot shown in Fig. 9(c), this model predicts that, on average, sand contents close to 950 kg/m3 are best in maximising the performance of fibres in terms of net increase in fR1 per unit volume of fibres. In fact, any of these contour plots can be used as double-entry graphs to obtain different pairs of values associated to the same predicted performance, as the contour lines correspond to constant fR1 values. Finally, the effect of the fibres dimensions is partially explored through the interaction between aspect ratio and volume fraction, in Fig. 9(d). This contour plot reflects the fact that residual flexural strength of SFRC can be improved by increasing either fibre content or aspect ratio, or a combination of both. The relationship between them is, however, not linear, and therefore different combinations of aspect ratio and volume fraction can be adequate depending on the fR1 value adopted as target. Of course, the contour plots presented in Fig. 9 are only partial examples of a preliminary model based on SFRC data, but they illustrate how the software tool resulting from the OptiFRC project will present the information and predictive models for optimisation purposes.

4 Summary and Conclusions This paper presents an overview of the ongoing project “Optimization of FiberReinforced Concrete using Data Mining” and its progress to date. The most salient aspects can be summarised as follows: • The main objectives of this ongoing research project are: to compile an exhaustive database of FRC mix proportionings and their properties; to analyse their variability using data mining techniques; to develop robust models for estimating the residual flexural strength; and to implement these developments in an optimisation tool (software OptiFRC). • Detailed information on hundreds of FRC mixes has been compiled from papers published in indexed journals and conference proceedings published since 1999. The database contains: relative amounts of mix constituents, maximum aggregate size; fibres aspect ratio, length, material, and fibre content; slump and compressive strength; and fL, fR1, fR2, fR3, and fR4 values. • A preliminary analysis has been carried out on a dataset of SFRC mixes consisting of 770 cases, extracted from more than 100 papers published over the last two decades. It was found that the fibre volume fraction was typically below 1%, and very rarely exceeded 1.6%. • Cements type CEM I and CEM II were found to be the most common in SFRC mixes. In almost half of the cases compiled, SCMs were used. In 90% of the cases, the total binder and cement contents were below 710 kg/m3 and 660 kg/m3,

Using Decades of Data to Rethink Proportioning and Optimisation

837

respectively. In 50% of the cases, total binder and cement contents were not higher than 450 kg/m3 and 400 kg/m3, respectively. • SFRC mixes with fibre contents up to 0.75% in volume were associated with gravel-to-sand ratios of 1.5 and above. Increasing fibre contents were linked with a decrease in both the gravel-to-sand ratio and the total aggregates content. • Regarding the average compressive strength at 28 days, it was not higher than 55 MPa in 50% of cases, and values higher than 110 MPa corresponded to only 10% of the SFRC mixes compiled. All residual flexural strength parameters followed a lognormal distribution, and it was confirmed that the linear correlation between them was very strong. • A preliminary analysis of the fR1 values with respect to the SFRC mix proportions revealed that, to optimise the efficiency of fibres in improving the residual flexural strength, a water-to-cement ratio of 0.40 and sand contents around 950 kg/m3 were optimal. Acknowledgements. The author wishes to thank the Concrete Research Council, the ACI Foundation and the American Concrete Institute (ACI) for the financial support awarded to the project “Optimization of Fiber-Reinforced Concrete using Data Mining” (2019–2021). The endorsement of fellow members of ACI Committee 544 as well as the support of industrial partners AECOM, OwensCorning and Banagher Concrete is also acknowledged.

References 1. ACI Committee 544: ACI 544.1R-96 Report on Fiber Reinforced Concrete. American Concrete Institute (1996) 2. ACI Committee 544: ACI 544.2R-17 Report on the Measurement of Fresh State Properties and Fiber Dispersion of Fiber-Reinforced Concrete. American Concrete Institute (2017) 3. ACI Committee 544: ACI 544.3R-08 Guide for Specifying, Proportioning, and Production of Fiber-Reinforced Concrete. American Concrete Institute (2008) 4. Ferrara, L., Park, Y.D., Shah, S.P.: A method for mix-design of fiber reinforced selfcompacting concrete. Cem. Concr. Res. 37(6), 957–971 (2007) 5. Habel, K., Viviani, M., Denarié, E., Bruhwiler, E.: Development of the mechanical properties of an ultra-high performance fiber reinforced concrete. Cem. Concr. Res. 36(7), 1362–1370 (2006) 6. ACI Committee 544: ACI 544.4R-18 Guide to Design with Fiber-Reinforced Concrete. American Concrete Institute (2018) 7. Cavalaro, S.H.P., Aguado, A.: Intrinsic scatter of FRC: an alternative philosophy to estimate characteristic values. Mater. Struct. 48(11), 3537–3555 (2014). https://doi.org/10.1617/ s11527-014-0420-6 8. European Committee for Standardization: EN 14651:2005. Test method for metallic fibre concrete. Measuring the flexural tensile strength (limit of proportionality (LOP), residual) (2005) 9. ASTM International: ASTM C1609/C1609M. Standard Test Method for Flexural Performance of Fiber-Reinforced Concrete (Using Beam with Third-Point Loading) (2012)

838

E. Garcia-Taengua

10. Garcia-Taengua, E., Marti-Vargas, J.R.: Multivariate analysis of the fresh state parameters of self-consolidating concrete. In: Proceedings of the 8th International RILEM Symposium on Self-Compacting Concrete SCC2016, pp. 221–231 (2016) 11. Garcia-Taengua, E.: Fundamental fresh state properties of self-consolidating concrete: a meta-analysis of mix designs. In: Advances in Civil Engineering, p. 5237230 (2018) 12. Hair, J.F., Black, W.C., Babin, B.J., Anderson, R.E.: Multivariate Data Analysis. Pearson Education Inc., Upper Saddle River (2010) 13. Button, K.S., Ioannidis, J.P.A., Mokrysz, C., Nosek, B.A., Flint, J., Robinson, E.S.J., Munafò, M.R.: Power failure: why small sample size undermines the reliability of neuroscience. Nat. Rev. Neurosci. 14, 365–376 (2013) 14. Jolliffe, I.T., Cadima, J.: Principal component analysis: a review and recent developments. Philos. Trans. Roy. Soc. A 374(2065), 20150202 (2016) 15. MacGregor, J.F., Kourti, T.: Statistical process control of multivariate processes. Control Eng. Pract. 3(3), 403–414 (2016)

Case Studies: Structural and Industrial Applications

Fire Resistance of Steel Fibre Reinforced Concrete Elevated Suspended Slabs: ISO Fire Tests and Conclusions for Design Xavier Destrée1(&), Andrejs Krasnikovs2, and Sébastien Wolf3 1

R&D ArcelorMittal Fibres, Bissen, Luxembourg [email protected] 2 Academy of Sciences, Riga, Latvia 3 ArcelorMittal Fibres, Bissen, Luxembourg

Abstract. Since 2004, a number of projects of multi-storey buildings have been completed in various countries where steel fibres are the only reinforcing of the cast in-situ elevated slabs. A number of full-scale tests at cold stage of these steel fibre reinforced concrete slabs have been conducted to show a very considerable safety margin between the most onerous service loading conditions and the ultimate cases at final collapse. The positive influence of steel fibre reinforcing on the material properties of the concrete under fire conditions is already well known and detailed in the literature dating back 25 years. In 2015, the ACI 544-6R 15 document, “Report on Design and Construction of Steel-Fiber Reinforced Concrete Elevated Slabs” outlines in detail the design method of such slabs but mentions also that further research is needed in the Fire Resistance area. The purpose of the paper here is precisely to address the Fire Resistance of S.F.R.C suspended elevated slabs and walls in full scale under service loadings. The paper outlines the setting-up of the tests of a number of elevated slabs and walls, subjected to the real loading conditions, the observations and then to conclude and confirm by the high positive impact on the design as well as the high safety of these slabs and walls. Keywords: Fire resistance


Steel fiber
Concrete
Slabs
Walls

1 Introduction The steel fibre reinforced concrete as a material has been investigated by many authors in the last 35 years under high temperature starting from 200 °C to 1100 °C. The A.C.I. report 544.5R-13 on “Physical properties and durability of fibrereinforced concrete” concludes that “Reinforcement in the form of steel fibres only,…, generally improves the performance of structural concrete members under extreme temperature and fire” and further “… fibres have been successful in extending the safe time of fire exposure for many practical and proven applications. Extending the safe time of fire exposure allows firefighters more time both to evacuate structures and to extinguish the fire safely” © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 841–851, 2021. https://doi.org/10.1007/978-3-030-58482-5_74

842

X. Destrée et al.

Kodur (1995) observed that “results indicate that the compressive strength at elevated temperatures of steel-fibre-reinforced concrete is higher than that of plain concrete. It is concluded that the mechanical properties of fibre-reinforced-concrete are more beneficial to fire resistance than those of plain concrete. The steel fibres prevented early cracking and also contributed to the compressive strength of concrete at elevated temperature” One reason is that the ductility of normal strength concrete with steel fibres is relatively stable despite the increase of temperature according to Widhianto et al. (2014). The presence of steel fibres increases also the ultimate strain as outlined also by the Canadian Steel Construction Institute in the Bulletin N° 26 where it is concluded that “With addition of steel-fibre reinforcing, fire endurance ratings up to 3 h are possible. Much like the bar-reinforcing, the steel fibres confine the concrete core by resisting the splitting forces generated as the concrete degrades and approaches its ultimate load carrying capacity”. Bednar and Wald (2012) concluded that “The material test of fibre concrete demonstrated that its behaviour of fibre concrete is ductile enough to create a membrane action if the cracks are developed at yield lines. The ductility of fibre concrete at 600 °C was higher than at 20 °C. The shear resistance of the fibre reinforced is significantly higher compared to concrete”. These effects give a great benefit, especially in redundant structural members such as slabs and walls. Nurchasanah and Massoud (2016) conclude it is expected that in future concrete having steel fibre will act as fire protective considerably (Ref. 5). Klingsch (2014) recommended that steel fibers should always be added to HPC and UHPC mixes since they increase the ductility of the concrete.

2 Purpose of the Tests A multi-storey typical residential building in Riga-Latvia, whose scaled model can be seen in Fig. 1, was taken as first case to design a slab and wall structure of twelve levels built in steel fibre reinforced concrete. The whole frame is built using tunnel forms where the walls and the slab of one level are cast in one single operation as shown in Fig. 2. After 18 h, the tunnel forms are removed and installed to form the next level up. Each level of the building is designed of 180 mm thick walls to carry the cast in situ slabs spanning 5,40 m, a span to depth ratio of 30 and where both slabs and walls are only reinforced by 50 kg/m3 of ArcelorMittal HE+ 1/60 steel fibres of 1 mm diameter, 60 mm length, 1500 N/mm2 steel wire tensile strength, provided with hooked ends. In the slab depth with 50 mm coverage from the bottom face, one rebar of 16 mm diameter is included at every 1,5 m distance apart in order to meet the Anti-Progressive Collapse condition following the ACI 544 6R15 provision J.1, p. 36. (ref 7) in case a wall should disappear by accident or explosion (Fig. 3). A full-scale test at cold temperature, reported by Kleinman and Destrée (2012) demonstrated that the first negative moment crack live load of such a slab was 6,5 kN/m2 and the ultimate loading was of 11 kN/m2 to be compared to the most onerous SLS service loading is of 3 kN/m2.

Fire Resistance of Steel Fibre Reinforced Concrete Elevated Suspended Slabs

Fig. 1. Overview of the project in Riga, Latvia

Fig. 2. Tunnel formed level installation

843

844

X. Destrée et al.

Fig. 3. Typical level cross sectional

The last question to resolve is about the fire resistance or rating of both slabs and walls and therefore the testing program has been defined, budgeted and decided.

3 Design of the Slab Under the Ambient Temperature The 180 mm slab of a 5,40 m multiple continuous span (center to center), under a 3 kN/m2 total variable loads, typical of residential or office applications, undergoes a maximum envelope positive Mmax = M. (A0—A1) of the edge span where p are the permanent loadings and s the variable loadings (Fig. 4).

Fig. 4. Maximum bending moment value of a multiple span system

The initial deflection is calculated to be of 2,4 mm as observed in reality under the test load as shown in Fig. 5, with a modulus of elasticity E = 31.500 N/mm2.

Fig. 5. Calculated deflection in the current loading system

Fire Resistance of Steel Fibre Reinforced Concrete Elevated Suspended Slabs

845

M max = 15,99 kN.m/m generates a flexion stress f SLS = 2.96 N/mm2 under unfactored loadings. The slab is also provided with a set of Anti-Progressive Collapse Rebars according to ACI544-6R15 and the Canadian CSA-A23,3-04, in order to prevent the slab to fall down in case of a supporting wall collapse, by accident, explosion or terrorist attack. The APC rebar, in the span-direction only, are 16 mm diameter rebars in the bottom of 16 mm diameter every 1.50 m distance apart. These APC rebars are bottom rebars with 50 mm coverage thus a very small of 0,07% of reinforcing or only of 3.6 kg per cubic metre of steel rebar, a very little proportion indeed as shown at installation in Fig. 6.

Fig. 6. SFRC slab pouring condition with APC rebars.

The principal reinforcing is obtained by the addition in the ready mixed concrete of 50 kg/m3 ArcelorMittal steel fibres of 1 mm diameter by 60 mm length made out of a 1500 N/mm2 constituent steel wire. The steel fibre (HE+ 1/60 type) are provided with hooked ends. The concrete matrix, supplied as a C25-30 with 280 kg CEM I and W/C < 0,50 with 16 mm maximum aggregate size was of 23,83 kN/m3 density but indeed showed a 52,6 N/mm2 average compressive strength at 28 days! As seen in Fig. 5, the flowability was such that no poker vibrating was needed. Suspended elevated slabs in steel-fibre reinforced concrete can be designed today following three available standard documents: • The ACI (American Concrete Institute) 544–6R15 “Report on Design and construction of Steel Fiber-Reinforced Concrete Elevated Slabs”. • The FIB Model Code 2012 (in its chapters 5 and 7) • The SS-EN 812310 in Sweden, “Design of Fibre Concrete Structure”.

846

X. Destrée et al.

4 Wall Design The walls were also of 180 mm thickness like the slabs, 3 m height and 4 m length, built in the same concrete as the slabs including 50 kg/m3 HE+1/60 steel fibres but without any APC rebars so that the walls were without any rebars, only steel fibres. The walls are subjected under the test fire to a 330 kN per lineal meter of vertical loading intensity, typical of a ten floors high building outer walls. The fire test program includes three identical walls subjected to a 330 kN/m lineal loading under fire.

5 Temperature Regime of the Fire Resistance Test The test itself has been organized and carried-out by the F.I.R.E lab in Poprad (Slovakia - SK) with a furnace set to follow the ISO 834, temperature vs. time diagram according to the following formula: T ¼ 345 log ð8t þ 1Þ þ 20

ð1Þ

As shown in Fig. 7, the ISO 834 and ASTM E119 standards are almost similar, the ISO being still of a slightly higher temperature from 60 min to 180 min fire duration and is thus slightly more stringent than the ASTM E119.

Fig. 7. ISO and ASTM fire curves

Such an ISO/ASTM fire is also called cellulosic fire and is typical of houses and condos. The hydrocarbon fire is much more severe but only applicable where oil and gas are involved like in petrochemical plants.

Fire Resistance of Steel Fibre Reinforced Concrete Elevated Suspended Slabs

847

6 The Fire Resistance Test Set-up The test set-up is according to EN 1365-2 and EN 13381.3.2015 standards, “Test methods for determining the contribution of the fire resistance of structural members” where a statically determinated 180 mm thick slab of 4 m net span, 3 m width is subjected to two line loadings so that the flexural moment is constant and maximum over 2,40 m length as shown in Fig. 8 (scheme) and in Fig. 9 (detail) including the location of the thermocouples and deflectometers.

Fig. 8. Statical system for the slab

Fig. 9. Location of the thermocouples and the deflectometers in the salb

848

X. Destrée et al.

The fire test program includes four identical slabs. The first slab has been fire tested under an initial bending stress of 2.5 N/mm2, indeed the same positive moment and stress to occur in a residential use condominium slab under its most onerous loading condition. The three next slabs are loaded even more up to 3.5 N/mm2 bending stress, thus 30% more than the most onerous loading in service condition in reality.

7 Fire Tests Observations and Results The first slab under 2.5 N/mm2 flexion strength could sustain the loading during more than four hours so that the test had to be stopped. The slab at 240 min fire duration was still stable and the deflection increase was still smaller than the 11 mm per minute as per the standards. Thus, no collapse was observed, even after 4 h! The slab was still flame tight at 240 min and its upper surface temperature below the 140 °C. The observed fire resistance has been exceeding far more the maximum standard requirement of 2 h for a slab! The three next slabs were fire tested under a flexion stress of 3,5 N/mm2 and met the fire resistance standard requirements up to respectively 192 min, 149 min and 204 min when the slabs collapsed, thus more than the 2 h limit although the flexion stress rate was 30% higher than any possible 3 kN/m2 variable loading in a real structure! It is important here to repeat that the tested slabs were tested as statically determinated as specified in the EN 1365-2 fire rating standards and this unlike in reality, where the slabs are statically indeterminated and can benefit of the plastic moment redistribution. Quite above any expectations for statically determinated slabs, a large number of bottom cracks (ca. 16 cracks) are observed and have grown up from the bottom as it is visible in the Fig. 10. Up to 192 min, 149 min and 204 min, the deflection was smaller than the 250 mm standard limit and its increase has been under 11 mm per minute limit as also required by the standard. The steel fibre concrete has been quite able to prevent the explosive spalling from the bottom so that a layer of burnt but insulating concrete remained attached around and by the fibres, indeed a very effective protection of the concrete core. The beneficial effect is shown in Fig. 11 and Fig. 12 where the temperatures at various depths from the top surface (30 mm depth) to the bottom (150 mm depth) so that it can be observed that at 150 mm depth from the top surface, at 4 h fire, the temperature remains as low 500 °C when the concrete offers still 35% of its initial strength at cold. In the top surface, the temperature did not exceed 70 °C, a lot colder than the maximum 140 °C according to the standard limit.

Fire Resistance of Steel Fibre Reinforced Concrete Elevated Suspended Slabs

849

Fig. 10. View of the slab over the heating room under at 192 min

Fig. 11. Recorded Temperatures at various depths during the fire test.

One wall only shown on Fig. 13. of the three walls, at the time of this report, could have been tested under fire but it nevertheless showed full compliance after 4 h fire to all performance criterions.

850

X. Destrée et al.

Fig. 12. Temperature vs. time from bottom in red (1150 °C at 4 h) to top (100 °C at 4 h) in blue and at 30 mm-orange); 60 mm-grey; 90 mm-yellow; 120 mm-blue; and 150 mm-green.

Fig. 13. The tested wall at 241 min fire under 330 kN/m vertical loading

The wall has also been stable and flame tight, with a cold surface temperature at 70 °C under the 140 °C limit. The wall met and exceeded as well all standard requirements of resistance to fire. Indeed, the fire test has been stopped at 242 min without any observed failure to meet the EN1365-3 requirement. More details are given in the Fig. 14 here below:

Fire Resistance of Steel Fibre Reinforced Concrete Elevated Suspended Slabs

851

Fig. 14. The wall official report table to report the 4 h resistance to fire.

8 Conclusions Three identical statically determinated steel fibre reinforced concrete slabs and one wall have been tested to confirm the full compliance under fire to EN 1365-2 and EN 1365-3 thus well over the 2 h regarding slabs and over 4 h regarding the only wall tested. Two more walls shall be tested in the same experimental programme. Steel Fibre Reinforced Concrete used as the main reinforcing of elevated suspended slabs and designed according to the provisions of the ACI 544 6R15, has fulfilled and exceeded the standard safety requirements of EN 1365-2. The thermal spalling has been prevented by the 50 kg/m3 steel fibres reinforcing, so that the concrete core remained insulated from the very high temperatures in the bottom. The prevention of spalling under fire is clearly one of the reasons why the steel fibre reinforcing is a viable and safe method for free suspended elevated slabs as designed following the ACI 544-6R15 report provisions.

References Lie, T.T., Kodur, V.K.R.: Mechanical properties of fibre-reinforced concrete at elevated temperatures. Internal Report N° 687, National Research Council Canada (1995) Kleinman, C., Destrée, X.: Steel fibre as only reinforcing in free suspended one-way elevated slab: design conclusions of a tunnel formed slab and walls based upon full scale testing results. In: BEFIB 2012, 8th RILEM International Symposium on fibre Reinforced Concrete: Challenges and opportunities (2012) Klingsch, E.W.: Explosive spalling of concrete in fire. Doctoral thesis, ETH Zürich (2014) Widhianto, A., Darmavadi, D.: Fire resistance of normal and high-strength concrete with contains of steel fibre. Asian J. Civ. Eng. 15(5), 655–669 (2014) Nurchasanah, Y., Massoud, M.A.: Steel fiber reinforced concrete to improve the characteristics of fire-resistant concrete. In: Applied Mechanics and Materials, vol. 845, pp. 220–225 (2016) Bednar, J., Wald, F.: Membrane action of composite fibre concrete slab in fire. Proc. Eng. 40, 498–503 (2012). Steel Structures and Bridges ACI 544-6R15: Report on design and construction of steel fiber-reinforced concrete elevated slabs (2015)

Structural Behavior of a Traditional Concrete and Hollow Tiles Mixed Floor Reinforced with HPFRC D. Sirtoli1(&), P. Riva1, and P. Girardello2 1

Department of Civil Engineering, University of Bergamo, Bergamo, Italy [email protected] 2 Kerakoll S.p.A., Sassuolo, Italy

Abstract. The use of high performance fiber reinforced concrete (HPFRC) in structural strengthening is widely accepted. The presence of steel fibers inside the cementitious matrix provides the hardened matrix enhanced resources in terms of strength and toughness, avoiding brittle crisis in favor of hardening/softening postpeak behavior. These performances open different scenarios in structural rehabilitation, e.g. RC column jacketing, shear strengthening of RC beam, slab reinforcement. However, the quality of the intervention is highly dependent on the technique considered for the element strengthening and on the HPFRC performance, which is strongly related to the mix design. This paper presents the results of an experimental campaign carried out on the flexural response of 4.5
1 m2 one-way floor elements reinforced with a commercial HPFRC material. The mechanical performances of the HPFRC adopted have been qualified following the Italian guidelines published in January 2019 for the identification and qualification of FRC products. The effect of the FRC strengthening was evaluated by comparing the performance of unreinforced and retrofitted floor elements, both using traditional r.c. and HPFRC topping. The effectiveness of the intervention is evaluated in terms of flexural strength, slab deflection and visible damage. keywords: Steel fiber-reinforced concrete

 Slab strengthening  Retrofitting

1 Introduction High performance fiber reinforced concrete (HPFRC) is a well-known composite material with good performance in compression thanks to the concrete matrix combined with the improved post-cracking behavior provided by the fibers crack opening control. Thanks to its performance, FRC has been used in several structural interventions for a multitude of purposes [1]. As a dispersed reinforcement it found application in industrial floors and pavements [2] or building slabs [3]. Its combination with steel bars grants resources against impact and fatigue, useful for building structural elements in seismic zones [4]. However, the effectiveness of the intervention is strongly dependent to the correlation between its design and the FRC mechanical properties. For both these aspects, codes of practice were produced, with specific indications on the definition, realization and application of FRC. In terms of design details and how to treat them, one of the reference codes is represented by the Model © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 852–863, 2021. https://doi.org/10.1007/978-3-030-58482-5_75

Structural Behavior of a Traditional Concrete and Hollow Tiles

853

Code 2010 [5], based on researches of Di Prisco [6] and others. On the other hand, in order to guarantee the quality of the material introduced into the market, each nation issues guidelines, as a grant for the final user and protection for the producer. Italy, in 2019, published its own reference document for the identification, qualification, certification of technical evaluation and acceptance criteria of fiber reinforced concrete [7]. This paper has the aim to investigate the mechanical behavior of a traditional oneway slab, made of a mixed reinforced concrete and clay brick void formers, reinforced with FRC which properties were determined following the recently published Italian guidelines. The performance of an un-reinforced sample is compared to four samples strengthened with different solutions. The obtained data are used to evaluate the effectiveness of the intervention, giving indications on the considered technique.

2 Materials and Methods 2.1

Samples Details

The structural elements subjected to the flexural strength test are traditional reinforced concrete and hollow-core bricks mixed floors of 4.5
1 m dimension in length and width respectively, reinforced with a topping layer of fiber reinforced concrete (FRC) of different thicknesses. The details of the traditional floor are reported in Fig. 1. A summary of the FRC mechanical performances is given in Table 1 with a typical graph for the flexural test performed in accordance with EN 14651 proposed in Fig. 2. These data are taken from an experimental campaign organized following the indications of the Italian guidelines for the identification, qualification, technical evaluation certification and acceptance control for FRC [7]. Five different samples were delivered to the laboratory, four strengthened and one plain. The basic sample is composed of hollow-core bricks of 38
16
25 cm dimensions, concrete with strength class C1215 and steel bars of B450C class and 12 mm diameter. The floor was strengthened following two techniques: the first is an FRC topping with three different thicknesses of 2, 3 and 4 cm while the second is composed of a welded mesh incorporated in a 4 cm topping layer made of HPC. Before casting the strengthening material, the upper surface of the basic sample was prepared by increasing its roughness in order to improve the adhesion between the old and new materials. No primer of shear stud connectors was used in any of the specimens to ensure shear stress transmission between the original topping and the retrofitting topping. The obtained results on the strengthened samples are compared with the one gathered from the un-strengthened floor sample in order to evaluate the efficacy of the strengthening technique by analysing the improvements on the structural element performances. The sample weight was considered in the calculation, adding 9.26, 11.42, 12.50, 13.58 and 13.58 kN to the total load for NR, R2, R3, R4 and RT respectively. The investigated samples of this experimental campaign are identified as follow: 1_NR: un-strengthened sample; 2_R2: strengthened sample with FRC topping of 2 cm; 3_R3: strengthened sample with FRC topping of 3 cm; 4_R4: strengthened sample with FRC topping of 4 cm; 5_RT: strengthened sample with HPC topping (Magma) of 4 cm combined with a welded mesh.

854

D. Sirtoli et al. Table 1. Mechanical performances of the FRC used in the experimental campaign.

Compressive strength [MPa] Rcm Rck 109.3 ± 1.71 106.50 Flexural strength [MPa] ffct,L fR,1 m c m 6.95±0.63 5.91 12.21±1.74 Tensile strength fFts [MPa] m c 7.40±1.20 5.32

fck 88.39

c 9.54

fR,3 m 9.88±1.62

Strength class C80/95

c 7.33

fR,3k/fR,1k

HC

0.768

8b

ε [%] m 0.0412

Fig. 1. Geometrical details of the floor section.

Fig. 2. Typical CMOD curve reported in the Italian guidelines [7].

Structural Behavior of a Traditional Concrete and Hollow Tiles

2.2

855

Samples Preparation and Storage

The floor samples were realized by an external company under the indication of the University of Bergamo. Once the floor samples were realized, the roughness of the top surface of the samples to be strengthened was improved by hydro-sand blasting in order to have better bond between old and new concrete topping material. Eventually, the topping FRC was casted and left as is until the day of the test, without particular curing attention, other than a sheet covering to avoid excessive drying for the first 24 h. The samples were sent to the laboratory before the 28 days of maturation, where they remained in laboratory environment conditions for, at least, 24 h before performing the test. 2.3

Set-up

The test done on the floor samples is a four-point bending test in deflection control. A general view of the used set-up is shown in Fig. 3. The two supports are realized using metallic cylinders of 7 cm diameter and length sufficient to cover the whole transverse section in order to sustain all the three joists. These cylinders are lodged in special metallic plates designed to block longitudinal movement while the sample is still free to rotate during the test, creating a hinge restraint. Between cylinders and joists a neoprene layer is placed to uniformly distribute the load on the sample transverse section. For the load application, a 500 kN electromechanical jack to apply the force vertically to the sample was used. A system composed of three metallic beams is used to distribute the single force of the jack on the two load lines defined for the test. Two beams lay transversally on the slab element, representing the load points and the supports for the third beam, which is placed longitudinally to the sample, distributing half the jack load to each single supporting beam. A ball bearing was placed between jack and longitudinal beam, in order to grant an equal load distribution. Between the two supporting beams and the upper floor surface a neoprene layer is placed in order to avoid any load concentration. The data acquisition has started in a configuration where the sample is completely unloaded. Afterwards, the two metallic supporting beams are positioned on the slab sample and loaded by the third. This own weight of the loading system is not recorded in the test, but considered on the final data elaboration. Considering the test protocol, just in the case of the un-retrofitted sample three cycles at incremental deflection level were considered in order to stabilize the floor response. For the other samples, a monotonic test was performed at a constant deflection rate of 0.05 mm/sec until collapse or complete plasticization.

856

D. Sirtoli et al.

Fig. 3. Set-up for the flexural test performed on the floor samples

2.4

Instrumentation

During the test, the displacements of the floor in seven different positions were recorded together with the applied force. In Fig. 4 the points where the instruments are applied with their identification name are shown.

Fig. 4. Instruments used during the test and their identification (De: Deflection; Dr: Drop; l: left; r: right)

Four potentiometric wire instruments with maximum stroke of one meter were used to register the deflections (De) in the mid floor section and under the two loading points. Considering “De mid” and “De r” the wire was placed at the mid-joist intrados. Instead, in Sect. 2, two points were used to record the deflection, one at each slab side,

Structural Behavior of a Traditional Concrete and Hollow Tiles

857

in order to analyse the likely rotation occurring on the transverse plane due to local failure or bad load redistribution. As under the two supports a series of neoprene layer were placed in order to balance the slab and equally redistribute the load to each of the three joists, three displacement transducers LVDT were placed at each floor end (Dr). Even in this case, on one side of the floor a single LVDT “Dr l” was used, placing it at the mid-section extrados over the support. On the other side of the floor, instead, two LVDT instruments were placed close to the lateral joists, symmetrically respect the mid-point, over the support on the slab extrados. As in the case of the deflection investigation, even here the two instruments are aimed to record bad redistribution or local failure of the lateral joists, besides the vertical settlement due to the compression of the neoprene layer. In both cases of deflection and support settlement investigation, the behavior of the sections in which two instruments were considered are described by their mean value, though unexpected behavior is identified. 2.5

Considered Formulations

2.5.1 Results Interpretation The set-up defined for the flexural tests on the floor elements yields to the static scheme reported in Fig. 5. The two extradoxal loads are redistributed on the whole width of the specimen, equal to one meter, while their distance from the supports is equal to one meter. Considering the maximum bending moment of the scheme to the left, setting it equal to the maximum moment of the right one, the equivalent distributed load is obtained. It represents that particular load which results in the same maximum bending moment and shear obtained with the test configuration. This operation is useful in order to compare the results obtained in the laboratory tests with the more realistic configuration of the buildings, where uniformly distributed loads are used for the structural design. Particular attention must be paid when comparing the results obtained in both configurations as the redistribution of the bending moment and shear is different. The reinterpretation of the acquired results must be done taking into account this aspect.

Fig. 5. Considered static schemes.

858

D. Sirtoli et al.

2.5.2 Theoretical Strength Theoretical flexural and shear strength of the analysed samples were obtained from the formulations reported in the main design codes and guidelines available. Being an experimental test on a real scale element, reduction coefficients in materials performance “c” are set equal to 1. Due to the slab transverse section dimensions, theoretical results for the single joist were doubled in order to take into account the presence of other two half joists at the floor sides. Considering the theoretical bending moment calculation for both strengthened and un-strengthened slab samples, the HPFRC Reluis guidelines were considered, though not published yet. Differently, for the shear strength, the indication reported in the Italian standard NTC 18 were considered. In the following extracts from the used codes are illustrated. Bending Moment The bending moment evaluation is done referring to the scheme of Fig. 6, where two different configurations are available, depending on whether the position of the neutral axis lays in the FRC topping (Fig. 6a) or in the old concrete layer (Fig. 6b). For both normal and fiber reinforced concrete, a stress-block distribution is considered in the compressed section. As the FRC topping layer acts mainly in the compressed zone and its thickness is limited, the possible tensile contribution is neglected. Firstly, the neutral axis position is evaluated, assuming its starting position inside the FRC topping layer, verifying thereafter its correctness. The neutral axis position is calculated through normal forces equilibrium: AS  fyd ¼ 0:8  x  fFcd  B ) x ¼

AS  fyd 0:8  fFcd  B

ð1Þ

If x < 1.25 hreinf the resistant bending moment is calculated with the next formulation: MRd ¼ 0:8  x  fFcd  B  d þ hreinf  0:4  x




ð2Þ

If x > 1.25 hreinf the neutral axis position must be calculated again with a new equilibrium: 
 AS  fyd ¼ hreinf  fFcd þ 0:8  x  hreinf  fcd  B

ð3Þ

Then, the reinforced floor resistant bending moment is calculated: MRd ¼ hreinf  fFcd

   
hreinf 0:8  x  hreinf B dþ þ 0:8  x  hreinf  fcd  B  d  2 2 ð4Þ

Shear All the calculation for the resistant shear are reported in chapter 4.1.2.3.5.1 of the NTC18 standard.

Structural Behavior of a Traditional Concrete and Hollow Tiles

859

Fig. 6. Stress and strain diagrams for the FRC reinforced section: (a) neutral axis inside the FRC thickness; (b) neutral axis inside the old concrete section.

2.6

Samples Real Conditions

For a better interpretation of the results reported hereafter and obtained from the test described above, the actual characteristics of the analyzed samples are provided. Considering floor R3, its collapse mechanism highlights the presence of a diagonal hole in longitudinal direction of diameter Ø5 cm and 50 cm length on one of the side joists, exactly under the point where the load is applied. Moreover, demolishing the midsection for a final inspection, just three of the four steel reinforcement bars were found, missing one of the two in the mid joist. Another aspect which was underlined by the collapse mechanism, even in sample R4, was the thin concrete cover to the longitudinal reinforcement, which turned out to be less than 1 cm in both extradoxal and lateral directions. Eventually, sample RT reported a mistake in the reinforcement overlap in one of the side joists where the two steel bars were placed head to head and not overlapped, creating a section with no flexural resources.

3 Results Results are reported in Fig. 7 in terms of shear/bending moment – mid deflection for some key-points describing the general behavior of the samples. The same results are summarized in Table 2, showing the deflections as a function of the sample length, while in Table 3 are reported data in terms of strength, derived from codes formulations (rd) and laboratory test (ex).

860

D. Sirtoli et al.

Figure 7 shows clearly the improvements due to the slab strengthening. Compared to the black line, which represents the un-reinforced sample, the other curves showed higher stiffness, maintained for higher load levels. Sample NR show a stiffness decrease at around 40–50 kN, which is the moment in which a longitudinal crack appears on the top surface of the sample from one of the two supports. The test was stopped before collapse, when the curve started to flatten, reaching 61.3 kN load which corresponds to 15.33 kN/m equivalent distributed load. Considering the curves representing samples strengthened with FRC topping layer, two of three samples evidenced a brittle collapse at a load of 68.56 and 68.26 kN. The peculiarity in this situation is that the two samples with thicker FRC layer evidenced this collapse, while sample R2 reached a maximum load of 72.64 kN deforming thereafter at a constant stress level. It must be underlined that sample R2 did not show any problem during the sample check-up; on the contrary, samples R3 and R4 were both characterized by thin concrete covers, with R3 characterized by many other defects. Considering the strength resources of the slab samples, they are composed by the sum of the basic resistance mechanisms, which remain constant between samples, and the additional resources provided by the strengthening technique, which improves mainly the flexural resistance. As can be seen by the failure mode shown by R3 and R4, it is characterized by the typical damage and crack pattern of a shear mechanism, thus, not much affected by the strengthening technique. With this assumption, it is not rational to have a sample R2 which do not show any failure during the whole test; however, taking into account all the defects recorded in samples R3 and R4 it is plausible that specimen R2 had enough resources to resist the shear collapse until it reaches its maximum flexural resistance, which proved to be close to the shear at which the other two samples collapsed (36 kN vs 34 kN for R2 and R3-R4 respectively). Considering that, it can be remarked that the shear resources of the basic floor on which a strengthening intervention will be applied are strongly connected with the entity of this intervention. Improving flexural strength of a floor element by increasing the thickness of the FRC topping layer must be considered together with the study of the shear resources of the un-strengthened system, requiring to improve the latter if it is necessary to increase further the resources in bending. The traditional strengthening technique used for sample RT showed a good performance, with stiffness lower than R4 but higher compared to the other samples. This can be related to the overlap defect evidenced in the sample, which can affect the floor performance in terms of strength and stiffness. In any case, the sample reached its maximum flexural capacity without showing any failure, performing the test until a mid-span deflection of 60 mm. Focusing the attention on the first 10 mm of deflection (Fig. 7), it is possible to relate the slope of each single curve to the sample tested. As expected, the lowest slope is found in sample NR (un-strengthened), which is the less stiff of the five samples, in agreement with the design data. It is followed by R2, which did not show any defect in the sample and during the test. The first consequence of the floor defects is visible in sample R3, where the curve slope is closer, nearly overlapping that of sample R2, while, from design data, we expected it a slightly stiffer response. The last two samples are characterized by the same thickness of the topping layer, thus, supposed to have similar stiffness. However, sample R4 showed a higher stiffness than RT, in which the reinforcement overlap mistake was found in one of the joists composing the floor.

Structural Behavior of a Traditional Concrete and Hollow Tiles

Shear [kN]; Flexural moment [kNm]

40

861

NR R2 R3 R4 RT

36 32 28 24 20 16 12 8 4 0

0

5

10

15

20 25 30 35 Mid deflection [mm]

40

45

50

55

60

Fig. 7. Shear/bending moment-mid deflection graph for the investigated floors

Table 2. Load and bending moment results corresponding to different mid-deflection levels. Mid deflection [-] L/1000

[mm] 4 [kN] [kN/m] L/500 8 [kN] [kN/m] L/400 10 [kN] [kN/m] L/300 13.33 [kN] [kN/m] L/200 20 [kN] [kN/m] L/150 26.67 [kN] [kN/m] L/100 40 [kN] [kN/m] Maximum force – [kN] [kN/m]

NR

R2

R3

R4

7.80 3.90 11.00 5.50 12.53 6.27 15.04 7.52 19.98 9.99 23.71 11.86 29.53 14.76 30.65 15.33

9.51 4.75 14.17 7.09 16.51 8.25 20.39 10.20 23.23 11.61 30.08 15.04 36.00 18.00 36.32 18.16

10.39 5.20 14.29 7.15 16.16 8.08 19.38 9.69 25.87 12.94 32.05 16.02

12.48 6.24 18.09 9.04 20.50 10.25 24.72 12.36 32.62 16.31

RT

11.45 5.72 16.17 8.08 18.30 9.15 21.70 10.85 28.09 14.05 33.82 16.91 38.95 19.47 34.28 34.13 40.54 17.14 17.07 20.27

862

D. Sirtoli et al.

Table 3. Theoretical and experimental shear and flexural strength results for the analyzed samples (c = 1). NR As Mrd Mex Vrd Vex

R2

R3

R4

RT

mm2 Design 452.4 452.4 452.4 452.4 452.4 Effective 452.4 452.4 339.3 452.4 339.3 kNm Design 28.81 34.38 36.41 38.45 38.45 Effective 28.81 34.38 27.35 38.45 28.88 kNm Test 31.00 35.95 34.42 34.13 40.79 kN Design 32.02 34.80 36.16 37.48 37.48 Effective 32.02 34.80 32.85 37.48 34.06 kN Test 31.00 35.95 34.42 34.13 40.79

4 Final Remarks In the present paper the flexural behavior of strengthened floor elements was investigated by means of a four-point bending test. Five different samples were analysed, four of which strengthened with different toppings and one left un-strengthened as reference. The technique used to reinforce the floors is represented by FRC topping with incremental thicknesses for three samples, while, for the fourth, a welded mash was drowned in a HPC topping layer. • The collected results report an increment of the stiffness with increasing topping thickness, with some exceptions due to samples construction defects. For instance, samples R3 and RT showed lower stiffness compared to what expected from design values; indeed, these samples were found with construction errors like hole presence in the joists, lack of steel reinforcement, wrong longitudinal bar overlap and thin concrete cover. • Considering the general behavior, it turns out that the maximum effective topping thickness is strongly dependent to the shear strength of the basic floor. As the used strengthening technique is not effective in shear but in bending, once the shear capacity of the basic floor is reached a further thickness increment of the FRC topping layer becomes useless. In this situation designers must consider shear strengthening intervention in combination to the FRC topping in case they want to increase the flexural capacity of the structural element. Acknowledgements. We acknowledge Kerakoll S.p.A. company for its financial support to the experimental campaign.

Structural Behavior of a Traditional Concrete and Hollow Tiles

863

References 1. Brandt, A.M.: Fibre reinforced cement-based (FRC) composites after over 40 years of development in building and civil engineering. Compos. Struct. 86(1–3), 3–9 (2008) 2. Sorelli, L.G., Meda, A., Plizzari, G.A.: Steel fiber concrete slabs on ground: a structural matter. ACI Struct. J. 103(4), 551 (2006) 3. Hedebratt, J., Silfwerbrand, J.: Full-scale test of a pile supported steel fibre concrete slab. Mater. Struct. 47(4), 647–666 (2013). https://doi.org/10.1617/s11527-013-0086-5 4. Beschi, C., Meda, A., Riva, P.: Column and joint retrofitting with high performance fiber reinforced concrete jacketing. J. Earthq. Eng. 15(7), 989–1014 (2011) 5. Code, C.F.M.: CEB-FIB model code 2010–Final draft. Thomas Thelford, Lausanne, Switzerland (2010) 6. Di Prisco, M., Colombo, M., Dozio, D.: Fibre-reinforced concrete in fib model code 2010: principles, models and test validation. Struct. Concr. 14(4), 342–361 (2013) 7. dei Lavori Pubblici, C.S.: Linea Guida per la identificazione, la qualificazione, la certificazione di valutazione tecnica ed il controllo di accettazione dei calcestruzzi fibrorinforzati FRC (Fiber Reinforced Concrete), Italy (2019)

Pedestrian Bridge over Las Vegas Avenue in Medellín. First Latin American Infrastructure in UHPFRC Joaquín Abellán-García1,2(&), Andrés M. Núñez-López3, and Samuel E. Arango-Campo3 1

Department of Civil Engineering, Polytechnic University of Madrid (UPM), Madrid, Spain [email protected] 2 Escuela Colombiana de Ingeniería Julio Garavito, Bogotá, Colombia 3 I+D+i Cementos Argos, Medellin, Colombia

Abstract. To introduce a new product into a new consolidated market is always a challenge, even for those products like ultra-high-performance fiber reinforced concrete (UHPFRC) which have proven their value, and it is even harder if this product does not count with local regulation that supports it, like the case of UHPFRC in Colombia. This is the reason that made local developments, both dosages and applications, of vital importance for UHPFRC wider use. That is why is important to emphasize the construction of the pedestrian bridge over Las Vegas Avenue in Medellin, the first infrastructure built in UHPC in Latin America, using a dosage with local materials developed by the company Argos SA. The new pedestrian bridge that connects the campus of the EAFIT University in Medellin with its language building is the first infrastructure work in Latin America in which UHPFRC was used. Initially the design included a metallic structure, however, Argos proposed to the University to change the steel for an UHPFRC dosage developed by Argos SA. The decision to opt for this material represented a saving of 34% in the total cost of the pedestrian bridge construction. Keywords: UHPFRC
Precast keystone
Pedestrian bridge
Fiber reinforced concrete

1 Introduction Over the past few decades, the advancements of the high-performance cementitious composites have made colossal progress in building sector worldwide, pointing to a new and high-tech material with enhanced material properties such as compressive strength, durability, toughness and ductility. According to that, a new type of ultrahigh-performance cement based composite material, known as ultra-high performance fiber reinforced concrete (UHPFRC), was defined by ACI committee 239R-18 as a “concrete that has a minimum specified compressive strength of 150 MPa with specified durability, tensile ductility and toughness requirements; fibers are generally included to achieve specified requirements” [1]. Such outstanding mechanical © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 864–872, 2021. https://doi.org/10.1007/978-3-030-58482-5_76

Pedestrian Bridge over Las Vegas Avenue in Medellín

865

properties can be ascribed to its low porosity and high packing density due to its low water to binder ratio, special mixture design and mixing procedure, leading to an extremely optimized grain-binder matrix [2–11]. Long-term durability is another outcome of the low porosity and high packing density [2, 9]. The applications of UHPFRC are increasing in the last years in Europe, North America, Japan, Korea, China and Australia. Among others, uses of UHPFRC include the construction of pedestrian bridges, liner tunnel segments, special pre-stressed and pre-cast concrete elements, concrete structure rehabilitation, precast deck panel bridge joints, urban furniture, overlay on damaged pavements and industrial floors, and architectural applications [2, 7, 9, 12–14]. Therefore, notwithstanding the superb performance in strength and durability, UHPFRC has not been widely applied in emerging construction markets mainly due to the lack of local codes for its design and construction, as well as to lack of knowledge and confidence local agents, builders and developers. The development of UHPFRC projects in those markets will demonstrate economic for departing from traditional practice under standard concrete. During the last years, the city of Medellin has stood out for being a leader in innovation in both Colombia and South America, both in the engineering and construction sector. In particular, significant efforts were made in the development of special concrete and construction processes that contribute significantly to the progress of the country. With this intention, the company Argos S.A, with its research group, has been a pioneer in the production of UHPFRC in Colombia; have already devoted some years to research and have at present a UHPFRC dosage patent with the commercial name of Advanced Concrete ®, developed from local raw materials available in Colombia. This paper presents the construction of the pedestrian bridge over Las Vegas avenue in Medellin (Colombia). This building was not only the first infrastructure project developed with UHPFRC in Colombia, but the first in all Latin America [2]. This pedestrian bridge, which connects the EAFIT University campus with its language building, was constructed by using Advanced Concrete ®.

2 General Characteristics of the Pedestrian Bridge The pedestrian bridge of EAFIT University, has double curvature - a horizontal and a vertical one - 110 linear meters, a main span of 43 m over Las Vegas avenue and is supported by five points: four columns and a head beam constructed in self-compacting concrete with compressive strength of 42 MPa. The section of the bridge is made up of a main drawer beam from which a cantilever emerges that together with the pre-slabs forms the support through which people walk over the avenue. It is important to denote that the asymmetric geometry of these sections implies efforts that a conventional concrete could not support. This project was carried out using the premanufactured keystone bridge construction process with non-adhered post-tensioning. This involved the construction of 29 prefabricated keystones, which were rise to their position with cranes, there spun with steel cables and tensioned to form the structural system. The structural system of the footbridge is composed of the main girder with a plug-socket joint system.

866

J. Abellán-García et al.

Each keystone weighs about ten tons and was built using 3.7 m3 of Advanced Concrete ®. In addition, in this work 120 bags were used for the construction of the joints, totalizing a consumption of 110 m3 of UHPFRC. The AFGC recommendations [15] were followed for the structural analysis and design of the UHPFRC elements. Table 1 summarizes the characteristics of the pedestrian bridge (Fig. 1). Table 1. Characteristics of the EAFIT pedestrian bridge. Total length Total weight Main span Number of structural supports Number of premanufactured keystones

110 m 300 ton 43 m 5 29

Fig. 1. Overall view of the pedestrian bridge.

Figure 2 depicts a cross section of the pedestrian bridge over the pile cap. The drawing includes dimensions and some reinforcement details for the main elements of the structure.

Pedestrian Bridge over Las Vegas Avenue in Medellín

867

Fig. 2. Cross section of the pedestrian bridge over the pile cap.

3 Material Advanced Concrete ®, an UHPFRC developed from local raw materials available in Colombia, was used. Table 2 summarizes the characteristics of the Advanced Concrete ® used in the construction of the pedestrian bridge. Table 2. Characteristics of UHPFRC at 28 days. Compressive strength Direct tensile strength Flexural strength Modulus of elasticity

150 MPa 10 MPa 36 MPa 40 GPa

4 Pedestrian Bridge Construction Using Precast UHPFRC Keystones The construction procedure consisted in several phases which are described below: 4.1

Foundation and Pillars

The foundation was executed by reinforced concrete piles excavated in situ. 4.2

Construction of the Keystones

Starting in December 2016, the construction of 29 keystones using Advanced Concrete ® was carried out at TITAN SA field office in Medellín, Colombia. Pre-mixed concrete was delivered through the concrete batch plant. To enhance the packing density of UHPC, a slight vibration was applied to the mold during the pouring, prolonging it until one minute after completed, although the concrete was visibly self-compacting and flowing easily as it seemed in Fig. 3. Besides, a to improve the aesthetical aspect of concrete's surface, a vibrating ruler was used as depicted in Fig. 4.

868

J. Abellán-García et al.

Fig. 3. UHPFRC flowing into the mould during construction of keystones.

Fig. 4. Pouring process and use of vibrating ruler on free surface.

After 24 h from pouring the keystones were demolding and stocked by using an overhead crane (see Fig. 5). As depicted in Fig. 5, an auxiliary steel structure was used to avoid damages during the stockage process.

Fig. 5. Keystone manipulation and displacement by overhead crane (left). Auxiliary structure to avoid damages in keystone during the stockage (right).

4.3

Construction of Piles and Bridge Stirrups

Taking advantage of the prefabrication, at the time that the keystones were built, the works of stacks and stirrups of the bridge were executed. Piles and bridge stirrups were constructed with poured-in-place reinforced concrete with compressive strength of 50 MPa (Fig. 6).

Pedestrian Bridge over Las Vegas Avenue in Medellín

869

Fig. 6. View of the construction of piles and stirrups.

4.4

Centring Placement

Until the keystones were inserted and post-tensioned the pedestrian bridge has no strength and needs the centring to keep the keystones themselves in their correct relative positions. Figure 7 depicts the porticoed formwork and steel beams used for this purpose.

Fig. 7. View of the placement of porticoed centring.

4.5

Deliver and Movement of Keystones

At the TITAN SA field shop in Medellin, the 29 keystones were delivered using 29 semi-truck (with a load of one keystone each truck as depicted in Fig. 8 left). At the pedestrian bridge site, the keystones were moved by crane as showed in Fig. 8 right.

Fig. 8. Delivering of keystones (left). Crane used to move the keystone (right).

870

J. Abellán-García et al.

4.6

Alignment of Keystones

After the keystones were placed in position, joints were executed with field-cast UHPC. After that, the post-tensioning process was carried out, creating a continuous girder (Figs. 9, 10 and 11).

Fig. 9. Alignment of the keystones

Fig. 10. Construction of the joints

Fig. 11. Post-tensioning process

4.7

Bridge Load Test

Finally bridge load test was performed using barrels filled with water as depicted in Fig. 12.

Pedestrian Bridge over Las Vegas Avenue in Medellín

871

Fig. 12. Bridge load test

Results of the proof load test performed compiled with the requirements of the Colombian regulation.

5 Conclusions From the construction of the first infrastructure in Latin America using UHPFRC, the following conclusions can be drawn: • The use of UHPFC instead the previously planned solution in steel meant a saving of 33% in the material cost of the project. • The experience gained during the construction of the pedestrian bridge encourages to continue with the development of infrastructure using UHPFRC. • Furthermore, this paper shows good practices, large applications of UHPC and attractive potential market in Colombia. Acknowledgements. Authors want to acknowledge R&D division of Cementos ARGOS SA, Escuela Colombiana de Ingeniería Julio Garavito, and EAFIT University for bringing us the confidence, support and opportunity to launch special concretes like UHPC in Colombia.

References 1. ACI Committe 239: ACI – 239 Committee in Ultra-High Performance Concrete (2018) 2. Abellan, J., Torres, N., Núñez, A., Fernández, J.: Ultra high preformance fiber reinforced concrete: state of the art, applications and possibilities into the latin american market. In: XXXVIII Jornadas Sudam. Ingeniería Estructural, Lima, Peru (2018) 3. Soliman, N.A., Tagnit-Hamou, A.: Using particle packing and statistical approach to optimize eco-efficient ultra-high-performance concrete. ACI Mater. J. 114, 847–858 (2017). https://doi.org/10.14359/51701001 4. Ghafari, E., Costa, H., Nuno, E., Santos, B.: RSM-based model to predict the performance of self-compacting UHPC reinforced with hybrid steel micro-fibers. Constr. Build. Mater. 66, 375–383 (2014). https://doi.org/10.1016/j.conbuildmat.2014.05.064

872

J. Abellán-García et al.

5. Schmidt, C., Schmidt, M.: ‘Whitetopping’ of asphalt and concrete pavements with thin layers of ultra-high-performance concrete - construction and economic efficiency. In: Fröhlich, M.S.E.F.C.G.S., Piotrowski, S. (eds.) 3rd International Symposium UHPC Nanotechnology High Performance Construction Materials, Kassel University, Kassel, Germany, pp. 921–927 (2012). ISBN online: 978-3-86219-264-9 6. Abbas, S., Nehdi, M.L., Saleem, M.A.: Ultra-High performance concrete: mechanical performance, durability, sustainability and implementation challenges. Int. J. Conc. Struct. Mater. 10, 271–295 (2016). https://doi.org/10.1007/s40069-016-0157-4 7. Nehdi, M., Abbas, S., Soliman, A.: Exploratory study of ultra-high performance fiber reinforced concrete tunnel lining segments with varying steel fiber lengths and dosages. Eng. Struct. 101, 733–742 (2015). https://doi.org/10.1016/j.engstruct.2015.07.012 8. Filho, R.D.T., Koenders, E.A.B., Formagini, S., Fairbairn, E.M.R.: Performance assessment of ultra high performance fiber reinforced cementitious composites in view of sustainability. Mater. Des. 36, 880–888 (2012). https://doi.org/10.1016/j.matdes.2011.09.022 9. Abellan, J., Torres, N., Núñez, A., Fernández, J.: Influencia del exponente de Fuller, la relación agua conglomerante y el contenido en policarboxilato en concretos de muy altas prestaciones. In: IV Congress International de Ingenieria Civil, Havana, Cuba (2018) 10. Abellán, J., Fernández, J., Torres, N., Núñez, A.: Statistical optimization of ultra-highperformance glass concrete. ACI Mater. J. 117, 243–254 (2020). https://doi.org/10.14359/ 51720292 11. Abellán-García, J., Núñez-López, A., Torres-Castellanos, N., Fernández-Gómez, J.: Effect of FC3R on the properties of ultra-high-performance concrete with recycled glass. Efecto del FC3R en las propiedades del concreto de ultra altas prestaciones con vidrio reciclado. Dyna 86, 84–92 (2019). https://doi.org/10.15446/dyna.v86n211.79596 12. Soliman, N.A., Tagnit-Hamou, A.: Using glass sand as an alternative for quartz sand in UHPC. Constr. Build. Mater. 145, 243–252 (2017). https://doi.org/10.1016/j.conbuildmat. 2017.03.187 13. Kalny, M., Kvasnicka, V., Komanec, J.: First practical applications of UHPC in the Czech Republic. In: Fehling, E., Middendorf, B., Thiemicke, J. (eds.) Proceedings of Hipermat 2016 - 4th International Symposium UHPC Nanotechnology Construction Mateials, Kassel, Germany, pp. 147–148 (2016) 14. Tagnit-Hamou, A., Soliman, N., Omran, A.: Green ultra - high - performance glass concrete. In: First International Interactive Symposium UHPC – 2016, vol. 3, pp. 1–10 (2016) 15. AFGC and SETRA: Ultra high performance fibre-reinforced concretes - recommendations, Associatio (2013)

First Experimental Full-Scale Elevated FRSCC Slab in South America Luis Segura-Castillo(&), Diego Figueredo, Iliana Rodríguez, and Nicolás García Facultad de Ingeniería, Universidad de la República, Montevideo, Uruguay [email protected] Abstract. A full-scale steel fibre reinforced self-compacting concrete elevated slab was built and tested up to failure. The experimental slab (without any conventional reinforcement), consisted of four continuous panels of 3.1  3.1 m2 each, supported by columns two meters above ground. The nominal thickness was 130 mm. 90 kg/m3 of steel fibres were used. For the slab testing, two opposite panels were punctually loaded. Simultaneously, displacements were recorded and crack pattern registered. Results show a fan type cracking pattern, corresponding to what is reported by literature. The maximum loads obtained were Fmax,1 = 156.4 kN and Fmax,2 = 211.8 kN, (dcentral  20 mm), and the ultimate loads FU,1 = 117.3 kN and FU,2 = 187.9 kN (dcentral  60 mm), i.e. for displacements approximately three times larger than the one registered at peak load, the slab could carry more than 75% of the peak load, showing good overall ductility. A simple software, capable of modelling yield line theory, and suitable for routine applications, was used to model the slab. The software input is the slab geometry, boundary conditions, and plastic moments. The results showed that using reasonable simplifications, a good agreement between theoretical models and experimental results can be obtained if characteristic material properties are used in the model. Keywords: Steel fibre


Self-compacting
Concrete
Test
Failure

1 Introduction Fiber-reinforced concrete (FRC) has been successfully used as the only reinforcement of elevated slabs, with dozens of multi-story buildings already completed [1]. The use of FRC in these elements simplifies the construction with cost and time savings associated to a reduction in labour and equipment while improving crack control and durability. These saving are further increased when self-compacting FRC (FRSCC) is used. For its development, elevated FRC slabs have been studied both experimentally and numerically by several authors. In general, a good response has been observed in the experimental slabs already tested, both in terms of load carrying capacity, crack patterns, and also in terms of ductility [2–6]. Different recommendations and guides exist for the design of elevated FRC slabs, most notably the ACI 544.6R report [1] and the fib Model-Code 2010 [7]. A notable difference between them is the procedure to obtain the constitutive equation for the FRC. While the first one recommends the use of small plates samples according to © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 873–882, 2021. https://doi.org/10.1007/978-3-030-58482-5_77

874

L. Segura-Castillo et al.

ASTM standard C1550 [8], the second one prefers the use of small beams, according to EN 14651 [9], and then applies coefficients to account for fibre distribution and structural redundancy of the element to be designed. For structural evaluation both guides accept the use of the Yield Line Method, developed by Johansen [10] in parallel with Gvozdev [11], for the estimation of the ultimate load-carrying capacity of slabs. These methods, originally developed to be performed by hand, have recently been successfully implemented in computer programs, which allows the analysis of more complex geometries and load cases. E.g. the software LimitState:SLAB [12], which is based on the discontinuity layout optimization (DLO) method [13]. One of the main focus of the Structural Concrete Group (GHE – Grupo de Hormigón Estructural) of the Universidad de la República, Uruguay, is the introduction of FRC for structural use in the country. Relevant recent developments by the group are; a simplified test for the FRC characterization in tension (Montevideo Test) [14], the study of FRC anisotropy [15], Precast FRC Sandwich Panels [16] or diaphragm walls strengthened by a layer of FRC [17]. More recently, in a collaboration project with the construction company ABENGOA-TEYMA, the group built and tested up to failure the first full-scale experimental elevated FRSCC slab in South America. The objective of this paper is to describe the main aspects of this experimental program and to show its preliminary results, mainly, the load–displacement behaviour under the point of loading and the general crack pattern. Also, the experimental results are compared with those obtained numerically by a computational program, based in the yield line theory.

2 Methodology 2.1

Experimental

A full-scale fibre reinforced self-compacting concrete (FRSCC) elevated slab was built and tested up to failure under Short-term loading. The construction site was located in Montevideo, Uruguay. The slab was built in June 2018 and tested between June and August 2019. The experimental slab, which did not have any type of conventional reinforcement, was 6.2  6.2 m2, dimensions in plant, and with a nominal thickness (t) of 130 mm (Fig. 1). It consisted of four continuous panels of 3.1  3.1 m2 each, supported on three lines of three cast in-situ reinforced concrete columns with a square cross-section of 0.2  0.2 m2. The slab was suspended two meters above ground level. A FRSCC raft foundation supported the columns. The materials used for the FRSCC mix were: cement CEM I, water, two natural river fine aggregates, granitic coarse aggregates with maximum dimension of 19 mm, superplasticizer and hooked-end steel fibres (Ferrofiber AR65) with a length (l) of 50 mm and a diameter (d) of 0.77 mm, resulting in an aspect ratio (k = l/d) of 65. A fibre content of 90 kg/m3 was used (equivalent to 1.14% by volume). The mix was designed to obtain both pumpable and self-compacting characteristics. The selfcompacting requisites of the mix were evaluated by executing the slump-flow [18] (obtaining a slump flow of 660 mm), and through visual inspection [19].

First Experimental Full-Scale Elevated FRSCC Slab

(a)

875

(b)

Fig. 1. Experimental program: a) slab plant; and b) sketch of the slab and foundation.

The FRSCC used for the elevated slab was evaluated under compression at 28 days by means of cylindrical specimens with a diameter of 150 mm and a nominal height of 300 mm, obtaining a mean compressive strength equal to 64.3 MPa. The post-peak material tensile behaviour was determined by testing 5 beams using a three-point bending test, also at 28 days, according to EN 14651 [9]. For all the specimens, the limit of proportionality fL, the residual flexural tensile strengths fR1, and fR3 for a crack mouth opening displacement (CMOD) of 0.5 and 2.5 mm, respectively, were calculated. Average values obtained were (coefficient of variation between brackets): fL = 7.02 MPa (8%); fR1 = 10.23 MPa (10%) and fR3 = 8.63 MPa (12%), with which the following characteristic residual tensile strengths were computed: fL,k = 6.10 MPa; fR1, k = 8.58 MPa and fR3,k = 7.00 MPa. The raft foundation, with a nominal thickness of 130 mm and the same FRSCC mix than the used in the elevated slab, was cast both to provide support for the structure, and also to serve as first full scale trial of the mix. It was directly resting on top of a 400 mm compacted gravel layer. Reinforcement Steel was left embedded to connect both the columns and also the cables used in the load tests. The columns were cast with conventional reinforced concrete. The slab was cast by one single pumping step, in approximately 6 min. The cast sequence is shown in Fig. 2a. The dotted line represents the path followed by the concreting hose, which was obtained after analyzing a recording of the process. It was indicated to the operators of the hose to launch the concrete mainly from the centre of each panel, following a clockwise direction. As can be seen, although the instructions were somewhat followed, the launch tended to lead a more random course. An image during the cast is shown in Fig. 2b. Due to the use of the self-compacting concrete, the slab was appropriately compacted without any poker vibrating. Two void plastic cylinders were embedded in the centre of the slabs designed for testing, foreseeing the passage of the testing rod. After casting the slab was water-cured for 5 days. For the slab testing, two opposite panels were punctually loaded, with loading/unloading cycles. The loading took place on one panel first (panel 1) and, after its rupture, the second loading occurred (panel 2). With this scheme, a certain loss of

876

L. Segura-Castillo et al.

(a)

(b)

Fig. 2. Cast of the slab: a) cast sequence; b) Image during concrete launching.

rigidity is expected for the second panel after testing the first panel. However, in previous experiences this influence was shown to be reduced, as the cracking take place in different areas of the slab. The load test of each panel was carried out in two stages: the first one was load controlled, recording every measure for constant load increments until the peak load was registered. After achieving the peak load, in the second stage, the test was controlled by the displacement of the central point of the panel, recording every measure for constant displacement increments. To introduce the load in the panel, a hydraulic hollow-jack, controlled by a hand pump, was used (Fig. 3). The jack was resting over a steel plate that was over the slab. To induce the load downward, the top of the jack reacted on another steel plate that was connected to a tie rod which, in turn, passed through a hole towards the bottom of the slab. Below the slab level, the rod was connected to four suspension cables which were connected to the bottom of the four columns around the panel under testing. One of these cables had a manual winch to ensure the tie rod was in a vertical position. The hydraulic pump was manually controlled to allow for a stepwise increase of the load by 18 kN (200 psi in the jack) and 9 kN (100 psi in the jack) increments for the slab 1 and 2 respectively. The load force was measured from a single load cell placed above the hydraulic jack, and between the jack and the plate connected to the tie rod. Simultaneously with the slab loading, seven points (Fig. 1a) of each panel were selected to record the vertical slab displacements by means of dial indicators. In this paper only the displacements of the main indicator of each panel, placed in its centre (Fig. 1a), are used. The dial indicators were installed on 5 steel beams fixed between columns (Fig. 3), which act as a type of yoke, to record the relative displacement between the slab and the support columns and minimize the influence of extrinsic deformations. The load cell and the dial indicators were measured after each load increment step. Finally, when the load tests were completed, the slab was surveyed for cracks and the crack pattern registered.

First Experimental Full-Scale Elevated FRSCC Slab

877

Fig. 3. Experimental FRSCC test set-up.

2.2

Numerical

The structural response of the slab was analyzed by the software LimitState:SLAB [12]. The software is capable of modelling slabs with the yield line theory and is suitable for routine applications. The ultimate load-carrying capacity and the yield lines associated can be estimated for a given slab geometry, boundary conditions, plastic moment and type and position of loads. The program is also able to consider the self-weight of the element. The experimental slab geometric parameters were used. Load was simplified to a point load in the position of the jack. To obtain the plastic moment a cross-section layer discretization analysis was used, as shown elsewhere (e.g. [20, 21]). Fib Model-Code 2010 [7] design rules were used to obtain the material constitutive law using the EN14651 test results. Mean and characteristic plastic moment of resistance per unit length obtained were mp,m = 26.33 kNm/m and mp,k = 21.67 kNm/m, respectively. In each case, the same plastic moment was used both for positive and negative yield lines, assuming an isotropic behaviour.

3 Results 3.1

Numerical Results

Collapse loads of Pc,m = 198.1 kN and Pc,k = 167.9 kN were obtained with the LimitState: Slab software for the mean and the characteristic residual tensile strengths, respectively. The associated collapse mechanism for both cases can be seen in Fig. 4. The yield lines are shown in blue and red for positive and negative moments, respectively. It can be seen that the yield-line pattern correspond to the previously described by other authors for centre point loads, with a fan type characterized by

878

L. Segura-Castillo et al.

several positive moment cracking underneath and negative moment cracking in a circular shape passing by the four footprints of columns at the top of the slab.

Fig. 4. LimitState: Slab collapse mechanism.

3.2

Experimental Results

The crack pattern observed after the test was finished are schematically represented, both for positive (Fig. 5a) and negative (Fig. 5b) moments. Thick lines indicate large (w  0.5 mm) and medium (0.5 mm > w  0.1 mm) cracks. Thin lines indicate smaller cracks (w < 0.1 mm). Examples of cracking in the slab are shown in Fig. 6. The positive macro crack at the edge of the slab (see Fig. 6a), was more than 3 mm wide at the end of the test. The results showed a strong resemblance with both the patterns reported by literature and with the numerical result previously described.

(a)

(b)

Fig. 5. Crack pattern observed in experimental test: a) bottom surface; b) top surface.

First Experimental Full-Scale Elevated FRSCC Slab

(a)

879

(b)

Fig. 6. Examples of cracking in the slab: a) bottom surface; b) top surface.

The load-displacement curves for the tested slab are shown in Fig. 7. The experimental results for the two panels are shown in black lines. Notable points are highlighted with circles and also the Plot is amplified in the left of the Figure, showing the first 10 mm of displacement. 250

Load - F [kN]

200

250 198,1

29,8; 211,8

200

167,9

150

150

100 50

19,5; 156,4

100

1,94; 86,7

58,5; 117,3

1,94; 86,7

Panel 1 Num_m

50

0

64,4; 187,9

Panel 2 Num_k

0 0

2

4

6

8

10

0

10

20

30

40

50

60

Mid-span displacement - δ [mm] Fig. 7. Experimental test results.

General behaviour shows a linear trend up to approximately 80 kN. This value approximately corresponds to the appearance of the first visible cracks in both panels. After cracking, a gradual loss of stiffness is registered, which continues until the peak load is reached. The maximum loads obtained were Fmax,1 = 156.4 kN and Fmax,2 = 211.8 kN, with a corresponding deflection in the centre of the panel of dFmax,1 = 20 mm and dFmax,2 = 29.8 mm. Suffix 1 and 2 indicate results for panel 1 and 2 respectively.

880

L. Segura-Castillo et al.

After peak load, a slowly descending branch developed until the test was interrupted. The tests were concluded when a central deflection of dmax,1 = 58.5 mm and dmax,2 = 64.4 mm was reached in each of the panels, which represents between two to three times the displacement for the maximum load (dFmax,i). For the maximum displacements, the recorded load were FU,1 = 117.3 kN and FU,2 = 187.9 kN, respectively, which represents approximately 75% and 89% of each maximum load. The large displacements observed, while still holding a large amount of the maximum load, indicates a good overall ductility. The difference between the two slabs may be attributable to several factors, e.g. possible difference in the fibre distribution and orientation, or the experimental error. In particular the real slab thickness was slightly different from the nominal values. The average of the thickness measured in different points of the slab was t1 = 124.8 mm and t2 = 139.3 mm for each slab. The numerical estimation of the ultimate load is also plotted in gray lines in Fig. 7, with the values obtained with the mean and characteristic values of the residual strength plotted in continuous and dashed lines, respectively. It can be seen that the numerical estimation, calculated with the mean values of residual strength fall above the experimental values, being 69% and 5% larger for each of the panels. However, the estimation made with the characteristic values fall between the two experimental results (43% above panel 1 and 11% below panel 2). This may indicate that, if the numerical method is adequate for the FRC slab analysis, the use of the mean results of residual strength is not a good predictor of the structural response of the slab. Still, a deeper analysis should be carried out in order to evaluate other possible factors that may be influencing the results.

4 Conclusions A full-scale steel fibre reinforced self-compacting concrete elevated slab was built and tested up to failure. Results showed: • A fan type cracking pattern, corresponding to what is reported by literature. The maximum loads obtained were Fmax,1 = 156.4 kN and Fmax,2 = 211.8 kN and the ultimate loads FU,1 = 117.3 kN and FU,2 = 187.9 kN. The ultimate load was registered for displacements (dcentral  60 mm) approximately three times larger than the one registered at peak load (dcentral  20 mm), still carrying more than 75% of the peak load, which implies a good overall ductility. • A simple software, capable of modelling yield line theory, and suitable for routine applications, was used to model the slab. The results also showed that using reasonable simplifications, a good agreement between theoretical models and experimental results can be obtained if characteristic material properties are used in the model. Future work includes the analysis of fibre distribution and structural redundancy of the slab, in order to improve the analysis of the slab.

First Experimental Full-Scale Elevated FRSCC Slab

881

Acknowledgements. The authors would like to thank ABENGOA-TEYMA for financial support, physical resources and the assistance of their staff (especially Mauricio Montaña, Ramiro Rodriguez and Ignacio Horta without whom this work could never have been completed). The authors also thank the collaboration of the following companies and institutions: Laboratorio de Vialidad, MTOP, where testing of small specimens was carried out; Ferrocement, for supplying the fibres; Atenko, for supplying the scaffolding; and Guardia Republicana – Ministerio del Interior, on whose premises the slab was built. The authors thankfully acknowledge the students, assistants and professors that helped during the project. Funding was also made available from the Agencia Nacional de Investigación e Innovación (ANII), Uruguay, through the Research Project Herramientas para la innovación - 2017 – “HPI_X_2017_1_141290”.

References 1. ACI Committee 544: ACI 544.6R-15. Report on Design and Construction of Steel FiberReinforced Concrete Elevated Slabs. American Concrete Institute, Farmington Hills (2018) 2. Mobasher, B., Destrée, X.: Design and construction aspects of steel fiber-reinforced concrete elevated slabs. In: SP-274 Fiber Reinforced Self-consolidating Concrete: Research and Applications, C. A. LF, Ed., pp. 95–107 (2010) 3. Salehian, H., Barros, J.A.O.: Prediction of the load carrying capacity of elevated steel fibre reinforced concrete slabs. Compos. Struct. 170, 169–191 (2017) 4. Colombo, M., Martinelli, P., di Prisco, M.: On the evaluation of the structural redistribution factor in FRC design: a yield line approach. Mater. Struct. 50(1), 1–18 (2017) 5. Maturana Orellana, A.: Estudio teórico-experimental de la aplicabilidad del hormigón reforzado con fibras de acero a losas de forjado multidireccionales. Universidad del País Vasco (2013) 6. di Prisco, M., Martinelli, P., Parmentier, B.: On the reliability of the design approach for FRC structures according to fib model code 2010: the case of elevated slabs. Struct. Concr. 17(4), 588–603 (2016) 7. FIB: Model Code 2010, Vol 1 & 2. International Federation for Structural Concrete (fib), Lausanne, Switzerland (2010) 8. ASTM: C1550-10a Standard Test Method for Flexural Toughness of Fiber Reinforced Concrete (Using Centrally Loaded Round Panel). ASTM Stand., pp. 1–14 (2010) 9. EN 14651: Test method for metallic fibre concrete—Measuring the flexural tensile strength (limit of proportionality (LOP), residual) (2005) 10. Johansen, K.: Yield-line theory. Cement and Concrete Association, London (1962) 11. Gvozdev, A.: The determination of the value of the collapse load for statically indeterminate systems undergoing plastic deformation. Int. J. Mech. Sci. 1, 322–335 (1960) 12. LimitState Ltd.: LimitState:SLAB, Sheffield, U.K 13. Gilbert, M., He, L., Smith, C.C., Le, C.V.: Automatic yield-line analysis of slabs using discontinuity layout optimization. Proc. R. Soc. 470(2168), 23 (2014) 14. Segura-Castillo, L., Monte, R., De Figueiredo, A.D.: Characterisation of the tensile constitutive behaviour of fibre-reinforced concrete: a new configuration for the wedge splitting test. Constr. Build. Mater. 192, 731–741 (2018) 15. Segura-Castillo, L., Cavalaro, S.H.P., Goodier, C., Aguado, A., Austin, S.: Fibre distribution and tensile response anisotropy in sprayed fibre reinforced concrete. Mater. Struct. 51(1), 1– 12 (2018). https://doi.org/10.1617/s11527-018-1156-5 16. Segura-castillo, L., García, N., Rodriguez Viacava, I., de Sensale, G.R.: Structural model for fibre-reinforced precast concrete sandwich panels. Adv. Civ. Eng. 2018, 11 (2018)

882

L. Segura-Castillo et al.

17. Segura-Castillo, L., Aguado De Cea, A.: Bi-layer diaphragm walls: evolution of concrete-toconcrete bond strength at early ages. Constr. Build. Mater. 31, 29–37 (2012) 18. EN 12350-8: Testing fresh concrete - Part 8: Self-compacting concrete - Slump-flow test. Comité Européen De Normalisation (2010) 19. ASTM C1611: Standard Test Method for Slump Flow of Self-Consolidating Concrete, pp. 1–6. ASTM International, USA (2005) 20. Segura-Castillo, L., Monte, R., De Figueiredo, A.D.: Analytical correlation between Montevideo test (MVD) and three-point bending test for fibre reinforced concrete (FRC). In: 4th FIB Congress, p. 8 (2018) 21. de la Fuente, A., Aguado, A., Molins, C., Armengou, J.: Numerical model for the analysis up to failure of precast concrete sections. Comput. Struct. 106–107, 105–114 (2012)

Masonry Walls Strengthened with Fiber Reinforced Concrete Subjected to Blast Load Salah Altoubat1(&), Abdul Saboor Karzad1, Moussa Leblouba1, Mohamed Maalej1, and Pierre Estephane2 1

Department of Civil and Environmental Engineering, Sustainable Construction Materials and Structural System Research Group, University of Sharjah, Sharjah, UAE [email protected] 2 GCP Applied Technologies, Dubai, UAE

Abstract. The aim and objective of this research project was to develop and investigate a rapid method of strengthening the Unreinforced Masonry walls (URM) for out-of-plane action. A total of 12 URM walls, built inside a precast fiber reinforced concrete boundary frame, were strengthened with a cement based concrete mix with or without steel reinforcement. The concrete mixes included Fiber Reinforced Concrete (FRC) (two dosage of fibers (4.6 & 6.0 kg/m3)), Engineered Cementitious Composite (ECC), and plain concrete shotcrete (PLS). Two out of 12 walls were strengthened with steel mesh reinforced shotcrete (MRS) such that a steel mesh was attached to the surface of the walls prior to spraying with concrete. In addition, one wall was plastered only with conventional cement mortar (RF) to serve as a reference. The strengthening materials were applied to both sides of URM walls using shotcrete technology. The walls were exposed to the air pressure of an actual blast from 10 kg of TNT explosive at a 1 m stand of distance. It was found that strengthening the walls with the proposed method significantly improved the out-of-plane behaviour of the URM walls. The results showed that the walls strengthened with ECCS exhibited the least damage followed by FRS-6, FRS-4.6, MRS, and PLS. Whereas, the reference wall (RF) did not survive the blast induced pressure and fully shattered into small pieces. Keywords: URM walls
Out-of-plane Synthetic fibers
Steel mesh


Strengthening
FRC
ECC


1 Introduction Rising threats of terrorism have boosted the need for engineers to build structures (such as embassies, airports, and other governmental buildings) capable of withstanding major impact loads, specifically air pressure generated by explosions. Unreinforced masonry (URM) walls have become a common building practice in the construction industry. URM walls compared to other building materials are most vulnerable to damage in accidental events such as blasts, earthquakes or any similar instances [1]. The destruction of masonry walls in the event of a blast, can lead to severe damage or full collapse of the building thus resulting in casualties. However, due to the ease of © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 883–894, 2021. https://doi.org/10.1007/978-3-030-58482-5_78

884

S. Altoubat et al.

construction, feasibility & availability of these materials, and the simplicity in building, it is still a preferred building material [2]. Masonry walls are primarily designed for in-plane forces (e.g. combined gravitational and horizontal (wind and/or seismic) loads) [3]. Due to its composite structure and inherent weak directions (along the head and bed joints) unreinforced masonry behaves in a highly anisotropic manner. This emphasizes the importance of knowing the parameters of masonry walls strength. Taking into consideration the action of the in-plane forces, walls exhibit the behaviour depicted in Fig. 1. As can be seen in figure, the wall is experiencing an in-plane force which results in a deflection in the wall. Adding to that, it is interesting to note that due to the deflection, there is a high amount of energy release, especially in the corners where deflection is greater, which ends up in diagonal cracks. Coming to the action of the out-of-plane forces, the walls experience severe damage since it is not designed to withstand these types of loads [4]. The masonry unit of the structure cannot resist the out-of-plane forces thus shattering the wall into pieces as illustrated in Fig. 1.

(a)

(b)

Fig. 1. Wall behaviour subjected to forces, a) in-plane, b) out-of-plane

To reduce the vulnerability of masonry walls under sever loading is to retrofit such walls. Many researchers around the globe have looked for methods of strengthening the walls to resist out-of-plane forces and suggested different methods of strengthening and their benefits [2, 5–10]. For instance, Sajal verma et al. [8] studied the influence of blast load on structures. In their study different methods of improving the out of plane behaviour is suggested such as increasing the mass of walls (as it would decrease the acceleration when subjected to external pressure), use of polyethylene fibers (which showed promising results), and using glass fiber reinforcement. In another study by Carney and Myers [5] the out-of-plane behaviour of URM walls strengthened with FRP and subjected to blast loading was investigated. Their results showed that the strengthening method using near surface mounted (NSM) FRP rods with anchorage is an effective method of strengthening. As mentioned in the literature review, there have been initiatives that have been conducted to find feasible solution for strengthening the URM walls against out-ofplane forces. Similarly, this research was conducted to study out-of-plane behaviour of the strengthened URM walls subjected to actual blast load which is scarce in the literature. This research focuses on showing the real effects of the blast on the URM

Masonry Walls Strengthened with Fiber Reinforced Concrete

885

walls with conventional cement mortar and walls strengthened with different types of concrete using shotcrete technology.

2 Air Blast Background Air pressure waves also called air blast waves is the resultant of sudden release of energy from detonation of an explosive material. The generated waves initiate with a single pulse of increased air pressure that last in milliseconds and travels faster than sound. The impact of blast waves on the object due to an explosion is known as overpressure and its magnitude & duration depends on two factors: the stand-off distance and the amount of charge exploded [11–14]. Figure 2 shows the types of explosive-induced shockwaves which are Friedlander wave, free-field wave, and complex wave. The idealized form (Friedlander wave) of the free-field blast wave can be altered considerably along its propagation by the medium encountered morphology thus can create multiple wave reflections (complex wave). Even the free-field wave characteristic can change due to the reflection of the blast waves from ground. For instance, if the blast wave is reflected by rigid boundary or surroundings, the overpressure can be enlarged by up to 14 times. Particularly, this situation can occur with more complexity in an urban environment or underground structures due to presence of multiple boundaries [11, 12, 15].

Fig. 2. Types of blast-induced shock waves; a) idealized form of wave, b) free-field or open-air blast wave, c) complex or closed environment blast wave (source: [16])

3 Experimental Program The objective of this research study was to investigate a rapid strengthening method to enhance the out-of-plane resistance of URM walls against lateral loads. For that a series of URM walls were built and strengthened with different types of concrete mixes with

886

S. Altoubat et al.

and without steel mesh reinforcement. The strengthening mixes were applied using shotcrete technology and the walls were tested subjected to an actual blast condition. 3.1

Geometry

A total of 12 masonry walls were constructed inside a precast Fiber Reinforced Concrete (FRC) boundary frame (fiber dosage of 2.3 kg/m3), each measuring 1.27  1.25  0.1 m made of hollow concrete blocks of size 400  200  100 mm. The FRC boundary frames (1.55  1.92 m) were reinforced using 16 mm diameter steel bars in both directions. The masonry units (hollow concrete blocks) were made of plain concrete with an average 28 days compressive strength of 7.5 MPa. The hollow concrete blocks were built inside the frame using conventional cement mortar between the layers and the vertical gaps. Figure 3 shows the dimensions of the walls and the reinforcement details of the frame constructed for the blast test. As depicted in Figure, the depth of the bottom side has been increased to provide an extra grip when buried into the ground during blast test.

Fig. 3. Geometry and reinforcement details (units are in cm unless specified in the drawing)

3.2

Strengthening Methodology

The constructed walls in this project consisted of duplicate walls strengthened in the out-of-plane direction. Each pair of strengthened walls was sprayed with a different

Masonry Walls Strengthened with Fiber Reinforced Concrete

887

type of concrete mix on both sides of the walls (with a thickness of around 4.5 cm on each side) using shotcrete technology (see Fig. 4). Four different types of concrete mixes were sprayed to the walls in addition to a reference wall that was plastered with cement mortar only. Table 1 shows the type of concrete mix used and the corresponding labels. For ease in identification of each type of wall, the walls were labelled according to the mix design sprayed on to it. Table 1. Types of concrete mixes applied on the surface of the walls Strengthening mixes Fibre Reinforced Shotcrete with fiber dosage rate of 4.6 kg/m3 Fibre Reinforced Shotcrete with fiber dosage rate of 6 kg/m3 Engineered Cementitious Composite with fiber dosage rate of 16 kg/m3 Mesh Reinforced Shotcrete (reinforced with steel mesh) Plain Concrete Shotcrete (Control) Reference Walls (Plastering mortar only)

Short form FRS-4.6 FRS-6 ECCS MRS PLS RF

As mentioned above, four different concrete mix was designed to be used for strengthening the walls. The mix design proportions of these concrete materials are presented in Table 2. Table 2. Mix design proportions of the mixes sprayed to the URM walls from both sides Type proportions (kg/m3) PLS & MRS FRS-4.6 OPC 420 420 Silica fume 25 25 0-5 mm aggregates 1480 1480 Dune sand 345 345 190 190 Water (L/m3) ADVA flow 480 (L/m3) 6 11.5 Strux 90/40 0 4.6 Spectra 900 0 0

FRS-6 420 25 1480 345 190 12.5 6 0

ECCS 1449 145 0 0 478 11.5 0 16

Fig. 4. Spraying the walls using shotcrete technology

888

S. Altoubat et al.

For FRS mixes the fibers used was synthetic macro-fibers commercially called Strux 90/40 with a length of 40 mm, a rectangular cross Sect. (1.4  0.105 mm), a modulus of elasticity of 9.5 GPa, and a tensile strength of 620 MPa. The ECCS mix which can be considered a special case of FRS, incorporates micro fibers commercially named Spectra 900 with a length of 12 mm, diameter of 0.039 mm, modulus of 66 GPa, and a tensile strength of 2610 MPa. Two walls were strengthened with mesh reinforced shotcrete. To achieve that, the A98 (Ø5 mm @200  200 mm) steel mesh was attached to the surface of the masonry walls on both sides before spraying the walls with plain concrete. To characterize the concrete mixes used, some standard tests were also conducted. Tables 3 and 4 present the test result of the cubes, prisms, and round panels samples tested according to ASTM C39, ASTM C293, and ASTM C1550. Table 3. Compressive and flexural strength (MPa) Mixes properties PLS & MRS FRS-4.6 FRS-6 ECCS Average Compressive Strength at 28 days 68 59 60 53 Average Flexural Strength at 28 days 4.7 4.6 5.3 4.3

Table 4. Energy absorption of the shotcrete mixes obtained by testing the round panels Mixes parameters PLS & MRS FRS-4.6 FRS-6 ECCS Deflection (mm) 39 39 39 39 Energy Absorption (J) 54 168 208 783

3.3

Test Setup and Instrumentation

From the observations in the trial tests, it was decided to place the wall at 1 m stand of distance from the explosive material and the walls were properly supported. Therefore, a rigid steel frame with bracing supports (see Fig. 5) was designed and fabricated. The rigid frame was used to support and hold the walls in its place during the explosion and to prevent the walls from tilting or flipping after the shock amid blast. The walls were clamped using steel clamps at four spots to the rigid steel frame to prevent separation of steel frame and the wall during the shock. For further support, the walls were also buried in the ground at the base. A total of four tests (first test was trial) were conducted with a set of 3 walls in each test. For each test the walls were placed around the explosive such that if looked from above, the walls were staggered in a triangular shape (see Fig. 5). An amount of 10 kg of TNT explosive was placed in between the three walls at a height of 60 cm above the ground to match the mid-height of the walls. In order to measure the acceleration and the air pressure exerted on the walls, specialized instruments (pressure sensor and accelerometer manufactured by PCB Piezotronics) were installed on the surface of the walls as shown in Fig. 5. The pressure sensor was fixed at mid-height facing the explosive material and accelerometer was fixed

Masonry Walls Strengthened with Fiber Reinforced Concrete

889

at mid-height on the opposite side or in other words at the back of the walls. The sensors attached to the walls recorded the peak air pressure and the acceleration of the wall with respect to time. The results obtained from the accelerometer and the pressure sensors were gathered for each wall and their behaviour was observed against the blast waves.

Fig. 5. Test setup (actual and schematic)

4 Results and Discussions 4.1

Damage Identification Through Cracking Pattern and Signal Processing

4.1.1 Cracking Pattern After the blast test the walls were visually inspected to determine the cracking pattern and ultimate damage. Figure 6 shows the photo of all the tested walls along with diagrams of signal processing implemented on acceleration data recorded during the tests. It should be noted that all the photos in Fig. 6 are captured from back side of the walls (opposite side to the side facing blast wave). Referring to the photo in Fig. 6a, the ECCS walls did not exhibit any noticeable damage except one horizontal crack in the mid height which is hardly visible. Figure 6b illustrates that the FRS-6 wall exhibited major diagonal cracks with one or two light horizontal cracks. The crack had small width and were not very severe to cause any damage, so the wall almost kept its initial shape. Similarly, the FRS-4.6 walls experienced multiple cracking pattern with a bit more intensity and wider cracks (see Fig. 6e) compared to the FRS-6 wall. However, the MRS walls had several much sever and deep cracks that resulted in noticeable damage shown in Fig. 6c. The control walls that were PLS (reference for shotcreted walls) and RF (presenting the conventional method of constructing walls) did not survive the blast waves as shown in Fig. 6d and 6f. As can be seen in Fig. 6d, the PLS wall exhibited sever damage associated with spalling of a big chunk of shotcrete from the back side (side opposite to the blast face). In addition, the side facing the blast force also had a

890

S. Altoubat et al.

significant and very deep crack like a punching action that even pushed the surface of the wall inward. This crack was at the bottom end of the wall only and was running horizontally. On the other hand, the RF wall completely shattered into pieces and disappeared from inside the precast boundary frame due to the pressure exerted from the blast (see Fig. 6f). 4.1.2 Signal Processing To gain more insight on the damage endured by the walls as a result of the blast, signal processing was conducted using Fast-Fourier Transform (FFT). Four types of walls were selected to conduct this analysis: ECC, FRS-6, MRS, and PLS. Given the acceleration record for each wall, the analysis is carried out in three steps: 1. Select a low-pass cut-off normalized frequency (0\fcut \1) through a visual analysis of the FFT curve; 2. Apply the Butterworth 2nd order filter to filter out all normalized frequencies higher than fcut ; 3. Perform FFT on the filtered acceleration record. The above steps were repeated for acceleration data obtained earlier through drop weight tests on duplicate walls. This step is important for that it gives a baseline FFT curves to compare with those obtained from blast tests. Figure 6 shows the amplitude spectra for each pair of walls after the impact and blast tests. It is clear from the spectra that frequency content of the blast-tested walls shifted towards the left indicating a decrease in the fundamental frequency of the walls. The decrease of the fundamental frequency means that the walls became flexible during and after the blast tests; this is an important indicator that the walls has experienced a damage that led to the reduction of its stiffness, thus, lengthened its period. Excluding the PLS wall, the amount of the frequency shift (i.e., reduction of the fundamental frequency) is the highest in the MRS wall, followed by FRS-6, then ECCS. This trend suggests that the among these three wall-types, the ECCS wall endured the least damage and the MRS wall experienced the highest damage. This has been also supported by the photos of the corresponding walls included in the same Figure. Now for the PLS wall, the impact test that should have served as a baseline resulted in a damage to the wall, thus, the results depicted in Fig. 6d are for a PLS wall with a certain degree of damage, hence, the frequency reduction in the duplicate wall after the blast is seen to be relatively small (the spectra in Fig. 6 are in logscale).

Masonry Walls Strengthened with Fiber Reinforced Concrete

(a) ECCS

(b) FRS-6

(c) MRS

Fig. 6. Amplitude spectra for the selected wall specimens and photo of tested walls.

891

892

S. Altoubat et al.

(a) PLS

(e) FRS-4.6

(f) RF

Fig. 6. (continued)

4.2

Pressure and Acceleration Time Histories

The results collected from blast test were the recorded air pressure, acceleration, cracking behaviour and the level of damage. The air pressure recorded during the blast test is plotted against the blast duration in Fig. 7a and shows that the maximum air pressure was around 130 MPa. The air pressure plot represents a complex wave due to the rebound of wave after collision on the surface of three surrounded walls. To confirm the arguments made above concerning the damage level caused by the blast, the diagrams of acceleration time histories are plotted in Fig. 7b and 7c. Figure 7b shows that the acceleration response of FRS-4.6 degrades quicker than the FRS-6 and ECCS walls. This indicates that the FRS-4.6 wall became more flexible due to the damage caused by the blast. Similarly, The FRS-6 wall that had higher level of cracking than the ECCS wall, exhibit higher rate of degradation in acceleration response compared to ECCS wall. However, the ECCS wall shows more stiff behaviour suggesting of less damage. Meanwhile, it can be inferred from Fig. 7c that the RF

Masonry Walls Strengthened with Fiber Reinforced Concrete

893

wall, which was fully shattered after the blast shows significant degradation of acceleration response compared to PLS wall and the other walls. Overall it should be noted that all the walls tested in this research had cracks on the opposite side to the blast side with different level of severity, depth, and width. The side facing the blast didn’t show any visible cracks except for the control walls. The behavioural comparison of the tested walls demonstrated that the ECCS and FRS strengthened walls had the highest level of energy absorption thus mitigated the level of damage caused by the air blast waves. While the MRS and PLS walls were severely damaged. In contrast the conventional method of constructing masonry walls had zero resistance against the air pressure caused by the blast force. TNT-BLAST INDUCED AIR-PRESSURE 140

PRESURE (MPA)

120 100 80 60 40 20 0 0

2

4

6

8

-20

10

12

14

16

18

20

TIME (MSEC)

(A) FRS-6

PLS

FRS-4.6 2200

1600

1600

ACCELERATION (G)

ACCELERATION (G)

ECCS 2200

1000 400 -200 -800 -1400

RF

1000 400 -200 -800 -1400

-2000

-2000

0

1

2

3

4

5

6

7

8

9

TIME (MSEC)

(B)

10

0

1

2

3

4

5

6

7

8

9

10

TIME (MSEC)

(C)

Fig. 7. Pressure exerted on the walls and acceleration time histories

5 Summary and Conclusions A total of 12 URM walls were built and then strengthened using cement based fiber reinforced shotcrete mixes with and without conventional steel reinforcement. Walls with conventional plaster were also constructed and used as a reference. The walls were subjected to 10 kg of TNT explosive at a stand-off distance of 1 m. The parameters measured during the blast test were the air pressure and acceleration.

894

S. Altoubat et al.

The acceleration measurement was used to calculate the frequency content of the walls using FFT and Filtering process. The walls were also visually examined for damage after the explosion shock. The results of frequency content, degradation in the acceleration time histories and visual inspection indicated that addition of fibers significantly improves the out-of-plane resistance of the URM wall even in the event of extreme pressure like blast load. The results revealed that the ECCS wall had the least damage due to blast load followed by FRS-6, FRS-4.6, MRS, and PLS walls. However, the walls strengthened with plain concrete materials with or without steel mesh reinforcement exhibited sever damage compared to the walls strengthened with fiber reinforced concrete. Moreover, the reference wall with cement mortar plaster (RF) could not survive the blast waves and was fully damaged.

References 1. Sameer, A., et al.: Out-of-plane Strengthening of masonry walls with reinforced composites. J. Compos. Constr. 5(August), 139–145 (2001) 2. Ahmad, S., et al.: Experimental study of masonry wall exposed to blast loading. Mater. Constr. 64(313) (2014) 3. Meisl, C.S.: Out-of-plane seismic performance of unreinforced clay brick masonry walls. The University of British Columbia (2006) 4. Valluzzi, M.R., da Porto, F., Garbin, E., Panizza, M.: Out-of-plane behaviour of infill masonry panels strengthened with composite materials. Mater. Struct. 47(12), 2131–2145 (2014) 5. Carney, P., Myers, J.J.: Out-of-plane static and blast resistance of unreinforced masonry wall connections strengthened with FRP. Aci Spec. Publ. 1(SP-230-14), 229–248 (2005) 6. Kalman, D.: Use of steel fiber reinforced cocrete for blast resistance design. Kansas State University (2010) 7. Maalej, M., Lin, V.W.J., Nguyen, M.P., Quek, S.T.: Engineered cementitious composites for effective strengthening of unreinforced masonry walls. Eng. Struct. 32(8), 2432–2439 (2010) 8. Verma, S., Choudhury, M., Saha, P.: Blast resistant design of structure. Int. J. Res. Eng. Technol. 04(13), 64–69 (2015) 9. Casadei, P., Agneloni, E.: Elastic systems for dynamic retrofitting (ESDR) of structures. In: Cost Action C26 Urban Habitat Constructions under Catastrophic Events – Proceedings of Final Conference, pp. 939–948 (2010) 10. Dizhur, D., Griffith, M., Ingham, J.: Out-of-plane strengthening of unreinforced masonry walls using near surface mounted fibre reinforced polymer strips. Eng. Struct. 59, 330–343 (2014) 11. Chakraborty, T., Larcher, M., Gebbeken, N.: Performance of tunnel lining materials under internal blast loading. Int. J. Prot. Struct. 5(1), 83–96 (2014) 12. I. of Medicine, Gulf war and health: long-term effects of blast exposure, vol. 9 (2014) 13. Goel, M.D., Matsagar, V.A.: Blast-resistant design of structures. Pract. Period. Struct. Des. Constr. 19(2) (2014) 14. Tai, Y.S., Chu, T.L., Hu, H.T., Wu, J.Y.: Dynamic response of a reinforced concrete slab subjected to air blast load. Theor. Appl. Fract. Mech. 56(3), 140–147 (2011) 15. Ning, Y.L., Zhou, Y.G.: Shock tubes and blast injury modeling. Chin. J. Traumatol. – Engl. Ed. 18(4), 187–193 (2015) 16. Mayorga, M.A.: The pathology of primary blast overpressure injury. Toxicology 121(1), 17– 28 (1997)

Smart FRCs

Interfacial Bond Quality in Functionally Graded Concretes Incorporating Steel Fibres and Recycled Aggregates Ricardo Chan1(&), Isaac Galobardes2, and Charles K. S. Moy1

2

1 Department of Civil Engineering, Xi’an Jiaotong-Liverpool University, Suzhou, China [email protected] School of Architecture, Planning and Design, Mohammed VI Polytechnic University, Ben Guerir, Morocco

Abstract. A functionally graded material (FGM) is a material presenting gradation in composition and structure, designed to attend to specific functions. FGM produced with concrete, known as functionally graded concrete (FGC), has been studied for several applications by combining layers of distinct types of concrete showing technical benefits. Due to the material discontinuity, an interfacial zone is created between layers, named as layer transition zone (LTZ). As the weakest link between different concrete layers, the bond quality of LTZ may influence the mechanical behaviour of FGC. In this paper, the quality of LTZ in FGC was assessed considering the type of aggregate, content of steel fibres and casting delay between layers. FGC were produced with a top layer of plain cement concrete (PCC) and a bottom layer of conventional fibre reinforced concrete (FRC) or fibre reinforced recycled aggregate concrete (FRRAC). The FGC were assessed for compressive strength and bond strength between layers. The results indicate that the bond quality of LTZ is affected by casting delay and compressive strength of each layer. Moreover, it was noticed that the impact of adding steel fibres was not significant to alter the bond quality in FGC. Overall, adequate bond strengths were obtained in FGCs with casting delays of up to 24 h. Keywords: Functionally graded concrete
Bond strength
Fibre reinforced concrete (FRC)
Fibre reinforced recycled aggregate concrete (FRRAC)

1 Introduction The concept of functionally graded material (FGM) consists of producing an optimised spatial gradation in material composition and structure, which results in tailored properties designed to attend specific functional requirements [1]. Most recently, FGM produced with concrete, known as functionally graded concretes (FGC), have been studied for specific applications, such as precast shield-tunnel segments [2], impact resistant panels [3], marine structures [4], and pavements [5, 6]; showing significant technical benefits. These FGC can be produced by combining two or more layers of distinct types of concrete, resulting in a material with a discontinuous gradation. © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 897–907, 2021. https://doi.org/10.1007/978-3-030-58482-5_79

898

R. Chan et al.

Due to this discontinuity, an interfacial zone is created between the layers, named hereafter as layer transition zone (LTZ). The LTZ can be considered equivalent to the overlay transition zone (OTZ) defined as the interfacial zone between concretes of different ages, which is studied similarly as the interfacial transition zone (ITZ) between the coarse aggregates and cement paste [7, 8]. As the weakest link between layers [9, 10], the bond quality of LTZ can influence the mechanical behaviour of FGC. Despite its importance, studies about the bond quality of LTZ in FGC are limited [11–13]. Therefore, this paper aims to study the quality of LTZ in FGC regarding the influence of several parameters. An experimental program was conducted to produce and test FGC specimens with natural and recycled aggregates, different contents of steel fibre (cf) and different casting delays between layers (Dt). Then, the results obtained from compressive strength and bond strength tests were presented and discussed. Finally, conclusions were drawn regarding the impact of those parameters in the bond quality of FGC.

2 Experimental Program 2.1

Experimental Program

Cement CEM I-42.5 N [14] and tap water at room temperature (20 °C) were used in this study. According to the manufacturer, cement initial and final setting times are 172 and 226 min, or 2.87 and 3.77 h, respectively. Natural and recycled aggregates were adopted for both coarse and fine aggregates. Natural aggregates were composed of limestone gravel and river sand, while recycled aggregates were produced from crushed demolition waste. Figure 1 shows the size particle distributions of aggregates, according to BS 812-103.1:1985 [15]. The main properties of aggregates are presented in Table 1. These are the oven dry density (qrd), the water absorption (WA) and the coefficients of uniformity (Cu) and curvature (Cc). 100

Cumulative passing (%)

Cumulative passing (%)

100 80 60 40 Natural Recycled

20 0

60 40 20

Natural Recycled

0 1

(a)

80

10

Size (mm)

100

0.1

(b)

1

10

Size (mm)

Fig. 1. Particle size distribution and grading limits for (a) coarse and (b) fine aggregates.

Interfacial Bond Quality

899

Table 1. Main characteristics of the aggregates. Type of aggregate Natural Recycled

Coarse Fine Coarse Fine

qrd (Mg/m3) 2.64 2.70 2.25 1.90

WA (%) 1.11 2.60 7.51 15.01

Cu

Cc

1.67 2.70 7.82 10.6

0.90 0.84 3.59 0.95

Natural aggregates present higher values of qrd and lower values of WA in comparison with recycled aggregates. These results are expected, due to the old mortar attached to recycled aggregates, which impacts qrd and WA [16, 17]. The values of Cu and Cc suggest that the natural aggregates are more uniformly graded than recycled aggregates. As reinforcement, hooked-end steel fibres were used. The fibres present 60 mm of length, 0.75 mm of diameter and tensile strength of 1150 MPa. 2.2

Mixes and Production

In this study, an FGC configuration optimised for the application in structures subjected to bending was used: conventional concrete in the top layer and reinforced concrete in the bottom layer [18]. Thus, to study the effect of type of aggregate, two FGC groups were considered: one with plain cement concrete (PCC) in the top layer and a bottom layer of fibre reinforced concrete (FRC), as shown in Fig. 2a; and other with PCC in the top layer and fibre reinforced recycled aggregate concrete (FRRAC) in the bottom layer, as shown in Fig. 2b. It should be noted that; for this study, each layer of concrete corresponded to half of the component’s height (h/2).

h/2

PCC

h/2

PCC

h/2

FRC

h/2

FRRAC

(a)

(b)

Fig. 2. FGC groups: (a) PCC+FRC and (b) PCC+FRRAC.

The mix proportions for PCC/FRC and FRRAC are presented in Table 2. The same cement content of 475 kg/m3 and free water/cement ratio (w/c) of 0.45 were adopted for both mixes. Due to the difference between the actual moisture content and water absorption of aggregates, the quantities of aggregates and water were adjusted according to [19].

900

R. Chan et al. Table 2. Mix proportions (kg/m3). Mix Cement Water Type of aggregate Fine aggregate Coarse aggregate PCC/FRC 475 215 Natural 790 890 FRRAC Recycled 805 715

The impact of steel fibres was assessed by adopting three values of cf: 0.25, 0.50 and 0.75%, in terms of concrete volume. To assess the influence of casting delay between layers, Dt of 0.50, 6.00 and 24.0 h were applied to produce FGC, representing the stages before, after and long after the setting time, respectively. A total of 18 FGC groups were produced, as listed in Table 3. Each group was identified by the following codification: Letter + First number + Second number. The letter corresponds to type of aggregate (N for natural and R for recycled), and the first and second numbers to cf and Dt, respectively. Furthermore, to simplify the analysis, FGC produced only with natural aggregates are hereafter referred to as N-concretes, and concretes containing recycled aggregates, as R-concretes. Homogenous concretes (PCC, FRC and FRRAC) were used to produce cylinders with 150 mm of diameter and 300 mm of height, while cubes with 150 mm of nominal size were produced with the FGC presented in Table 3. The cylindrical specimens were produced according to BS EN 12390-2:2009 [20]. On the other hand, cubic specimens were cast as follows: first, the bottom layer was casted, compacted and covered with a polyethylene sheet until the moment of casting the top layer; then, after waiting for the

Table 3. FGC groups considered in this study. Mix PCC+FRC

Group Type of aggregate N0.25-0.50 Natural N0.25-6.00 N0.25-24.0 N0.50-0.50 N0.50-6.00 N0.50-24.0 N0.75-0.50 N0.75-6.00 N0.75-24.0 PCC+FRRAC R0.25-0.50 Recycled R0.25-6.00 R0.25-24.0 R0.50-0.50 R0.50-6.00 R0.50-24.0 R0.75-0.50 R0.75-6.00 R0.75-24.0

cf (%) Dt (h) 0.25 0.50 6.00 24.0 0.50 0.50 6.00 24.0 0.75 0.50 6.00 24.0 0.25 0.50 6.00 24.0 0.50 0.50 6.00 24.0 0.75 0.50 6.00 24.0

Interfacial Bond Quality

901

respective time Dt, the polyethylene sheet was removed and the top layer was casted on top of the bottom layer. After casting, the specimens were covered again with polyethylene sheet. After 24 h, the specimens were demoulded and cured in water at a temperature of approximately 20 °C for 28 days according to BS EN 12390-2:2009 [20]. 2.3

Test Methods

The compressive strength of each layer of FGC was used as quality control of concrete. The compressive strength was taken as the average of three cylindrical specimens assessed according to BS EN 12390-3:2009 [21], as shown in Fig. 3a. The compressive strength (fcm) is calculated using Eq. (1), which depends on the maximum compressive load at failure (F) and the specimen’s cross-section area on which the compressive load acts (Ac).

(a)

(b)

Fig. 3. Setups for (a) compressive test and (b) splitting test.

fcm ¼

F Ac

ð1Þ

The bond strength in FGC was assessed according to the splitting test described in BS EN 12390-6:2009 [22] and illustrated in Fig. 3b. For each FGC group, three cubic specimens were used. The bond strength (fctm) is calculated using Eq. (2), which depends on the maximum load (F) and the length and width of the specimen (L and d, respectively). Furthermore, the bond quality of LTZ in FGC was categorized according to the classification suggested in [23] and presented in Table 4. fctm ¼

2F pLd

ð2Þ

902

R. Chan et al. Table 4. Bond quality classification according to bond strength [23]. Bond quality Excellent Very good Good Fair Poor

fctm (MPa)  2.1 1.7–2.1 1.4–1.7 0.7–1.4 0.0–0.7

3 Results and Discussion The compressive strength of the concrete used in each layer of FGC are presented in Fig. 4. The values are referred to the average compressive strength of three specimens evaluated at the age of 28 days, with the respective standard deviation. As previously indicated, the specimens from top layer are not reinforced while bottom layer specimens are reinforced with steel fibres. 50.0 Top layer

Bottom layer

fcm (MPa)

40.0 30.0 20.0 10.0 0.0

Fig. 4. Compressive test results.

The variation in the results for each layer is below 12.5%, assuring the quality of the concrete produced. The PCC used in the top layer presents fcm equal to 30.3 MPa. On the other hand, fcm of FRC and FRRAC is enhanced with the increase of cf, as observed in previous studies [24, 25]. Also, FRRAC presents, on average, 30% lower fcm than FRC, as expected [26]. The bond strength of 18 FGC groups are presented in Fig. 5. The values are related to the average of three results and the respective standard deviation, obtained from FGC cubes evaluated at the age of 28 days.

Interfacial Bond Quality

903

4.00 0.50 h

6.00 h

24.0 h

fctm (MPa)

3.00 2.00 1.00 0.00

Fig. 5. Bond strength results.

The results from Fig. 5 indicate that N-concretes present higher fctm than Rconcretes, for Dt = 0.50 h. This is expected, since fctm is affected by fcm of each layer, as pointed in other studies [13, 27]. However, for Dt > 0.50 h, N-concretes present a decrease in fctm, while the opposite is verified in R-concretes. This trend difference can be related to transition in failure mode presented in the splitting test with the increase of Dt, shown in Fig. 6. The classification adopted for the failure modes observed in this study is described as follows: (Type A) interfacial failure combined with concrete fracture; (Type B) interfacial failure combined with partial concrete fracture; and (C) interfacial failure combined with minimum or non-visible concrete fracture. For Dt = 0.50 h, N-concretes present type A failure mode, with a main crack in the LTZ and fracture of one or both of the layers, as indicated in Fig. 6a. This strong bond may be related to the fact that, at that moment, the concrete did not reach the initial setting time [8]. However, when Dt is increased to 6.00 h and 24.0 h, the failure mode changes to type B (see Fig. 6b) and type C (see Fig. 6c) in N-concretes, respectively. This transition in failure mode suggests that the bond strength of LTZ becomes lower than the tensile splitting strength of the concrete layers as Dt increases. This is expected, since the final setting time of 3.76 h was already passed when Dt  6.00 h [8]. On the other hand, type A failure mode is observed in all R-concretes specimens, independently of Dt, as indicated in Fig. 6d–f. This failure behaviour suggests that the bond strength of LTZ in R-concretes is not reduced by the increase of Dt. In fact, as previously pointed out, the bond strength in R-concretes is enhanced with the increase of Dt. The differential stiffness between layers may contribute to this behaviour, as shown in other study [28]. Nevertheless, further investigation is required to better identify other factors affecting the bond strength of LTZ in FGC produced with PCC and FRRAC. Furthermore, the results presented in Fig. 5 suggest that the bond strength of LTZ in FGC is not affected by cf, contrasting with other study [29]. This difference can be related to the size of steel fibres used in each study. Long fibres (60 mm of length) were used in this study in contrast to the short fibres (13 mm of length) used in [29], which are more effective in the control of microcracking than long fibres [30]. In addition, the fibre orientation could have also played a role on the results difference between studies.

904

R. Chan et al.

The wall effect in moulded cubes induces certain fibre orientation which is different from the one observed in larger concrete elements, such as slabs and panels [31]. As a result, the stress distribution provided by fibres could be different between moulded cubes and extracted cubes from large elements. Thus, further investigation should be done regarding the effect of fibre orientation in the bond strength of FGC.

(a) Type A

(b) Type B

(c) Type C

(d) Type A

(e) Type A

(f) Type A

Fig. 6. Failure modes in N-concrete and R-concrete cubes for Dt equal to: (a, d) 0.50 h; (b, e) 6.00 h; and (c, f) 24.0 h.

Finally, Table 5 presents the classification of bond quality of LTZ in FGC, along with bond strength results and coefficient of variation, in brackets. According to Table 5, N-concretes present LTZ with bond quality lowering from “excellent” to “very good”, as Dt increases. On the other hand, the bond quality of LTZ in R-concretes enhances from “good” or “very good” to “excellent” with the increase of Dt. Overall, despite the difference in the test results, the bond quality of LTZ in FGC assessed in this study can be considered adequate. Thus, in this study, FGC behaves as a monolithic element, increasing its durability.

Interfacial Bond Quality

905

Table 5. Classification of bond quality of LTZ in FGC. Group N0.25-0.50 N0.25-6.00 N0.25-24.0 N0.50-0.50 N0.50-6.00 N0.50-24.0 N0.75-0.50 N0.75-6.00 N0.75-24.0

fctm (MPa) 3.07 (9.03%) 1.85 (9.63%) 2.09 (9.56%) 2.45 (12.72%) 1.77 (6.95%) 1.84 (26.30%) 2.43 (5.51%) 2.00 (10.79%) 1.97 (11.98%)

Bond quality Excellent Very good Very good Excellent Very good Very good Excellent Very good Very good

Group R0.25-0.50 R0.25-6.00 R0.25-24.0 R0.50-0.50 R0.50-6.00 R0.50-24.0 R0.75-0.50 R0.75-6.00 R0.75-24.0

fctm (MPa) 1.81 (8.25%) 1.82 (1.73%) 2.17 (7.94%) 1.57 (2.82%) 2.31 (2.87%) 2.40 (5.89%) 1.77 (8.17%) 1.98 (4.07%) 2.17 (22.37%)

Bond quality Very good Very good Excellent Good Excellent Excellent Very good Very good Excellent

4 Conclusions The bond quality of LTZ in FGC was assessed by means of an experimental program, considering the impact of type of aggregate, content of fibre and casting delay. As a general conclusion, the experimental results suggest that the bond quality is strongly affected by type of aggregate and casting delay but is not influenced by content of fibre. Furthermore, the results also indicate that an adequate bond quality can be achieved for casting delays up to 24 h, expanding the potential of producing FGC in large-scale. Further conclusions are drawn, as follows: • FRC presented higher compressive strength than FRRAC, showing that the type of aggregate is more important than content of fibre; • N-concretes presented higher bond strength than R-concretes, for casting delay of 0.50 h, indicating the influence of type aggregate. Nevertheless, for longer casting delays, the bond strength of LTZ decreased in N-concretes but increased in Rconcretes; • The behaviour of bond strength with the increment in casting delay can be related to the failure mode observed in the FGC specimens. While N-concretes presented a transition of failure mode, R-concretes failed according to the same type of failure mode; • The long steel fibres used in this study did not significantly affect the bond strength of LTZ in FGC, possibly due to the low contribution in microcracking control provided by this type of fibre; • Adequate bond quality is achieved in all FGC groups studied, regardless of type of aggregate, content of fibre or casting delay. Acknowledgements. The authors would like to acknowledge the Xi’an Jiaotong-Liverpool University (XJTLU) Research Development Fund for the financial support received from the project with reference RDF-16-02-42.

906

R. Chan et al.

References 1. Kawasaki, A., Watanabe, R.: Concept and P/M fabrication of functionally gradient materials. Ceram. Int. 23, 73–83 (1997) 2. Zhang, N., et al.: Support performance of functionally graded concrete lining. Constr. Build. Mater. 147, 35–47 (2017) 3. Mastali, M., Ghasemi Naghibdehi, M., Naghipour, M., Rabiee, S.M.: Experimental assessment of functionally graded reinforced concrete (FGRC) slabs under drop weight and projectile impacts. Constr. Build. Mater. 95, 296–311 (2015) 4. Wen, X., Tu, J., Gan, W.: Durability protection of the functionally graded structure concrete in the splash zone. Constr. Build. Mater. 41, 246–251 (2013) 5. Rao, S., et al.: Composite Pavement Systems, Volume 2: PCC/PCC Composite Pavements, vol. 2 (2013). https://books.google.com/books?id=o1zn2oqvQmEC&pgis=1 6. Hu, J., Fowler, D.W., Siddiqui, M.S., Whitney, D.: Feasibility study of two-lift concrete paving: technical report (2014). https://texashistory.unt.edu/ark:/67531/metapth638590/ 7. Beushausen, H., Alexander, M.G.: Bond strength development between concretes of different ages. Mag. Concr. Res. 60, 65–74 (2008) 8. Qian, P., Xu, Q.: Experimental investigation on properties of interface between concrete layers. Constr. Build. Mater. 174, 120–129 (2018) 9. Rashid, K., Ueda, T., Zhang, D., Miyaguchi, K., Nakai, H.: Experimental and analytical investigations on the behavior of interface between concrete and polymer cement mortar under hygrothermal conditions. Constr. Build. Mater. 94, 414–425 (2015) 10. Beushausen, H., Höhlig, B., Talotti, M.: The influence of substrate moisture preparation on bond strength of concrete overlays and the microstructure of the OTZ. Cem. Concr. Res. 92, 84–91 (2017) 11. Bajaj, K., Shrivastava, Y., Dhoke, P.: Experimental study of functionally graded beam with fly ash. J. Inst. Eng. Ser. A 94(4), 219–227 (2014) 12. Cao, Y., Li, P., Brouwers, H.J.H., Sluijsmans, M., Yu, Q.: Enhancing flexural performance of ultra-high performance concrete by an optimized layered-structure concept. Compos. Part B Eng. 171, 154–165 (2019) 13. Hussein, L., Amleh, L.: Structural behavior of ultra-high performance fiber reinforced concrete-normal strength concrete or high strength concrete composite members. Constr. Build. Mater. 93, 1105–1116 (2015) 14. BSI: BS EN 197-1:2011, Cement - Composition, specifications and conformity criteria for common cements. BSI Standards Limited, London (2011). https://doi.org/10.3403/30205527 15. BSI: BS 812-103.1:1985, Testing aggregates - Method for determination of particle size distribution - Sieve tests. BSI Standards Limited, London (1998). https://doi.org/10.3403/ 00139627 16. Mehta, P.K., Monteiro, P.J.M.: Concrete: Microstructure, Properties, and Materials. McGraw-Hill Education (2006). https://doi.org/10.1036/0071462899 17. Neville, A.M.: Properties of Concrete. Pearson Education Limited, Harlow (2011) 18. Chan, R., Liu, X., Galobardes, I.: Parametric study of functionally graded concretes incorporating steel fibres and recycled aggregates. Constr. Build. Mater. 242, 118186 (2020) 19. Teychenné, D.C., Franklin, R.E., Erntroy, H.C.: Design of Normal Concrete Mixes. Construction Research Communications Ltd., Watford (1997) 20. BSI: BS EN 12390-2:2009, Testing hardened concrete - Making and curing specimens for strength tests. BSI Standards Limited, London (2009). https://doi.org/10.3403/30164903 21. BSI: BS EN 12390-3:2009, Testing hardened concrete - Compressive strength of test specimens. BSI Standards Limited, London (2011). https://doi.org/10.3403/30164906

Interfacial Bond Quality

907

22. BSI: BS EN 12390-6:2009, Testing hardened concrete - Tensile splitting strength of test specimens. BSI Standards Limited, London (2010). https://doi.org/10.3403/30200045 23. Sprinkel, M.M., Ozyildirim, C.: Evaluation of high performance concrete overlayers placed on route 60 over lynnhaven inlet in virginia (2000).. http://www.virginiadot.org/vtrc/main/ online_reports/pdf/MicrosoftWord-VTRC01-R1_Sprinkel&Ozyildirim_pdf 24. Senaratne, S., Gerace, D., Mirza, O., Tam, V.W.Y., Kang, W.H.: The costs and benefits of combining recycled aggregate with steel fibres as a sustainable, structural material. J. Clean. Prod. 112, 2318–2327 (2016) 25. Carneiro, J.A., Lima, P.R.L., Leite, M.B., Toledo Filho, R.D.: Compressive stress–strain behavior of steel fiber reinforced-recycled aggregate concrete. Cem. Concr. Compos. 46, 65– 72 (2014) 26. Chan, R., et al.: Analysis of potential use of fibre reinforced recycled aggregate concrete for sustainable pavements. J. Clean. Prod. 218, 183–191 (2019) 27. Costa, H., Carmo, R.N.F., Júlio, E.: Influence of lightweight aggregates concrete on the bond strength of concrete-to-concrete interfaces. Constr. Build. Mater. 180, 519–530 (2018) 28. Duarte Santos, P.M., Santos Júlio, E.N.B.: Factors affecting bond between new and old concrete. ACI Mater. J. 108, 449–456 (2011) 29. Huang, H., Yuan, Y., Zhang, W., Gao, Z.: Bond behavior between lightweight aggregate concrete and normal weight concrete based on splitting-tensile test. Constr. Build. Mater. 209, 306–314 (2019) 30. Bentur, A., Mindess, S.: Fibre Reinforced Cementitious Composites. Taylor & Francis, London and New York (2007) 31. Torrents, J.M., et al.: Inductive method for assessing the amount and orientation of steel fibers in concrete. Mater. Struct. 45, 1577–1592 (2012)

Towards Rebar Substitution by Fibres – Tailored Supercritical Fibre Contents Katharina Look1(&), Peter Heek2, and Peter Mark1 1

2

Ruhr University Bochum, Bochum, Germany [email protected] HOCHTIEF Engineering GmbH, Essen, Germany

Abstract. Structural application of steel fibre reinforced concrete (SFRC) progressively increases. To create load-bearing components without additional steel reinforcement, mixtures with supercritical fibre contents tailored to the structure and application must be devised. An innovative formwork concept is developed that enables for fibre contents up to 1.0 Vol.-% the steered alignment of fibres. Based on the observation that fibres orient parallel to formwork edges it applies internal formworks. The efficiency is verified by measurements of spatial fibre orientations and flexural tensile strengths. Results on single span beams indicate that favourable fibre orientations perpendicular to cracks can be achieved which is even more effectively with increasing fibre content. The impact does not go along with enhancements of flexural tensile strength. To investigate the effect of steered fibres on bearing capacities of spatial elements like slabs, the SFRC’s composition is optimized with regard to fibres’ content and orientation. The aim is to achieve two-dimensional fibre orientations in directions of principle tensile stresses by artificially limiting the specimen’s height, so that fibres align horizontally in concrete’s flow direction without additional steering. For a supercritical fibre content of 1.8 Vol.-%, it is possible to substitute conventional (mesh) reinforcements of approximate 6.7 cm2/m (fyk = 500 N/mm2). Keywords: Steel fibre reinforced concrete
Fibre alignment
Rebar substitution
Strain-hardening
Formwork concept
Steering of fibres Prefabrication




1 Introduction Gain of knowledge in the field of SFRC is opening up new areas of application for the material. The fibres’ effect in e.g. fire [1], fatigue [2] or under creep [3] is subject of current scientific investigations. Numerous standards (e.g. [4, 5]) and design tools (e.g. [6–8]) for SFRC are available in the meantime. Although fibres are still mainly used in non-load-bearing structures or in combination with conventional reinforcement. The main issue is the ductile load-bearing behaviour, which is - taking into account also the high scatter in the SFRC’s strength - hard to fulfil with common subcritical fibre contents. As fibre contents with about 0.5 Vol.-% usually yield to strain-softening after cracking and thus low calculative residual strength, refined concrete mixtures with high © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 908–919, 2021. https://doi.org/10.1007/978-3-030-58482-5_80

Towards Rebar Substitution by Fibres

909

Fig. 1. Concept of substituting crosswise reinforcement with a prefabricated slab of SFRC.

strength fibres are subject of current research to enable reliable strain-hardening behaviour and ductile tensile failure without additional rebars. Thereby, effects of composition mixture [9], structural dimensions for optimum fibre distributions [10], innovative casting methods and sequences [11] to establish desired fibre orientations as well as scattering in the post-cracking strength [12] are in the focus of interest. As the individual parameters are strongly interacting, structural and material interrelationships are itemised subsequently. Objective is the complete replacement of conventional (bar) reinforcement by macro steel fibres while ensuring ductile structural behaviour. This should be achieved by inserting as many fibres as feasible into the concrete for maximum performance. In addition, the alignment of the fibres is steered in order to achieve the most effective orientation. This is obtained differently for one- (beams) and two-dimensional structures (slabs, foundations). In the first case, additional interior formworks apply, as fibres tend to orient parallel to formwork edges. In two-dimensional members, the cross-section’s height is artificially limited within the casting sequence to steer fibre orientations in direction of the fresh concrete’s flow. To assess the accuracy of the innovative formwork and casting approach, experiments are executed that provide information on achieved spatial fibre orientations. Resulting load-bearing capacities as well as scattering of strength values in the post-cracking domain are verified with beam tests. A tailored fibre mixture for substituting conventional reinforcement is derived that also takes into account workability of fresh concrete with increasing fibre content. Starting with beam elements, the findings are transferred in a second step to spatial structures by optimizing the material and structural parameters. This becomes necessary, as a forced alignment with interior formwork does not work here. To steer fibres’ alignment in the horizontal plane, the specimen’s height is limited. Applications are seen in hybrid systems, where a prefabricated SFRC-slab with supercritical fibre content replaces conventional mesh reinforcement, e.g. in foundations or wall-elements (Fig. 1). The precasted SFRC-slab is transported to the project site and then the structure is completed with cast-in-place concrete. In order to ensure the transmission of shear forces in the joint between SFRC and normal concrete, the surface of the SFRC-slab is roughened.

910

K. Look et al.

2 One-Directional Steering Of Fibres 2.1

General

For beams, which are one-dimensional structures, the fibres should be steered mainly into the direction of the tensile stress trajectories resulting from bending. Therefore, an internal formwork is designed consisting of metal segments dividing the beam formwork into several horizontal sections. The influences of the mechanical alignment in combination with different fibre quantities on the fibre orientation, workability and load-bearing capacity are investigated here. For this purpose, typical fibre contents up to 1.0 Vol.-% apply. Beams are tested in four-point-bending to determine residual flexural tensile strengths. To investigate the efficiency of the internal formwork concept, fibre orientation of the beams is measured subsequently. 2.2

Experimental Program

The test series (M1) covers normal strength concrete of class C30/37 and conventional fibre contents of 0.5 Vol.-% (M1-1) and 1.0 Vol.-% (M1-2). The utilized fibre type FF60/75–1450 (lf = 60 mm, df = 0.75 mm) has a tensile strength of 1450 MPa and a straight shape with two anchor knots at each end (cf. Fig. 4). Due to this type of anchorage a better orientation factor and less agglomerations should be achieved [12, 13]. The internal formwork concept consists of a wooden frame in which metal segments are fixed, so that the formwork of a beam (w  h  l = 150  150  700 mm3) is divided into five 30 mm wide sections during casting and compaction (Fig. 2). Three beams with (notation a) and without internal formwork (notation b) are casted for each fibre quantity. After casting and compacting, the internal formwork is slowly removed by hand to prevent delamination of the concrete’s matrix (see grips in Fig. 2). No compaction energy is added afterwards. Before testing, the specimens are rotated by 90°. Discontinuities among the layers of FRC cannot be observed (Fig. 3).

Fig. 2. Practical realization of the internal formwork.

Fig. 3. Concept of the internal formwork for a beam.

After testing, the beams are cut into cubes. Fibre orientations in all spatial directions are then measured based on ferromagnetic induction with the measurement device

Towards Rebar Substitution by Fibres

911

BSM 100 as described in [14]. The fibre orientation factor ηi for each spatial direction (i = x, y, z) follows from the measured induction voltage Ui using Eq. (1) [14]. Ui gi ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Ux2 þ Uy2 þ Uz2

ð1Þ

Defining ηi as the ratio of the steel fibre length projected in the respective axis i to the actual steel fibre length lf [15], the angle a between the horizontal plane (x-y-plane) and the fibre can be determined with Eq. (2). a ¼ arcsinðgz Þ

2.3

ð2Þ

Results

2.3.1 Flexural Tensile Strengths Beams are tested in four-point-bending according to the German DAfStb-Guideline “Steel Fibre Reinforced Concrete” [4]. Figs. 4, 5, 6 and 7 show deflection-stress diagrams for the individual beams (black lines, primary ordinate) as well as mean values (black dashed lines, primary ordinate) and deflection-dependent coefficients of variation (COV) as a measurement of the flexural strength’s scatter (grey lines, secondary ordinate). Mean values of residual flexural tensile strengths for deflections of d = 0.5 mm (ffcflm,L1) and d = 3.5 mm (ffcflm,L2) are presented in Table 1. Mean values of the coefficient of variation (COVm) are also given. Deflection-hardening behaviour for all specimens with 1.0 Vol.-% of fibres is achieved (Figs. 5 and 7). For beams with 0.5 Vol.-% (Figs. 4 and 6), the post-cracking behaviour stabilizes after a stress loss to the pre-cracking initiation state. For test series M1, the residual flexural tensile strengths can be increased up to 167% by doubling the fibre content - regardless of additional internal formwork. No essential difference in the residual flexural tensile strength can be determined with or without internal formwork, i.e. with steered fibres no enhancement in performance is gained. However, it can be observed that scattering decreases significantly by using the internal formwork. This phenomenon is much more pronounced for the lower fibre content (M1-1) than for the higher fibre content (M1-2). There, the COV for mechanically oriented fibres even slightly increases. Nevertheless, it remains significantly below the value of M1-1b and a value of 25%, which is typical for SFRC beam tests [16]. Considering only those specimens without internal formwork (b), it is noticeable that the scattering in the postcracking behaviour decreases with increasing fibre content. This observation can be traced back to the fact, that at very low intended fibre contents the number of fibres in the cracked cross-section may even be zero, which exerts a much stronger impact on the post-cracking strength than scattering fibre numbers at high intended contents. Due to the high quantity of fibres, they interfere with each other in their alignment. This effect can be used positively combined with the known correlations between casting direction and fibre orientation. For example, together with the preferred orientation of the fibre perpendicular to the casting direction [17], the fibre orientation in twodimensional components can be specifically steered and predicted under constant fabrication conditions.

912

K. Look et al.

Fig. 4. Deflection-stress curves, mean value and COV for M1-1a.

Fig. 5. Deflection-stress curves, mean value and COV for M1-2a.

Fig. 6. Deflection-stress curves, mean value and COV for M1-1b.

Fig. 7. Deflection-stress curves, mean value and COV for M1-2b.

Table 1. Mean values of residual flexural tensile strengths for deflections of 0.5 mm (L1) and 3.5 mm (L2) and mean value of COV for M1. Mixture f fcflm;L1 [MPa] f fcflm;L2 [MPa] COVm [%] M1-1a M1-1b M1-2a M1-2b

5.51 4.94 7.57 7.82

4.40 4.05 6.66 6.78

4.7 15.8 7.6 8.1

2.3.2 Fibre Orientation By means of the aforementioned BSM 100 the percentages of orientation for the x-, yand z-direction are measured. Therefore, the tested SFRC beams of series M1 are cut into five (A–E) cubes (Fig. 8). The results are shown in Fig. 8 for M1-1 and in Fig. 9 for M1-2. In both figures, the dashed lines represent orientation for conventionally casted beams. Solid lines correspond to beams casted with an internal formwork.

Towards Rebar Substitution by Fibres

Fig. 8. Percentages of orientation for M1-1.

913

Fig. 9. Percentages of orientation for M1-2.

In SFRCs with high and low fibre contents, the proportions of fibres oriented in the spatial directions do not significantly differ. Lower percentages for the y-axis (casting direction) in cube B and cube D result from the geometry of the internal formwork: A cross stud stabilizes the lamellas in the middle of the beam (cf. Fig. 2), so that direct filling is not possible here and the fibres cannot orient optimally perpendicular to the casting direction [17]. The fibres in the end cubes are also predominantly aligned in the y-direction, as the fibres straighten up at the edge during casting (“wall-effect” [18]), since orientation in z-direction is obstructed by the internal formwork. Table 2 lists the mean values of the orientation percentages in each spatial direction with the corresponding COV in brackets as well as the orientation factor ηz (Eq. (1)) and the angle to the horizontal plane a (Eq. (2)). The mean percentage of fibres that do not contribute to the load transfer, i.e. those parallel to cracks (z-direction), can be reduced by 65% compared to conventionally casted beams. Since the orientation is strongly limited in the height of the beam, the factors scatter in a small range. However, the scatter for the high fibre content is lower than for the lower fibre content if there is no internal formwork used - one more indication for correlations between fibre orientation and high fibre contents. In addition, the angle a - limited to 30° by the internal formwork - is observed with a mean angle of 9° (M1-1a) or 11° (M1-2a), respectively. This underlines the very good efficiency of the developed formwork concept. Table 2. Mean values of fibre orientations, orientation factor for z-direction and fibre angle to the horizontal plane for M1 (COV [%]). Mixture M1-1a M1-1b M1-2a M1-2b

x [%] 60.6 (7.2) 49.1 (11.6) 61.9 (7.0) 45.6 (9.6)

y [%] 28.3 (15.9) 18.8 (20.1) 25.0 (15.6) 17.5 (16.0)

z [%] 11.1 (7.5) 32.1 (8.1) 13.1 (2.9) 36.9 (5.8)

ηz [-] 0.16 0.52 0.19 0.60

a [°] 9° 31° 11° 37°

914

K. Look et al.

Comparing the results of the flexural tensile tests from the previous section with the measured fibre orientations, the lower scatter for series M1-1b in contrast to M1-2b can be traced back to a lower percentage of aligned fibres in the z-direction. This confirms the assumption that larger contents of fibres in fresh concrete impair their free orientation. Most fibres straighten nearly oriented in the direction of principle tensile stress trajectories which results in lower scattering of flexural tensile strengths, but not in higher load-bearing capacities (cf. Table 1).

3 Two-Directional Steering of Fibres 3.1

General

The presented findings for the steering of fibres are now transferred to spatial structures. Two objectives are pursued: First, the performance class is maximized by adding as many fibres as possible to the concrete. Second, fibre orientations are optimized in term of two-dimensional alignment in directions of principle tensile stresses. Criteria to which the concrete must comply with regard to the aspired performance are among other aspects a very good workability and a homogeneous fibre distribution. To achieve these requirements, a small maximum grain size of 8 mm, a fine grading curve and a high amount of cement and fly ash are used. For verification of performance classes, beams are casted without internal formwork and tested according to Sect. 2. To investigate the fibre orientation in two-dimensional structures, slabs are casted. Thereby, the height of the slab is limited to 10 cm in order to minimize the fibre angle to the horizontal plane [18]. As usual, the width and depth are greater than its height [19]. For verification of fibre orientations, the slabs are cut into cubes and measured by means of the aforementioned BSM 100 as well. 3.2

Experimental Program

In the second series (M2), the fibre content is maximized. Beginning from an initial content of 1.0 Vol.-% (cf. M1), 0.06 Vol.-% are successively added resulting in a final content of approx. 1.8 Vol-%. The dosage of the superplasticizer is adjusted after each fibre addition. The fibre type is also adapted to maximize residual tensile strengths. For this purpose, the high-performance fibre Dramix 5D-65/60BG (lf = 60 mm, df = 0.9 mm) with a tensile strength of 2300 MPa and a triple deformed hooked end (cf. Fig. 10) serves. Although straight fibres are recommended to obtain a more pronounced orientation factor, this series focuses on reachable performance classes. In order to account for the high tensile strength of the fibres -which require sufficient bond strength with the concrete’s matrix to activate full bearing capacities- a high strength concrete of strength class C60/75 is applied. Six beams are tested for validation of performance. Additionally, a slab with dimensions of 50  50  10 cm3 (cf. Fig. 12) is casted to evaluate the desired two-dimensional fibre orientation for the intended design of a prefabricated SFRC-slab.

Towards Rebar Substitution by Fibres

3.3

915

Results

3.3.1 Flexural Tensile Strengths The beams of the second series (M2) are also tested in four-point-bending. Corresponding deflection-stress curves are presented in Fig. 10, which show a deflectionhardening behaviour with the exception of beam 2 with a load-bearing capacity many times lower compared to the other specimens. After cutting, some irregularities in the form of matrix imperfections are detected. The beam can bear less force because fibres are not sufficiently anchored in the concrete and therefore cannot transmit forces. Additionally, the beam could be eliminated as an outliner with the Grubbs test [20]. However, mean values and COV with and without beam 2 are highlighted in Fig. 10. A load-bearing capacity with maximum flexural tensile strength of about 9.0 MPa is reached.

Fig. 10. Deflection-stress curves, mean value and COV (with and without beam 2) for M2-1.

Table 3 presents the mean values of residual flexural tensile strengths for deflections of d = 0.5 mm (ffcflm,L1) and d = 3.5 mm (ffcflm,L2) and COVs for M2 excluding beam 2. In contrast to the conventionally casted beams of series M1-2b, scattering is greater for series M2, although the fibre content is maximized. The reason behind may be the applied fibre type as the hooked ends are attributed to exert a dual impact. On the one hand, they improve load-bearing capacities due to their plastic deformation capacities. On the other hand, they promote adverse fibre distributions and orientations in fresh concrete due to fibre agglomerations [13]. It is also important to point out that the load-bearing capacity for SFRC with 1.8 Vol.-% of the second fibre type (M2) equals the capacity for SFRC with 1.0 Vol.% of the first fibre type (M1–2). Since fibre sedimentation can be excluded, stagnation of performance is mainly seen as a result of maximum interlocking forces reached in the fibre-matrix zone. However, other effects, e.g. due to aggregate composition or concrete strength, have to be part of further investigations.

916

K. Look et al.

Table 3. Mean values of residual flexural tensile strengths for deflections of 0.5 mm (L1) and 3.5 mm (L2) and mean value of COV for M2. Mixture f fcflm;L1 [MPa] f fcflm;L2 [MPa] COVm [%] M2

7.98

6.76

12.7

3.3.2 Fibre Orientation The moulds of the slabs are casted from the centre to the edges while continuous external vibrations prevail. The aim is to keep the flow direction of the concrete constant in this direction (tangential). Thereby, the fibres are aligned perpendicular to the flow direction [21]. In beams, fibres tend to align parallel to the flow direction of fresh concrete as described by the so called “tunnel-effect” [10]. In contrast, slabs allow for spatial dispersal of fresh concrete and thus fibres predominately align radially in a circle around the filling location and perpendicular to the casting direction, i.e. in the horizontal x-y-plane [21]. The slab is cut into 16 cubes, each with an edge length of 10 cm (Fig. 11). Fig. 11 shows the scheme of dividing the slab and the notation of the cubes. To eliminate areas with “wall-effect” (cf. Sect. 2.3.2) [18], the first 5 cm on each side of the specimen are cut off (Fig. 11). Related values for the individual cubes are shown in Fig. 12. Mean values of fibre orientation for

Fig. 11. Measuring concept of the slab (dimensions in [cm]).

Fig. 12. Percentages of orientation for M2.

the whole slab, COV, orientation factor for the z-direction and calculated fibre angle to the horizontal plane are presented in Table 4. Mean values of orientation percentages of the slab are only slightly higher than those of the beams with an internal formwork concept of M1, although no internal formwork was used here. By reducing the specimen’s height and applying a very high fibre content, it is possible to achieve a mean fibre angle a of 15° to the horizontal plane. Nevertheless, it should be noted that the measured orientation factors for all directions scatter more than those of the beams do as no additional internal formwork is used here. If the fibre orientation value for the

Towards Rebar Substitution by Fibres

917

z-direction (16.2%) is subtracted from the possible 100%, 41.9% of the fibres should be aligned in both, the x- and y-direction respectively for an ideal two-dimensional fibre distribution. With an absolute deviation of ±3.2%, an equally, two-dimensional orientation of the fibres is given here. Table 4. Mean values of fibre orientations, orientation factor for z-direction and fibre angle to the horizontal plane for M2 (COV [%]). Mixture x [%] y [%] z [%] ηz [-] a [°] M2 45.1 (20.3) 38.7 (22.4) 16.2 (19.2) 0.26 15°

4 Practical Application In order to make reliable predictions on load-bearing capacities, the scatter in postcracking behaviour has to be minimized. Since the idea of substituting conventional reinforcement by a SFRC layer with a bidirectional fibre orientation for twodimensional load-bearing behaviour, multiple statically indeterminate systems are particularly appropriate. Steel fibres transfer tensile forces over crack edges. Plastic hinges or yield lines arise, which allow the redistributions of stresses. Loads can be further increased after cracking and load-bearing capacity reserves are activated. In contrast to statically determined systems, like beams in flexural tensile tests, statically indeterminate structures with a large cross-sectional area under tension scatter less [22]. Standardization over the whole production process reduces scattering in fibre orientation and load-bearing capacities. Tensile strengths and finally the design can be statistically predicted and used for dimensioning. Prefabrication with lean construction methods is a suitable option for continuous quality control and compliance of the components. Based on the results derived in the experiments, the replacement of a conventional (mesh) reinforcement by a prefabricated SFRC-slab is presented. With a maximum flexural tensile strength of 9.0 MPa (cf. Sect. 2 and 3) and a conservative conversion factor of 0.37 [4], a calculated tensile stress of 3.33 MPa results. Smearing this stress over a cross-section with a supposed slab height of 10 cm the slab has a resistance of 333 kN/m in both directions. Assuming a standard reinforcement B500 (fyk = 500 MPa), reinforcement of 6.7 cm2/m is required. This corresponds to a mesh reinforcement of about 2  Q 335 (each with 3.35 cm2/m in both spatial directions), which are possible to substitute with the presented SFRC mixture. This amount of reinforcement is used in typical foundations of building construction.

5 Conclusions In this contribution, an innovative casting concept in terms of location of formwork and casting sequence is developed to optimize fibre orientations in beams and slabs. The main conclusions are:

918

K. Look et al.

• For beams, additional interior formworks are suitable since fibres tend to orient perpendicular to formwork edges. The developed internal formwork concept can also be adapted to other load distributions, such as the alignment to the inclined tensile struts under shear load. • In case of slabs, two-dimensionally alignment of fibres can be achieved by artificially limiting the cross-sections height. Fibre distributions can be steered by means of the fresh concrete’s flow direction. • Residual flexural tensile strengths do not increase by the systematic alignment of steel fibres for one-dimensional structures. Nevertheless, a mechanical alignment for low fibre contents can considerably reduce scattering. • In spatial structures with aligned fibres, a critical fibre content seems to limit maximum flexural tensile strengths as the further increase of the fibre content do not significantly enhance load-bearing capacities. The critical fibre content depends on several conditions and is here for hooked end and macro fibres of high strength steel about 1.8 Vol.-%. The here achieved maximum post-cracking flexural strength of 9.0 MPa exceeds calculative fibre strength for design purposes according to valid standards like [4] considerably. • Tests to determine residual tensile strengths should represent the structural conditions of the actual application. As standard-beam tests just represent one-directional fibre orientations, uniaxial tensile tests on specimens cut from the component (thin SFRC-slab) provide the most accurate results of the residual tensile strength and should therefore be performed. Acknowledgements. The authors would like to thank BASF Construction Solutions GmbH, BauMineral GmbH, Bekaert GmbH, Dyckerhoff GmbH, Feel Fiber GmbH and StraTec GmbH for their friendly provision of the test materials.

References 1. Heek, P., Tkocz, J., Mark, P.: A thermo-mechanical model for SFRC beams or slabs at elevated temperatures. Mater. Struct. 51, 87 (2018) 2. Heek, P., Ahrens, M.A., Mark, P.: Incremental-iterative model for time-variant analysis of SFRC subjected to flexural fatigue. Mater. Struct. 50, 62 (2017) 3. Plizzari, G., Serna, P.: Structural effects of FRC creep. Mater. Struct. 51, 167 (2018) 4. DAfStb: DAfStb-Guideline Steel Fibre Reinforced Concrete, Beuth (2015) 5. Fédération Internationale du Béton (fib): Model Code 2010, Ernst & Sohn (2013) 6. Gödde, L., Strack, M., Mark, P.: M-N-Interaktionsdiagramme für stahlfaserverstärkte Stahlbetonquerschnitte – Anwendung am Beispiel von Tübbingen. Beton- und Stahlbetonbau 105(5), 318–323 (2010a) 7. Gödde, L., Strack, M., Mark, P.: ‘Bauteile aus Stahlfaserbeton und stahlfaserverstärktem Stahlbeton – Hilfsmittel für Bemessung und Verformungsabschätzung nach DAfStbRichtlinie Stahlfaserbeton’. Beton- und Stahlbetonbau 105(2), 78–91 (2010b) 8. Look, K., Heek, P., Mark, P.: ‘Stahlfaserbetonbauteile praxisgerecht berechnen, bemessen und optimieren‘. Beton- und Stahlbetonbau 114(5), 296–306 (2019) 9. Marković, I.: High-Performance hybrid-fibre concrete – development and utilisation. Dissertation, TU Delft (2006)

Towards Rebar Substitution by Fibres

919

10. Stähli, P., Custer, R., van Mier, J.G.M.: On flow properties, fibre distribution, fibre orientation and flexural behaviour of FRC. Mater. Struct. 41(1), 189–196 (2008) 11. Hadl, P., Tue, N.V.: Einfluss der Faserzugabe auf die Streuung im Zugtragverhalten von Stahlfaserbeton. Beton- und Stahlbetonbau 111(5), 310–318 (2016) 12. Hadl, P., Gröger, J., Tue, N.V.: ‘Experimentelle Untersuchungen zur Streuung im Zugtragverhalten von Stahlfaserbeton‘. Bautechnik 92(6), 385–393 (2015) 13. Huß, M.: ‘Neuer Stahlfasertyp eröffnet vielfältige Möglichkeiten für die Betonindustrie‘. BWI BetonWerk Int. 5, 72–76 (2018) 14. Wichtmann, H.-J., Holst, A., Budelmann, H.: ‘Ein praxisgerechtes Messverfahren zur Bestimmung der Fasermenge und -orientierung im Stahlfaserbeton‘. Beton- und Stahlbetonbau 108(12), 822–834 (2013) 15. Lin, Y.-z.: Tragverhalten von Stahlfaserbeton, Dissertation, Universität Kaiserslautern (1999) 16. Heek, P., Lingemann, J., Mark, P., Schnütgen, B., Schulz, M., Zilch, K.: Sicherheitskonzept der DAfStb-Richtlinie Stahlfaserbeton, Deutscher Ausschuss für Stahlbeton (DAfStb) Heft 614 Erläuterungen zur DAfStb-Richtlinie Stahlfaserbeton, pp. 62–69. Teil B: Allgemeine Erläuterungen zur Richtlinie Stahlfaserbeton (Autorenbeiträge), Beuth (2015) 17. Empelmann, M., Teutsch, M.: Faserorientierung und Leistungsfähigkeit von Stahlfasersowie Kunststofffaserbeton. Beton 59(6), 254–259 (2009) 18. Soroushian, P., Lee, C.-D.: Distribution and orientation of fibers in steel fiber reinforced concrete. Mater. J. 87(5), 433–439 (1990) 19. Abrishambaf, A., Barros, J., Cunha, V.: Relation between fibre distribution and postcracking behaviour in steel fibre reinforced self-compacting concrete panels. Cem. Concr. Res. 51, 57–66 (2013) 20. Grubbs, F.E.: Sample criteria for testing outlying observations. Ann. Math. Stat. 21(1), 27– 58 (1950) 21. Barnett, S.J., Lataste, J.-F., Parry, T., Millard, S.G., Soutsos, M.N.: Assessment of fibre orientation in ultra high performance fibre reinforced concrete and its effects on flexural strength. Mater. Struct. 43(7), 1009–1023 (2010) 22. di Prisco, M., Plizzari, G., Vandewalle, L.: Fibre reinforced concrete: new design perspectives. Mater. Struct. 42(9), 1261–1281 (2009)

A Constitutive Model for Steel-FibreReinforced Lightweight Concrete Hasanain K. Al-Naimi and Ali A. Abbas(&) University of East London, London, UK [email protected]

Abstract. A method is proposed to derive and validate material properties for lightweight fibrous concrete using experimental and numerical data. The coarse lightweight material tested (produced by LYTAG) is recycled and offers an alternative to gravel and quarry resources which are at risk of depletion in future. Also, this material can lead to reduction in the mass of the structure which results in economical designs. However, in comparison to normal weight aggregate concrete (NWAC), lightweight aggregate concrete (LWAC) tends to be more brittle as it typically shears through the aggregates leading to instantaneous drop in peak load in both compression and tension tests. Hence, to address this brittleness for LWAC, modern hooked-end DRAMIX fibres with different geometry (hooks), dosages (Vf) and bond strengths (sb) are added to mixes with different strengths (fck). This paper focuses on both tensile properties using a direct pullout and indirect notched beam tests, and compressive properties (fck, E, l) using the conventional compression test for cylinders of plain and steel-fibre reinforced lightweight concrete (SFRLC). A tensile semiempirical multilinear r-x relation was derived besides compressive r-e and validated against available steel fibre reinforced concrete (SFRC) constitutive models for the tested fibrous notched beams using ABAQUS. Keywords: Recycled lightweight concrete
Hooked-end fibres
Tension
Compression
Ductility
Pullout
Notched beam
NLFEA
ABAQUS
r-x

1 Introduction The use of structural lightweight aggregate concrete brings several advantages as compared to the conventional normal weight concrete such as thermal insulation and fire resistance. Besides, the lightweight aggregate Lytag used in this work is recycled and offers reduction in CO2 emissions as well as being an alternative to the depleting gravel and quarry resources (Gerritse 1981). The coarse aggregate Lytag is made from fly ash which is a by-product of coal-fired power stations, by the process of palletisation. The UK generates about 10% of its electricity from coal. Structural Lytag has been around since the 1960s, available in the UK and Europe and can be used to produce concrete strengths of up to 60 N/mm2. Also, the improved strength-to-weight ratio of LWAC results in smaller cross sections which in turn leads to a decrease in area of reinforcement and savings in material transport costs due to lower inertial and gravity loads. The usage of lightweight concrete can therefore be ideal and competitive © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 920–937, 2021. https://doi.org/10.1007/978-3-030-58482-5_81

A Constitutive Model for Steel-Fibre-Reinforced Lightweight Concrete

921

in industry for the growing need for taller and longer span structures, especially in seismic and dynamic zones (Libre et al. 2011; Campione 2014; Dias-Da-Costa 2014; Mo et al. 2017). Structural applications of LWAC include bridges, towers and slabs, with notable modern structures such as the Acton Swing bridge, the Gherkin, the Shard and the 103 Colmore Row building. A study on Lytag in 2014 (Lytag 2014) showed that lightweight concrete can bring about 34% savings in CO2 as well as a reduction of up to 48% of concrete and reinforcement when compared to conventional gravel concrete allowing a reduction in foundation sizes and increase in spans and building space. These advantages however come as a trade-off for the increased brittleness of lightweight concrete. concrete usually being more porous and having a poor aggregate interlock mechanism as aggregates tend to be as weak as cement, which translates into lacking a natural toughening mechanism post-crack such as in the case for normal weight concrete. This causes lower tensile to compressive strength ratio which ultimately lowers shear resistance in structures such as beams and slabs and causes brittle failure. Also, the porous nature of lightweight concrete leads to a lower modulus of elasticity (Chen et al. 2010; Badogiannis and Kotsovos 2013) which causes excessive deflections and cracking (Lim et al. 2006; Wu et al. 2011). In addition, another disadvantage highlighted by Lim et al. (2006) is that lightweight aggregate concrete usage is reduced when compared to normal weight concrete due to a combination of lack of confidence and guidance for designers (although some exist, they are adapted from research in the past century on normal weight concrete), and especially the lack of understanding regarding how this increasingly brittle material behaves on the mesoand macro scales. The brittle nature of lightweight concrete can be addressed by incorporating traditional reinforcement. However, the latter solution can become inherently counterproductive and impractical when reduction in structural elements is sought by employing lightweight concrete especially at critical zones such as joints. Therefore, fibre reinforcement which has long proven its effectiveness in controlling and bridging tensile and shear cracks in the past for both lightweight and normal weight concretes can become an adequate solution (Gao et al. 1997; Campione and La Mendola 2004; Abbas et al. 2014a; Di Prisco et al. 2013; Grabois et al. 2016; Mo et al. 2017). For over 40 years, the usage of steel fibres in concrete mixes has been experimented with and used (Ritchie and Kayali 1975), however, comprehensive studies on fibrous lightweight concrete (especially on the material level which is key in understanding concrete behaviour) are still scarce with most work being merely theoretically linked to SFRC, carried on the structural level only and involve other lightweight aggregates types uncommon in the UK such as pumice stone and oil palm aggregates (Swamy et al. 1993; Kang and Kim 2010; Di Prisco et al. 2013; Iqbal et al. 2015; Grabois et al. 2016; Mo et al. 2017). It should be noted that, at present, there is no international standards specific for steel fibre reinforced lightweight concrete (SFRLC) with current guidelines being usually adapted from fibrous normal weight concrete (SFRC). Hence, it is of environmental, economic and engineering design benefit and need to study the behaviour of steel fibre reinforced recycled lightweight aggregate concrete within an adaptable methodology to different types of fibres and aggregates and derive a generic material and structural relationships capable of guiding designers to become more confident in using SFRLC for different structural elements, thus reducing its underutilization and paving the path for future researchers. This work

922

H. K. Al-Naimi and A. A. Abbas

suggests a methodology to test and derive the complete material constitutive relation in compression (r-e) including Young’s modulus of Elasticity (E) and Poisson’s ratio (l), and in tension (r-x) for SFRLC. The compression properties are based on the conventional uniaxial compression test while the tension properties are derived based on the pullout tests. The parameters used in this work include the type of fibres (3D and 5D), fibre dosage (Vf), interfacial bond strengths (sb) and concrete strengths (flcm). Fibre embedded length (LE), and fibres with hooks being cut off are also parameters in pullout tests. Nonlinear finite element analysis using ABAQUS (Habbit et al., 2000) is employed and discussed to validate notched beams tested according to RILEM TC 162-TDF (2002). The proposed constitutive model derived is also compared to material models based on RILEM TC 162-TDF (2003), Barros et al. (2005) and fib Model Code 2010 (FIB 2013).

2 Methodology The methodology of the work done consists of two main parts. The first involves experimental testing which includes the study of hardened properties of plain Lytag and fibrous concrete. The focus on this paper will be on compressive tests of cylinders, and tensile tests on pullout prisms and notched beams. The second part covers nonlinear numerical testing and finite element analyses of notched beams using ABAQUS which includes validating the material models derived and comparison with available constitutive SFRC models. Studying the flexural and shear behaviour of reinforced concrete beams is shown in an accompanying paper.

3 Experimental Study 3.1

Experimental Programme

As previously noted, the experimental programme includes a variety of specimens. For uniaxial compression tests, cylinders are tested to study the effect of fibres on short columns. For tensile tests, direct pullout prisms are tested to evaluate the direct tensile behaviour of plain and fibrous lightweight concrete while indirect tensile tests on notched beams are carried out to study the material flexural and tensile behaviour of SFRLC. For all the specimens tested, fck = 30 MPa being the minimum strength for structural concrete and hooked-end 3D fibres being the most commonly used in industry (Abdallah et al. 2018a), are considered to be the control parameters. Other specimens include different compressive strengths: fck = 35 MPa and fck = 40 MPa, fibre type 5D and fibre dosages of 0%, 1% and 2%. The choice to skip lower dosages such as Vf = 0.5% is so since most common structures such as tunnels and pavements require at least a dosage of Vf = 1% for efficient crack control and thickness reduction (The concrete Society 2007).

A Constitutive Model for Steel-Fibre-Reinforced Lightweight Concrete

923

3.1.1 Materials Portland-Limestone cement (CEM 11) according to the specification supplied in EN 197-1 was used. Coarse aggregate Lytag, also known as Sintered Pulverised Fuel Ash Lightweight Aggregate (LYTAG) was provided by LYTAG Ltd. The loose dry density of LYTAG was calculated in the lab to be approximately 760 kg/m3 while the water absorption was estimated to be around 15% per mass of LYTAG. Natural river sand with a 4.75 mm maximum size was used as the fine aggregate of the concrete. The sand had a water absorption coefficient of 0.09% and specific gravity of 2.65 complying with BS EN 12620. The properties of the fibres used are summarized in the table below. It should be noted that to prevent the possibility of balling, fibres were collated from the manufacturer. Table 1. Properties of fibres Fibre Type 3D 65/60 4D 65/60 5D 65/60

ru (MPa) 1160 1500 2300

lf (mm) 60 60 60

df (mm) 0.9 0.9 0.9

3.1.2 Mix Design The mix design used is summarized in Table 1 below. The mix design of the Lytag concrete for the characteristic cylinder and cube compressive strengths used are summarised below. These were directly adopted from Lytag (2011) manuals (Table 2). Table 2. Mix design used (fck/fck, cube) LC30/33 LC35/38 LC40/44

Cement (kg/m3) 370 420 480

Sand (kg/m3) 592 546 485

Loose bulk Lytag (kg/m3) 668.8 668.8 668.8

Effective water (kg/m3) 175 175 175

Calculating the water content of Lytag was of high importance as Lytag aggregates were found to absorb water of approximately 15% of their weight which is also confirmed by Lytag manual. For this reason, Lytag aggregates were added 24 h after mixing to obtain a saturated surface dry (SSD) state. The mixing process is shown in Fig. 1 below.

924

H. K. Al-Naimi and A. A. Abbas

Fig. 1. Mixing process for plain and fibrous lightweight concrete

3.2

Experimental Tests

As mentioned previously, the experiments include a uniaxial tensile pullout, compression cylinder and flexural notched beam tests. A direct uniaxial tensile fibre pullout test shown in Fig. 2 was designed and used to investigate the influence of the embedded fibre onto the tensile behaviour of the concrete. Unlike Robins et al. (2002) pullout test, this test was designed with the fibre being completely embedded in Lytag concrete mimicking to a large extent the real behavior of concrete at the crack in a structure. To ensure breakage of the specimen in tension and to introduce the embedded length LE as a variable at the monitored middle section, this section of the concrete was reduced to a diameter corresponding to the area a single DRAMIX fibre of diameter 0.9 mm is predicted to occupy in the concrete (the diameter = 12 mm). This area was determined based on numerical and statistical models (Krenchel et al. 1975; Romualdi and Mendel 1964; Soroushian and Lee 1990). Hence, a single fibre pullout test is equivalent to Vf = 1% while embedding 2 fibres is equivalent to Vf = 2%. The fibre was placed at the notch of the pullout mould for the chosen LE using fixed suspended cables mechanically attached to the fibre during casting. These were removed once concrete was cast to avoid any interference with the results. Thus, this test can be regarded as a truer and more realistic representation of a crack being bridged by a fibre than the classical pullout test with the fibre embedded on one side while the other end of the fibre is clamped and pulled by a tensile machine (Abdallah et al. 2018a). While one end of the tensile machine is fixed, the other was gradually pulled in tension at a displacement-controlled loading of 1 mm/min (Fig. 2). The tensile machine was calibrated and fitted with a sensitive displacement transducer capable of accurately measuring the slip once crack was initiated. Both concrete blocks embedding the two hooks of the fibre were assumed to be rigid. This assumption was proven to be correct. The pullout specimen was designed to only deform between the two gripped carbon steel bars as shown in Fig. 3.

Fig. 2. A pullout specimen during the uniaxial tensile test

A Constitutive Model for Steel-Fibre-Reinforced Lightweight Concrete

925

Fig. 3. Pullout mould and dimensions of the specimen

The dimensions of the beam (150 mm, 150 mm, 500 mm) and the notch were identical to RILEM TC 162 TDF (2002). The LVDTs were glued using high strength epoxy after the concrete surface in touch with the LVDT was ground in the mid-section at the front of the beam to enable the LVDTs to fully adhere onto the concrete. The LVDT’s were connected to a computer software. For the purpose of estimating the vertical deflection accurately, a steel bar inspired by a technique similar to JSCE-SF4 recommendations was made. The beams were placed onto two frictionless steel supports. This was deemed adequate as the loading was symmetrical (Fig. 4). A classic three-point displacement-controlled loading of 0.2 mm/min was adopted using the hydraulic machine which had a load capacity of 500kN. It should be noted that the notch was introduced using an incompressible plastic material with width 4 mm (depth of the notch), length 150 mm and height 25 mm. The plastic material was pre-coated with a spray to prevent it from binding to concrete in order to remove it before testing.

Fig. 4. Notched beam during testing

Cylinders with 200 mm depth and 100 mm diameter are tested in compression to generate the complete compressive stress-strain behaviour for lightweight plain and fibrous concrete. A calibrated compressometer-extensometer steel ring designed according to ASTM C 469 and fitted with LVDT’s is clamped onto the concrete cylinders as shown in Fig. 5. A loading rate of 1.2 mm/min was deemed adequate to study the post-peak behaviour of fibrous concrete in compression.

926

H. K. Al-Naimi and A. A. Abbas

Fig. 5. Uniaxial compression cylinder test

4 Numerical Study Three-dimensional nonlinear finite element analyses will be carried out using ABAQUS (Zienkiewicz and Taylor 2005). This software has shown to be successful at predicting the behaviour of SFRC in tension, compression, flexure and shear as well as the cracking pattern and mode of failure of plain and reinforced elements to a good level of accuracy (Tlemat et al. 2006; Syed Mohsin 2012; Abbas et al. 2014a, 2014b; Behinaein et al. 2018). The approach adopted in this work will involve modelling both plain and fibrous concrete. The fibres will not be modelled discretely, instead they will be introduced directly into the constitutive tensile and compressive models of the concrete in ABAQUS. This methodology was opted for since modelling fibres discretely can be time consuming and will produce a nonflexible FE-based model difficult to be adopted by designer engineers. Moreover, although modelling fibres discretely using probabilistic techniques such as Monte Carlo is aimed to account for the random distribution of fibres (Cunha et al. 2010), its usage can itself be unrealistic as the prediction might as well be likely to completely miss the actual distribution and location of fibres in a particular structural element. Hence, modelling fibres as part of the concrete matrix as shown in a number of design guidelines can offer an easier and perhaps safer prediction of the behaviour of composite material (Lok and Xiao 1999; RILEM TC 162-TDF 2002; Barros et al. (2005); Di Prisco et al. 2013). However, unlike the discrete 3D modelling, it should be noted that the homogenous fibrous concrete modelling is not aimed to detect local failures on the mesoscale level explicitly such as fibre rupture and concrete fracturing at the IFZ (Zhang et al. 2018). Generally, there are two approaches that can be used in FEM to predict the tension stiffening behaviour of fibrous concrete: the discrete crack approach (r-x) (Ngo and Scordelis 1967) or smeared crack approach (r-e) (Rashid 1968). Although more accurate at post-crack, the discrete crack approach can be impractical and numerically intensive to use (Tlemat et al. 2006), while the more accepted smeared crack approach that assumes the crack is smeared over an element can be more useful for design. In this work, a smeared crack approach with (r-x) in tension is adopted using ABAQUS option to define cracking displacement rather than cracking strain. ABAQUS derives the strains based on the characteristic length lc which is defined as the mesh size for hex

A Constitutive Model for Steel-Fibre-Reinforced Lightweight Concrete

927

elements using e = x/lc. The values of crack width were input into ABAQUS directly from the pullout tests using the constitutive model depicted in Fig. 9. As compared to the strain, the r-x approach offers a number of advantages since it represents the actual behaviour of the fibrous material, is member size-effect independent and can be directly applied to FEM (ABAQUS) from the pullout tests used (De Montaignac et al. 2011). In ABAQUS, mesh sensitivity is not an issue for r-x relation as compared to r-e. Also, r-x relationship can provide the necessary information needed to design for service limit state including fatigue and shrinkage. In compression, the direct results from the compression r-e tests were adopted using Eq. 1, 2 and 3. Concrete damaged plasticity (CDP) was calibrated and chosen to model the notched beam specimens to check the validity of the r-x model. The calibration of CDP on the material level for both cylinder compression test and pullout test can be found with more detail elsewhere (Al-Naimi and Abbas 2019). Table 3 below summarises the CDP parameters used.

Table 3. Parameters usage for CDP Dilation angle Eccentricity fb0/fc0 K Viscosity 25 0.1 1.16 0.666 0

The explicit solver is adopted. The explicit dynamic analysis was found to be a more computationally efficient tool at solving the problems used in this work as compared with the implicit solver which had a tendency to terminate. To ensure a quasi-static solution the ratio between kinetic energy and internal energy was kept below 1% throughout the analysis. The analysis was ran using a displacement loading rate of 0.1 mm/step in a smoothed step with a mass scaling of 50. The finite element model investigated to validate the notched beams is shown below. Due to the symmetrical setting only half the beam was modelled with a symmetry boundary condition along the Z direction (Fig. 6). The beam was restricted from moving in the Y direction by applying a displacement rotation boundary condition (in the initial step) along the middle line of both supports. The displacement-induced load was applied (in the explicit dynamic step) on the surface of the loading steel plate in a similar manner to the experiment. To estimate the load, the reaction force was calculated by summing up the load along the boundary condition line on the supports then multiplied by 2. The displacement, however, was calculated by taking the average Y displacement of the surface or line of the structure in middle point of the span of the beam. It should be noted that hexagonal brick element C3D8R with reduced integration and hourglass control was used with a size of 20 mm with the exception of notch where the size was 4 mm by 20 mm. The supports and load blocks were made to be rigid and a tie coupling to the concrete was applied. The meshing, element size, solver, loading rate and mass scaling adopted were based on a comprehensive convergence study.

928

H. K. Al-Naimi and A. A. Abbas

Fig. 6. Finite element model used for the notched beam and boundary conditions (ABAQUS)

5 Results and Discussion 5.1

Experimental Results

5.1.1 Compression The average compressive stress-strain relationship of cylinders from 6 mixes is shown in Fig. 7 below.

Fig. 7. Average compressive stress-strain of cylinders from 6 mixes

The figure shows that once the plain cylinders i.e. Vf = 0% with different strengths reach their compressive peak strength, an instantaneous shear failure takes place. Following visual inspection of the cylinders, it was evident that the Lytag aggregates were sheared through which explains the lack of the strain toughening aggregate interlock mechanism in the lightweight specimens and the occurrence of the brittle failure. When 3D fibre dosages of 1% and 2% are added to the 30 MPa mix, the compressive strength remains unaffected, however, it is clear that the failure becomes ductile with the specimens having Vf = 2% developing a higher compressive ductility due to the increased lateral confinement. Nonetheless, the specimens reinforced with Vf = 1% of 5D fibres develop approximately 10% higher peak compressive strength than that of Vf = 0% although with decreased ductility as compared with the fibrous specimens having lower compressive strength and similar fibre volume fraction. This can be explained by the stronger mix naturally being more brittle. It can be concluded

A Constitutive Model for Steel-Fibre-Reinforced Lightweight Concrete

929

therefore that the more extensive hooks of the 5D fibres render it more difficult for the fibres to pull out as compared to the 3D fibres causing strength increase prior to peak. Table 4 below summarises the results of the compression test. It should be noted that the modulus of elasticity ranged between 19.6 to 22.9GPa depending on the concrete strength with fibres having generally lower values due to creating air voids in the lattice. Tests of lightweight concrete in compression with similar fibres agree with these findings (Li et al. 2018). Poisson’s ratio ranged between 0.16 to 0.19 (which agrees with Lambert (1982)) with no pattern relating to neither flcm, Vf or fibre type. Also, the strain for fibrous lightweight concrete at 85% peak varies with the lower strength specimens developing larger strains. For design calculations, the strain at 85% should be taken as shown in Eq. (3). Table 4. Summary of findings from the compression tests Vf (%) flcm (MPa) E (GPa) l 0 1 2 0 0 1

31.84 32.20 32.9 35.8 41.75 45.41

20.7 19.6 19.9 21.0 22.9 21.3

0.16 0.16 0.16 0.16 0.17 0.18

Normal Strain ‰ eel epeak eult 0.51 2.13 2.13 0.49 2.33 4.43 0.64 3.36 12.1 0.67 2.26 2.26 0.72 2.53 2.53 0.85 2.61 3.85

Ductility (@85% fcm) 0 0.9 2.7 0 0 0.48

Based on a more comprehensive study of the compression behaviour of over 100 plain and fibrous lightweight concrete cylinders, the following equations of static modulus of elasticity and strain at peak were derived: 0:43 Estatic ¼ 4:54flcm ðGPaÞ with R2 ¼ 0:92

ð1Þ

ð2Þ

ð3Þ Using all the above it can be deduced that the constitutive r-e model in compression for SFRLC is identical to that of plain concrete with the exception of eult Eq. (3). 5.1.2 Tension Figure 8 below illustrates the pullout load-slip of a few plain and fibrous concrete specimens with different parameters.

930

H. K. Al-Naimi and A. A. Abbas Pullout Load-Slip

700

fck=30MPa, Vf=0% fck=30MPa, Vf=1%, 3D, LE=24mm fck=30MPa, Vf=1%, 3D, LE=19mm fck=30MPa, Vf=2%, 3D fck=30MPa, Vf=1%, 5D, LE=26mm fck=30MPa, Vf=1%, 5D, LE=12mm fck=40MPa, Vf=1%, 5D, LE=14mm fck=30MPa, Vf=1%, 3D hooks cut off

Pullout Load (N)

600 500 400 300 200 100 0 0

5

10

Slip (mm) 15

20

25

Fig. 8. Pullout load-slip of some of the specimens tested

It can be seen that, once plain concrete reaches its maximum strength, it fails abruptly in a brittle manner. As was mentioned previously, this behaviour illustrates the absence of any tension stiffening mechanism in the form of the aggregate interlock present in normal weight concrete. For all fibrous specimens, the behaviour becomes ductile with specimens with higher embedment length LE developing a more increased ductility while the peak strength remains unaffected (solid and dashed green lines). The latter remains correct provided that the embedded length is large enough to fully bond the hook end length LH, otherwise this can give rise to premature concrete fracture near the hook preventing the fibre from developing full strength (dashed blue line). For the hook length to be fully bonded to develop the maximum pullout strength Eq. (1) must be satisfied. LE [ LH þ 5df

ð4Þ

with df as the fibre diameter in (mm). This equation also agrees with similar findings in Abdullah et al. (2018) on DRAMIX hooked-end fibres. Moreover, for the 3D fibres with hooks being cut-off (dotted black line), it can be seen that the mechanical hook contribution directly contributes to the peak strength while the remaining length of the fibre only provides frictional pullout for lightweight concrete. Depending on the fibre type, the pullout strength varies with 5D resulting in the highest post-peak pullout strength and ductility (so long as LE is sufficient and identical). The reason behind the differences in strength is stemmed from the fact that the more extensive mechanical hooks (for 5D) are able to develop a better bond with the concrete. The ductility differed due to the longer extensive hooks requiring more work to straighten and eventually pull out. Increasing Vf from 1% to 2% leads to generating nearly double the pullout strength at Vf = 1%. Reducing the W/C ratio i.e. increasing the concrete compressive strength which means having less air voids in the concrete lattice, leads to a higher pullout strength due to developing a better fibre concrete bond (dashed and long dashed blue lines). Since the

A Constitutive Model for Steel-Fibre-Reinforced Lightweight Concrete

931

contribution of the concrete is negligible post-crack due to the absence of natural tension stiffening mechanism, the max residual stress of SFRLC due to fibre reinforcement can be written as follows. rSFRLC ¼ rtr1 ¼ rf Vf g0 ¼

Pmax Ac

ð5Þ

with Ac (mm2) as area of concrete that pertains to the fibre dosage (Soroushian and Lee, 1990), rf (MPa) fibre stress from the test and Pmax (N) is the maximum pullout. The orientation factor η0 to account for the randomness of fibres in concrete is defined differently in literature, and is taken as 0.5 (Hannant 1978) if no size-effects are involved, or determined using Lee et al. (2011) chart when size-effects are influencial. Another alternative to calculate the residual strength is by using bond strength sb from available pullout tests: rtr1 ¼ 4

sb L E with df as the fibre diameter in ðmmÞ and sb as the bond strength ðMPaÞ df ð6Þ

with rtp (MPa) as the tensile plain concrete strength. Using the tensile tests on plain concrete of LC30, 35 and 40, the plain concrete strength can be calculated using Eq. (7). rtp ¼ 0:065flcm

ð7Þ

Based on the above, a semi-empirical r-x is derived in the following figure:

Fig. 9. Constitutive model in tension

etp is calculated by dividing rtp by Young’s modulus of elasticity assuming Ec = Et. The crack width xtr1 is the slip needed for the fibre to develop maximum tensile stress and is calculated using Eq. (8) adapted from Lok and Xiao (1999), with Ef as the modulus of Elasticity of fibre (GPa) and LT as the characteristic length that converts x to e and is an indication of crack length. LT varies depending on specimen size, Vf, fibre type, reinforcement, loading level and matrix strength. No agreement has yet been

932

H. K. Al-Naimi and A. A. Abbas

achieved to determine LT, however for beams, this paper suggests LT as the minimum of srm (mean distance between cracks or hsp/2 for all phases of SFRLC deformation with hsp as the unnotched depth (Barros et al. 2005; Tlemat et al. 2006; De Montaignac et al. 2011; FIB 2013). xtr1 ¼ etr1  LT ¼ LT

rtr1 E f g0 V f

ð8Þ

Also, rtr2 ¼ 0:6rtr1 at xtr2 ¼ L

ð9Þ

rtr3 ¼ 0:3rtr1 at xtr3 ¼ LH

ð10Þ

rtu ¼ 0:1rtr1 at xtu ¼ LH þ 5df

ð11Þ

It should be noted that, the choice of the residual crack widths also agrees with the pulley pullout model of Alwan et al. (1999) and its recent revision by Abdallah et al. (2018b) on hooked-end fibres and is based on the geometry of the mechanical hooks, while the choice of rtr2 and rtr3 is empirical based on the lower bound results from the pullout tests. A similar approach can be used to derive constitutive models for any particular concrete, fibre type and geometry. Figure 10 below compares the constitutive models based on the notched beam tests up to a x of 2.5 mm (ULS) according to fib Model Code 2010 (FIB 2013). On ABAQUS, in order to avoid mesh sensitivity for unreinforced concrete, r-x method was used. To obtain r-x relation from r-e relation for these constitutive models, the characteristic length LT was chosen as hsp/2 (Barros et al. 2005). For Model Code 2010, hsp = 125 mm was recommended (Blanco et al. 2013).

Fig. 10. Comparison between the available constitutive models

Figure 11 below shows the Load-CMOD and Load-Deflection from the notched beam test. As expected, the plain concrete notched beam failed once the flexural load reached its peak about 7.5 MPa. Both notched beams reinforced with Vf = 1% and Vf = 2% exhibited increased load capacity and ductility. The first developed a flexural load capacity almost 3 times higher than that of plain concrete while the second reached

A Constitutive Model for Steel-Fibre-Reinforced Lightweight Concrete

933

a load capacity about 4 times higher. The fracture energy Gf was calculated using the area under graph. Gf for fibrous specimen of Vf = 2% was 65% higher than that of Vf = 1%.

Fig. 11. Notched beams behaviour

5.2

Numerical Results

Figure 12 shows half the notched beam simulated using the r-x constitutive model on ABAQUS at displacement of 4 mm. It is clear that the failure pattern was concentrated at the notch as expected. The simulation remained quasi-static throughout the analysis.

Fig. 12. Notched beam test at displacement 3 mm from ABAQUS

Figure 13 below illustrates the 4 constitutive models discussed in Fig. 10. It can be seen that the SFRLC proposed constitutive model used predicted the notched beam behaviour to a good accuracy from both a design and analysis point of view for Vf = 1% and 2% while other SFRC models overestimated the strength and underestimated the ductility of fibrous lightweight concrete.

934

H. K. Al-Naimi and A. A. Abbas

Fig. 13. Load-deflection predictions of notched beams

6 Conclusions • The plain concrete fails in compression once it reaches its peak strength due to the lack of aggregate interlock mechanism. • The peak compressive strength is unaffected by the addition of conventional 3D fibres. The more extensive hooks in 5D fibres however enhance flcm slightly. • The addition of fibres leads to increased confinement and compressive ductility leading the concrete failing in a ductile manner. • The r-e behaviour of SFRLC in compression remains identical to that of plain concrete with the exception of enhanced ductility. • The pullout test designed was effective at predicting the post-peak behaviour of the fibrous lightweight concrete. • Due to the absence of natural tension stiffening, plain concrete fails in a brittle manner once it reaches its peak tensile strength. • An increase in Vf, number of hooks and reduction in W/C leads to an increase in both strength and fracture energy. • Only the hook length influences the peak strength for fibrous lightweight concrete while the remaining embedded length of the fibre increases ductility due to frictional pullout. • A unique semi-empirical r-x constitutive model was derived and compared to other available SFRC models which appeared to differ in peak load and ductility. By applying FEA using ABAQUS’ concrete damaged plasticity model in a stresscracking displacement method which offers good mesh insensitivity, it was revealed that the proposed constitutive SFRLC model is successful at predicting the experimental load-deflection behaviour of the notched beams of different Vf to a good accuracy and is hence deemed superior than the SFRC based models used in this work.

A Constitutive Model for Steel-Fibre-Reinforced Lightweight Concrete

935

References Gerritse, A.: Design considerations for reinforced lightweight concrete. Int. J. Cem. Compos. Lightweight Concr. 3(1), 57–69 (1981) Libre, N., Shekarchi, M., Mahoutian, M., Soroushian, P.: Mechanical properties of hybrid fiber reinforced lightweight aggregate concrete made with natural pumice. Constr. Build. Mater. 25 (5), 2458–2464 (2011) Campione, G.: Flexural and shear resistance of steel fiber-reinforced lightweight concrete beams. J. Struct. Eng. 140(4), 04013103 (2014) Dias-da-Costa, D., Carmo, R., Graça-e-Costa, R., Valença, J., Alfaiate, J.: Longitudinal reinforcement ratio in lightweight aggregate concrete beams. Eng. Struct. 81, 219–229 (2014) Mo, K., Goh, S., Alengaram, U., Visintin, P., Jumaat, M.: Mechanical, toughness, bond and durability-related properties of lightweight concrete reinforced with steel fibres. Mater. Struct. 50(1), 46 (2017) Lytag: Ramboll Frame comparison study (2014). https://www.aggregate.com/our-businesses/ lytag. Accessed 31 Dec 2019 Chen, H., Huang, C., Tang, C.: Dynamic properties of lightweight concrete beams made by sedimentary lightweight aggregate. J. Mater. Civ. Eng. 22(6), 599–606 (2010) Badogiannis, E., Kotsovos, M.: Monotonic and cyclic flexural tests on lightweight aggregate concrete beams. Earthquakes Struct. 6(3), 317–334 (2014) Lim, H.S., Wee, T.H., Mansur, M.A., Kong, K.H.: Flexural behavior of reinforced lightweight aggregate concrete beams. In: Asia-Pacific Structural Engineering and Construction Conference, pp. 68–82. APSEC, Kuala Lumpur (2006) Wu, C., Kan, Y., Huang, C., Yen, T., Chen, L.: Flexural behavior and size effect of full scale reinforced lightweight concrete beam. J. Mar. Sci. Technol. 19(2), 132–140 (2011) Gao, J., Sun, W., Morino, K.: Mechanical properties of steel fiber-reinforced, high-strength, lightweight concrete. Cem. Concr. Compos. 19(4), 307–313 (1997) Campione, G., La Mendola, L.: Behavior in compression of lightweight fiber reinforced concrete confined with transverse steel reinforcement. Cem. Concr. Compos. 26(6), 645–656 (2004) Abbas, A., Syed Mohsin, S., Cotsovos, D.: Seismic response of steel fibre reinforced concrete beam–column joints. Eng. Struct. 59, 261–283 (2014a) Di Prisco, M., Colombo, M., Dozio, D.: Fibre-reinforced concrete in fib model code 2010: principles, models and test validation. Struct. Concr. 14(4), 342–361 (2013) Grabois, T., Cordeiro, G., Filho, R.: Fresh and hardened-state properties of self-compacting lightweight concrete reinforced with steel fibers. Constr. Build. Mater. 104, 284–292 (2016) Ritchie, A., Kayali, O.: The effects of fiber reinforcement on lightweight aggregate concrete. In: Neville A (ed.) Proceedings of RILEM Symposium on Fiber Reinforced Cement and Concrete, pp. 247–256. The Construction Press Ltd. (1975) Swamy, N., Jones, R., Chiam, A.: Influence of steel fibers on the shear resistance of lightweight concrete i-beams. ACI Struct. J. 90(1), 103–114 (1993) Kang, T., Kim, W.: Shear strength of steel fiber-reinforced lightweight concrete beams. Korea Concrete Institute, Oklahoma, pp. 1386–1392 (2010) Iqbal, S., Ali, A., Holschemacher, K., Bier, T.: Mechanical properties of steel fiber reinforced high strength lightweight self-compacting concrete (SHLSCC). Constr. Build. Mater. 98, 325–333 (2015) RILEM TC 162-TDF: Bending test: final recommendation. Mater. Struct. 35, 579–582 (2002) RILEM TC 162-TDF: r-e design method: Final Recommendation. Mater. Struct. 36, 560–567 (2003)

936

H. K. Al-Naimi and A. A. Abbas

Barros, J., Cunha, V., Ribeiro, A., Antunes, J.: Post-cracking behaviour of steel fibre reinforced concrete. Mater. Struct. 38(275), 47–56 (2005) Fédération Internationale du béton: fib model code for concrete structures 2010, pp. 147–150. Ernst and Sohn, Berlin (2013) Sadoon, A., Rees, D.W.A., Ghaffar, S.H., Fan, M.: Understanding the effects of hooked-end steel fibre geometry on the uniaxial tensile behaviour of self-compacting concrete. Constr. Build. Mater. 178, 484–494 (2018a) The Concrete Society: Guidance for the design of steel-fibre-reinforced concrete. Technical report No. 63. Cement and Concrete Industry (2007) Lyag: Technical Manual. Lytag ltd, London (2011) Robins, P., Austin, S., Jones, P.: Pull-out behaviour of hooked steel fibres. Mater. Struct. RILEM 35, 4343–4442 (2002) Krenchel, H.: Fiber spacing and specific fiber surface. In: Neville, A.M. (ed.) RILEM Symposium on Fiber Reinforced Cement and Concrete, pp. 69–79. The Construction Press, London (1975) Romualdi, J.P., Mandel, J.A.: Tensile strength of concrete affected by uniformly distributed and closely spaced short length of wire reinforcement. In: AC1 Journal Proceedings vol. 61, no. 6, pp. 657–671, June 1964 Soroushian and Lee: Tensile strength of steel fiber reinforced concrete: correlation with some measures of fiber spacing. ACI Mater. J. 87(6), 542–546 (1990) Zienkiewicz, O.C., Taylor, R.L.: The Finite Element Method for Solid and Structural Mechanics, vol. 2, 6th edn. Butterworth-Heinemann, Oxford (2005) Tlemat, H., Pilakoutas, K., Neocleous, K.: Stress-strain characteristic of SFRC using recycled fibres. Mater. Struct. 39, 365–377 (2006) Syed Mohsin, S.M.: Behaviour of fibre-reinforced concrete structures under seismic loading. Ph. D. thesis, Imperial College London, London, UK (2012) Abbas, A., Syed Mohsin, S., Cotsovos, D., Ruiz-Teran, A.: Shear behaviour of steel-fibrereinforced concrete simply supported beams. Proc. Inst. Civ. Eng. Struct. Build. 167(9), 544– 558 (2014b) Cotsovos, B.D.M., Abbas, A.A.: Behaviour of steel-fibre-reinforced concrete beams under highrate loading. Comput. Concr. 22(3), 337–353 (2018) Cunha, V., Barros, J., Sena-Cruz, J.: Tensile behavior of steel fiber-reinforced self-compacting concrete. In: Fiber-Reinforced Self Consolidating Concrete: Research and Applications (ACI SP-274), pp. 51–68. American Concrete Institute, Detroit (2010) Lok, T.-S., Xiao, J.R.: Flexural strength assessment of steel fiber reinforced concrete. J. Mater. Civ. Eng. 11(3), 188–196 (1999) Zhang, Y.J., Huang, Z.J., Yanga, S.L., Xua, X.W.C.: A discrete-continuum coupled finite element modelling approach for fibre reinforced concrete. Cem. Concr. Res. 106(2018), 130– 143 (2018) Ngo, D., Scordelis, A.C.: Finite element analysis of reinforced concrete beams. J. ACI 64(3), 152–163 (1967) Rashid, Y.R.: Ultimate strength analysis of prestressed concrete pressure vessels. Nucl. Eng. Des. 7(4), 334–344 (1968) De Montaignac, R., Massicotte, B., Charron, J.-P., Nour, A.: Design of SFRC structural elements: post-cracking tensile strength measurement. Mater. Struct. 45(4), 609–622 (2012) Al-Naimi, H., Abbas, A.: Ductility of steel-fibre-reinforced lightweight concrete, Eccomas Proceedia. In: COMPDYN, pp. 4009–4023, Crete, Greece, 24-26 June 2019 (2019). viewed 31 Dec 2019. www.eccomasproceedia.org Lambert, G.: Properties and behaviour of structural lightweight (Lytag-sand) concrete. PhD thesis, University of Sheffield (1982)

A Constitutive Model for Steel-Fibre-Reinforced Lightweight Concrete

937

Hannant, D.J.: Fibre Cements and Fibre Concretes. Wiley, Hoboken (1978) Lee, S.-C., Cho, J.-Y., Vecchio, F.J.: Diverse Embedment Model for Steel Fiber-Reinforced Concrete in Tension: Model Verification. ACI Mater. J. 108(5), 526–535 (2011) Alwan, J., Naaman, A., Guerrero, P.: Effect of mechanical clamping on the pull-out response of hooked steel fibers embedded in cementitious matrices. Concr. Sci. Eng. 1, 15–25 (1999) Sadoon, A., Rees, D.W.A., Ghaffar, S.H., Fan, M.: Predicting pull-out behaviour of 4D/5D hooked end fibres embedded in normal-high strength concrete. Eng. Struct. 172, 967–980 (2018b) Blanco, A., Pujadas, P., de la Fuente, A., Cavalaro, S., Aguado, A.: Application of constitutive models in European codes to RC–FRC. Constr. Build. Mater. 40, 246–259 (2013) Li, F.-Y., Cao, C.-Y., Cui, Y.-X., Wu, P.-F.: Experimental study of the basic mechanical properties of directionally distributed steel fibre-reinforced concrete. Adv. Mater. Sci. Eng. 3, 1–11 (2018)

UV-C Treatment to Functionalize the Surfaces of Pet and PP Fibers for Use in Cementitious Composites. Adherence Evaluation María E. Fernández(&), María E. Pereira, Fernando Petrone, Claudia Chocca, and Gemma Rodríguez Faculty of Architecture, Design and Urbanism, University of the Republic (Udelar), Montevideo, Oriental Republic of Uruguay, Uruguay [email protected]

Abstract. The objective of this work is to present an evaluation of the adhesion of synthetic fibers with cementitious matrices when the surface is functionalized by exposure to UV-C light. The synthetic fibers used were obtained from post-consumer containers of polyethylene terephthalate (PET). Their performances were compared with commercial macro-fibers of polypropylene (PP). Tensile strength and adhesion in two matrices—one of Portland cement and one with a partial replacement of Portland cement by ceramic waste—were evaluated at ages of between 7 days and 6 months. The obtained values were compared with equal samples made with fibers whose surfaces were not exposed to this radiation. The results show that surface functionalization by this procedure does not produce the expected effects in terms of adhesion with either of the two matrices used. In commercial polypropylene fibers, functionalization is detrimental to its tensile strength by decreasing it in the environment from 70% to 80%; this loss of resistance of the polymer causes the start-up test to be interrupted by the failure of the fiber. This does not occur in PET fibers; functionalization does not substantially affect their tensile strength. Keywords: Fibro-reinforced concrete functionalization


Synthetic fibers
UV-C

1 Introduction Waste management of plastics continues to be difficult because plastic production does not decrease at the same rate as its consumption. This situation generates the need to look for new ways to value plastics through recycling or reuse, and there are several investigations that seek solutions to the final destination of synthetic waste; many of these focus this valorization on their use as reinforcement of cementitious matrix materials [1–3], particularly fibers obtained from difficult-to-manage industrial and domestic waste products [4–8]. The adhesion between both materials is affected by the hydrophobic behavior of the polymer surface and there are possible techniques that can be used to improve this property [9–12]. One of the techniques investigated is the exposure of the surface of the © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 938–948, 2021. https://doi.org/10.1007/978-3-030-58482-5_82

UV-C Treatment to Functionalize the Surfaces

939

polymers to UV radiation [13, 14], particularly to the wavelength fraction between 100–280 nm (UV-C) [15, 16]. The use of polyethylene terephthalate (PET) residues in cementitious matrices has been shown to produce a superficial degradation of the polymeric material, which decreases when a mixed matrix with ceramic residue is used [17]. This hypothesis was used to investigate the performance of the polymer and the interface through tensile and adhesion tests using PET and PP fibers for both matrices. The results obtained at an early stage did not allow us to draw conclusive results regarding the loss of resistance of the polymers, although they demonstrated that the adhesion of PET fibers is much lower than that of commercial fibers (PP). In trying to improve that adhesion, the surface functionalization was raised through its exposure to UV-C radiation. Through the evaluation of the tensile performance of the polymeric materials exposed to both matrices, and of the fiber-matrix adhesion, this work shows the results obtained in ages up to 6 months, comparing them with those of the nonfunctionalized fibers.

2 Experimental Procedure For the preparation of the matrices, Portland CPN-40 cement was used. The mixed matrix was made with a 25% replacement of cement with ground ceramic waste whose pozzolanic capacity was previously evaluated. The chemical characteristics of the cement can be seen in Table 1 and the particle size distribution of both materials in Fig. 1. In both matrices the water/binder ratio was 0.5. For the pull-out samples for both matrices, a plastic mortar was prepared as indicated in the UNE-EN 196-1 standard [18]. Table 1. Chemical characteristics of the cement used % CaO 63.89 3.30 Fe2O3 SiO2 22.19 Al2O3 3.83 MgO 2.97 SO3 1.86 K2O 0.22 Na2O 0.05 Na2O eq 0.19

% Soluble residue 0.64 P.P.C. 1.44 Compounds C3S C2S C3A C4AF

55.70 21.60 4.56 10.04

The fibers used were obtained from post-consumer containers of polyethylene terephthalate (PET) and commercial macro-fibers of polypropylene (PP). Samples were made from both polymers which had surfaces both with and without functionalization. For the fibers with functionalized surfaces, an exposure of PET films and PP fibers to UV-C radiation was performed. The exposure to this radiation was

940

M. E. Fernández et al.

Fig. 1. Distribution of particle size of Portland cement and ceramic waste.

carried out in a box made for this purpose (Fig. 2) with two 18 W lamps emitting at a wavelength of 254 nm. Samples exposed between 1 to 7 days were obtained on each side, to analyze the variation of the surface tension of the PET by microscope (Fig. 3). By this method it was determined that at 7 days the surface showed a hydrophilic behavior, defining the age of exposure for the present work.

Fig. 2. Equipment developed for the exposure of polymers to UV-C radiation.

The conditioning of the PET fibers for the evaluation of tensile strength was carried out by cutting samples in the form of a halter, as indicated in Standard UNE-EN ISO 527-3 [19]. For the adhesion test (pull-out) the residue was cut using a document shredder, obtaining fibers 4 mm wide, 0.21 mm thick and 35 mm long. Commercial fibers were used with their original shapes and sections, 1.07 mm diameter, and their lengths were conditioned to 35 mm, equal to PET fibers, for pull-out tests. The samples used to evaluate the evolution of the tensile strength of the fibers were made in such a way that the degradation produced by the matrix was located in the central area of the fiber (Fig. 4). They were demolded at 24 h and cured and submerged in water at laboratory temperature (20 ± 1 °C) until the test date: 7, 28, and 56 days; 3 and 6 months. Prior to the test, the matrix was removed and subsequently washed with a 1 molar hydrochloric acid solution. Eight samples were tested for each age, 4 without

UV-C Treatment to Functionalize the Surfaces

941

Fig. 3. Variation of surface tension using optical microscope images.

functionalization (PET with cementitious matrix, CMPET; PET with mixed matrix, MMPET; PP with cementitious matrix, CMPP; PP with mixed matrix, MMPP) and 4 with functionalization (CMPET-UVC; MMPET-UVC; CMPP-UVC; MMPP-UVC). The test procedure used was that indicated in standards UNE-EN 14889-2 [20] and UNE-EN ISO 6892-1 [21].

Fig. 4. Fibers with matrix in central part for subsequent tensile testing. Left: PET. Right: PP.

For the evaluation of the adhesion by means of a double pull-out test, 25 mm  25 mm  100 mm specimens were prepared whose molds were specially designed for this purpose (Fig. 5). Screws were left in the samples for fastening to the test equipment. After 24 h of filling, they were demolded and kept wrapped in film paper at laboratory temperature (20 ± 1 °C) until the test dates mentioned above. There were 8 samples for each test age with the same denomination. The test equipment consisted of a universal ZPM tester equipped with a 1000 N load cell, 1  45° hardened pyramid jaws, and software that allows the speed to be regulated continuously and facilitates data acquisition. In both types of trial, the speed was 1 mm/min, the same as that used in previous studies in order to compare results.

942

M. E. Fernández et al.

Fig. 5. Molds used to make pull-out samples

3 Results Obtained and Discussion 3.1

Tensile Strength

As mentioned earlier, tensile strength was determined in fiber samples with and without surface functionalization and with exposure to two matrices in each case. The obtained values were compared with the control sample, without immersion in any matrix (PET and PP). Maximum tensile strength and elastic modulus were determined. Figures 6, 7, and 8 show the tensile strength values obtained.

Fig. 6. Tensile strength results of the fibers studied without functionalization

According to these results, it can be observed that the tensile stress behavior of these fibers exposed to both matrices until 6 months of age does not show great variation with respect to the control sample. Although values that differ with the standard samples were obtained, they do not present a clear pattern of behavior, with those values included in the dispersion ranges of the results. The tensile strengths of PET fibers in all cases were lower than those of PP, comprising approximately 50% of their value. Considering the two matrices studied with both polymers, no statistically significant differences were found between the results obtained, and the F values were much lower than the critical value for F (FPET = 0.0164; FPP = 0.079; FCRIT = 4.965) and the highest p-value than 0.05 for PET and PP (0.901 and 0.785, respectively).

UV-C Treatment to Functionalize the Surfaces

943

Fig. 7. Results of tensile strength of PET fibers with and without functionalization

Fig. 8. Results of tensile strength of PP fibers with and without functionalization

Analyzing the effect of surface functionalization, we can see that in PET fibers the exposure to UVC radiation generates a small increase in the tensile strength of the material before exposing it to the matrices. However, when the polymer is exposed to the cementitious and mixed matrix, this variation is no longer observed at all ages, tending to stabilize with age. Unlike PET fibers, PP fibers with a functionalized surface have tensile strengths with values of between 20% and 30% of those reached by PP fibers that have not been exposed to UV-C radiation. Regarding the standard fibers with and without surface functionalization, they also have a similar behavior, with no observable large differences in the ages analyzed. 3.2

Elasticity Module

From the results obtained in the tensile tests described above, the modulus of elasticity was determined for each situation. The results obtained can be seen in Figs. 9, 10, and 11. In PET fibers embedded in the Portland cement matrix it can be observed that the modulus of elasticity increases as age progresses, implying greater rigidity of the material. However, when they are embedded in the mixed matrix, they have a more variable behavior; their value stabilizes after 3 months of age. PP fibers embedded in Portland cement matrix have an elastic modulus similar to the standard sample; the one that is observed when the fibers are embedded in the mixed matrix is more variable.

944

M. E. Fernández et al.

Fig. 9. Results of the elastic modulus of the fibers studied without functionalization

Fig. 10. Results of the elastic modulus of PET fibers with and without functionalization

Fig. 11. Results of the elastic modulus of PP fibers with and without functionalization

If we analyze the effect of functionalization of the surfaces in the modulus of elasticity we can observe that in PET fibers, when they are embedded in the Portland cement matrix for 28 days, the exposure UV-C radiation produces a small increase and the fibers have a variable behavior from 56 days of age. When PET fibers with a functionalized surface are exposed to the mixed matrix, they have a variable behavior that does not allow determination of their pattern. In commercial fibers, surface functionalization does not produce variation in the elastic modulus of the standard samples prior to immersion in the different matrices. PP fibers embedded in the Portland cement matrix, up to 56 days of age, show that the functionalization of the fibers produces a small decrease in the modulus of elasticity. From 3 months onwards, the values obtained do not show differences between fibers with an un-functionalized surface and those that were exposed to UV-C radiation.

UV-C Treatment to Functionalize the Surfaces

945

However, in commercial PP fibers embedded in the mixed matrix, except for the results obtained at 7 days, the modulus of elasticity is similar when fibers with and without functionalized surfaces are used. 3.3

Adherence Resistance

The adhesion strength was determined by the double pull-out test. The samples were made with fibers of both polymers, PET and PP, embedded in plastic mortars made with the same matrices used to determine bond stress. In turn, samples were made with fibers with and without surface functionalization by exposure to UV-C radiation. The results obtained can be seen in Figs. 12, 13, and 14.

Fig. 12. Bond stress results of the fibers studied without functionalization

The fibers with PET embedded in the Portland cement matrix (CMPET) have an increasing adhesion resistance over time, except at 3 months of age, at which time the result is approximately 8% lower than that obtained at 56 days of age. The samples embedded in mixed matrix (MMPET) have lower adhesion tensions than those obtained with the Portland cement matrix, except at 3 days of age. However, in PP fibers, the behavior is variable and the dispersion of results much greater than those obtained with PET fibers. At all ages, the tear resistance of PET fibers is lower with values between 20% and 50% of those obtained with PP fibers. These tests ended when the fibers detached from the matrix across its length. PET fibers with functionalized surfaces embedded in the cement matrix (CMPETUVC) have an adhesion resistance without a clear pattern of behavior. At 6 months of age, the tear resistance is the same if fibers with surfaces with and without UV-C exposure are used. The same variable behavior occurs when the fibers used have been exposed to the mixed matrix. In these fibers, at 6 months, the adhesion obtained with fibers exposed to UV-C (MMPET-UVC) is 20% lower than that for fibers that did not undergo functionalization (MMPET). In all samples the test was completed when the fiber was completely extracted from the matrix.

946

M. E. Fernández et al.

Fig. 13. Bond stress results of PET fibers with and without functionalization

Fig. 14. Bond stress results of PP fibers with and without functionalization

The results shown in Fig. 14 show that the PP fibers whose surfaces were exposed to UV-C radiation and were embedded in the Portland cement matrix (CMPP-UVC) mostly present tear tensions lower than fibers without functionalization. However, the tests of the samples with ages greater than 7 days, whose values are indicated by an asterisk, were completed before the detachment of the fiber; in these cases, the test ended when the fiber broke due to tensile stress. In samples with mixed matrices, this type of rupture occurred only in samples with a radiation-exposed surface (MMPPUVC) at ages 3 and 6 months; for other ages, the adhesion tension presented by the phases was lower when the fiber was functionalized with light radiation.

4 Conclusions From the results obtained in this work, the following conclusions can be drawn: • The functionalization of PET fiber surfaces through exposure to UV-C radiation does not produce significant changes in tensile strength or the modulus of elasticity. • Exposure to UV-C radiation of commercial PP fibers produces a decrease in tensile strength of between 70% and 80%. However, these changes do not affect the modulus of elasticity.

UV-C Treatment to Functionalize the Surfaces

947

• In most of the results, the functionalization of the surfaces by means of UV-C radiation does not show improvements in the adhesion between the fibers and the different matrices. In the case of PP fibers with ages greater than 28 days, it causes a significant decrease in joint work, causing the fiber to break due to its degradation, before its total detachment. Therefore, it is necessary to continue investigating other techniques for the functionalization of PET surfaces that allow improvement of their adhesion, such as the use of physical methods that generate a greater roughness on the surface of the material.

References 1. Naik, T.R., Singh, S.S., Huber, C.O., Brodersen, B.S.: Use of post-consumer waste plastics in cement-based composites. Cem. Concr. Res. 26, 1489–1492 (1996) 2. Sharma, R., Bansal, P.P.: Use of different forms of waste plastic in concrete—a review. J. Clean. Prod. 112, 473–482 (2015) 3. Gu, L., Ozbakkaloglu, T.: Use of recycled plastics in concrete: a critical review. Waste Manag 51, 19–42 (2016) 4. Pelisser, F., Montedo, O.R K., Gleize, P.J.P., Roman, H.R.: Mechanical properties of recycled PET fibers in concrete. Mater. Res. 15, 679–686 (2012) 5. Borg, R.P., Baldacchino, O., Ferrara, L.: Early age performance and mechanical characteristics of recycled PET fiber reinforced concrete. Constr. Build. Mater. 108, 29– 47 (2016) 6. Vijaya, G.S., Ghorpade, V.G., Sudarsana Rao, H.: The behavior of self-compacting concrete with waste plastic fibers when subjected to chloride attack. Mater. Today Proc. 5, 1501–1508 (2018) 7. Meza, A., Siddique, S.: Effect of aspect ratio and dosage on the flexural response of FRC with recycled fiber. Constr. Build. Mater. 213, 286–291 (2019) 8. Mohammed, M.-K., Al-Hadithi, A.I., Mohammed, M.H.: Production and optimization of eco-efficient self-compacting concrete SCC with limestone and PET. Constr. Build. Mater. 197, 734–746 (2019) 9. Awaja, F., Gilbert, M., Kelly, G., Fox, B., Pigram, P.J.: Adhesion of polymers. Prog. Polym. Sci. 34, 948–968 (2009) 10. Drobota, M., Persin, Z., Zemljic, L.F., et al.: Chemical modification and characterization of poly (ethylene terephthalate) surfaces for collagen immobilization. Cent. Eur. J. Chem. 11, 1786–1798 (2013) 11. Cazan, C., Cosnita, M., Duta, A.: Effect of PET functionalization in composites of rubber– PET–HDPE type. Arab. J. Chem. 10, 300–312 (2017) 12. Michael, F.M., Khalid, M., Walvekar, R., Siddiqui, H., Balaji, A.B.: Surface modification techniques of biodegradable and biocompatible polymers. Biodegrad. Biocompatible Polym. Compos. 33–54 (2018). Elsevier 13. Mathieson, I., Bradley, R.H.: Improved adhesion to polymers by UV/ozone surface oxidation. Int. J. Adhes. Adhes. 16, 29–31 (1996) 14. Moyano, M.A., Martín-Martínez, J.M.: Surface treatment with UV-ozone to improve adhesion of vulcanized rubber formulated with an excess of processing oil. Int. J. Adhes. Adhes. 55, 106–113 (2014) 15. Ossola, G., Wojcik, A.: UV modification of tire rubber for use in cementitious composites. Cem. Concr. Compos. 52, 34–41 (2014)

948

M. E. Fernández et al.

16. Coopamootoo, K., Masoero, E.: Cement pastes with UV-irradiated polypropylene: fracture energy and the benefit of adding metakaolin. Constr. Build. Mater. 165, 303–309 (2018) 17. Fernández Iglesias, M.E., Payá, J., Borrachero, M.V., Soriano, L., Mellado, A., Monzó, J.: Degradation process of postconsumer waste bottle fibers used in portland cement – based composites. J. Mater. Civ. Eng. 29(10) (2017). Content ID 04017183 18. AENOR: UNE-EN 196-1 Methods of testing cement. Part 1: Determination of strength (2018) 19. AENOR: UNE-EN ISO 527-3 Plastics. Determination of tensile properties. Part 3: Test conditions for films and sheets (1996) 20. AENOR: UNE-EN 14889-2 Fibers for concrete. Part 2: Polymeric fibers. Definitions, specifications and compliance (2008) 21. AENOR: UNE-EN ISO 6892-1 Metal materials. Tensile test Part 1: Test method at room temperature (2017)

Potential of Using Recycled Carbon Fibers as Reinforcing Material for Fiber Concrete Magdalena Kimm1(&), Amna Sabir1,2, Thomas Gries1, and Piyada Suwanpinij2 1

Institut für Textiltechnik of RWTH Aachen University, Aachen, Germany [email protected] 2 The Sirindhorn International Thai-German Graduate School of Engineering, King Mongkut’s University of Technology North Bangkok, Bangkok, Thailand Abstract. Carbon fibre reinforced polymers (CFRP) are taking over the aerospace, automotive, and wind energy sector. The growing demand for CFRP leads to a growing mass of CFRP waste. Landfilling and incineration are not effective and environmentally friendly routes for CFRP waste. The primary aim of this research paper is to determine the potential of using short pyrolyzed recycled carbon fibres (rCF) from CFRP waste to enhance the mechanical properties of fibre reinforced concrete (FRC). The secondary objective is to enhance the mechanical properties of FRC by optimizing fibre-matrix bonding in the interphase region. Recycled carbon fibre reinforced concrete (rCFRC) 60 specimens were prepared. These specimens consist of different rCF volume content of 0.00, 0.25, 0.5, 0.75 and 1 vol.-%. 4-point bending test procedure, along with visual analysis was performed for the characterization of the specimens. Oxygen (O2) plasma treatment has been used to improve the rCF (reinforcement) and concrete (matrix) adhesion. A maximum gain of 31% was achieved in flexural strength at fibre volume content of 0.5 vol.-% as compared to plain concrete. O2 plasma-treated rCFRC has a much higher value for elongation at break as compared to untreated rCFRC series. KEYWORDS: Carbon fibre reinforced polymer Fibre reinforced concrete


Recycled carbon fibres


1 Introduction Modern industry is interested in materials with high strength-to-weight ratio and stiffness, e.g. to produce fuel-efficient aircrafts and automobiles, massive wind energy blades with high-efficiency rates, robust and lightweight pressure vessels or materialefficient building structures. Carbon fibres (CFs) and CFRP are durable, lightweight, and possess the ability to mold into complex shapes. These properties make them potential and primary candidates for manufacturers [1, 2]. Global demand for CFRP in the year 2018 was approximately 75.5 kt. This demand is forecasted to reach 120.5 kt in the year 2022 [3]. Increasing demand for CFRP directly leads to an increase in waste too. By estimation, a sum of 5 kt of recycled CFRP is annually being wasted in Europe. Incineration is not the solution to CFRP waste because 50% to 70% of slag of incinerated fibre reinforced polymers are minerals and ashes which still need to be © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 949–960, 2021. https://doi.org/10.1007/978-3-030-58482-5_83

950

M. Kimm et al.

landfilled [4]. Industry-specific legislations, such as the End of Life Vehicle Directive (ELVD), pose high demands to producers to design vehicles towards recyclability [1]. The government’s current primary goal is also to focus on circular economy and sustainable built environment [6]. Therefore, a sustainable, economic, and naturefriendly path for CFRP waste recycling is need of the hour. An attractive approach is the use of rCF from CFRP waste as reinforcement for concrete. Mechanical properties using conventional fibres made of steel and alkaliresistant glass can be achieved by rCF. On the other hand, rCF can easily overcome the disadvantages of both steel and AR-glass fibres. Steel rust easily, which leads to deterioration of the concrete structure. Additionally, the mechanical properties of steel FRC are limited. As the workability of concrete becomes lower with increasing fibre volume content leading to balling effects [7, 8]. In particular, conventional glass fibres and also alkali-resistant glass fibres become brittle due to alkalinity of cement mortar. Moreover, alkali-resistant glass fibres show a high price compared to conventional glass and steel fibres [8]. CFs are high-performance and expensive fibres commonly used in composite structures for the aerospace and automotive industries. Production process of vCFs is energy-intensive, which leads to high fibre cost. Due to this reason, vCFs have not been established in the construction market [9]. Therefore, from an environmental and economic perspective, it is important to test the potential application of cost-reduced recycled CFs. There are three main types of recycling, namely mechanical, chemical, and thermal recycling routes. Mechanical recycling is used in two steps, these steps are known as primary and secondary processes. Mechanical recycling targets size reduction by shredding and cutting. Thermal (pyrolysis) and chemical (solvolysis) processes are identified under the category of tertiary processes. Pyrolysis process is most widely and commercially used for recovery of fibres from glass and carbon fibre reinforced composite [2, 5]. Recycling by pyrolysis reduces cost from 57 €/kg (vCFs) to 5-10 €/kg (rCF) and the required energy from 55–166 kWh/kg to 3–10 kWh/kg [9]. rCF recovered through pyrolysis retain 98–99% of young’s modulus of vCFs and only lose 4–10% of the virgin fibre’s tensile strength. The only disadvantage is the loss in fibre length during the recovery process. In research and current composite material market, rCF from pyrolysis have not established yet; just reuse of cut-off from vCFs is taking place, mostly in the companies themselves. Therefore, new markets for the application of rCF have to be focused. Thus, in this research the influence of short pyrolyzed recycled carbon fibres as reinforcement for FRC for the construction industry is studied [9, 10]. Some research has been conducted on the integration of recycled carbon fibres in concrete by Mastali et al. and Saccani et al.: • In work done by Mastali et al., the fresh and hardened properties of FRC were assessed, considering rCF at four different fibre volume contents. Maximum fibre content was up to 2 vol.-% along with three different fibre lengths of 10, 20, and 30 mm [4, 7]. Compressive strength and impact resistance were observed to be maximum with fibre length of 30 mm and flexural strength was maximum with 20 mm fibre length. CFs were recycled from the remained unusable CFs sheets and were mechanically shredded into different lengths without any thermal treatment.

Potential of Using Recycled Carbon Fibers as Reinforcing Material for Fiber Concrete

951

Therefore, these results are comparable to the use of chopped vCFs not comparable to rCF gained by pyrolysis as in this work. • Saccani et al. investigated the possibility of recycling composites made of carbon fibres and epoxy resin in different inorganic matrixes without any previous thermal and chemical treatment in concrete mixture [12]. Short chopped CFRP sticks (5– 8 mm), including resin, were used in different volume fractions up to 5 vol.-%. The CFRP sticks shift the failure mode of composite from brittle to semi ductile. • Kimm et al. have investigated the effect of surface modification and volume content of pyrolyzed rCF in concrete. Fibre volume contents up to 2 vol.-% with an average length of 14.9 mm were used as reinforcement in concrete [9]. The maximum gain (+111%) in flexural strength for rCFRC samples in comparison to plain concrete was observed at 1 vol.-%. Although, with an increase in volume percentage, the workability of concrete mixer decreases. An enhanced bonding between concrete and fibres by a silane treatment could not be achieved. The material properties of FRC are generally influenced by fibre volume content, fibre arrangement, length, fibre-matrix bonding, and concrete mixture, which predominantly have been investigated with focus on steel polymer and glass fibres [11]. In general, rCF show a chemically inert surface, and therefore, it does not react strongly to any matrix materials unless a sizing is applied. Sizing is a microscopically thin coating of the fibre surface, usually using polymer-based materials, which intends to improve adhesion to the surrounding matrix in a composite material. During the thermal recovery process of carbon fibres (pyrolysis), the sizing layer is completely burned down. Therefore, a treatment after recycling of fibres is desired to improve the bonding to the concrete matrix. Only limited work has been done on vCFs, while no significant work is done in case of rCF. For surface treatment, the main approach is using a plasma, which is a mixture of particles at the atomic molecular level consisting of partially charged components, ions, and electrons. Depending on the plasma functional groups, e.g., carboxyl, carbonyl, or hydroxyl groups, which accumulates on the fibre’s surface and alter fibre properties. These functional groups modify the chemical or physical structure of fibres, thus tailoring fibre-matrix bonding strength but without influencing their bulk mechanical properties. • Schneider et al. investigated the incorporation of plasma-treated CFs yarns along with mineral coating in concrete matrix [13]. Sized carbon fibre yarns were treated by plasma. Three different gases (O2, O2 and Ar, O2 and CF4) were used with a treatment time of 1.6 min and 3.3 min. Pure O2 plasma-treated CFs showed the most improvement in mechanical properties (increase of pull out energy up to four times in comparison to the untreated reference specimen). • In the Ph.D. thesis of Hambach flexural and compressive strength of FRC with plasma-treated vCFs (2–3 mm) using three different process gases (O2, CO2, NH3) were investigated [14]. The results show that +6% gain in flexural strength was achieved using plasma treatment. This gain in flexural strength was achieved at relatively short exposure (2 min), at low energy (50 W), and using O2 as plasma gas. Based on Schneider et al. and Hambach’s research work, one can conclude that O2 plasma treatment is a suitable choice for rCF surface treatment. This research work

952

M. Kimm et al.

aims to determine the potential of using recycled carbon fibres as reinforcement in concrete considering the following aspects: • Influence of different dosage (0, 0.25, 0.5, 0.75 and 1 vol.-%) of pyrolized rCF on flexural strength and ductility of rCFRC. • Effect of O2 plasma-treated rCF on flexural strength and ductility of rCFRC.

2 Experimental 2.1

Materials

Polyacrylonitrile-based rCF obtained by a pyrolysis process was used in the preparation of FRC and are presented along with their technical data in Table 1 and Fig. 1 (a). The fibre material was ordered as per company specification in the range of 10–30 mm. Additionally, length analysis was carried out using a random number of 300 rCF filaments out of the bulk, according to DIN 53808-1. The fibre length distribution is shown in Fig. 2.

Fig. 1. a) rCF from pyrolysis b) plasma treated rCF from pyrolysis Table 1. Properties of rCF Properties Manufacturer Material specification Product commercial name Density Length

Values CFK Valley Stade Recycling GmbH & Co. KG, Wischhafen, Germany Carbon (>95%) CarboNXT pure

1.77 g/cm3 Mean: 27 mm Standard deviation: 12 mm (details in Fig. 2(a)) Diameter 7 µm Tensile strength* 3,620 N/mm2 Young’s modulus* 207,750 kN/mm2 *Values measured at Institut für Textiltechnik of RWTH Aachen University

Potential of Using Recycled Carbon Fibers as Reinforcing Material for Fiber Concrete

953

Relative frequency by number of fibers in [%]

20%

13,9% 11,7%

11,0%

10%

8,9% 8,5% 6,4% 5,7%

0%

7,1%

6,4%

7,1% 3,9% 4,3% 3,9% 2,1% 1,8%

1,4% 1,1% 0,4% 0,0%

1,1% 0,0%

Length interval in [mm]

Fig. 2. Fibre length distribution of rCF measured based on DIN 53808-1

A part of rCF was treated with O2 plasma in the Atto Plasma System of Diener electronic GmbH & Co KG, Germany. These fibres are shown in Fig. 1 (b) after treatment. Process parameters and specifications are presented in Table 2. The concrete matrix is composed of cement CEM I 42.5 R (490 kg/m3), fly ash (175 kg/m3), micro silica powder (Elkem Microsilica® 940U, 70 kg/m3), quartz flour (500 kg/m3), sand 0.2–0.6 mm (713 kg/m3), water (245 kg/m3) and Polycarboxylate Ether-based plasticizer in different contents varying with the fibre volume content (Master Glenium ACE 460, 7 kg/m3 for 0, 0.25 and 0.5 vol.-%, 7.5 kg/m3 for 0.75 and 8.125 kg/m3 for 1 vol.-%). Table 2. Process parameters and specifications of plasma treatment Process parameters Values rCF quantity of single batch 1.79 kg Time 5 min/each side of 3–5 cm thick layer of rCF Pressure 0.3–0.5 mbar Generate power 50 W Gas flow rate 5–10 cm3/min Process gas O2

954

2.2

M. Kimm et al.

Specimen Preparation

The tests were performed at fibre volume contents of 0, 0.25, 0.5, 0.75 and 1 vol.-% on ten specimens per series with dimensions of 32.5 cm  10 cm  2 cm. In case of plasma-treated fibres, only one specific fibre volume content 0.5 vol.-% was selected. During preparation, the solid and liquid components of concrete were first pre-mixed separately. Liquid components were gradually added into the solid component mixture and were mixed with the help of a hand mixer. The concrete mixture was placed on vibrating table for 1 min. rCF were afterwards added in small quantities to the concrete mixture to avoid fibres agglomeration. The plasma-treated fibres were kept in a vacuum storage bag after treatment and were used for FRC specimen preparation within 48 h. A decrease in the workability of concrete mixture with an increase in fibre volume content was observed. Therefore, at 0.75 and 1 vol.-%, an extra amount of plasticizer was added (Subsect. 2.1). Molds were filled with the FRC mixture and were manually leveled with the help of a trowel. FRC specimens in fresh state were placed on a vibrating table for 20 s. Specimens were covered with plastic sheets and left to dry for 1 day. Then, the specimens were placed in a water bath for 7 more days and 20 more days to dry in a room climate of 22 ± 3 °C. In total, 60 specimens were prepared. 2.3

Testing Method

For material characterization, EN 1170-5 was followed, which is the standard to determine the flexural strength of glass FRC using 4-point bending test, see Fig. 3 (a). A standard climate according to EN ISO 139 was maintained during the tests.

3 Results 3.1

Flexural Strength of RCFRC

In Fig. 3 (b), typical stress elongation curves of rCFRC series are shown. The flexural strength r and elongation e were calculated based on a formula according to EN 1170-5: r¼

FL 27 Dl  2 d and e ¼ 5 L b  d2

ð1Þ

with: force F and the geometrical relations L = Span (300 mm), Δl = deformation, b = width of specimen [mm], and d = Thickness of specimen [mm], see Fig. 3 (a).

Potential of Using Recycled Carbon Fibers as Reinforcing Material for Fiber Concrete

955

Fig. 3. a) Test setup according to EN 1170-5 b) average flexural strength of tested series c) average elongation at flexural strength d) representative stress elongation curve of tested series

The trend obtained for the tested series resemble the statement of Kimm et al. [9] that under flexural load, rCF exhibit different behavior in comparison to steel and glass fibre. Increasing the fibre content up to 0.5 vol.-% increases flexural strength from rMOR = 7.43 to 10.32 MPa and elongation at break from 0.13% to 0.19%. At 0.75 vol.-% (see Table 3 and Fig. 4 (c)), there is a sudden decrease in LOP, while at 1 vol.-% a broader plateau with eMOR = 0.2% can be observed (see Fig. 4 (d)). The strength at 1 vol.-% is higher than at 0.75 vol.-% but lower than both at 0.25 and 0.5 vol.-%. The MOR decreases with an increase in volume. The broad plateau indicates the shift from brittle to ductile behavior with increase in fibre volume content. It can be observed in Fig. 3 (d) that the elongation at break for content of 0.25 vol.-% and 0.5 vol.-% is increasing in a steady manner, but as the volume contents are increased to 0.75 vol.-% and 1 vol.-% a substantial gain in elongation at break can be observed. The result of plasma-treated rCFRC series are being represented by grey triangles in Fig. 3 (c and d). Comparing plasma-treated rCFRC series to plain concrete, there is an increase of 13% in flexural strength. But, there is a reduction of 18.5% in rMOR as compared to rMOR obtained for untreated short rCFRC series at 0.5 vol.-%. Plasma treatment also has an improving effect on eMOR, which increases by about 43%. The

956

M. Kimm et al.

value obtained for eMOR at 0.5 vol.-% with plasma-treated rCFRC along with standard deviation is comparable to the content of 1 vol.-% of untreated rCFRC. Table 3. rCFRC series’ properties under flexural load Volume content of rCF [vol.-%] 0 0.25 0.5 0.5 (plasma treated) 0.75 1

3.2

Flexural strength rMOR [MPa] 7.43 ± 0.44 10.12 ± 1.06 10.32 ± 0.38 8.42 ± 0.55 9.41 ± 0.6 9.81 ± 0.29

Elongation at break eMOR [%] 0.13 ± 0.01 0.14 ± 0.01 0.14 ± 0.02 0.20 ± 0.01 0.19 ± 0.02 0.20 ± 0.02

Visual Analysis of RCFRC

In Fig. 4, the cracked cross-section area of a plain concrete specimen is shown. Air voids are minimal in number and in range of 0.05 cm to 0.07 cm in diameter.

Air voids

1 cm

Fig. 4. Plain concrete crack cross section

In general, it can be observed that increase in fibre volume content leads to more air voids and with a greater diameter as compared to plain concrete. As shown in Fig. 5 at 0.25 vol.-%, the cracked cross section area highlights a few prominent air voids while fibres are not noticeable here. As the volume percentage is increased, rCF can be easily seen in the crack cross section area with the naked eye. Air voids are prominent at 0.75 vol.-% and 1 vol.-%. As shown in Fig. 1 (a) rCF are presented in bundles form rather than individual fibres after the pyrolysis recovery process. Under the action of shearing during mixing in the concrete matrix, some rCF are broken into individual shorter fibres but bundles of intact rCF can also be seen in Fig. 5 (c, d).

Potential of Using Recycled Carbon Fibers as Reinforcing Material for Fiber Concrete a) 0.25 Vol.-%

Air voids

1 cm c) 0.75 Vol.-%

1 cm

b) 0.5 Vol.-%

957

Air voids

1 cm

Air voids

d) 1 Vol.-%

Air voids

1 cm

Fig. 5. Crack cross section area of rCFRC specimen

4 Discussion As compared to plain concrete, rCFRC at 0.5 vol.-% shows a maximum gain in flexural strength of +39%. A slight decrease of flexural strength can be observed with increasing volume percentage beyond 0.5 vol.-%. According to general composite mechanics, an increase in aspect ratio (length to diameter ratio) of fibres or volume percentage of fibres will lead to an increase in mechanical properties. Nevertheless, there is a practical limit for the fibre volume resulting in a saturation point beyond which the increase in fibre volume content will deteriorate the mechanical properties of concrete composite. The decrease in mechanical properties with increasing fibre volume content can be due to difficulty in mixing, air entrapment, or poor bonding between fibre and matrix [15]. It can be deduced that plasma treatment dramatically increases the ductility of the FRC as observed from the stress-elongation behavior of plasma-treated rCFRC series in Fig. 3(d). Interestingly, the O2 plasma did not contribute to a higher flexural strength as compared to untreated fibres which is different from the finding of Hambach and Schneider [13, 14]. An explanation for this behavior is still anonymous. Considering the uneven pattern of the standard variation of flexural strength, shown in Table 3, the highest value of standard deviation was observed in case of the lowest fibre content at 0.25 vol.-% in the contrary, a minimum value was observed in case of the highest fibre content at 1 vol.-%. Generally, fibre length analysis (see Fig. 2) has shown a considerable deviation in fibre range from the mean value and an uneven pattern of fibre length distribution. Hence, fibre length analysis proves that fibre range differs significantly from the ordered range of 10–30 mm. The maximum fibre length measured was 64 mm, and the minimum was 4 mm due to the reason that rCF comes from different origins and have different tensile and young modulus values. This might have led to the diverging standard deviations of flexural strength between the series. Especially for low fibre contents, a small amount of fibres are picked from bulk for testing, which can vary strongly in fibre origin and length. So, the homogenous mixing of rCF after recovery from CFRP waste is an important factor. In order to avoid variation of fibre characteristics and to gain isotropic mechanical properties batches of rCF should already be mixed at recycling companies [9]. In Fig. 6, the relationship between flexural strength and fibre volume content is illustrated by linear regression analysis for all untreated series. The coefficient of

958

M. Kimm et al.

determination (R2 = 0.3031) is considerably below one, which indicates a strong divergence of the real values from the regression curve. The value of R2 in the work of Kimm et al. is 0.8868 for the same interval, but in contrast to this work, Kimm et al. used a fibre material with shorter mean length and lower length variation [9]. The coefficient of determination found by Mastali et al. is >0.93 [4, 7].

Fig. 6. Linear regression analysis between flexural strength rMOR and fibre volume content V

An important point in Mastali et al. work is that rCF were originated from a single source, one fixed type of carbon fibre sheet, and fibres were mechanically shredded to smaller pieces (no irregularities on fibre surface or deformation on fibre surface from thermal processing for recovery of fibres) [4, 7]. Hence a good dispersion of fibres can be easily obtained. So, Mastali obtained a higher coefficient of determination. In the case of the present work, the fibre’s length had a broad range with twice the value of standard deviation (Table 1) as compared to Kimm et al. where fibre length mean value is 14.9 mm and the standard deviation is 7.2 mm [9]. On the other hand in current research work fibres went through a thermal process of recovery and single individual fibres are not obtained rather than that bundles of fibres are obtained (see Fig. 2). Homogenous mixing of closely adhered fibres in form of bundles is much difficult to achieve as compared to mechanically shredded fibres. Considering the difference in origin of fibres, source of fibres, and the recovery process of fibres, the vast difference in coefficient of determination values of R2Mastali [ 0:93; R2Kimm ¼ 0:8868, and R2currentresearch ¼ 0:3031 can be understood.

5 Conclusion This work has been conducted to determine the influence of short recycled carbon fibres (rCF) on the mechanical properties of fibre reinforced concrete (rCFRC). rCF were added in concrete in different volume percentages of 0, 0.25, 0.75, 1 vol.-% and analysed in terms of their flexural strength. Visual analysis was also conducted thoroughly to connect visual effects with the mechanical properties of the tested specimens.

Potential of Using Recycled Carbon Fibers as Reinforcing Material for Fiber Concrete

959

O2 plasma treatment was carried out for short pyrolyzed rCF to enhance fibre-matrix bonding. The results concluded based on mechanical and visual analysis are following: • Concrete mechanical properties are dependent on fibre volume percentage. With the increase in fibre volume content flexural strength improves, but to a specific volume percentage limit. • Maximum flexural strength was achieved at 0.5 vol.-% of rCF. This value is 39% higher than the flexural strength of plain concrete. • Maximum elongation at break value was achieved at maximum volume percentage of conducted series (1 vol.-%). The same value was obtained in case of plasmatreated rCFRC series at 0.5 vol.-%. • Based on regression analysis, it can be concluded that origin, dispersion of fibres and source of rCF effects the coefficient of determination value and further the predictability of FRC. • Optimized mechanical method of mixing of rCF in concrete matrix is required and should be further investigated. • Plasma treatment influenced the ductility of FRC composite but did not increase the flexural strength as compared to untreated rCF. • An increase in the fibre volume leads to decrease in workability of the concrete paste inducing air voids and surface irregularities on the dried rCFRC test specimen. Generally, the investigations have shown, that rCF from carbon fibre reinforced polymers have a very great potential for a high-performance, non-hazardous reuse as short fibre reinforcement in concrete. In particular, taking proper volume contents, fibre material characteristics and mixing procedures into account. Also, in comparison to vCFs, rCF offer exclusive economic benefits due to their lower costs and are therefore considered to be competitive to conventional fibres in concrete [16].

References 1. Das, S., Warren, J., West, D.: Global Carbon Fiber Composites Supply Chain Competitiveness Analysis, Clean Energy Manufacturing Analysis Center, Oak Ridge National Laboratory. University of Tennessee, Knoxville (2016) 2. Melendi-Espina, S., Morri, C., Turner, T., Pickering, S.: Recycling of Carbon Fibre Composites, pp. 1–4 (2016) 3. AKV-Industrial Association Reinforced Plastics eV. Composites Market Report 2018, Market developments, trends, outlooks and challenges, Frankfurt, Deutschland (2018) 4. Mastali, M., Dalvand, A.: The impact resistance and mechanical properties of selfcompacting concrete reinforced with recycled CFRP pieces. Compos. Part B Eng. 92, 360– 376 (2016) 5. Fernández, A., Lopes, C.S., González, C., López, F.A.: Characterization of Carbon Fibers Recovered by Pyrolysis of Cured Prepregs and Their Reuse in New Composites. IntechOpen (2018). http://doi.org/10.5772/intechopen.74281 6. Pimenta, S., Pinho, S.T.: Recycling carbon fibre reinforced polymers for structural applications: technology review and market outlook. In: Waste Management, New York, N. Y., vol. 31, pp. 378–392 (2011)

960

M. Kimm et al.

7. Mastali, M., Dalvand, A., Sattarifard, A.: The impact resistance and mechanical properties of the reinforced self-compacting concrete incorporating recycled CFRP fiber with different lengths and dosages. Compos. Part B Eng. 112, 74–92 (2016) 8. Rai, A., Joshi, Y.P.: App. Prop. Fibre Reinfor. Concr. 4, 124–127 (2014) 9. Kimm, M.: Investigation of surface modification and volume content of glass and carbon fibres from fibre reinforced polymer waste for reinforcing concrete. J. Hazard. Mater. (2019). http://doi.org/10.1016/j.jhazmat.2019.121797 10. Russell, D.: Recycled Carbon Fibre: A New Approach to Cost Effective Lightweighting. https://www.igcv.fraunhofer.de/content/dam/igcv/de/docs/Travelling_Conference. Accessed 27 Oct 2019 11. Triantafillou, T.C.: Blast Protection of Civil Infrastructures and Vehicles Using Composites. Woodhead Pub Ltd., Boca Raton (2010) 12. Saccani, A., Manzi, S., Lancellotti, I., Lipparini, L.: Composites obtained by recycling carbon fibre/epoxy composite wastes in building materials. Constr. Build. Mater. 204, 296– 302 (2019) 13. Schneider, K., Lieboldt, M., Liebscher, M., Fröhlich, M., Hempel, S., Butler, M., Schröfl, C., Mechtcherine, V.: Mineral-based coating of plasma-treated carbon fibre rovings for carbon concrete composites with enhanced mechanical performance. Materials 10, 1–5 (2017) 14. Hambach, M.: High Strength Multifunctional Composites Based on Portland Cement and Carbon Short Fibres. University of Augsburg, Germany (2016) 15. Naaman, A.E.: Fiber Reinforced Cement and Concrete Composite. Techno press 3000, Florida (2017) 16. Kimm, M.: Recycling von Carbonbeton - Wie kann eine hochwertige Wiederverwendung gelingen? In: ‘11. Carbon - und Textilbetontage, Dresden, 24–25 September 2019, pp. 34– 35 (2019)

Textile Reinforced Concrete (TRC)

Development of Textile Reinforced UHPC with Reduced Steel Fiber Contents Mengchao Zhai1, Yiming Yao1(&), Jingquan Wang1, and Barzin Mobasher2 1

School of Civil Engineering, Southeast University, Nanjing, China [email protected] 2 School of Sustainable Engineering and the Built Environment, Arizona State University, Tempe, USA

Abstract. Textile reinforced ultra high performance concrete (TR-UHPC) is developed to further improve the tensile strength and ductility with reduced steel fiber contents. Alkali resistant glass textiles are used to partially replace steel fibers in a hybrid manner. The effects of different steel fiber content on the tensile and bending properties is investigated for four short fiber content levels. It is found that the synergistic effects of the hybrid reinforcements can be used to enhance of strengthening and toughening mechanisms with reduced cost. Pronounced strain hardening and deflection hardening are observed under tension and bending tests. The highest ductility of the TR-UHPC specimens are obtained when the steel fiber content is 1.0% vol with enhanced strength. Keywords: UHPC
Textile
Low fiber content
Tensile behavior
Flexural behavior
Strain hardening

1 Introduction Ultra high performance concrete (UHPC) is a new class of cement-based material with ultra-high mechanical properties and durability, which has been extensively investigated in recent years [1, 2]. However, the high dosage of steel fibers affects the workability of UHPC negatively and increases the cost. Textile reinforced concrete (TRC) is a composite material composing of textiles and fine grained concrete, which has the characteristics of high strength, large strain capacity and excellent durability [3]. A combination of high performance textiles and UHPC could be an promising way to further enhance the tensile properties of UHPC and reduce the use of steel fibers, especially for thin-walled structures. At present, there are few researches on textile reinforced ultra high performance concrete (TR-UHPC). Zhou et al. [4] investigated the effects of different textile types and textile processing methods on the tensile properties of TR-UHPC; Jiang et al. [5, 6] studied the tensile properties of basalt textile reinforced engineering cement composites (ECC) and the bonding behavior between fiber and matrix, the results show that ECC can greatly improve the reinforcement effect of textile and the crack width can be effectively controlled. In this study, a TR-UHPC reinforced with alkali-resistant (AR) glass textiles is developed with different dosages of steel fiber. Mechanical properties of TR-UHPC © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 963–970, 2021. https://doi.org/10.1007/978-3-030-58482-5_84

964

M. Zhai et al.

under tension and flexure were investigated and the effects of short steel fiber content were studied.

2 Material A proprietary UHPC mix produced by Jiangsu Subote New Materials Co., Ltd was used. The short steel fibers with length of 6 mm were adopted. Alkali-resistant (AR) glass textile was used as the reinforcement of the TR-UHPC specimens, and the specific specification parameters are shown in Table 1. Table 1. Properties and geometry of yarns made up the textiles. Tensile Strength (MPa) 1700

Young’s Modulus (GPa)

Density (g/cm3)

Linear density (tex)

Linear density of monofilament (tex)

72

2.68

1200

19

3 Preparation of Specimens and Test Setup The 300 mm  300 mm  18 mm square TR-UHPC thin plate was made by using a laminating technique. The process started with pouring a thin layer of UHPC at the bottom of the mold. Then a layer of textile was laid on this fresh concrete layer and pressed properly to ensure the contact between the textile and the matrix. After the first layer of textile was arranged, the second layer of UHPC was poured. The operations were repeated until the last layer of textile was laid. All plates were demolded after 1 day, and cured for 28 days in a curing chamber (20 °C, 95% RH). In this experiment, two layers of textiles with equal spacing were set for all specimens, in other words, the thicknesses of UHPC layers were all 6 mm. Six plate specimens with dimension of 50 mm  300 mm  18 mm were cut from each of the large plates mentioned above for uniaxial tension and four-point bending tests. The specimens were divided into four groups and labeled as 0.5%-T, 1.0%S-T, 1.5%T and 2.0%-T for tension, and 0.5%-B, 1.0%S-B, 1.5%B and 2.0%-B for bending. The volume fraction of steel fibers ranged from 0.5% to 2.0%. Aluminum sheet of 75 mm long were glued on both ends of all the direct tensile test specimens to strengthen the clamping zone. The gauge length was 150 mm, and the extensometer was used to measure the tensile deformation (Fig. 1). The uniaxial tensile test was carried out with a controlled deformation rate of 0.5 mm/min.

Development of Textile Reinforced UHPC

965

Fig. 1. Dimension of TR-UHPC tensile specimen.

In four-point bending test, all specimens were tested with a span of 270 mm (see Fig. 2). LVDT was placed to measure the mid span deflection of the specimen. The displacement control was used in the test to ensure that the loading rate of the machine is 1 mm/min.

Fig. 2. Bending test set up.

4 Results and Discussion 4.1

Tensile Behavior of TR-UHPC

The tensile stress-strain curves of TR-UHPC are shown in Fig. 3. The addition of short fiber has apparent influence on the first-crack stress and tensile strength of the specimens. When 1.5%S-T is compared with 0.5%S-T, in the first crack stress increases for as much as 50%. Strain hardening stage of the stress-strain curves is observed after cracking. The post-peak responses are characterized as extended softening behavior as only two layers of glass textiles were used. While the ductility of was improved prominently in comparison with normal UHPC. However, as the steel fiber content increased to 2.0%, tensile strength was decreased. The tensile responses showed that the hybrid reinforcements can achieve equivalent effects of strengthening and toughening by partially replacing short steel fibers.

966

M. Zhai et al. 14

14

12

0.5%S-T1 0.5%S-T2 0.5%S-T3

10

Tensile Stress, MPa

Tensile Stress, MPa

12

8 6 4 2 0

1.0%S-T1 1.0%S-T2 1.0%S-T3

10 8 6 4 2

0

0.02

0.01

0

0.03

0

(a)

0.03

14

1.5%S-T1 1.5%S-T2 1.5%S-T3

10

2.0%S-T1 2.0%S-T2 2.0%S-T3

12

Tensile Stress, MPa

12

Tensile Stress, MPa

0.02

(b)

14

8 6 4

10 8 6 4 2

2 0

0.01

Strain, mm/mm

Strain, mm/mm

0

0.02

0.01

Strain, mm/mm (c)

0.03

0

0

0.01

0.02

0.03

Strain, mm/mm

(d)

Fig. 3. Tensile stress - strain curves of TR-UHPC with different fiber volume fractions.

The effects of different steel fiber content on the tensile properties of TR-UHPC are compared in Fig. 4. The strain values corresponding to the ultimate bearing capacity of all fiber contents had little change. In contrast, when the fiber content was 1.5% vol, the ultimate bearing capacity of the test specimens was the highest. While when the fiber content was 2.0% vol, the ultimate bearing capacity was reduced by a certain extent. It might be explained by higher matrix porosity due to high fiber dosage and interaction between steel fiber and glass textiles [5]. The laying of textiles under high fiber content increased the difficulty of the uniform fiber dispersion, which lead to more micro initial defects.

Development of Textile Reinforced UHPC 0.5 0.4

12

Strain, mm/mm

Tensile Strength, MPa

14

10 8 6 4

0.3 0.2 0.1

2 0 0

967

0.5

1

1.5

2

2.5

Steel Fiber Content, vol% (a)

0 0

1

0.5

1.5

2

2.5

Steel Fiber Content, vol% (b)

Fig. 4. (a) Average tensile strength and (b) average strain at ultimate bearing capacity at varying steel fiber content.

4.2

Flexural Behavior of TR-UHPC

Figure 5 shows the experimental curves of bending stress and the midspan deflection, where the bending stress was the maximum normal stress at the midspan section calculated according to the plane section assumption. It can be seen that the flexural strengths are higher than the direct tensile strength of TR-UHPC for same fiber volume fraction. This can be explained by the differences in the stress distribution profiles of the two test methods. In the tension test, the entire volume of the specimen is a potential zone for crack initiation. Comparatively, in the flexural test, only a small fraction of the tension region is subjected to an equivalent ultimate tensile stress. With the increase of short fiber content, the bending strength of the specimens showed a remarkable increase. However, when the fiber content was increased from 1.5% to 2%, the bending strength was slightly decreased, which is consistent with the trends found in direct tension tests. When the fiber content was 1% vol, and the highest ductility was obtained and the bending strength was also improved (see Fig. 6).

968

M. Zhai et al. 30

30

20 15 10 5 0

0

5

10

15

20

25

1.0%S-B3 1.0%S-B4 1.0%S-B5

25

0.5%S-B4 0.5%S-B5

Bending Stress, MPa

Bending Stress, MPa

25

20 15 10 5 0

30

0

5

Deflection, mm

(a)

20

25

30

30

1.5%S-B4 1.5%S-B5 1.5%S-B6

20 15 10 5

0

5

10

15

20

Deflection, mm

(c)

25

2.0%S-B4 2.0%S-B5 2.0%S-B6

25

Bending Stress, MPa

25

Bending Stress, MPa

15

(b)

30

0

10

Deflection, mm

30

20 15 10 5 0

0

5

10

15

20

25

30

Deflection, mm

(d)

Fig. 5. Bending stress versus mid-span deflection curves of TR-UHPC with different fiber volume fractions.

30

10

25

8

Deflection, mm

Flexural Strength, MPa

Development of Textile Reinforced UHPC

20 15 10

6 4 2

5 0 0

969

0.5

1

1.5

2

2.5

0 0

0.5

1

1.5

2

Steel Fiber Content, vol%

Steel Fiber Content, vol%

(a)

(b)

2.5

Fig. 6. (a) Average flexural strength and (b) average mid-span deflection at ultimate bearing capacity at varying steel fiber content.

5 Conclusions TR-UHPC reinforced by AR-glass textile and steel fiber were developed in this study. The use of textiles in UHPC can effectively reduce the amount of steel fibers and thus reduce the cost. The synergistic effects of hybrid fibers were ensured the strengthening and toughening mechanisms of TR-UHPC with reduced steel fibers. Experimental tests showed an increasing trend of tensile and flexural strength with increasing content of steel fibers, while the highest strengths were obtained with 1.5% vol. of steel fiber. The highest ductility was observed at fiber dosage of 1.0% with enhanced strength. This new type of materials may be ideal for the application in thin-walled structures with enhanced mechanical properties and pouring workability with reduced fiber contents. Acknowledgements. This study was supported by the Natural Science Foundation of China (51908120) and the Natural Science Foundation of Jiangsu Province (BK20180383). The financial supports are gratefully appreciated.

970

M. Zhai et al.

References 1. de Larrard, F., Sedran, T.: Optimization of ultra-high-performance concrete by the use of a packing model. Cem. Concr. Res. 24(6), 997–1009 (1994) 2. Chen, B., Ji, T., Huang, Q., et al.: Review of research on ultra-high performance concrete. J. Architect. Civ. Eng. 31(3), 1–24 (2014) 3. Liu, S., Zhu, D., Li, A.: Research and application progress of textile reinforced concrete. J. Architect. Civ. Eng. 34(5), 134–146 (2017) 4. Zhou, Z., Zhang, Y., Wang, Y., et al.: Experimental study on tensile mechanical property of grid reinforced UHPC plates. J. SE Univ. 4, 1 (2019) 5. Jiang, J., Jiang, C., Li, B., et al.: Bond behavior of basalt textile meshes in ultra-high ductility cementitious composites. Compos. Part B Eng. 174, 107022 (2019) 6. Li, B., Xiong, H., Jiang, J., et al.: Tensile behavior of basalt textile grid reinforced engineering cementitious composite. Compos. Part B Eng. 156, 185–200 (2019)

Reinforcement of Concrete with Glass Multifilament Yarns: Effect of the Impregnation on the Yarn Pull-Out Behaviour A.-C. Slama(&), J.-L. Gallias, and B. Fiorio CY Cergy Paris Université, L2MGC, 95000 Cergy, France [email protected]

Abstract. The impregnation of the yarn by the cementitious matrix is a key parameter to predict the mechanical behaviour of textile reinforced concrete. The mechanism of this impregnation is specific because of the heterogeneous nature of both materials. Since there is no method to observe and quantify this impregnation, all the models defined to describe the mechanical failure of this type of composite are based on statistics and incomplete observations. To adjust those models to the experimental results and to understand the influence of the matrix on the impregnation, several pull-out tests were performed on samples composed of a glass multifilament yarn and several matrices. After pull-out test, some samples were impregnated with resin in a three steps process that includes cementitious matrix dissolution and enable to visualize the remaining filaments and to assess their impregnation degree by the cementitious matrix. Keywords: Multifilament yarn Confocal microscopy


Mortar
Impregnation
Pull-out test


1 Introduction TRC, or textile reinforced concrete, is a promising material for multiple applications such as prefabricate and structural building [1, 2]. Its advantages, compared to steel reinforcements, are no corrosion, lightness and flexibility that enable the manufacture of thin structure with various innovative shapes. However, to fully use its potential, it is necessary to understand its mechanical behaviour [3, 4]. This mechanical behaviour depends mainly on the bond between those two assembled materials [5]. On the contrary to the impregnation of textile reinforcement in most composites with resin polymer matrix, the impregnation of the textile reinforcement in composite with cement matrix is not complete, due to the specific structure of the mortar and the multifilament yarn [6]. A very large proportion of the solid particles of the mortar has a bigger size than the spaces in between the filaments of the yarn [7], leading to a random penetration of the matrix inside the bundle. Therefore, there is a discontinuous impregnation in the transverse direction of the yarn but also along it [8]. In the non-impregnated areas, the mechanical behaviour of the free length of the filaments is also difficult to describe because it depends on the waviness of those filaments and the inter-filament friction [9, 10]. All those elements were © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 971–982, 2021. https://doi.org/10.1007/978-3-030-58482-5_85

972

A.-C. Slama et al.

modelled by several successive models that take into account all those specificities [11–14]. However, since the visualization of the impregnation is complicated without destroying the sample, the impregnation degree of the filaments is based on indirect observation methods and statistical parameters [15, 16]. As a result, the modelling of the bundle as concentric layer with impregnation level increasing from the outside to the core is a simplification of the real situation and the impregnation degree is not defined from experimental data [17]. In order to study the impregnation of a yarn embedded in different cementitious matrices and with different embedded lengths, several pull-out tests have been conducted. The matrices were characterized, and then pull-out samples were manufactured and tested. The results of the pull-out test have been linked to matrix characteristics and to the impregnation of the yarn by the cementitious matrix thanks to an innovative visualization method.

2 Materials and Methods 2.1

Multifilament Yarns and Cement Matrices

The multifilament yarn used in this study is an Alkali Resistant glass direct roving of 2400 Tex from Owens Corning Cem-l 5325. Its characteristics are given in Table 1 according to [18]. Table 1. Characteristics of the yarn (supplier data) Linear weight (Tex) 2400

Diameter of the filaments (µm)

Sizing (%)

Modulus of elasticity of the filaments (GPa)

Tensile Strength of the filaments (MPa)

27

0.8

72

1000–1700

Matrices with various rheological and mechanical properties were used, in order to understand the effect of the matrix characteristics on the composite properties. They are all based on standard mortar composition (NF EN 196-1 standard) with normalized sand, CEM I 52.5 R cement where various quantities of Sika viscocrete Tempo 653 superplasticizer were introduced (expressed by the S/C ratio, where S is the amount of the superplasticizer’s dry material) in order to obtain matrices with higher workability and the same W/C ratio or matrices with the same workability and lower W/C ratios and higher than the standard mortar’s compressive strength. The compositions of the matrices used in this study are given in Table 2, with their workability (NF P18-452 standard) and compressive strength (NF EN 196-1 standard).

Reinforcement of Concrete with Glass Multifilament Yarns

973

Table 2. Matrices composition and properties Mortar M50-00 M50-05 M50-10 M47-20 M45-20 M45-30

2.2

W/C ratio S/C ratio (%) workability (s) Compressive strength (MPa) 0.50 0 10 58 + 2 0.50 0.05 4 55 + 3 0.50 0.1 2 59 + 3 0.47 0.2 2 64 + 2 0.45 0.2 10 68 + 2 0.45 0.3 4 61 + 4

Samples Manufacturing and Pull-Out Test

Samples are of prismatic shape (4  4  16 cm), made in steel moulds, with a yarn placed along the central axis of the prism, as shown in Fig. 1. After placing the yarns, the mortar is made and cast according to the NF EN 196-1 standard. The samples are removed from the mould after 24 h and stored in water for 27 days. Then, the edges of the samples are sawed at the desired embedded length, called Le, to obtain the pull-out samples. After drying, the pull-out samples are prepared for the pull-out tests by sticking the end of the free length, Lf (after 10 cm of yarn), in-between two 2.5  5 cm epoxy plates. After 24 h of drying in room conditions, the pull-out samples are ready for the pullout tests. The six matrices are tested with two different embedded lengths: 0.7 and 1 cm. Six pull-out samples are manufactured for each matrix and each embedded length. However, since the free length of some samples was too damaged during manufacturing, several samples were excluded from the study. The pull-out tests are performed on an Instron testing machine at 2 mm/min until displacement reaches the value of the embedded length. A special device to place the pull-out sample is designed as shown in Fig. 1. The free length end is hold in the jaw of the press, with the previously stuck plate. The obtained load/displacement curves are then smoothed using weighted averages.

Fig. 1. Manufacturing process of the pull-out samples (left) and device for the pull-out test (right)

974

2.3

A.-C. Slama et al.

Characterization of the Yarn/Mortar Interface

As described in the literature [13, 19, 20], the pull-out failure of a multifilament yarn embedded in a cementitious matrix is characterized by the formation of two groups of filaments with distinct behaviour, according to their impregnation degree: outer filaments of the yarn that are well impregnated in the matrix fail in tension and filaments of the core of the yarn that are not impregnated enough by the matrix, and that slip and are extracted from the embedded length. In order to differentiate the filaments that were well impregnated from those which were sparsely or not impregnated, a new method of double resin impregnation joined with confocal microscopy observations was developed. It is based on two successive impregnations of samples after pull-out test with resin with two different markers, separated by a step of cement matrix dissolution (presented Fig. 2). The first step is an impregnation of the yarn inside the cementitious matrix with a rhodamine (a red coloured marker) marked resin using a syringe to inject it in the hole left by the extracted filaments. The second step is the cementitious matrix complete dissolution in acid. The third step is a second resin impregnation with a fluorescein (green/yellow coloured marker) marked resin of the initially obtained moulding of the yarn into the cementitious matrix after matrix dissolution.

Fig. 2. Process of double resin impregnation

Thanks to this double marking, it is possible to identify different zones in the yarn after pull-out test on several considered sections. The first zone is the void left by filaments extraction on this section, which contains only the rhodamine marked resin and no filament (zone 1 in Fig. 3). The second zone contains the filaments sparsely impregnated on this section by the cementitious matrix before its dissolution (zone 2 in Fig. 3). During the first impregnation, those filaments were not impregnated enough by the cementitious matrix to prevent the rhodamine resin to reach them. On the contrary, the third zone contains filaments on this section that were fully impregnated by the cementitious matrix during the first impregnation by the resin with rhodamine (zone 3 in Fig. 3). After dissolution of the cementitious matrix, they were set free and impregnated during the second resin impregnation with fluorescein. Despite they were

Reinforcement of Concrete with Glass Multifilament Yarns

975

not reached by the rhodamine marked resin injection due to their complete inclusion in the cementitious matrix, those filaments are not lost after dissolution of the cementitious matrix because they stay connected to the rhodamine injected part of the yarn through the excess of rhodamine that remain at the top face of the sample (on the yarn’s free length side) after the injection. It can therefore be considered that almost no filaments are lost after the two successive impregnations of the sample.

Fig. 3. Identification of three zones in a section of an impregnated yarn with double marking observed with confocal microscopy

To study the variability of the filaments/ matrix interaction along the embedded length of the yarn, different sections of those impregnated samples, made by successive polishing with a polishing device Buelher, are observed using a confocal inverse microscope ZEISS LSM 710 with a 10 lens. The areas of zones 1, 2 and, to some extent, 3 are computed using ImageJ software. The Trainable Weka plugin is used to count the filaments by machine learning. It is then possible to evaluate the areas of the different zones and the number of filaments in zone 2 and zone 3. Averages are then computed on the data obtained for each section.

3 Pull-Out Behaviours 3.1

Two Pull-Out Modes

Two pull-out mechanical behaviours were encountered: failure of the yarn in the embedded length, followed by the slippage and extraction of some filaments on a displacement equal to the embedded length or failure of the yarn in the free length of the sample. In the following, these two failure modes are respectively referred as mode 1 and mode 2. Examples of the behaviour associated to each of these modes are presented in Fig. 4. It is also noticeable that some filaments failures along the free length of the yarn are observed even for mode 1.

976

A.-C. Slama et al.

Mode 1 is differentiable from mode 2 by the appearance of a hole on the face of the sawed surface after the pull-out test, which means that a number of filaments has been extracted from the mortar. Those filaments were not impregnated or not impregnated enough by the mortar along the embedded length so, during pull-out test, instead of failure, they experience slippage along the embedded length after failure of some potential links between the matrix and those filaments [21]. This sliding phenomenon is visible on the load/displacement curve as a post-peak residual phase, which ends at a displacement equal to the embedded length. For mode 2, the mechanical behaviour is similar to the one of the yarn during the tensile test. However, the pre-peak stiffness, post-peak stiffness and maximum load of the load/displacement curve of mode 2 pull-out behaviour are lower than the ones of the load/displacement curve of the yarn tensile test results. It can be due to the yarn damage during manufacturing of the pull-out samples and the occurrence of some failures in the embedded length. The load reaches zero for a displacement very inferior to the embedded length and most maximum load values for this type of pull-out behaviour are superior to the ones for the first type. 180

Pmax

160

mode 1

Load (N)

140

mode 2

120 100 Pres

80 60 40 20 0

0

2

4 Displacement (mm)

6

8

Fig. 4. Two load/displacement curves obtained for pull-out tests of two samples from M50-10 Le = 1 cm and M50-00 Le = 0.7 cm

The results show that the mode 1 pull-out behaviour occurs generally more frequently for lower compressive strength matrices but the correlation between these two parameters is not strong (R2 = 0.55). On the other hand, the fluidity of the matrix has no significant influence on the pull-out behaviour.

Reinforcement of Concrete with Glass Multifilament Yarns

977

For all the samples, Pmax, the maximum load, is computed. It is the load value at the peak. When a number of filaments slipped and is extracted during the mode 1 pull-out behaviour, the residual load, Pres, is computed as the load value of the intersection point of the regression line associated to slope of the post-peak stiffness, and those of the slope of the residual post-peak stiffness. The post-peak stiffness is the slope of the regression line computed between 30% and 50% of Pmax in the post-peak phase. The residual post-peak stiffness is the slope of the regression line computed between 30% and 50% of the sliding length of the sample. 3.2

Influence of the Matrix on the Pull-Out

Despite a high variability, the results show that matrices with high fluidity and compressive strength lead more frequently to high maximum load values and a failure in the free length of the pull-out sample (mode 2). On the other hand, low fluidity and compressive strength lead more frequently to the slippage and the partial extraction of the filaments (mode 1). It was established also that the embedded length does not have a significant influence on the pull-out mechanical results. More precisely, in this study, a relatively strong correlation is found between the compressive strength of the matrices and the average Pmax of all the tested samples as well as the average Pres values of samples with mode 1 pull-out test independently of the embedded length (Fig. 5). It is in accordance with Butler et al. [22] whose study shows that a densification of the matrix (so a higher compressive strength) leads to more filaments failure and less slippage. However, in his case, the maximum load values decrease with this densification since it is the result of ageing and since other mechanisms can take place during this ageing. In this present study, all the samples are tested at the same age, so the maximum load values increase with matrix densification. This correlation does not exist if the fluidity of the matrix is considered (see Fig. 6), on the contrary to what was shown in the study of Peled and Mobasher [23], but this contradiction can be explained by the fact that the fluidity has varied by changing the W/C and S/C ratios in the present study and by adding silica fume and fly ashes in their case that modified also the particle size distribution of the matrix. For samples with mode 1 pull-out behaviour, the residual load value for each sample is only weakly related to the maximum load value (see Fig. 7) so, even if the compressive strength has a strong influence on both the maximum and the residual load, different phenomena might interfere in the peak phase and in the residual phase, leading to their weak link and high variability for each sample.

978

A.-C. Slama et al.

Average Pmax (N)

250

R² = 0.7557

200

M50-05

150

M50-10

100

M47-20

50

M45-20

0 50.00

M45-30 60.00 70.00 Compressive strength (MPa)

50 Average Pmax (N)

M50-00

M50-00

40

M50-05

R² = 0.7291

30

M50-10

20

M45-20

10 0 50.00 -10

55.00

60.00

65.00

70.00

M45-30 75.00

Compressive strength (MPa)

Average Pmax (N)

Fig. 5. Influence of the compressive strength of the matrix on the pull-out maximum load (up) and residual load (down)

250

M50-00

200

M50-05

150

M50-10

100

M47-20

50 0 0.00

M45-20 5.00 10.00 workability (s)

M45-30

Fig. 6. Influence of the rheological properties of the matrix on the maximum load

Reinforcement of Concrete with Glass Multifilament Yarns

979

250 R² = 0.2667

200 Pmax (N)

M50-00

150

M50-05

100

M50-10 M45-20

50

M45-30

0 0

20

Pres (N)

40

60

Fig. 7. Link between the maximum load and the residual load for the samples with mode 1 pullout behaviour

4 Influence of the Impregnation on the Pull-Out To explain the different phenomena influencing the maximum load and the residual phase, images obtained as described in Sect. 2.3 are analysed. Between the samples considered previously, only the M50-00 and M50-10 samples, matrices with a majority of mode 1 pull-out behaviour and the same W/C ratio are observed. Considering the determined zones by double resin impregnation (Fig. 3), the results show (Fig. 8) that there is no link between Pres values and the areas of the void left by filaments extraction (zone 1), on the contrary to what was found by Banholzer [8]. There is no link also between Pmax values and the areas of the zones 1 and 2 or their sum. However, there is a link between load parameters and the number of filaments embedded in the cementitious matrix. It is important to note that the area of the zone 3 is difficult to estimate because the dissolution of the matrix releases the fully embedded filaments that could be reorganized during the second impregnation, leading to an incorrect estimation of this area. So, the number of sparsely impregnated filaments in zone 2 and fully impregnated filaments in zone 3 is then considered to establish the correlation between the level of impregnation and the pull-out behaviour. The most noticeable links observed are those between the average number of filaments in zone 3 and the Pmax values (Fig. 9), and to a lesser extent the Pres values. Therefore, it seems that those values depend essentially on the number of filaments that are fully impregnated in the matrices. The number of extracted and sparsely impregnated filaments does not have a direct influence on those values. However, some other mechanisms that cannot be observed with this post pullout test visualisation method (such as friction between the filaments and the matrix particles during extraction [24], surface roughness of the filaments due to the formation of hydrate products on them [25] or successive extractions of several groups of filaments at different displacements [8]) might influence the Pres values. Those other mechanisms need to be investigated in further studies.

980

A.-C. Slama et al.

30

Pres (N)

25 20 15 10 5 0 0

0.5

1 1.5 2 2.5 Void left by filaments extraction (mm2)

M50-00 L=0,7

M50-10 L=0,7

3

3.5

M50-00 L=1

Fig. 8. Link between the void left by filaments extraction (area of zone 1) and the residual load

200 R² = 0.6366 Pmax (N)

150 100 50 0 -100

100 300 500 700 Number of impregnated filaments M50-00 Le=0.7 M50-10 Le=0.7 M50-00 Le=1

900

Fig. 9. Link between the filaments fully impregnated and the maximum load

5 Conclusions By comparison of the relative influence of the different matrices used in the study, the paper sheds a new light on the impact of the matrix properties and of the yarn impregnation on the pull-out behaviour of this yarn: • Two pull-out modes are encountered. The first mode is characterized by the slippage followed by the extraction of a number of filaments, it is mostly found for stronger matrices, and the second mode is characterized by the failure of most of the filaments, without slippage, mostly found for weaker matrices.

Reinforcement of Concrete with Glass Multifilament Yarns

981

• The matrix compressive strength influences the values obtained for the two mechanical parameters computed on the pull-out load/displacement curve: the maximum load for all the samples and the residual load for samples with filaments slippage. The average maximum load and residual load both increase with the compressive strength of the matrix. However, the fluidity of the matrix seems to have no significant influence on the pull-out results. • For the first mode with filaments slippage and extraction, a new visualization method is developed, using double resin impregnation and matrix dissolution, and it enables to conclude that the maximum load values can be explained by the number of fully impregnated filaments. The residual load, however, is not related to the extracted filaments.

References 1. Papanicolaou, CG.: 10 - Applications of textile-reinforced concrete in the precast industry. In: Triantafillou, T., Textile Fibre Composites in Civil Engineering, pp. 227–244. Woodhead Publishing, Cambridge (2016) 2. Kulas, C.: Actual applications and potential of textile-reinforced concrete in GRCA. In: Proceedings of an international conference, Dubai, 1–11 April 2016 3. Häußler-Combe, U., Jesse, F., Curbach, M.: Textile reinforced concrete-overview, experimental and theoretical investigations in Fracture mechanics of concrete structures. In: Proceedings of an international conference, Vail, 12–16 April 2004 4. Hegger, J., Will, N., Rüberg, K.: Textile reinforced concrete — a new composite material in advances in construction materials. In: Proceedings of an international conference, Stuttgart, pp. 147–156, July 2007 5. Xu, S., Krüger, M., Reinhardt, H.-W., Ožbolt, J.: Bond characteristics of carbon, alkali resistant glass, and aramid textiles in mortar. J. Mater. Civ. Eng. 16(4), 356–364 (2004) 6. Peled, A., Zaguri, E., Marom, G.: Bonding characteristics of multifilament polymer yarns and cement matrices. Compos. Part Appl. Sci. Manuf. 39(6), 930–939 (2008) 7. Butler, M., Hempel, S., Mechtcherine, V.: Modelling of ageing effects on crack-bridging behaviour of AR-glass multifilament yarns embedded in cement-based matrix. Cem. Concr. Res. 41(4), 403–411 (2011) 8. Banholzer, B.: Bond of a strand in a cementitious matrix. Mater. Struct. 39(10), 1015–1028 (2006) 9. Chudoba, R., Vořechovský, M., Konrad, M.: Stochastic modeling of multi-filament yarns. I. random properties within the cross-section and size effect. Int. J. Solids Struct. 43(3), 413– 434 (2006) 10. Vořechovský, M., Chudoba, R.: Stochastic modeling of multi-filament yarns: II. random properties over the length and size effect. Int. J. Solids Struct. 43(3), 435–458 (2006) 11. Chudoba, R., Konrad, M., Mombartz, M., Vorechovskỳ, M., Meskouris, K.: ‘Multiscale modeling of textile reinforced concrete within a consistent modeling framework’ in ‘Computation of Shell and Spatial Structures’. In: Proceedings of an international conference, Salzburg, June 2005 12. Zastrau, B., Lepenies, I., Richter, M.: On the multi scale modeling of textile reinforced concrete. Tech. Mech. 28(1), 53–63 (2008)

982

A.-C. Slama et al.

13. Banholzer, B., Brockmann, T., Brameshuber, W.: Material and bonding characteristics for dimensioning and modelling of textile reinforced concrete (TRC) elements. Mater. Struct. 39(8), 749–763 (2006) 14. Lepenies, I., Meyer, C., Schorn, H., Zastrau, B.: modeling of load transfer behavior of ARglass-rovings in textile reinforced concrete. Spec. Publ. 244, 109–124 (2007) 15. Banholzer, B., Brameshuber, W., Jung, W.: Analytical simulation of pull-out tests—the direct problem. Cem. Concr. Compos. 27(1), 93–101 (2005) 16. Banholzer, B., Brameshuber, W., Jung, W.: Analytical evaluation of pull-out tests—the inverse problem. Cem. Concr. Compos. 28(6), 564–571 (2006) 17. Hegger, J., Will, N., Bruckermann, O., Voss, S.: Load–bearing behaviour and simulation of textile reinforced concrete. Mater. Struct. 39(8), 765–776 (2006) 18. Cem-FIL® Roving 5325 - Owens Corning Composites [Internet]. https://www. owenscorning.com/composites/product/cem-fil-roving-5325 19. Ohno, S., Hannant, D.J.: Modeling the stress-strain response of continuous fber reinforced cement composites. Mater. J. 91(3), 306–312 (1994) 20. Homoro, O., Michel, M., Baranger, T.N.: Pull-out response of glass yarn from ettringite matrix: Effect of pre-impregnation and embedded length. Compos. Sci. Technol. 170, 174– 182 (2019) 21. Badanoiu, A., Holmgren, J.: Cementitious composites reinforced with continuous carbon fibres for strengthening of concrete structures. Cem. Concr. Compos. 25(3), 387–394 (2003) 22. Butler, M., Mechtcherine, V., Hempel, S.: Durability of textile reinforced concrete made with AR glass fibre: effect of the matrix composition. Mater. Struct. 43(10), 1351–1368 (2010) 23. Peled, A., Mobasher, B.: Properties of fabric-cement composites made by pultrusion. Mater. Struct. 39(8), 787–797 (2006) 24. Laws, V., Langley, A.A., West, J.M.: The glass fibre/cement bond. J. Mater. Sci. 21(1), 289– 296 (1986) 25. Vrijdaghs, R., di Prisco, M., Vandewalle, L.: Short-term and creep pull-out behavior of polypropylene macrofibers at varying embedded lengths and angles from a concrete matrix. Constr. Build. Mater. 147, 858–864 (2017)

Influence of Fibres Impregnation on the Tensile Response of Flax Textile Reinforced Mortar Composite Systems Giuseppe Ferrara1, Marco Pepe1,2(&), Enzo Martinelli1,2, and Romildo D. Tolêdo Filho3 1

3

Department of Civil Engineering, University of Salerno, Salerno, Italy [email protected] 2 TESIS srl, Fisciano, SA, Italy Civil Engineering Department, COPPE, Federal University of Rio de Janeiro, Rio de Janeiro, Brazil

Abstract. Textile Reinforced Mortar (TRM) composite systems as a technique to retrofit and reinforce existing structures represents, nowadays, an efficient application. The use of plant fibres textile as reinforcement, instead of the most employed industrial ones, resulted a promising solution as response to the sustainability criteria more and more required in the construction sector. However, some issues have been manifested as well related to the use of such natural reinforcements in cement- and lime-based matrices mainly lying in the fibre-tomatrix interaction and in the durability of both the overall composite and the reinforcement embedded within the mortar. In this context, the present study proposes an experimental activity aimed at investigating the efficiency of an impregnation treatment (by using styrene butadiene rubber latex) of flax fabrics on the mechanical behaviour of TRMs. The study confirms that the use of impregnated textile leads to an improvement of the overall behaviour of the composite system and paves the way for further investigations aimed at verifying the efficiency also in terms of durability. Keywords: Plant fibres


Composite
TRM
FRCM

1 Introduction The use of Textile Reinforced Mortar (TRM) composites as strengthening system for masonry structures in the last decade became a well-established reinforcement technique [1]. Such composites are performed by embedding high strength synthetic fibres in inorganic matrices, typically cementitious or lime-based mortars. The use of plant fibres, instead of the most conventional high strength synthetic textiles, as reinforcement in TRMs emerged as a smart solution to reduce the environmental impact of the strengthening system. Several studies focused the attention on this type of application by using jute, sisal, coir, flax, hemp fibres in mortar based composites [2–4]. These studies emphasised promising mechanical properties of plant fibres, and a good compatibility with mortars in terms of strength and bond behaviour. © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 983–990, 2021. https://doi.org/10.1007/978-3-030-58482-5_86

984

G. Ferrara et al.

However, also emerged some drawbacks mainly due to durability issues, a large deformability of the fibres at low strains, and variability of the properties of the textile due to non-standardised industrial production processes [5, 6]. Fibre treatments by means of impregnation procedures, came up as an efficient technique to address these issues. The application of resins on the textile external surface may lead to fibre stiffness increase and improvement of the adherence with the mortar [7–10]. In this context, the present study aims at investigating the mechanical behaviour of TRM composite performed by means of impregnated flax textile. With the purpose of analyse the influence of the treatment on the composite mechanical behaviour, two series of specimens, characterised by non-impregnated and impregnated flax textiles, were performed and tested in tension. The study, highlighting a much better behaviour of the coupons characterised by the impregnated textile, shows promising results and paves the way for further investigations aimed at optimising the adopted treatment and at analysing its durability performance.

2 Materials and Methods The textile consists of a bidirectional flax fabric (Fig. 1a) whose physical and mechanical properties are summarized in Table 1. The impregnated configuration (Fig. 1b) was obtained by coating the textile by means of a carboxylated styrene butadiene rubber latex, and by drying it for 24 h at a controlled temperature of 38 °C. The mechanical characterisation of the textile was carried out by means of tensile tests performed on non-impregnated and impregnated flax threads, constituting the elementary element of the fabric. The main mechanical properties are showed in Table 1.

Fig. 1. a) Flax textile; b) Impregnated flax textile; c) Mortar bending strength characterisation; d) Implementation of the TRM composite; e) TRM tensile test.

Influence of Fibres Impregnation on the Tensile Response

985

The matrix consists of a hydraulic lime-based mortar whose mechanical characterisation was carried out according to the EN 196-1 [11] (Fig. 1c). The mortar was characterised by a compressive strength of 9.81 MPa and a flexural strength of 4.33 MPa. The mechanical characterisation of the composite was carried out by means of tensile tests performed on TRM coupons. Two series of specimens were considered: – Non-impregnated Flax TRM: 5 specimens characterised by a non-impregnated flax textile; – Impregnated Flax TRM: 5 specimens characterised by the impregnated flax textile. The specimens were realized by alternating the first layer of mortar, then the textile ply, making sure the mesh voids to be fully penetrated by the mortar (Fig. 1d), and the second layer of mortar. The samples were characterised by a thickness of 7 mm, a width of 60 mm and a length of 500 mm, of which 300 mm in the middle represents the gauge length. The two edges of the specimens were clamped by gluing aluminium plates for a length of 100 mm. The gripping system was the clevis type (Fig. 1e). Tensile tests were carried out by means of universal testing machine with a load capacity of 10 kN, in displacement control (0.3 mm/min). Table 1. Physical and mechanical properties of the flax textile (mean values). Linear density n° threads/cm thread cross section Non-impregnated Young’s modulus strain to failure tensile strength Impregnated Young’s modulus strain to failure tensile strength

[Tex] [−] [mm2] [GPa] [%] [MPa] [GPa] [%] [MPa]

302 4.3 0.25 9.06 6.09 341 9.11 3.63 296

3 Results and Discussion Figure 2 shows the typical response of the TRM composite subjected to tensile load. Specifically, the tensile behaviour can be divided in three phases. The so-called Stage I represents the elastic response of the composite. Then, when the first crack in the matrix occurs, the Stage II begins: this stage is characterised by a development of cracks through the specimen length. Finally, the Stage III is characterised by a linear behaviour, mainly governed by the textile response, up to the rupture of the textile.

986

G. Ferrara et al.

Fig. 2. Load – displacement curve of a representative specimen.

An overview of the whole results obtained herein are summarized, in terms of Load-Displacement, in Fig. 3 and Fig. 4 for both Non-impregnated-Flax TRM and Impregnated-Flax TRM series, respectively.

Fig. 3. Load – displacement curves for non-impregnated Flax TRM specimens.

The tensile response of Non-impregnated Flax TRM specimens is characterised by a linear branch up to the occurrence of the first crack, and, in some cases a second crack occurred as well (see Fig. 3). The failure occurred due to the slipping of the textile through the mortar in the gripping edge of the specimen. On the other hand, the Impregnated-Flax TRM specimens show a response characterised by the development of several cracks (see Fig. 4), and by a drop of the load, corresponding to each crack

Influence of Fibres Impregnation on the Tensile Response

987

Fig. 4. Load – displacement curves for impregnated Flax TRM specimens.

Fig. 5. Average value of the main mechanical parameter for both non-impregnated and impregnated TRM series of specimens.

occurrence, much lower than the one observed in the reference series (i.e., Non-impregnated Flax TRM). A more comprehensive analysis of the results is proposed in Fig. 5 in which the results obtained from the two series are summarized in terms of stress and strain registered in correspondence of the key transition points between: the Stage I and the Stage II (i.e. r1 and e1), the Stage II and the Stage III (i.e., r2 and e2), as well as in terms of maximum stress and the corresponding strain (i.e., rmax and emax). Concerning the point of transition 1, it is clear that the average value of the stress attained in the reference series is higher than the one observed when the impregnated

988

G. Ferrara et al.

textile was adopted. Such experimental evidence may be due to a less participation to the strength of the non-impregnated textile in the elastic phase, causing a more uniform distribution of the stress within the mortar, letting it to achieve a higher value of the load. With respect to the point of transition 2 the figure highlights that the impregnated configuration is characterised by a lower value of the parameter e2. Such behaviour may be explained by referring to the tensile behaviour of the textile. As a matter of the fact, the non-impregnated textile is characterised by a less stiff response when subjected to low values of strain. Such phenomenon, typical of plant fibres, tends to disappear in the impregnated textiles, that are, in fact, characterised by a lower failure strain (see also Table 1). As a matter of fact, the use of impregnated fibres within the composite reduces its deformability in the Stage II. Moreover, Fig. 6 and Fig. 7 show the cracks development during the tensile test of a representative specimen of each series. It can be observed that the Impregnated-Flax TRM specimens are characterised by a much higher number of cracks with respect to the reference series. Specifically, for the Non-impregnated Flax TRM series, one or two cracks, characterized by an average width of 13 mm (ranging between 9 mm and 18 mm), were observed. Contrarily, the Impregnated Flax TRM series presented a significant higher number of cracks (from 7 to 9) with an average width of 2.8 mm (ranging between 2.6 mm and 3.1 mm). The crack development is mainly related to the matrix strength, textile stiffness, and textile-to-matrix bond capacity. Being the two series of specimens characterised by the same mortar, it can be asserted that the use of impregnated textile resulted in a textile stiffness, and in textile-to-matrix bond behaviour much satisfying of those occurred in the reference series. As a matter principle, the textile impregnation conferred to the material a significant improvement leading to a tensile response characterised by a

Fig. 6. Cracks development of a representative specimen of the series non-impregnated Flax TRM (spec. 3).

Influence of Fibres Impregnation on the Tensile Response

989

Fig. 7. Cracks development of a representative specimen of the series impregnated Flax TRM (spec. 1).

more distributed crack pattern through the specimen length, hence allowing it to work as an actual composite system. Moreover, this response led to a failure mode characterised by the rupture of the fibres, allowing to exploit the entire strength of the adopted textile.

4 Conclusions The study deals with the influence of the textile impregnation on the tensile behaviour a Flax TRM system. The main findings of the research are reported as follow: – the impregnation of the textile, although resulting in a slight reduction of strength, conferred to the textile a lower deformability; – the tensile response of the Non-Impregnated-Flax TRMs was characterised by the development of as maximum 2 cracks before the failure occurring due to the slipping of the textile through the mortar; – unlike the reference series, the Impregnated-Flax TRMs showed a tensile response characterised by the three stages behaviour, and by a failure due to the rupture of the textile; – the series of specimens reinforced by impregnated textile showed a crack pattern developed through the entire length of the specimen, allowing the material to work as an actual composite system. The study confirmed the potential in the use of plant fibres as reinforcement in TRMs, and showed that it is possible to improve its mechanical behaviour by means of fibres impregnation treatment. Moreover, it paves the way for further investigations

990

G. Ferrara et al.

aimed at defining the optimum amount of textile and at optimising the impregnation treatments, in order to improve even more the mechanical behaviour. Finally, further investigations are needed in order to study the influence of the impregnation treatment on the durability of the system, an aspect, the latter, of fundamental importance in view of long terms applications of the reinforcement system. Acknowledgements. The present study is part of the activities carried out by the Authors within the “SUPERCONCRETE Project (www.superconcrete-h2020.unisa.it) funded by the European Union’s Horizon 2020 Research and Innovation Programme under Grant Agreement No 645704 (H2020-MSCA-RISE-2014), whose financial support is gratefully acknowledged. The authors gratefully acknowledge the company INNOVATIONS s.r.l. for providing the materials tested in the experimental research presented in this paper.

References 1. Giacomin, G.: Innovative strengthening materials for the post-earthquake reconstruction of L’Aquila masonries. In: Proceedings of the 10th International Conference on Structural Analysis of Historical Constructions, SAHC 2016, LEUVEN, Belgium, 16–26 September 2016 2. Codispoti, R., Oliveira, D.V., Olivito, R.S., Lourenço, P.B., Fangueiro, R.: Mechanical performance of natural fiber-reinforced composites for the strengthening of masonry. Compos. Part B Eng. 77, 74–83 (2015) 3. Olivito, R.S., Codispoti, R., Cevallos, O.A.: Bond behavior of Flax-FRCM and PBO-FRCM composites applied on clay bricks: experimental and theoretical study. Compos. Struct. 146, 221–231 (2016) 4. Ghiassi, B., Razavizadeh, A., Oliveira, D.V., Marques, V., Lourenço P.B.: Tensile and bond characterization of natural fibers embedded in inorganic matrices. In: Proceedings of 2nd International Conference on Natural Fibers, Azores/Portugal, 27–29 April 2015 5. Olivito, R.S., Cevallos, O.A., Carrozzini, A.: Development of durable cementitious composites using sisal and flax fabrics for reinforcement of masonry structures. Mater. Des. 57, 258–268 (2014) 6. Cevallos, O.A., Olivito, R.S.: Effects of fabric parameters on the tensile behaviour of sustainable cementitious composites. Compos. Part B Eng. 69, 256–266 (2015) 7. Menna, C., Asprone, D., Durante, M., Zinno, A., Balsamo, A., Prota, A.: Structural behaviour of masonry panels strengthened with an innovative hemp fibre composite grid. Constr. Build. Mater. 100, 111–121 (2015) 8. de Carvalho Bello, C.B., Boem, I., Cecchi, A., Gattesco, N., Oliveira, D.V.: Experimental tests for the characterization of sisal fiber reinforced cementitious matrix for strengthening masonry structures. Constr. Build. Mater. 219, 44–55 (2019) 9. Mercedes, L., Gil, L., Bernat-Maso, E.: Mechanical performance of vegetal fabric reinforced cementitious matrix (FRCM) composites. Constr. Build. Mater. 175, 161–173 (2018) 10. Ferrara, G., Pepe, M., Martinelli, E., Toledo Filho R.D.: Influence of an impregnation treatment on the morphology and mechanical behaviour of flax yarns embedded in hydraulic lime mortar. Fibers 7(4), 30 (2019) 11. EN 196-1:1994 “Methods of testing cement – Part 1: Determination of strength” European committee for standardization

Experimental Investigation of Mechanical Properties of Smart Textile Reinforced Concrete Pipes Gozdem Dittel(&), Michelle Wangler, Bastian Maiworm, and Thomas Gries Institut fuer Textiltechnik of RWTH Aachen University, Aachen, Germany [email protected]

Abstract. Leakages in pipes results in a 35% loss of the total water supplied worldwide, which is a critical issue given the impact of climate change and global warming. Therefore, early leakage warning systems have to be developed in order to reduce the water losses occurring due to cracks and leakages in pipes. However, conventional pipes available today do not contain any integrated leakage detection mechanism. Hence, the proposed solution, of using conductive carbon fibres in the reinforcement as leakage sensors allowing for the fault and leakage determination, is being developed. This principle has paved the way for research into sustainable hybrid textile reinforced concrete (TRC) pipe systems. With the aim of realizing an industrial production method for TRC pipes, different grid-shaped textile reinforcement structures, with integrated sensory rovings, are developed for concrete pipes at the Institut fuer Textiltechnik (ITA) of RWTH Aachen University. This work forms the future basis of an automated pipe production. The aim of this study is to characterise the mechanical properties of these new age TRC pipes. For this purpose, lab scale TRC pipes with a length of l = 500 mm, an outer diameter of do = 300 mm and a wall thickness of d = 25 mm are casted using these smart hybrid textile reinforcement structures made by using alkali-resistant (AR) glass and carbon rovings. Thereafter, mechanical tests for compressive strength of the TRC pipes are carried out according to the DIN EN 1916 standards. The results are evaluated and compared with each other. Keywords: Textile reinforced concrete pipe compressive strength


Smart pipe
Peak value of

1 Introduction Due to the ever increasing global population, the need for building new settlements is constantly rising every year. Therefore, in order to meet this growing need, additional infrastructural materials and products are required. One such construction element is a reinforced concrete pipe, which is primarily used for the transportation of fluids. Reinforced concrete pipes are used in a variety of ways, including inlet and outlet pipes

© RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 991–1000, 2021. https://doi.org/10.1007/978-3-030-58482-5_87

992

G. Dittel et al.

to households and all other demands, including those wherein the transport of industrial fluids is required. In Germany, there exist approximately 600,000 km of sewage pipes, 45% of which are made using concrete material [1, 2]. According to the LAWA guidelines, sewage pipes are designed for a service life of 50-80 years and were mostly built in the 1950s [3]. Hence, given that these sewage pipes are nearing their end of life, frequent and more severe damages are being observed. The main reason of failure can be attributed to corrosion, root in-growth and deformations resulting from operational loading. Damage caused by corrosion accounts for 10% of all damage types [4]. The high moisture content inside and outside of the sewage pipes causes corrosion of the reinforcement and thus the corrosion of the concrete. During corrosion, the steel reinforcement reacts with the elements in its environment and this chemical reaction causes an increase in volume of the steel reinforcement and as a result spalling of the concrete structure occurs. The assembly and maintenance of water pipelines involves high costs in terms of transport, handling, logistics and monitoring. There are different methods to control damages of sewage pipes, which depend on the use and size of the sewage pipes. A common method is the wireless TV-inspection, in which a camera mounted robot is remotely driven through the pipe and the recorded images are then evaluated using a software algorithm. However, it should be noted that only 5% of the sewage pipes are regularly monitored, because continuous monitoring is cost intensive and time-consuming [5]. The geometry of steel reinforced concrete pipes varies depending on the static load requirements. The cross-sectional shapes of sewage pipes are defined in DIN 4045 as circular, egg-shaped or mouth-shaped cross-sections [3]. The diameter of steel reinforced concrete pipes ranges from DN 300 to DN 1200 [6]. For protection of the steel reinforcement from corrosion, a minimum thickness of concrete material must be deposited over the reinforcing bars. The required wall thickness depends on the nominal diameter and ranges from 71 to 141 mm. In order to maintain the high wall thicknesses, a large amount of cement is required for concrete production. During cement production, lime is burned in kilns to produce hardened cement paste and is therefore responsible for 4-8% of the global CO2 emission [7]. This in turn is an additional challenge that presents itself and has adversarial effects on the environment. In order to find a solution for the problems mentioned above, smart textile reinforced concrete pipes are realized in the project “Smart Pipe: Development of a textile reinforced pipe system with integrated monitoring functions”. The hybrid textile reinforcement consists of alkali-resistant (AR) glass rovings and carbon rovings. Due to the non-corrosive fibre-based reinforcement system, a large amount of the concrete material deposition requirements over the reinforcement structure can be reduced by up to 70%. Thus, the wall thickness of these TRC pipes is scaled down by 70% and thus a significant decrease in the CO2 emission per concrete pipe is achieved. At the same time, the electrically conductive carbon rovings act as a leakage detection system in the concrete structure. Through electrical measurements, it is possible to detect hairline cracks in these TRC based concrete pipes. The use of Carbon rovings both as reinforcement as well as leakage sensors enables an integrated early warning system which has never been developed before.

Experimental Investigation of Mechanical Properties

993

At the Institute for Textile Technology (ITA) of RWTH Aachen University, various prototypes of textile reinforced concrete pipes with a length of 50 cm and an outer diameter of 30 cm have been realized. Two different production processes are developed to implement the optimal textile reinforcement for the smart TRC pipe; biaxial warp knitted textile grid and wound textile grid. The optimal smart textile reinforcement has to ensure a leakage detection over the entire pipe surface and at the same time the required structural performance. In this study, the mechanical properties of the smart TRC pipes are determined by measuring the peak value of compressive strength.

2 Properties of the Reinforcement Materals and the Concrete Mixture Both types of textile reinforcement consist of alkali-resistant (AR) glass and carbon rovings coated with a styrene-butadiene rubber (SBR) polymer (50% material concentration) through a dispersion process. The type of the coating material is Lefasol VL 90/1 and the added cross linker is Lefasol VP 4-5LF, both produced by Lefatex Chemie GmbH. The PES 167f48 yarns are used in the biaxial warp knitted structure as knitting thread and in the wound structure as wrapping yarn to ensure an oval shaped cross section for the AR-glass and carbon rovings. AR-glass rovings of Cem-FIL® 5325 produced by Owens Corning for cement based matrices are used for this study. The type of the carbon rovings used is SIGRAFIL® C50 T024 EPY 382 produced by SGL Group. The physical and mechanical properties of roving materials are listed in Table 1.

Table 1. Properties of the roving materials. Property Filament Diameter [µm] Filament tensile strength [MPa] Modulus of elasticity [GPa] Elongation at break [%] Linear density of roving [tex] Density [g/cm3]

AR-glass roving Carbon roving 19 7 1700 5000 72 270 2.4 1.9 2400 1600 2.68 1.81

The composition of the fine grain concrete material used as a matrix has a compressive strength of 74.2 N/mm2 and a flexural strength of 7.6 N/mm2 [8]. The properties of this material are given in Table 2.

994

G. Dittel et al. Table 2. Composition of the fine grain concrete mixture [8]. Content Quantity [kg/m3] Cement CEM I 52,5 N 490 Fly ash 175 Silica fume (Elkem Microsilica® 940 U) 35 Quartz flour 500 Sand 0,2–0,6 mm 713 Water 280 Superplasticizer 7

3 Production of the TRC Pipe Specimens The fibre based reinforcing structures are manufactured using two textile production methods: the biaxial warp knitting process and the filament winding process. In the following sections, the production methodology for each textile grid is presented. Both textile structures are given in Fig. 1.

sensory carbon rovings

ARglass rovings 30 mm

100 mm

Fig. 1. Biaxial warp knitted (left) and wound (right) textile reinforcing structures.

3.1

Biaxial Warp Knitted Textile Reinforcing Grid Structure

The optimal geometry, material design and coating procedure of the biaxial warp knitted reinforcement for the smart concrete pipe is determined in the preliminary research [9–14]. AR-glass rovings are used as the main reinforcement base and eight of them were replaced by using four carbon roving pairs in the warp direction. The carbon fibres are providing the reinforcement and are also functional as leakage sensors. The mesh opening in warp and weft direction is approximately 8 mm. The knitting type is counterlaid tricot. As discussed above, to ensure the required load transmission between the individual fibre filaments of the roving and to prevent the reinforcement from shifting, the textile grid is coated with styrene-butadiene rubber (SBR) with 50% solid content. The two dimensional textile grid is shaped into a cylindrical form with a

Experimental Investigation of Mechanical Properties

995

diameter of d0 = 270 mm before the concrete matrix material is added. The overlap area of the two edges is 50 mm wide. The four sensory carbon rovings are positioned in the longitudinal direction at every 90° across the entire pipe surface [15, 16]. 3.2

Wound Textile Reinforcing Grid Structure

The wound reinforcing grid structure consists of AR-glass rovings in the longitudinal direction and two parallel wound sensory carbon rovings in the circular direction. All rovings are entwined using the same knitting thread as in the biaxial warp knitted structure to achieve a comparable elliptical cross-sectional roving shape. The AR-glass rovings are positioned longitudinally on a winding core with a diameter of d0 = 270 mm and thereafter are coated. A sensory carbon roving couple is wound on top of the AR glass rovings in the transverse direction by a single filament machine and is also coated to stabilize the reinforcement cage. The mesh opening in both directions is approximately 8 mm. The winding process is developed to eliminate the overlap area in the biaxial warp knitted structure and to ensure a leakage detection across the full-surface area of the concrete pipe [17]. 3.3

TRC Pipe Production

The cylindrical reinforcement is positioned in a mould with an inner diameter of di = 250 mm and outer diameter of do = 300 mm using spacers to ensure the spatial accuracy of the reinforcement. Concrete is then poured into the vertical mould by applying vibration. The smart TRC pipes are demoulded after a 24-hour cure cycle and placed in a water bath for 6 days at room temperature. The specimens are stored for further 21 days at room temperature to complete the hydration process and achieve the final strength of the concrete. Three smart TRC pipes with a length of l = 500 mm and a wall thickness of d = 25 mm are produced for each textile configuration. Figure 2 shows the schematic design and actual prototype images of both configurations (warp knitted and wound) of the smart TRC pipes [15–17].

4 Determination of the Peak Value of Compressive Strength In order to test the mechanical properties of the textile reinforced concrete pipes, peak compressive strength tests are carried out at IKT- Institut für Unterirdische Infrastruktur gGmbH, Gelsenkirchen according to DIN EN 1916 [18]. This particular standard is an industrial norm for the testing of steel reinforced concrete pipes. In principle, the smart TRC pipes will be subjected to the same loading conditions. 4.1

Experimental Setup

The loading unit applying the required force is located centrally above the test specimen. It comprises of an elastomeric uniform beam which is gradually lowered to the pipe at a predefined rate. Thus, applying a steadily increasing load without any impact

996

G. Dittel et al. sensory carbon rovings

ARglass rovings

20 cm

Fig. 2. Smart TRC pipes with biaxial warp knitted (right) and wound (left) textile grid reinforcement.

or shock. The pipe movement constrainers are positioned at an angle of 30° below the test piece (see Fig. 3). The test is carried out as it is proposed in the methodology in Annex C of DIN EN 1916. The concrete pipe must achieve a minimum apex compressive strength Fn, which corresponds to its nominal diameter and strength class. The concrete pipe should withstand a crack force of Fc = 0.67 * Fn, whereby in the concrete tensile zone, a surface crack must not exceed a crack width of 0.3 mm over a length of 300 mm or more. As soon as the first cracks appear on the textile reinforced concrete pipes, it is measured and documented. Thereafter, the test load is further applied, and the crack widths are continuously monitored and measured. The test is carried out until a deformation of 25 mm is reached.

Experimental Investigation of Mechanical Properties

997

elastomeric beam

TRC pipe

pipe supports

30 cm

Fig. 3. Test setup for compressive loading to failure measurement in accordance to the Annex C in DIN EN 1916

4.2

Results and Analysis

Due to the low wall thickness of 25 mm the loading speed, specified in DIN EN 1916, is reduced by 50% to 0.1 kN/s. Usually, steel reinforced concrete pipes are loaded up to the initial crack and then the load is increased further to a stabilized longitudinal crack of 0.3 mm. However, the pipes tested in this study already developed a crack width of 0.3 mm at the initial crack. Hence, further loading in force control mode was aborted. A change was made to the displacement control mode and the load was continued at a speed of 1.0 mm/s. In this way, the maximum loading case could be achieved. In the pipes with biaxial reinforcement structure, after a force of 12 kN, the first crack is observed at an average load of 23.2 kN/m in the area of the impost, which after a few seconds reaches a crack width of 0.3 mm. The pipes are further loaded up to a deformation of 25 mm. The average maximum load recorded is 35.2 kN/m. In the pipes with wound reinforcement structure, the first crack occurs at an average load of 23.4 kN/m in the impost and shortly thereafter a crack width of 0.3 mm is observed. Up to the average maximum force of 42.8 kN/m, further cracks occur in the area of the impost. In both test series, cracks occur in the apex area and concrete spalling occurs in the sole area. The results of both test series are shown in Fig. 4.

5 Discussion and Conclusions In principle, crack widths of 0.3 mm already determined during the initial crack are be classified as ‘too high’. Here, there are already risks regarding the load-bearing capacity and also the tightness of the pipes. This means that the textile reinforcement merely

G. Dittel et al. 50

50

40

40

30

23,2

23,4

20 10 0 Warp knitted structureWound structure

M aximum load [kN/m]

Cracking Load [kN/m]

998

42,8 35,2

30 20 10 0 Warp knitted structure Wound structure

Fig. 4. Comparison of mechanical properties for TRC pipes with bi-axial warp knitted and wound reinforcement

prevents the complete collapse of the pipes. With respect to the crack pattern, it should be noted that the crack distribution in pipes with wound reinforcement appears to be more favorable as compared to pipes with biaxial warp knitted reinforcement. The pipes with wound reinforcement demonstrated a significantly increased crack behavior with smaller crack widths than the pipes with biaxial warp knitted reinforcement (see Fig. 5).

Fig. 5. Average cracking load and maximum load of the pipes with different textile reinforcement

The resulting crack pattern of the wound reinforcement structure is comparable to the crack pattern of steel reinforced concrete pipes. Cracks occur in the apex and bottom area as well as in the area of the impost. Due to the applied load, high moment and bending loads occur in the apex. After exceeding the concrete tensile strength, the first crack occurs. The concrete spalling in the sole is caused due to the high shear stresses between the two support points. This is also a typical behaviour in peak compressive strength tests for steel reinforced concrete pipes. Due to the load transfer from apex and sole into the impost, the tensile strength of the textile-reinforced concrete is exceeded, and a full joint is formed there. This becomes visible by a main crack. Until the maximum force is reached, a few secondary cracks occur in the impost area.

Experimental Investigation of Mechanical Properties

999

The first cracks occurred during the determination of the peak value of the compressive strength represent the mechanical properties of the fine grained concrete mixture. For TRC structures, it is unusual that the cracks reach the maximum allowed crack width of 0.3 mm after only a few seconds of the load application. One reason for this is the low wall thickness of the tested TRC pipes. The pipes were not sized to match their steel reinforced counterparts. The wall thicknesses of equivalent steel reinforced pipes range between 70 to 80 mm which is three times thicker than the tested TRC pipes. Thus if a smart TRC pipe is designed for operational loads comparable to their steel reinforced counterparts, a significantly different cracking pattern will be observed. It is very unusual that reinforced concrete pipes only show one crack in the impost area, as in the test of the biaxial reinforcement structure. This manner can be explained by a possible inhomogeneous force absorption of the textile reinforcement. Twodimensional biaxial warp-knitted grids are shaped to a three dimensional cylindrical form, which built an overlapping area of 5 mm. This change in the reinforcement crosssection generates a weak point due to the force transmission. The continuous roving placement in the wound reinforcement structure ensures a homogeneous force absorption by the rovings and thus an increase of 21% in the maximum load absorption. The goal of this paper was to perform a destructive test on textile reinforced concrete pipes for determining the maximum loading conditions and to compare two different reinforcement typologies, biaxial warp knitted structure and wound structure. Given the results of the current experiment, it can be conclusively stated that wound reinforcement structures are more suitable to concrete pipes than bi-axial warp knitted structures. Concurrently, from the manufacturing point of view, the textile winding process corresponds to the manufacturing process of the steel reinforcement for concrete pipes. Future work will consist of sizing the pipes to defined operational loads. The wound reinforcement structure will be further developed regarding the textile architecture such as the mesh opening size, the pre-stress of the rovings and the optimal position of the different roving materials in the grid structure. A comparison to steel reinforced pipes will be accurately made. Acknowledgements. The authors would like to thank the Federal Ministry of Education and Research (BMBF) - Germany for funding the project “SmartPipe - Development of a textile reinforced pipe system with integrated monitoring functions” and the Project Management Agency Karlsruhe (PTKA) for the project coordination. We would like to acknowledge that the compressive tests were carried out at the IKT - Institut für unterirdische Infrastruktur gGmbH, Gelsenkirchen.

References 1. Breitkopf, A.: Länge des Kanalnetzes in Deutschland im Jahr 2016. Accessed 24 Jan 2020. https://de.statista.com/statistik/daten/studie/152743/umfrage/laenge-des-kanalnetzes-in-deut schland-im-jahr-2007/. (20th December 2018)

1000

G. Dittel et al.

2. Stein, D., Stein, R.: Instandhaltung von Kanalisationen, Aufbau und Randbedingungen von Kanalisationen, Rohrwerkstoffe und Ausbildung der Rohrverbindungen, Übersicht über Werkstoffe und Rohrverbindungen (1998). Accessed 24 Jan 2020. https://www.unitracc.de/ know-how/fachbuecher/instandhaltung-von-kanalisationen/aufbau-und-randbedingungenvon-kanalisationen/rohrwerkstoffe-und-ausbildung-der-rohrverbindungen/uebersicht-ueberwerkstoffe-und-rohrverbindungen 3. Stein, D., Stein, R.: Instandhaltung von Kanalisationen, Aufbau und Randbedingungen von Kanalisationen, Querschnittsformen und–abmessungen (1998). Accessed 24 Jan 2020. https://www.unitracc.de/know-how/fachbuecher/instandhaltung-von-kanalisationen/aufbauund-randbedingungen-von-kanalisationen/querschnittsformen-und-abmessungen 4. Statista Research Department: Abwasserkanäle - Verteilung der Schäden in Deutschland 2013, 7th November 2019. Accessed 24 Jan 2020. https://de.statista.com/statistik/daten/ studie/456149/umfrage/verteilung-der-festgestellten-schaeden-an-abwasserkanaelen-indeutschland/ 5. Bütow, E.: Umweltbundesamt, November 2001. Accessed 24 Jan 2020. https://www. umweltbundesamt.de/sites/default/files/medien/publikation/long/2052.pdf 6. Betonwerk Bieren GmbH, Datenblatt_1.3_Stahlbetonrohre_ohne_Fuss. Accessed 24 Jan 2020. https://betonwerk-bieren.de/produkte/?gclid=EAIaIQobChMIwu3r0dmc5wIVhIxRC h2lnQNsEAAYASAAEgLuH_D_BwE 7. Kretschmer, A.: Klimabilanz der Zementindustrie, 25th March 2019. Accessed 24 Jan 2020. https://www.chemietechnik.de/klimabilanz-der-zementindustrie/ 8. Brockmann, T.: Mechanical and fracture mechanical properties of fine-grained concrete for TRC structures. In: Gross, C.U., (ed.) Advances in Construction Materials, pp. 119–129. Springer, Berlin (2007) 9. Perry, G., Dittel, G., Gries, T., Goldfeld, Y.: Mutual effect of textile binding and coating on the structural performance of TRC beams. Journal of Materials in Civil Engineering (2020, in print) 10. Perry, G., Dittel, G., Gries, T., Goldfeld, Y.: The effect of textile configuration on the monitoring capabilities of smart carbon-based TRC elements to detect water infiltration, submitted for publication (2020) 11. Quadflieg, T., Goldfeld, Y. Dittel, G., Gries, T.: New age advanced smart water pipe systems using textile reinforced concrete. In: 15th Global Conference on Sustainable Manufacturing, Technion-IIT, Haifa, Israel, 25–27 September 2017 12. Goldfeld, Y., Perry, G., Dittel, G., Gries, T.: Development of a textile reinforced pipe system with integrated monitoring function (SmartPipe). In: German-Israeli Cooperation in Water Technology Research Status Conference 2019, Dresden, Germany, 24th–25th September 2019 13. Goldfeld, Y., Perry, G.: Electrical characterization of smart sensory system using carbon based textile reinforced concrete for leakage detection. Mater. Struct. 51(17), 1–17 (2018) 14. Goldfeld, Y., Perry, G.: A-R glass/carbon-based textile reinforced concrete elements for detection water infiltration within cracked zones. Struct. Health Monitor. (2018). https://doi. org/10.1177/1475921718808223 15. Dittel, G., Heins, K., Gries, T.: Development and design of smart textile reinforcement for concrete pipes. In: Proceedings of the ACI Convention, Ohio, USA, 20–24 October 2019 16. Heins, K.: Entwicklung und Realisierung von Konzepten zur Herstellung intelligenter textilverstärkter Betonrohre, Aachen (2019) 17. Maiworm, B.: Filament wound smart textile reinforcement for leakage detection in concrete pipes, Aachen (2019) 18. DIN1916: Rohre und Formstücke aus Beton, Stahlfaserbeton und Stahlbeton; Deutsche Fassung EN 1916:2002, April 2003

UHPFRC, SHCC and ECC

Full-Scale Construction Test for Improvement of RC Void Slab Bridges Using UHPFRC – Part 1: Experimental Test Plan Tohru Makita1(&), Yuji Watanabe2, Shuji Yanai2, and Hirokazu Kitagawa1 1

Central Nippon Expressway Company Limited, Nagoya, Japan [email protected] 2 Kajima Corporation, Tokyo, Japan

Abstract. Kajima Corporation and NEXCO Central started a joint research and development (R&D) project to develop a method for improving existing RC void slab bridges using cast-in-situ UHPFRC in 2016 and as the final investigation of the R&D project a full-scale construction test was conducted. This paper is the first one of the two papers dealing with the construction test. UHPFRC mix used in the construction test is called AFt-UHPFRC because it is characterised by its matrix densified by controlled ettringite (AFt) formation. UHPFRC was produced in a batching plant built in a testing field of that production capacity is 3.0 m3 per hour. A slab-on-ground (SOG) structure modelling top part of RC void slab bridge decks was built in the testing field and UHPFRC was cast on top of the SOG structures where UHPFRC was transported and placed by a wheel loader and spread/compacted/finished by newly developed construction equipment. Keywords: UHPFRC


Strengthening
Durability enhancement
Bridges

1 Introduction In Japan, the first expressway was put in service in 1963 and since then about 9,500 km long expressways have been built so far. Central Nippon Expressway Company Limited (NEXCO Central) operates and manages about 2,000 km long expressways in the central part of Japan, 40% of which have been in service for more than 40 years and average service period of the expressways is now approximately 30 years. Although inservice period of 30 years is not so long considering the fact that expected service life of civil engineering structures is over 100 years, recent years have seen growing number of damaged and deteriorated bridges in expressways of Japan and the number of bridges in such conditions will be getting larger in the coming decades. In order to address this issue, extensive expressway renewal project was launched in 2015. In the project, replacement and improvement (strengthening and durability enhancement) of bridge decks and girders are planned to be undertaken among others and total of approximately one trillion yen will be spent for the renewal of NEXCO Central’s expressway structures during 15 years. © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 1003–1011, 2021. https://doi.org/10.1007/978-3-030-58482-5_88

1004

T. Makita et al.

Reinforced concrete (RC) void slab bridge (Fig. 1) is the most common type of concrete bridges in expressways managed by NEXCO Central, constituting approximately 25% of the NEXCO Central’s bridges. About 95% of the RC void slab bridges were designed according to old design codes and strengthening is necessary if those bridges don’t fulfil today’s traffic load requirements. In addition, waterproof membranes had not been applied as standard to the NEXCO Central’s bridges until 1998 and concrete bridge decks built before that time have been seriously exposed to deleterious environmental influences (e.g. de-icing salts). Top surface of old RC void slab bridge decks are often found to be chloride contaminated reaching the depth of top steel rebars. The above-mentioned conditions motivated the improvement of damaged and deteriorated RC void slab bridges in the expressway renewal project where chloride contaminated top surface concrete of RC void slab bridge decks is removed and refilled with cementitious material followed by application of waterproofing membrane on the top surface. 12,500 mid-span

over pier

900

600

2,205

7,740

2,555 [unit: mm]

Fig. 1. Cross sections of an RC void slab

As the refilling material, normal strength concrete (NSC) or fibre reinforced concrete (FRC) made of NSC containing 100 kg/m3 (1.3vol.-%) of steel fibres has been conventionally used. However, in order to improve the load carrying capacity of RC void slab bridges (Fig. 2a), the deck thickness often has to be increased with additional rebars (Fig. 2b), resulting in increase of superstructure self-weight. Consequently, it is usually necessary to strengthen substructures especially for seismic action. Furthermore, expansion joints need to be replaced for increased height of bridge decks. By using Ultra-High Performance Fibre Reinforced cement-based Composites (UHPFRC) instead of conventional NSC or FRC, unwanted increase of the deck thickness can be avoided due to its excellent mechanical properties: replacement of top surface concrete of bridge decks with UHPFRC is sufficient for improvement of the load carrying capacity (Fig. 2c). Moreover, it is not necessary to apply waterproofing membrane on top surface of bridge decks because UHPFRC function as protective layer due to its very low permeability. In order to develop a method for improving existing RC void slab bridges using cast-in-situ UHPFRC, Kajima Corporation and NEXCO Central started a joint research and development (R&D) project in 2016. All laboratory tests planned in the R&D project were completed at the end of 2018 and as the final investigation of the R&D project a full-scale construction test was carried out.

Full-Scale Construction Test for Improvement of RC Void

Void RC

FRC

UHPFRC

Void

Void RC

RC

(a)

1005

(b)

(c)

Fig. 2. Schematic comparison of cross sections of (a) RC void slab and RC void slab strengthened with (b) FRC and (c) UHPFRC

This paper is the first one of the two papers dealing with the construction test, presenting the experimental test plan where motivation/objectives and program of the test are detailed.

2 Motivation and Objectives of Full-Scale Construction Test In laboratory tests, UHPFRC was manufactured with small mixer (100 litre twin-shaft forced action mixer or pan type mixer) and poured manually in specimen moulds. Besides, UHPFRC specimens were cured in perfectly controlled constant temperature and humidity rooms. However, when UHPFRC is applied to real bridges, in particular multi-span and large bridges which comprise a majority of expressway bridges in Japan, UHPFRC is supposed to be manufactured with larger mixer and placed/spread/compacted/finished with construction equipment for work efficiency enhancement. Therefore, it is necessary to understand material properties and behaviour of UHPFRC that is manufactured and cast in the same condition as real construction work. In order to meet this need, the full-scale construction test was planned.

3 UHPFRC Densified by Controlled Ettringite Formation (AFt-UHPFRC) UHPFRC mix used in the construction test is characterised by its matrix densified by controlled ettringite (AFt) formation; thus, it is called AFt-UHPFRC. Microstructure of the AFt-UHPFRC matrix is basically formed by decreasing water/binder ratio using spherical pozzolan particles and superplasticiser and packing ultrafine particles optimally. In addition to that, numerous needle-shaped ettringite crystals of 1 to 2 µm length (Fig. 3) fill micropores of hydration structure together with inert and

Fig. 3. Needle-shaped ettringite crystals

1006

T. Makita et al. Table 1. Composition of AFt-UHPFRC

Component

Remarks

Mass [kg/m3] 927 360

Portland cement Premixed material

pozzolanic materials, ettringite formation additives crushed sand, dmax  2.5 mm 3.0 vol.%, l = 15 mm, d = 0.2 mm

Sand 905 Steel fibre 235.5 Superplasticiser 36 Shrinkage reducing 12.9 admixture Defoaming agent 6.4 Water* 195 * including water in superplasticiser

W/B = 0.152

reactive fine fillers. The AFt-UHPFRC mix composition is shown in Table 1. The premixed material contains silica fume (mean particle diameter of 0.2 lm), fly ash (mean particles diameter of 3 lm) and additives allowing ettringite to form in a controlled manner. The mix has crushed sand of maximum diameter of 2.5 mm and 3.0vol.-% of steel fibres with length of 15 mm and diameter of 0.2 mm. 20 18

10-60μm 17.0

1-10μm

16

0.5-1μm

Porosity (%)

100-500nm

13.5

14

12.3

12

12.2

10-100nm 10.6

10

6-10nm 9.28

3-6nm

7.94

8

6.07

6

4.38

4

2.66

2 0

18 MPa

44 MPa

64 MPa

73 MPa

15 h

18 h

21 h

24 h

99 MPa 115 MPa 133 MPa 165 MPa 193 MPa 215 MPa 2 days

20 °C

3 days

7 days

28 days

20 h

5 years

85 °C Steam

Cast in-situ

Fig. 4. Evolution of porosity and pore size distribution of AFt-UHPFRC

Figure 4 shows a result of long-term measurement of porosity and pore size distribution of the AFt-UHPFRC cured under room conditions (20°C and 60% RH). As favourable effect of ettringite crystal growth, porosity and pore size gradually reduces as time proceeds: in 28 days the porosity decreases from 17.0% to 6.07% and the mode

Full-Scale Construction Test for Improvement of RC Void

1007

of pore size distribution changes from 10–100 nm pores to 3–6 nm pores. As a reference the porosities and pore size distributions of steam-cured Aft-UHPFRC and AftUHPFRC cured outdoors are indicated in Fig. 3, from which it is understood the porosity of the Aft-UHPFRC is lowered significantly in five years even without steam treatment.

4 Test Program 4.1

Slab-on-Ground Structure Modelling Top Part of an RC Void Slab Bridge Deck

Since building a mock-up RC void slab bridge is impractical, a slab-onground (SOG) structure modelling top part of RC void slab bridge decks was built in a testing field where NSC of 30 MPa strength was used (Fig. 5). The width and length of the SOG structure were 5 m and 62 m, respectively. The SOG structure was divided into six zones and construction conditions were varied for each zone (Table 2). Top surface of the SOG structure was roughened with Fig. 5. SOG structure for construction test water jets for interfacial bonding between UHPFRC and concrete. Assuming that top 10 cm surface concrete of RC void slab bridge decks is replaced with UHPFRC, it was planned that 10 cm thick UHPFRC layer is cast on top of the SOG structure except zone 4 where UHPFRC layer thickness was increased to 15 cm in order to investigate the influence of layer thickness on the compactability of thixotropic UHPFRC. Steel rebars were arranged on top of the SOG structure, which, however, did not model fully the arrangement of steel rebars of real bridges; yet, densely arranged steel rebars were partly modelled in order to check if thixotropic UHPFRC properly fill narrow rebar spacing and gaps between rebars and concrete substrate. On the basis of the expressway design manual [1] in that combined slope of expressway roads is prescribed to range from 2% to 9%, top surface of the SOG structure was sloped transversely either 2% or 9% (Fig. 6). An epoxy bonding adhesive was applied to some part of top surface of the SOG structure to investigate the performance of the adhesive for interfacial bonding between UHPFRC and concrete. The amount of the adhesive was 1.2 kg/m2.

1008

T. Makita et al. Table 2. Test parameters and points of interest in construction test

Zone Construction conditions UHPFRC Slope Adhesive* thickness (%) (kg/m2) (cm) 0 10 2 1.2 1 2 1.2

Mix Points of interest Compaction Fibre Energy (vol.-%) (Trial) Minimum Maximum

–

3 4

3.0 2.0 3.0

15

5 10 9 * applied to part of top surface

250 74 74 135 165 102 100165 135

8×150=1,200

3.0

5,500 5,000 135 135 135 165 165 165

• Compactability at various compaction energy • Fillability • Interfacial bonding Assume loss of slump flow (limit of fluidity/fillability) Influence of thickness on compactability Slope torelance and workability

14×150=2,100 φ13

250 74 74 100 102

[unit :mm]

φ13 φ32 φ13

φ13

120 80 200

650 570 80

3.0

φ13 φ13

φ13

φ13

φ13

Fig. 6. Cross section of the slab-on-ground structure transversely sloped 9%

4.2

Production and Casting of UHPFRC

A batching plant was built at the testing field (Fig. 7a) and a twin-shaft forced action mixer with a capacity of 3.0 m3 was used for producing UHPFRC. Except sand and steel fibre, UHPFRC components needed for each batch were measured and put in the mixer automatically; sand and steel fibre were measured and put in the mixer manually (Fig. 7b). 1.5 m3 of UHPFRC was manufactured for each batch, taking 30 min. Steel fibre content was 3 vol.-% except for UHPFRC cast in part of zone 2 whose fibre content was reduced to 2 vol.-% in order to see the influence of fibre content on the consistency of fresh UHPFRC. Thixotropy was conferred on UHPFRC by adding mineral-based inorganic powder the amount of which was varied depending on the top surface slope of the SOG structure and the consistency of fresh UHPFRC. UHPFRC was transported and placed by a wheel loader (Fig. 7c) and spread/compacted/finished by new construction equipment that was developed by remodelling a concrete paving

Full-Scale Construction Test for Improvement of RC Void

(a)

(b)

(c)

(d)

1009

Fig. 7. Production and casting of UHPFRC: (a) batching plant, (b) steel fibres manually put in the mixer, (c) wheel loader and construction equipment, (d) application of sheet membrane

spreader (Fig. 7c). Moving speed of the construction equipment was kept constant to be 0.5 m/min. Vibration frequency of screed mounted on the construction equipment was changed so as to understand proper vibration for compacting thixotropic UHPFRC. In order to prevent plastic shrinkage cracking after casting, sheet membrane curing was applied on top of UHPFRC immediately after the surface finish (Fig. 7d) and the sheet membrane was removed the next day. 4.3

Evaluation of Construction Test

The construction test was evaluated during and after the construction test by performing visual observation, measuring the behaviour of UHPFRC and the SOG structure and examining properties of UHPFRC. Table 3 lists UHPFRC property tests performed as part of the evaluation of the construction test.

1010

T. Makita et al. Table 3. List of UHPFRC property test

Property Fresh state properties Temperature Flow Air content Mechanical Compressive strength properties Tensile cracking strength Tensile strength Behaviour of Deformation composite structure Interfacial bond strength Transport properties Air permeability

Test method JIS JIS JIS JIS JIS – – –

B 7411 R 5201 A 1128 A 1108 A 1113

Remarks

Modulus of elasticity Splitting tensile test Direct tension test Direct tension test Mould strain gauge embedded in UHPFRC In-situ test

JIS A 6909 ASTM C 1583 SIA 262/1 Annex Torrent method E Alternate immersion test in Chloride diffusivity JSCE recommendation salt solution [2] Porosity and pore size Mercury distribution intrusion method

During the construction test, vibration of the SOG structure was measured using piezoelectric accelerometers installed on top surface of the SOG structure in order to investigate the influence of frequency of screed vibration and UHPFRC layer thickness on vibration transmission which was considered to index the degree of compaction of thixotropic UHPFRC. In addition, shrinkage deformation of UHPFRC cast on top of the SOG structure was measured with mould strain gauges, which was started immediately after casting of UHPFRC. Fresh UHPFRC properties were checked per batch by measuring temperature, slump flow and air content and optimal quality control of UHPFRC on site is also investigated. In addition to material-related measurement and testing, in order to understand the labour cost, the basic production rate of all works performed for casting UHPFRC, from producing to curing, was examined. Slope tolerance of thixotropic UHPFRC was evaluated qualitatively by visual observation. After the construction test, the compressive strength, tensile cracking strength and tensile strength of UHPFRC were tested where test specimens were fabricated using UHPFRC from one batch of each day (the construction test was four days long). Interfacial bonding strength between UHPFRC and concrete was also tested by performing in-situ pull-off tests at two locations for each zone. Protective function of UHPFRC was evaluated by investigating air permeability, chloride diffusivity and porosity/pore size distribution of UHPFRC. Air permeability testing was carried out at one location per zone using Torrent method on site. Chloride diffusivity was determined according to Japanese recommendations for design and construction of

Full-Scale Construction Test for Improvement of RC Void

1011

UHPFRC [2] where alternate immersion tests in salt solution were conducted on a UHPFRC core drilled from zone 0. Porosity and pore size distribution were measured for UHPFRC cores drilled from each zone except zone 0 and 3 by using a mercury intrusion porosimetry.

5 Conclusions This paper presents the program of the full-scale construction test conducted as the final investigation of an R&D project to develop a method for improving existing RC void slab bridges using cast-in-situ UHPFRC. In the construction test, a batching plant was built in a testing field and 1.5 m3 of UHPFRC was produced per batch every 30 min. UHPFRC was cast on top of a slab-on-ground structure modelling top part of RC void slab bridge decks where UHPFRC was transported and placed by a wheel loader and spread/compacted/finished by newly developed construction equipment. Construction conditions were varied and the construction test was evaluated during and after the test by measurement of specimen behaviour and investigation of material properties. Evaluation results of the construction test are described in the second one of the two papers dealing with the construction test [3].

References 1. Central Nippon Expressway Company Limited: Design Manual Volume 4: Road Geometry – Mainline Geometry, Nagoya, Japan (2017) 2. Japan Society of Civil Engineers: Recommendations for design and construction of ultra high strength fiber reinforced concrete structures – draft, Tokyo, Japan (2004) 3. Watanabe, Y., Yanai, S., Makita, T., Kitagawa, H.: Full-scale construction test for improvement of RC void slab bridges using UHPFRC – part 2: test results. In: RILEM-fib X International Symposium on Fibre Reinforced Concrete, BEFIB2020, Valencia, 21–23 September 2020

Full-Scale Construction Test for Improvement of RC Void Slab Bridges Using UHPFRC – Part 2: Test Results Yuji Watanabe1(&), Shuji Yanai1, Tohru Makita2, and Hirokazu Kitagawa2 1

2

Kajima Corporation, Tokyo, Japan [email protected] Central Nippon Expressway Company Limited, Nagoya, Japan

Abstract. In Japan, a growing number of damaged and deteriorated bridges on expressways have been seen in recent years, and the number of bridges in such conditions is set to increase in the coming decades. In order to address this issue, an extensive expressway renewal project was launched in 2015. Research and development have begun upgrading bridge decks by utilizing UHPFRC where either overlaying UHPFRC or replacing top surface concrete with UHPFRC is conducted, which leads to an increase in bridge deck stiffness. A full-scale UHPFRC casting test was carried out at the final stage of the research and development project. This paper is the second of the two papers regarding the UHPFRC casting test. Several evaluations of the construction test are presented regarding the compaction of thixotropic UHPFRC, the bonding properties of UHPFRC with existing slabs, and the transport properties of UHPFRC. Keywords: UHPFRC


Strengthening
Durability enhancement
Bridges

1 Introduction Over the last decade, damaged and deteriorated bridges have been increasingly observed in Japan’s expressways. Typical examples of those damaged and deteriorated bridges include reinforced concrete (RC) void slab bridges in which top surface concrete is deteriorated due to de-icing agents. Bridge deck overlays are most often used on existing bridges when their decks require rehabilitation. Overlays using Steel Fibre Reinforced Concrete (SFRC) can also be employed as a preventative measure for deterioration on existing decks that are in good structural condition; however, there is a possibility that the top surface of bridge decks may perhaps deteriorate again due to the repeated loading of heavy vehicles, severe environmental conditions and improper construction methods. Moreover, the SFRC overlay thickness could reach to more than 100 mm in order to attain the required load bearing capacity. Furthermore, it is necessary to strengthen the other members of bridges in order to carry the increased selfweight of overlaid bridge decks.

© RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 1012–1021, 2021. https://doi.org/10.1007/978-3-030-58482-5_89

Full-Scale Construction Test for Improvement of RC

1013

Following the above-mentioned situation, the development of a new bridge deck overlay method has begun for the expressway renewal project, where the strengthening of bridge decks is carried out. The new method uses Ultra-High Performance Fibre Reinforced cement-based Composites (UHPFRC) as overlay material [1]. UHPFRC has very low permeability against liquid/gas and high resistance to freezing and thawing action. In addition, UHPFRC is known for its high strength and high elastic modulus. By using UHPFRC instead of SFRC, the overlay thickness can be made thinner, and an increase of the concrete bridge deck thickness can even be made unnecessary by replacing a certain depth of the top surface concrete with UHPFRC (Fig. 1) [2]. Moreover, UHPFRC functions, such as the protective layer and application of waterproofing material, is not needed. Thus, the application of UHPFRC to concrete bridge decks for upgrading is an efficient and effective method because the increase of the load bearing capacity and enhancement of the durability are achieved simultaneously.

(a) RC void slab and RC void slab strengthened with (b) SFRC and (c) UHPFRC

Fig. 1. Schematic comparison of cross sections

Fig. 2. Full-scale construction test

Table 1. Test conditions Zone Construction UHPFRC thickness (mm) 0 100 1 2

conditions Slope Adhesive* (%) (kg/m2) 2

1.2 1.2 –

3 4

150

5

100

9

* applied to part of top surface

Compaction energy (Trial) Minimum Maximum

Mix Fibre (vol. %) 3.0 3.0 2.0 3.0

3.0

Points of interest

• Compactability at various compaction energy • Fillability • Interfacial bonding Assume loss of slump flow (limit of fluidity/fillability) Influence of thickness on compactability Slope torelance and workability

1014

Y. Watanabe et al. Table 2. List of UHPFRC tests

Property

Test method

Remarks

Fresh state properties

JIS JIS JIS JIS

Modulus of elasticity

Temperature Flow Air content Mechanical Compressive properties strength Tensile cracking strength Tensile strength Behaviour of Shrinkage composite structure strain Interfacial bond strength Transport Air properties permeability Chloride diffusion coefficient Porosity

B 7411 R 5201 A 1128 A 1108

This paper – 〇 〇 〇

JIS A 1113 Splitting tensile test New method [4] (cylinder) Static tensile tests

〇

New method [4] Static tensile tests

〇

–

Mold strain gauge embedded 〇 in UHPFRC On site 〇

JIS A 6909 ASTM C 1583 SIA 262/1 Torrent method Annex E Alternate immersion test in JSCE recommendation salt solution Mercury intrusion method

Core specimen

– –

〇

Table 3. Mix proportion of AFt-UHPFRC Component Portland cement Premixed materials

Mass (kg/m3) 927 360

Sand 905 Steel fibre** 235.5 Superplasticiser 36 Shrinkage reducing 12.9 admixture Defoaming agent 6.4 Water* 195 * including water in superplasticizer **not including in the unit volume

Remarks

Pozzolanic material, ettringite formation additives Crushed sand, dmax < 2.5 mm 3.0 vol.%, d = 0.2 mm, l = 15

W/B = 0.152

Full-Scale Construction Test for Improvement of RC

1015

The authors have conducted various detailed laboratory studies after confirming the feasibility of the overlay method and identifying issues of the construction method through small-scale construction tests. However, when UHPFRC is applied to real bridges, in particular multi-span and large bridges that constitute most of expressway bridges in Japan, UHPFRC is supposed to be manufactured with a larger mixer and placed/spread/compacted/finished with construction equipment for the enhancement of work efficiency. Therefore, it is necessary to understand material properties and the behaviour of UHPFRC that is manufactured and cast under the same conditions as real construction. In order to meet this need, a full-scale construction test was planned and carried out (Fig. 2) [2]. This paper is the second of the two papers dealing with the construction test and presents evaluation results of the construction test.

2 Evaluation Results of Construction Test The outline of the experimental plan and a slab-on-ground (SOG) structure is as shown in the first paper “PART1: TEST PLAN” [2]. The SOG structure was divided into six zones, and construction conditions were varied for each zone (Table 1). Table 2 shows material testing that is mainly performed after the construction test. In this paper, results of (1) fresh state property tests, (2) mechanical property tests, (3) deformation measurement, (4) vibration/compaction energy measurement, (5) bond strength and (6) porosity measurement are presented. 2.1

Fresh State Properties

The UHPFRC mix used in the test is characterized by its matrix, which is densified by controlled ettringite (AFt) formation; thus, it is called AFt-UHPFRC [3]. Table 3 shows the mix proportion of AFt-UHPFRC. The fresh state properties of AFt-UHPFRC were adjusted according to the slope of the SOG slab and the compaction energy of the paving machine, and a flow value suitable for thixotropic AFt-UHPFRC was determined. Figure 2 and Fig. 3 show the results of the mortar flow test performed in compliance with JIS R 5201. It was found that when the flow value of UHPFRC falls within a range of 200 ± 25 mm, it is slope tolerant. UHPFRC temperature was varied depending on atmospheric temperature. The lowest was recorded to be 20.5 °C in the morning, and the highest was recorded to be 31.5 °C at noon. Air content in UHPFRC was almost constant, being approximately 4 vol.% where the maximum and minimum value was 4.4 vol.% and 3.0 vol.%, respectively. 2.2

Mechanical Properties

Table 4 shows the strength test results of specimens cured at an outdoor environment as with the SOG structure. At the age of 28 days, the compressive strength reached over 150 MPa, and the cracking strength reached over 8 MPa. It is notable that the compressive strength reached over 120 MPa even at seven days. Rapid hardening property is conferred to the AFt-UHPFRC by controlled ettringite formation, and the property is considered favorable for construction with time constraints.

1016

Y. Watanabe et al.

In order to investigate the tensile property of the UHPFRC, direct tension tests were performed. Figure 4 shows specimen geometry and the test set-up configuration. The specimen was 400 mm long and 100 mm wide with varying thickness (dog-bone shaped) to make fractures occur within the 100 mm long tapered central part of the specimen. The thickness of the central and end parts of the specimen were 40 mm and 100 mm, respectively; moreover, there were 90 mm long transitional zones between the central and end parts. The test set-up used for the tests was developed in the previous study [4]. Stress deformation relationships obtained from the static tensile tests are shown in Fig. 4.

Fig. 3. Thixotropic formulation (Flow)

Fig. 4. Fresh state property (Flow)

The averages of elastic limit strength and ultimate tensile strength were 12.1 MPa and 12.8 MPa, respectively. In the average stress-deformation curve of UHPFRC, strain-hardening is hardly observed. This is probably because the UHPFRC specimens didn’t have sufficient amount of fibres allowing for strain-hardening. This result is different from the result of the direct tension tests using specimens fabricated from UHPFRC produced by a small mixer and cured in perfectly controlled constant temperature and humidity rooms. It is also possible that the method of sampling the specimen affected the fibre orientation. This discrepancy will be investigated further. Prestressing bar (φ23)

100 60

100

90

Jig 2 Jig 3

40

60

90

400 100

Jig 1

[unit: mm]

Fig. 5. Static tensile tests (Fibre 3.0 vol. %)

Full-Scale Construction Test for Improvement of RC

1017

Fig. 6. Strain measurement results (shrinkage strain)

2.3

Deformation

Figure 5 shows UHPFRC strain measurement results. Strain measurement commenced immediately after casting. The contraction of UHPFRC was caused by autogenous shrinkage and dry shrinkage. When comparing the strain values of the UHPFRC cast on top of the SOG structure with the ones of prism specimens, the former is smaller than the latter. This is thought to be because contraction of the UHPFRC was restrained more significantly by concrete substrates and steel rebar (higher degree of restraint) and mitigated by creep of the UHPFRC. The strain values and evolution of the UHPFRC layers on top of the SOG structure were the same regardless of thicknesses between 100 mm and 150 mm.

Table 4. Strength test results (Fibre 3.0 vol.%) Property

Material age 7 days 28 days Compressive strength (MPa) Ave. 131 164 Max. 137 177 Min. 120 144 Elastic modulus (GPa) Ave. 41 45 Cracking strength Splitting tensile test (cylinder) Ave. 7.6 8.7 (MPa) Max. 9.3 9.9 Min. 5.8 6.4 Static tensile tests [4] Ave. Max. Min. Tensile strength Static tensile tests [4] Ave. (MPa) Max. Min. -

91 days 184 196 166 47 9.3 10.1 8.5 -

257 days 189 198 181 46 11.3 11.8 10.7 12.1 12.6 11.6 12.8 14.6 11.7

1018

2.4

Y. Watanabe et al.

Compaction Energy

Vibration of the SOG structure was measured for the purpose of understanding that the construction equipment can compact thixotropic UHPFRC so that proper interfacial bonding between the UHPFRC and concrete is achieved. The accelerometer was installed at the end of the SOG structure farthest from the screed vibrator so that the energy transmitted from the vibrator was minimum. Figure 6 shows the measurement results of acceleration. In the SOG structure, vibration frequency of vibrating screed was increased by the same percentage as the UHPFRC layer thickness: a 100 mm thick UHPFRC layer was screeded and consolidated by 2,000 (Minimum) vibrations per minute, while a 150 mm thick UHPFRC layer was screeded and consolidated by 3,000 (Maximum) vibrations per minute. Acceleration of the specimen with a 150 mm thick UHPFRC layer was larger than the specimen with a 100 mm thick UHPFRC layer from which it might be said that the compaction energy produced by screed vibration transmits to UHPFRC-concrete interface irrespective of UHPFRC layer thickness.

Fig. 7. Acceleration measurement result

Fig. 8. Pull-off test of the UHPFRC cast on top of the SOG structure

2.5

Bond Strength

The interface bonding strength between UHPFRC and concrete was assessed on site using the pull-off test method specified in ASTM C 1583 (Fig. 7). A partial core was drilled through the bonded material and into the substrate material; furthermore, a steel pull-off disc that is approximately 80 mm in diameter was bonded at the desired test location.

Full-Scale Construction Test for Improvement of RC

1019

Table 5 shows the bond test results. Failure occurred mostly at the substrate concrete irrespective of application of adhesive on the UHPFRC-concrete interface. In addition, the bond strength determined from all tests was almost the same (about 2.8 MPa) regardless of the UHPFRC thickness, tested location in the SOG structure (center or edge) and imposed compaction energy. 2.6

Porosity of Core Specimen

Figure 8 shows measurement results of porosity and pore size distribution of the top, medium and bottom part of the UHPFRC layer. The specimens were made from cores taken from the SOG structure 91 days after UHPFRC casting and cured at 20 °C and 60% RH for about 270 days. Porosity of all specimens was lower than approximately 6% due to the effect of ettringite crystal growth over time (Fig. 9).

Table 5. Results of bond tests

In all zones, the porosity of the bottom part of the UHPFRC layer was smaller than that of the top part of the UHPFRC layer, which is brought about by the decrease of pores between 10 to 100 nm in diameter. This is probably explained by the fact that more extensive hydration occurred in the bottom part of the UHPFRC layer than in the top part because the SOG structure was subjected to long-term drying conditions after sheet curing at the early age, and water was lost at the top part of the UHPFRC layer.

1020

Y. Watanabe et al.

Fig. 9. Porosity evolution of UHPFRC (After 270 days)

From the porosity of UHPFRC of Zone 1, 2 and 5 where UHPFRC layer thickness is 100 mm, it might be said that the higher the flow value of UHPFRC or the frequency of the vibrating screed, the lower the UHPFRC porosity will be. However, further investigation is necessary regarding this tendency.

3 Conclusions This paper is the second of the two papers dealing with the UHPFRC casting test where test results are presented. Test results are summarised as follows: • The fresh state properties of UHPFRC were adjusted according to the slope of the GOG structure and the compaction energy of the construction equipment and a flow value (200 ± 25 mm) suitable for thixotropic UHPC was determined. • At the age of 28 days, the compressive strength reached over 150 MPa, and the cracking strength determined by splitting tensile tests reached over 8 MPa. Averages of the cracking strength and ultimate tensile strength were determined to be 12.1 MPa and 12.8 MPa, respectively, from the direct tension test. • The shrinkage strain of UHPFRC cast on top of the SOG structure was smaller than that of the prism specimen. This is thought to be because contraction of the UHPFRC was restrained more significantly by concrete substrates and steel rebar (higher degree of restraint) and mitigated by creep of the UHPFRC. • Acceleration of the SOG structure with a 150 mm thick UHPFRC layer was larger than that with a 100 mm thick UHPFRC layer. It might be said that the compaction energy of vibrating screed transmits to UHPFRC-Concrete interface irrespective of UHPFRC layer thickness. • Failure of specimens of UHPFRC-concrete interfacial bonding tests occurred mostly in the substrate concrete irrespective of application of adhesive on the interface. All bond strengths were almost the same (about 2.8 MPa) regardless of the UHPFRC thickness and compaction energy.

Full-Scale Construction Test for Improvement of RC

1021

• The porosity of the bottom part of the UHPFRC layer was smaller than that of the top part of the UHPFRC layer. It might be said that the higher the flow value of UHPFRC or the frequency of the vibrating screed, the lower the UHPFRC porosity; however, further investigation is necessary for this tendency.

References 1. Brühwiler, E.: Structural UHPFRC: welcome to the post-concrete era. In: Proceedings of the First International Interactive Symposium on Ultra-High Performance Concrete, Des Moines, Iowa, 18–20 July 2016 2. Makita, T., Watanabe, Y., Yanai, S., Kitagawa,H.: Full-scale construction test for improvement of RC void slab bridges using UHPFRC – PART 1: experimental test plan. In: RILEM-fib X International Symposium on Fibre Reinforced Concrete, BEFIB2020, Valencia, 21–23 September 2020 3. Watanabe, Y., Ichinomiya, T., Yanai, S., Iriuchi-jima, K., Suhara, K.: Development of cast-inplace method of ultra high strength fiber rein-forced concrete. In: Proceedings of the Fifth International Conference on Construction Materials, CONMAT15, Whistle, British Columbia, 19–21 August 2015 4. Makita, T., Watanabe, Y., Yanai, S., Ichinomiya, T.: Upgrading of existing bridge decks using UHPFRC densified by ETtringite Formation (AFt-UHPFRC): preliminary investigation. AFGC-ACI-fib-RILEM International Symposium on Ultra-High Performance FibreReinforced Concrete, UHPFRC 2017, Montpellier, 2–4 October 2017

Influence of Fiber Type on the Tensile Behavior of High-Strength Strain-Hardening Cement-Based Composites (HS-SHCC) During and After Exposure to Elevated Temperatures Iurie Curosu(&), Sarah Burk, Marco Liebscher, and Viktor Mechtcherine Institute of Construction Materials, Faculty of Civil Engineering, Technische Universität Dresden, Dresden, Germany [email protected]

Abstract. The paper summarizes selected results of an extensive experimental investigation, in which high-strength strain-hardening cement-based composites (HS-SHCC) made with different high-performance polymer fibers were investigated in terms of mechanical behavior under and after exposure to elevated temperatures of 105 °C, 150 °C and 200 °C. Besides the ultra-high molecularweight polyethylene (UHMWPE) fibers, which are commonly used in HS-SHCC, high-modulus poly(p-phenylene-2,6-benzobisoxazole) (PBO-HM) fibers have been analyzed, since they exhibit a considerably higher temperature resistance in comparison to UHMWPE fibers. In contrast to the expectations, the in-situ and residual tension experiments at temperatures of up to 150 °C showed that the highstrength SHCC reinforced with UHMWPE fibers yielded considerably superior performance and less pronounced decrease of the mechanical properties compared to the composites made with PBO-HM fibers. Furthermore, the SHCC made with UHMWPE fibers showed a significant recovery after being cooled down, while the SHCC made with PBO-HM fibers exhibited a limited recovery; the degradation was proportional to the temperature increase. The 200 °C treatment led to brittle failure of both composites with dramatically reduced tensile strength and with low recovery after specimen cooling in the residual experiments. Keywords: SHCC
UHMWPE fiber
PBO-HM fiber
Elevated temperature
Tension tests

1 Introduction Strain-hardening cement-based composites (SHCC) represent a novel type of fiber reinforced concrete with a notably high tensile ductility, which results from the formation of multiple fine cracks under increasing tensile load [1]. Despite their numerous advantageous features, the range of possible applications of SHCC is restricted by the high sensitivity of their tensile strength and ductility to elevated temperatures. This limitation is mainly determined by the polymer fibers suitable for SHCC, which are made of polyvinyl alcohol (PVA) [2] or ultra-high molecular weight polyethylene (UHMWPE) [3]. © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 1022–1033, 2021. https://doi.org/10.1007/978-3-030-58482-5_90

Influence of Fiber Type on the Tensile Behavior of HS-SHCC

1023

Various studies were performed to evaluate the residual mechanical properties of PVA-SHCC after exposure to elevated (  300 °C) and high temperatures (>300 °C). These investigations also aimed to analyze the mitigation effect of the PVA fibers on the explosive spalling of the composites at temperatures of up to 800 °C [4–11]. It was shown that the residual tensile ductility of PVA-SHCC reduces already at 100 °C, while the residual tensile strength reduces considerably at temperature of 200 °C and higher due to fiber melting. The most pronounced reduction of tensile strength and ductility occurs between 200 °C and 300 °C, which corresponds to the range in which the fibers suffer the most pronounced degradation. At temperatures above 400 °C, not only the complete decomposition of fibers but also the pronounced deterioration of the matrix by thermally induced cracks, decomposition of the hydrated phases, increase in porosity, etc. lead to a brittle behavior in tension and compression with dramatically reduced mechanical strength. Above 600 °C, the thermal deterioration is similar to that of concrete [4], while at temperatures of 800 °C severe spalling occurs. As opposed to the studies on residual properties, only few in-situ investigations on PVA-SHCC have been reported [12–14]. These studies showed that already at temperatures of 60 °C the first crack stress and tensile strength of PVA-SHCC decrease, but ductility increases due to the thermally reduced stiffness of the fibers, which leads to wider crack openings. At 100 °C, the ductility reduces dramatically, while above 150 °C the composites show strain-softening behavior only. At temperatures higher than 150 °C, the crack bridging of the fibers vanishes completely. No published studies are known so far on the effect of elevated temperatures on SHCC reinforced with UHMWPE (short: PE) fibers. While PE fibers exhibit excellent mechanical properties and high chemical stability in cement-based environments, their low melting point of approximately 150 °C [15, 16] may impose even more stringent limitations on their applicability for SHCC compared to PVA fiber. Various researchers proposed using additionally steel fibers to improve the residual behavior of SHCC under high temperatures [10, 11, 17]. While the melting of the polymer fibers provides pathways for vapored water to escape, steel fibers may improve the ductility at failure localization. However, compared to high-performance polymer fibers, the steel fibers do not ensure a pre-peak tensile ductility in cementitious matrices typical for SHCC. Moreover, the addition of steel fibers may have an adverse effect on the fresh-state properties of SHCC and may affect their applicability by lamination or spraying. In a previous research by the authors it was demonstrated that such high-performance polymer fibers as PBO (poly(p-phenylen-2,6-benzobisoxazol) and para-aramid exhibit desirable mechanical and geometrical properties for highstrength SHCC, yielding composites with an enhanced first crack stress, tensile strength and considerably reduced crack width compared to those reinforced with PE fibers [3]. Moreover, according to the producers, the decomposition temperature of the PBO and aramid fibers is as high as 500 °C and 650 °C, respectively, which makes them promising for SHCC exposed to elevated temperatures [18, 19]. With the purpose of a detailed assessment of the performance of these fibers in highstrength SHCC subjected to elevated temperatures, an extensive experimental investigation was performed involving analytical and mechanical experiments at the composite level as well as on plain cementitious matrix and single fibers. The paper at hand aims to present briefly some representative results from the mentioned study. Emphasis is put on

1024

I. Curosu et al.

the composites’ tensile behavior during and after exposure to temperatures of up to 200 °C. Besides the reference PE fibers, the high-modulus PBO (PBO-HM) fibers are discussed in this paper, since they are also representative for the other temperature resistant polymer fibers, i.e. as-spun PBO (PBO-AS) and para-aramid.

2 Materials 2.1

Fibers

The PE fibers Dyneema® SK62 from DSM, the Netherlands, are gel spun, multi-filament fibers with high tensile strength and low elongation at break [15]; see Table 2. These polyolefin fibers exhibit excellent resistance to acids, alkalis and most other chemicals, including water. Due to their nonpolar molecular structure, they are hydrophobic and yield weak interfacial bonding towards water-based systems like hardened cement paste. The PE fibers melt at temperatures of approximately 150 °C and, according to Liu and Yu [16], the critical temperature for their safe use is around 70 °C. The alternative to PE fiber presented in this work is the high-modulus p-phenylene2,6-benzobisoxazole (PBO-HM) fiber, known under the brand name Zylon®, produced by Toyobo, Japan. The PBO fibers have high resistance to various organic solvents, acids and bases [18, 20] and a high decomposition temperature [20–23]. Furthermore, they have a very high tensile strength, more than twice that of the Dyneema PE fibers, and a very high tensile modulus of elasticity [18]; see Table 1. As opposed to the PE fibers, the PBO fibers do not melt, but decompose, and they exhibit a weak hydrophilicity, which leads to a considerably stronger fiber-matrix bond compared to PE [3]. Table 1. Mechanical, geometrical and physical properties of the fibers under investigation [15, 18]. Producer DSM Brand Dyneema® Fiber type UHMWPE Average diameter* [µm] 20 Length [mm] 6 3 Density [g/cm ] 0.97 Tensile strength [MPa] 2500 Tensile modulus of elasticity [GPa] 80 Elongation at break [%] 3.5 Decomposing temperature [°C] n.a. Coefficient of linear thermal expansion [10–6 1/K] −12 Melting temperature [°C] 150

Toyobo Zylon® PBO-HM 13 6 1.56 5800 270 2.5 650 −6 n.a.

Influence of Fiber Type on the Tensile Behavior of HS-SHCC

2.2

1025

Cementitious Matrix

The pronounced hydrophobicity and the high tensile strength of PE fibers determine their optimal crack bridging behavior in high-strength rather than in normal-strength cementitious matrices [3, 24–28]. Depending on their composition, high-strength SHCC can yield tensile and compressive strength values comparable to those of steel fiber reinforced UHPC, but with considerably higher tensile strain capacity prior to failure localization [3, 27]. The high-strength cementitious matrix presented in the paper at hand is identical to that analyzed in the previous study by the authors, in which the reinforcing performance of PE, PBO and aramid fibers was investigated [3]. It has a high cement content and a high amount of silica fume as partial cement replacement; see Table 2. Furthermore, it has a relatively small amount of quartz sand as fine aggregates and no additional binders or fillers. The fine-grained nature of the matrix is imposed by the small diameters of the reinforcing fibers and should facilitate proper fiber dispersion on one hand and a low fracture toughness of the matrix on the other. Table 2. Composition of the high-strength SHCC under investigation. Components CEM I 52.5 R-SR3/NA Silica fume Quartz sand 0.06–0.2 mm Superplasticizer Glenium ACE 460 Water UHMWPE fiber (2% by vol.) PBO-HM fiber (2% by vol.)

Content [kg/m3] 1460 292 145 35 315 20 – – 31

To compensate for the low water-to-binder ratio and negative effect of the fibers on the workability of SHCC, a high dosage of superplasticizer is used. Note that adequate fresh-state properties of SHCC are essential for an appropriate homogeneity, fiber dispersion and robustness of their mechanical properties in hardened state [29]. In the paper at hand, the SHCC under investigation will be named according to the reinforcing fiber, i.e. M-PE and M-PBO-HM, in which M stays for matrix.

3 Experimental Program 3.1

Applied Temperature Treatments

The treatment conditions in the presented study were defined in accordance with the previous in-situ studies [12–14] by taking into consideration the melting temperature of PE fibers and the higher temperature resistance of the PBO and aramid fibers. To achieve an effective 1 h treatment of the specimens at maximum temperature, benchmark measurements were performed on heated SHCC specimens with embedded

1026

I. Curosu et al.

thermoelements (temperature sensors). The specimens with thermoelements were positioned in the furnace on an isolating pedestal and had no contact with any conductive elements. Figure 1 shows the treatment profile of a 105 °C experiment.

120

Temperature [°C]

100 80 60 40 specimen core gauge portion

20

furnace 0 0

20

40

60 80 Time [min]

100

120

Fig. 1. Temperature profile in the electric furnace and in the specimen core in a 105 °C experiment.

As shown in Fig. 1 based on a 105 °C control test, the temperature evolution in the specimen yields a considerable delay compared to the increase of the air temperature in the electric furnace and the phase prior to reaching maximum temperature in the specimen core only slowly converges to the plateau of the air temperature. The time needed to reach the target temperature inside the specimens was approximately 65 min for 105 °C, 85 min for 150 °C and 105 min for 200 °C. Thus, the total treatment duration with a heating rate of 5 K/min was chosen to be 120 min for 105 °C, 150 min for 150 °C, and 180 min for 200 °C. 3.2

Testing Configuration

The uniaxial tension experiments were carried out in an Instron hydraulic testing machine on dumbbell shaped SHCC specimens with a geometry commonly used by the authors in previous studies [3]. Note that the in-situ experiments did not imply testing inside the electric furnace. For ensuring a short duration of the testing process after specimen extraction from the furnace, the SHCC specimens were gripped mechanically in specially fabricated elements made of heat resistant stainless steel; see Fig. 2. The gripping only ensured axial fixation and imposed almost no restriction to specimen rotation. Given the short testing duration of approximately 3 min for each specimen, it was assumed that no negative effects of temperature gradients and no significant specimen cooling would affect the tensile behavior of the composites. The electric furnace was positioned in the immediate vicinity of the testing machine as shown in Fig. 2. Upon reaching the target temperature in the specimens and the required treatment duration, the first specimen was extracted from the electric furnace, weighed,

Influence of Fiber Type on the Tensile Behavior of HS-SHCC

1027

mounted in the gripping elements, and tested. The same procedure was carried out for four subsequent specimens, which were tested within a time span of approximately 15 min, this being the difference between the treatment duration of the first and last tested specimens in a series.

a)

b)

cameras for DIC specimen

furnace

Fig. 2. a) Testing machine with the electric furnace and DIC equipment and b) specimen mounted in the gripping elements.

Upon the extraction of the last (usually fifth) specimen to be tested, the furnace was shut down, the door was left ajar and the specimens intended for residual experiments were left to cool down naturally in the furnace. The cooled down specimens were tested in the next day in the same testing configuration. The main reason for testing the heated specimens outside of the electric furnace (which is otherwise suitable for in-situ tests in the Instron machine) was to enable optical measurements of specimen deformation, crack formation and their evaluation using Digital Image Correlation (DIC). Moreover, such a configuration presented a significant advantage of allowing series of multiple experiments for one parameter variation, which was an important aspect considering the large number of investigated material parameters and temperatures in the overall study. For the optical measurements, all the specimens were sprayed with speckle pattern using heat resistant paint. The specimens were first provided with a black coat and a cloud of silver dots was subsequently sprayed on top. A stereo-camera system VIC-3D from Correlated Solutions was used for the optical measurements. At the beginning of every test series, the system was calibrated and reference experiments were performed

1028

I. Curosu et al.

at room temperature. The DIC evaluation of the specimens’ tensile behavior assumed defining two virtual calipers on the sides of the resolved surfaces in the gauge portion of the specimens as shown in Fig. 3. The strains were calculated as caliper elongation, averaged for two calipers, and used for plotting the stress-strain behavior of the composites under investigation.

Fig. 3. Virtual calipers positioned on the DIC resolved surface for deriving the global elongation (strain) of the specimen in the gauge length.

4 Results and Discussion Figures 4 and 6 show representative stress-strain curves obtained in in-situ and residual experiments on M-PE and M-PBO-HM, respectively. The influence of increased temperatures on the tensile strength and ductility of the composites is illustrated comparatively in Figs. 5 and 7. Note that the used gripping elements had the disadvantage of inducing pronounced stress concentrations and premature failure of some specimens in the gripping region. The corresponding specimens were not evaluated in terms of mechanical properties, which can be seen by the varying number of presented data points in Figs. 5 and 7. The temperature treatment at 105 °C did not yield a significant effect on the tensile strength of M-PE, whereas the corresponding strain capacity (strain at peak load) yielded a marked increase compared to the non-treated specimens; see Fig. 4a and 5b. This is a result of a reduced fiber stiffness leading to wider crack openings prior to failure localization. Furthermore, besides the reduction in fiber stiffness, the temperature of 105 °C seems to cause a moderate decrease in matrix strength, which led to a lower first crack stress, thus, to a higher strain-hardening modulus and to a more pronounced multiple cracking. Based on optical analysis of representative specimens, the number of cracks in the gauge portion of M-PE at 105 °C was approximately 83, at room temperature it was approximately 56, while at 150 °C it was 40. The average crack width increased with temperature from approximately 45 µm at 20 °C to 100 µm at 150 °C. Whereas at 150 °C the composites still showed a ductile strain-hardening tensile behavior, at 200 °C M-PE exhibited brittle failure with a dramatically reduced tensile strength, this being the reason why the in-situ curve was not included in Fig. 4a.

Influence of Fiber Type on the Tensile Behavior of HS-SHCC

M-PE in-situ

8

Stress [MPa]

Stress [MPa]

6

4 2

room temperature 105°C in-situ 150°C in-situ

0 0 (a)

M-PE residual

8

6

1029

2

4

6 8 Strain [%]

10

4 2

105°C residual 150°C residual 200°C residual

0

12

0 (b)

2

4 6 8 Strain [%]

10

12

Fig. 4. Representative (a) in-situ and (b) residual stress-strain curves of M-PE at different temperature treatments.

Comparing the residual and in-situ tensile behavior of M-PE treated at 105 °C, it seems that the fibers can partly regain their stiffness, as indicated by the equal tensile strength but reduced elongation capacity.

8

M-PE tensile strength in-situ residual

6 4 2 0 0 (a)

10 Strain at peak load [%]

Tensile strength [MPa]

10

8

M-PE tensile ductility in-situ residual

6 4 2 0

25 50 75 100 125 150 175 200 Temperature [°C]

0 25 50 75 100 125 150 175 200 (b) Temperature [°C]

Fig. 5. (a) In-situ and residual tensile strength values of individual M-PE specimens and (b) corresponding strains at peak load.

At 150 °C the residual tensile strength recovered considerably, see Figs. 4 and 5a. Furthermore, even after 200 °C treatment, M-PE showed some recovery in terms of tensile strength, while failure exhibited a weak softening; see residual curves in Figs. 4 and 5. It seems that the melting fibers are confined inside their channels, which enables

1030

I. Curosu et al.

a partial but beneficial recovery after specimen cooling after 150 °C and even after 200 °C treatment. Whereas protruding fibers could be observed on the fracture surfaces of specimens treated at 150 °C, the specimens treated at 200 °C showed smooth fracture surfaces with no visible fibers. In addition, the lateral surfaces of the latter showed a spider web of fine cracks, demonstrating the severe degradation of the matrix. As opposed to the previous study [3], the tensile strength and ductility of M-PBOHM in the current investigation were lower and the composites yielded a lower robustness of their mechanical properties. This was be traced back to the rotatable specimen gripping in the current work. The relatively low crack toughness of M-PBOHM especially at higher temperatures augmented the negative effect of rotatable specimen gripping on the composite performance and led to an uncontrolled failure with a sudden release of the accumulated strain energy. The influence of boundary conditions on the tensile performance of these composites will be assessed in a future work, since this is an important aspect with regard to material testing. The increase in temperature had a considerable negative effect on both the in-situ and residual tensile strength and ductility of M-PBO-HM; see Figs. 6 and 7. As shown in Fig. 7, the tensile strength reduction was proportional with the temperature increase and the recovery in the residual experiments was lower compared to M-PE. Also the reduction in strain capacity of M-PBO-HM was proportional to the temperature increase. M-PBO-HM in-situ

10

105°C residual 150°C residual 200°C residual

8 Stress [MPa]

8 Stress [MPa]

M-PBO-HM residual

10

6 4 room temperature 105°C in-situ 150°C in-situ

2 0 0.0

0.5

1.0 1.5 Strain [%]

2.0

6 4 2 0 0.0

0.5

1.0 1.5 Strain [%]

2.0

Fig. 6. Representative (a) in-situ and (b) residual stress-strain curves of M-PBO-HM at different temperature treatments.

The optical analysis of the specimens showed that the crack width in the corresponding SHCC was not significantly affected by the increase in temperature, the average values being 10.6 µm at 20 °C, 8.6 µm at 105 °C and 11.1 µm at 150 °C. At the same time, the number of cracks in the gauge length decreased from 38 at 20 °C to 14 at 105 °C and to 9 at 150 °C. Same as in the case of M-PE, the fracture surfaces of MPBO-HM treated at 200 °C showed no protruding fibers, explaining the brittle failure of the in-situ tested specimens and the weak softening in the residual experiments.

Influence of Fiber Type on the Tensile Behavior of HS-SHCC

8

M-PBO-HM tensile strength in-situ residual

6 4 2 0 0 (a)

25 50 75 100 125 150 175 200 Temperature [°C]

1.8 Strain at peak load [%]

Tensile strength [MPa]

10

1.5

1031

M-PBO-HM tensile ductility in-situ residual

1.2 0.9 0.6 0.3 0.0 0 25 50 75 100 125 150 175 200 (b) Temperature [°C]

Fig. 7. (a) In-situ and residual tensile strength values of individual M-PBO-HM specimens and (b) corresponding strains at peak load.

The reason for a poor mechanical performance of M-PBO-HM at elevated temperatures is most probably related to the degradation of the high-strength matrix and to the fact that the cementitious matrix yields a positive thermal expansion, whereas the polymer fibers yield a negative one; see Table 1. It could be assumed that the high stiffness of the PBO-HM fibers might lead to a delamination from the matrix already during the heating process, i.e. prior to crack formation, drastically reducing their crack bridging effectiveness. In the case of the PE fibers, the thermal incompatibility might be compensated by their low Young’s modulus and high elongation capacity. These aspects will be clarified in a systematic future study involving appropriate testing methods.

5 Conclusions The experimental results of high-strength SHCC tested under and after exposure to elevated temperatures indicated clearly the importance of fiber type with regard to the in-situ and residual tensile behavior of the composites at elevated temperatures. • The high-strength SHCC made with UHMWPE fibers yielded no reduction in tensile strength at 105 °C, while the tensile strain capacity doubled due to the more pronounced multiple cracking and increased crack width. The residual experiments yielded a reduced strain capacity compared to the in-situ 105 °C tests but still higher than of the non-heated specimens. This was traced back to the partial recovery of fibers’ stiffness, restraining the crack openings. At 150 °C, the SHCC made with UHMWPE fibers showed a pronounced reduction in tensile strength and a moderate reduction of strain capacity compared to the results obtained at 105 °C. However, despite this, the composites showed a strain capacity higher than 3% and a tensile strength of approximately 3 MPa.

1032

I. Curosu et al.

• In contrast to the expectations, the composites reinforced with PBO-HM fibers yielded a pronounced reduction in tensile strength and ductility already at 105 °C, this decline being proportional to the temperature increase up to 200 °C, at which the composites yielded brittle failure with no multiple cracking. This effect was partly traced back to the negative thermal expansion of the fibers. However, further investigations are necessary for a sound clarifications of the responsible phenomena, which will be a matter of interest in an upcoming study. Acknowledgements. The authors express their gratitude to the German Research Foundation (Deutsche Forschungsgemeinschaft - DFG) for the financial support within the Research Training Group GRK 2250 “Mineral-bonded composites for enhanced structural impact safety”. Furthermore, the authors express their acknowledgement to Mr. Syed Fasih Mohiuddin, Mr. Kai Uwe Mehlisch, and Mr. Tilo Günzel for their valuable support in preparing and performing the experimental investigations. Credit is given to the Institute of Timber Structures of the Technische Universität Dresden for providing the stereo DIC system.

References 1. Li, V.C.: On engineered cementitious composites (ECC): a review of the material and its applications. J. Adv. Concr. Technol. 1(3), 215–230 (2003) 2. Drechsler, A., Frenzel, R., Caspari, A., Michel, S., Holzschuh, M., Synytska, A., Curosu, I., Liebscher M., Mechtcherine, V.: Surface modification of poly(vinyl alcohol) fibers to control the fiber-matrix interaction in composites. Colloid Polymer Sci. 297(7), 1079–1093 (2019) 3. Curosu, I., Liebscher, M., Mechtcherine, V., Bellmann, C., Michel, S.: Tensile behavior of high-strength strain-hardening cement-based composites (HS-SHCC) made with highperformance polyethylene, aramid and PBO fibers. Cem. Concr. Res. 98, 71–81 (2017) 4. Sahmaran, M., Lachemi, M., Li, V.C.: Assessing mechanical properties and microstructure of fire-damaged engineered cementitious composites. ACI Mater. J. 107(3), 297–304 (2010) 5. Bhat, P.S., Chang, V., Li, M.: Effect of elevated temperature on strain-hardening engineered cementitious composites. Constr. Build. Mater. 69, 370–380 (2014) 6. Magalhaes, M.S., Toledo Filho, R.D., Fairbairn, E.M.R.: Thermal stability of PVA fiber strain hardening cement-based composites. Constr. Build. Mater. 94, 437–447 (2015) 7. Yu, J., Lin, J., Zhang, Z., Li, V.C.: Mechanical performance of ECC with high-volume fly ash after sub-elevated temperatures. Constr. Build. Mater. 99, 82–89 (2015) 8. Li, X., Wu, L., Yan, Q., Ma, H., Chen, G., Zhang, H.: Thermal and mechanical properties of high-performance fiber-reinforced cementitious composites after exposure to high temperatures. Constr. Build. Mater. 157, 829–838 (2017) 9. Du, Q., Wei, J., Lv, J.: Effects of high temperature on mechanical properties of polyvinyl alcohol Engineered Cementitious Composites (PVA-ECC). Int. J. Civil Eng. 16(8), 965–972 (2017) 10. Liu, J.-C., Tan, K.H.: Fire resistance of strain hardening cementitious composite with hybrid PVA and steel fibers. Constr. Build. Mater. 135, 600–611 (2017) 11. Liu, J.-C., Tan, K., Fan, S.: Residual mechanical properties and spalling resistance of strainhardening cementitious composite with Class C fly ash. Constr. Build. Mater. 181, 253–265 (2018)

Influence of Fiber Type on the Tensile Behavior of HS-SHCC

1033

12. Mechtcherine, V., Silva, F.A., Müller, S., Toledo Folho, R.D.: Coupled strain rate and temperature effects on the tensile behavior of strain-hardening cement-based composites (SHCC) with PVA fibers. Cem. Concr. Res. 42(11), 1417–1427 (2012) 13. Purfalah, S.: Behavior of engineered cementitious composites and hybrid engineered cementitious composites at high temperatures. Constr. Build. Mater. 158, 921–937 (2018) 14. de Oliveira, A.M., Silva, F.A., Fairbairn, E.M.R., Filho, R.D.T.: Coupled temperature and moisture effects on the tensile behavior of strain hardening cementitious composites (SHCC) reinforced with PVA fibers. Mater. Struct. 51(3), 1–13 (2018) 15. Dyneema Fact Sheet, Ultra high molecular weight polyethylene fiber from Dyneema, Eurofibers (2010). https://issuu.com/eurofibers/docs/name8f0d44 16. Liu, X., Yu, W.: Evaluating the thermal stability of high performance fibers by TGA. J. Appl. Polymer Sci. 99, 937–944 (2006) 17. Deshpande, A.A., Kumar, D., Ranade, R.: Influence of high temperatures on the residual mechanical properties of a hybrid fiber-reinforced strain-hardening cementitious composite. Constr. Build. Mater. 208, 283–295 (2019) 18. Technical Information, PBO Fiber Zylon, Toyobo CO., LTD. http://www.toyobo-global. com/seihin/kc/pbo/zylon-p/bussei-p/technical.pdf 19. Technora, ‘High Tenacity Aramid Fiber, Technical Information’, Teijin Techno Products Limited, Aramid division, Technora section, June 2004 20. Afshari, M., Kotek, R., Chen, P.: High performance fibers. In: Mittal, V. (ed.) High Performance Polymers and Engineering Plastics’. Wiley, Hoboken (2011) 21. Ghae, H.G., Kumar, S.: Rigid-rod polymeric fibers. J. Appl. Polymer Sci. 100(1), 791–802 (2006) 22. Kuroki, T., Tanaka, Y., Hokudoh, T., Yabuki, K.: Heat resistance properties of poly(pphenylene-2,6-benzobisoxazole) fiber. J. Appl. Polymer Sci. 65(5), 1031–1036 (1997) 23. Liu, X., Weidong, Y.: Degradation of PBO fiber by heat and light. Res. J. Text. Appar. 10 (1), 26–32 (2006) 24. Curosu, I., Mechtcherine, V., Millon, O.: Effect of fiber properties and matrix composition on the tensile behavior of strain-hardening cement-based composites (SHCCs) subjected to impact loading. Cem. Concr. Res. 82, 23–35 (2016) 25. Kanda, T., Takaine, Y., Tomoe, S., Takahashi, M., Yamamoto, Y., Kawano, K., Kunieda, M., Mizobuchi, T.: Development of coupling beam elements utilizing UHP-SHCC for highrise R/C building. In: Schlangen, E., Sierra Betran, M.G., Lukovik, M., Ye, G. (Eds.) Proceedings of the 3rd RILEM Conference on Strain-Hardening Cementitious Composites, pp. 409–416 (2014) 26. Kunieda, M., Denarié, E., Brühwiler, E., Nakamura, H.: Challenges for strain hardening cementitious composites – deformability versus matrix density. In: Reinhardt, H.W., Naaman, A.E. (eds.) Proceedings of the Fifth International RILEM Workshop on HPFRCC, pp. 31–38 (2007) 27. Ranade, R., Li, V.C., Heard, W.F.: Tensile rate effects in high strength-high ductility concrete. Cem. Concr. Res. 68, 94–104 (2015) 28. Kamal, A., Kunieda, M., Ueda, N., Nakamura, H.: Evaluation of crack opening performance of a repair material with strain hardening behavior. Cem. Concr. Compos. 30(10), 863–871 (2008) 29. Li, M., Li, V.C.: Rheology, fiber dispersion, and robust properties of Engineered Cementitious Composites. Mater. Struct. 46(3), 405–420 (2013)

Tensile and Compressive Performance of HighStrength Engineered Cementitious Composites (ECC) with Seawater and Sea-Sand Jing Yu1(&), Bo-Tao Huang2, Jia-Qi Wu1, Jian-Guo Dai2, and Christopher K. Y. Leung1 1

Department of Civil and Environmental Engineering, The Hong Kong University of Science and Technology, Kowloon, Hong Kong, China [email protected] 2 Department of Civil and Environmental Engineering, The Hong Kong Polytechnic University, Hung Hom, Hong Kong, China

Abstract. Marine infrastructures play an important role in the social-economic development of coastal cities. However, the shortage of river/manufactured sand and fresh water is a major challenge for producing concrete on site, as the transportation of these materials is not only costly but also environmentally unfriendly, while desalination of sea-sand and seawater is also pricey. Seawater sea-sand Engineered Cementitious Composites (SS-ECC) have a great potential for marine/coastal applications; but the present knowledge on SS-ECC is extremely limited. This study aims to explore the feasibility of producing highstrength SS-ECC. The effects of key composition parameters including the length of polyethylene (PE) fibers (6 mm, 12 mm, and 18 mm) and the maximum size of sea-sand (1.18 mm, 2.36 mm, and 4.75 mm) on the mechanical performance of SS-ECC were investigated. SS-ECC with compressive strength over 130 MPa, tensile strength over 8 MPa and ultimate tensile strain about 5% were achieved. Test results also showed that the tensile strain capacity increased with increasing fiber length, while sea-sand size had limited effects on the tensile performance of SS-ECC. The findings provide insights into the future design and applications of ECC in marine infrastructures for improving safety, sustainability, and reliability. Keywords: Fiber-reinforced concrete
Engineered cementitious composite
Strain-hardening cementitious composite
Marine infrastructures
Seawater
Sea-sand
Tensile performance

1 Introduction Marine infrastructures play an important role in the social-economic development of coastal cities. A major challenge in the development of marine infrastructures is the shortage of fresh water and river/manufactured sand for producing concrete on site, as the transportation of these materials is not only costly but also environmentally unfriendly, while desalination of seawater and sea-sand is also pricey. Additionally, direct use of

© RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 1034–1041, 2021. https://doi.org/10.1007/978-3-030-58482-5_91

Tensile and Compressive Performance of High-Strength ECC

1035

seawater and sea-sand is generally unsuitable for conventional steel-reinforced concrete structures, as chloride ions can lead to significant steel corrosion [1–3]. Engineered Cementitious Composite (ECC) is a family of advanced fiberreinforced concrete material, which exhibits strain-hardening and multiple-cracking with fine cracks width (typically  100 lm) under tension [4, 5]. Generally, the tensile strain capacity of ECC materials is over 3%; the compressive strength is about 20– 80 MPa for normal-strength ECC [6–10] and is about 80–160 MPa for high-strength ECC [11–16]. Compared to conventional concrete, ECC shows much better durability performance [17] and mechanical performance under static, cyclic, fatigue and impact loadings [18–26]. From the perspectives of shortage of fresh water and river sand, ECC made of seawater and sea-sand (i.e., seawater sea-sand ECC, or SS-ECC) has a great potential for marine/coastal applications; but the present knowledge on SS-ECC is extremely limited. This study aims to explore the feasibility of producing high-strength SS-ECC. An experimental program was conducted to investigate the influence of the length of fiber (6 mm, 12 mm, and 18 mm) and the maximum size of sea-sand (1.18 mm, 2.36 mm, and 4.75 mm) on the compressive strength and tensile performance of highstrength SS-ECC.

2 Experimental Program 2.1

Mix Proportion and Raw Materials

Table 1 shows the SS-ECC mixes studied. Ultra-high-molecular-weight polyethylene (PE) fiber (Table 2) was used for achieving high tensile strength in SS-ECC [27]. In the mix ID, taking PE12-2.0-S1 for example, “PE12” stands for the length of PE fiber (12 mm), “2.0” stands for the dosage of PE fiber (2.0 vol.%) and “S1” stands for the maximum particle size of sea-sand (1.18 mm). The particle size distribution of sea-sand is shown in Table 3. According to the design theory of ECC, fine sand is preferred [4, 5, 28]. Hence, the sea-sand with the size of  1.18 mm was used for the fiber-length series in Table 1. However, the raw sea-sand contained over 77% particles with a size >1.18 mm (Table 3). To improve the utilization ratio of the sea-sand, another two groups with the maximum particle sizes of 2.36 and 4.75 mm (PE12-2.0-S2 and PE12-2.0-S4 in Table 1) were prepared. Table 1. Mix proportions (in weight ratio) of high-strength SS-ECC Mix ID

PE06-2.0-S1 PE12-2.0-S1 PE18-2.0-S1 PE12-2.0-S2 PE12-2.0-S4

Seawater HRWR PE fiber

Cement Silica fume

Sea-sand  1.18 mm

 2.36 mm

 4.75 mm

0.8 0.8 0.8 0.8 0.8

0.3 0.3 0.3 / /

/ / / 0.3 /

/ / / / 0.3

0.2 0.2 0.2 0.2 0.2

0.18 0.18 0.18 0.18 0.18

0.0135 0.0135 0.0135 0.0135 0.0135

Length (mm)

Vol. (%)

6 12 18 12 12

2.0 2.0 2.0 2.0 2.0

1036

J. Yu et al. Table 2. Nominal physical properties of PE fiber Length (mm) Diameter (µm) Tensile strength (MPa) Elastic modulus (GPa) 6/12/18 24 3000 120

Table 3. Particle size distribution of sea-sand Particle size (mm) Weight ratio (%)

2.2

2.36– 4.75 16.13

1.18– 2.36 61.20

0.6– 1.18 13.83

0.3– 0.6 8.42

0.15– 0.3 0.39

0.075– 0.15 0.01

130 MPa for all the SS-ECC mixes (Table 4). The fiber length (from 6 to 18 mm) had a very little effect

Tensile and Compressive Performance of High-Strength ECC

1037

on the compressive strength, which is reasonable as these mixes had the same fiber dosage. For the sea-sand size, the compressive strength of SS-ECC slightly decreased as the sea-sand size increased (from 1.18 to 4.75 mm), which may be related to the higher seashell content in coarser sea-sand. The broken seashell is not only in plate shape but also generally weaker than silica sand; and therefore sea-sand with higher seashell content can lead to lower compressive strength in the resulting SS-ECC. It should also be pointed out that the sand/binder ratio for SS-ECC in this study was relatively low (only 0.3 as shown in Table 1), and hence the influence of sea-sand size on the compressive strength was limited. Table 4. Summary of 28-day compressive strength and tensile performance of high-strength SS-ECC Mix ID

PE06-2.0-S1 PE12-2.0-S1 PE18-2.0-S1 PE12-2.0-S2 PE12-2.0-S4

3.2

Compressive strength (MPa) Average Deviation 132.87 6.23 136.82 6.16 134.00 8.22 133.37 3.16 130.48 3.61

Tensile strength (MPa) Average Deviation 7.45 0.64 7.92 0.67 7.13 0.73 7.35 0.56 7.03 0.54

Ultimate tensile strain (%) Average Deviation 2.43 0.70 5.14 1.42 7.05 2.46 5.03 1.19 4.98 0.90

Tensile Performance

The tensile strain capacity and strength of the high-strength SS-ECC are summarized in Table 4, while the tensile stress-strain curves are presented in Fig. 2. It can be found that with increasing fiber length, the strain capacity increased from 2.43% for 6-mm fibers to 7.05% for 18-mm fibers, while the strength changed little. According to the design theory of ECC, with the precondition of limited fiber rupture, fibers with a high aspect ratio is effective to improve the fiber-bridging effectiveness [27]. On the other hand, the increase in sand size had very little effect on the strain capacity, but slightly decreased the tensile strength (Table 4). This phenomenon is similar to the trend in the compressive strength of sand-size group and it may be related to the seashell content in sea-sand. Additionally, it can be found in Fig. 2 that the tensile strain capacity of SSECC materials showed considerable scatters, which is widely observed in ECC materials and is due to the random nature of the fiber distribution and flaw size. Typical crack patterns of the tensile specimens of SS-ECC after testing are shown in Fig. 3. All the ECC specimens showed multiple-cracking behaviors. As the fiber length increased, the crack number significantly increased, while the sea-sand size had a limited effect on the crack number.

1038

J. Yu et al.

Fig. 2. Influence of fiber length and sea-sand size on the tensile stress-strain curves of SS-ECC. All mixtures show strain-hardening behavior.

Tensile and Compressive Performance of High-Strength ECC

1039

Fig. 3. Influence of fiber length and sea-sand size on the crack pattern of SS-ECC. All mixtures show multiple cracking behavior.

4 Conclusions This study explores the feasibility of producing high-strength seawater sea-sand Engineered Cementitious Composites (SS-ECC). According to the materials used and results obtained, the following conclusions can be drawn: • SS-ECC achieved compressive strength >130 MPa, tensile strength >8 MPa and ultimate tensile strain about 5% at 28 days, which fulfil the mechanical requirements of many infrastructures. • The tensile strain capacity increased with increasing fiber length from 6 mm to 18 mm, while the size of sea-sand up to 4.75 mm had limited effects on the tensile performance. • The compressive strength of SS-ECC slightly decreased as the sea-sand size increased (from 1.18 to 4.75 mm), which may be related to the higher seashell content in coarser sea-sand. These findings provide insights into the future design and applications of SS-ECC in marine/coastal infrastructures. Acknowledgements. This study was financially supported by the Hong Kong Research Grants Council (No.: T22-502/18-R) and the National Key Research Program of China (No.: 2017YFC0703403). The authors also thank Dr. Yu Xiang, Mr. Ji-Xiang Zhu and Mr. Ke-Fan Weng for their assistance in the experiment.

1040

J. Yu et al.

References 1. Xiao, J., Qiang, C., Nanni, A., Zhang, K.: Use of sea-sand and seawater in concrete construction: current status and future opportunities. Constr. Build. Mater. 155, 1101–1111 (2017) 2. Teng, J.-G., Xiang, Y., Yu, T., Fang, Z.: Development and mechanical behaviour of ultrahigh-performance seawater sea-sand concrete. Adv. Struct. Eng. 22(14), 3100–3120 (2019) 3. Ahmed, A., Guo, S., Zhang, Z., Shi, C., Zhu, D.: A review on durability of fiber reinforced polymer (FRP) bars reinforced seawater sea sand concrete. Constr. Build. Mater. 256, 119484 (2020) 4. Li, V.C., Leung, C.K.Y.: Steady-state and multiple cracking of short random fiber composites. J. Eng. Mech. 118(11), 2246–2264 (1992) 5. Li, V.C.: Engineered Cementitious Composites (ECC) - Bendable Concrete for Sustainable and Resilient Infrastructure. Springer, Heidelberg (2019) 6. Yu, J., Leung, C.K.Y.: Strength improvement of strain-hardening cementitious composites with ultrahigh-volume fly ash. J. Mater. Civil Eng. 29(9), 05017003 (2017) 7. Yu, J., Li, H., Leung, C.K.Y., Lin, X., Lam, J.Y.K., Sham, I.M.L., Shih, K.: Matrix design for waterproof engineered cementitious composites (ECCs). Constr. Build. Mater. 139, 438– 446 (2017) 8. Huang, B.-T., Li, Q.-H., Xu, S.-L., Zhou, B.: Strengthening of reinforced concrete structure using sprayable fiber-reinforced cementitious composites with high ductility. Compos. Struct. 220, 940–952 (2019) 9. Yu, J., Yao, J., Lin, X., Li, H., Lam, J.Y.K., Leung, C.K.Y., Sham, I.M.L., Shih, K.: Tensile performance of sustainable Strain-Hardening Cementitious Composites with hybrid PVA and recycled PET fibers. Cem. Concr. Res. 107, 110–123 (2018) 10. Yu, J., Wu, H.-L., Leung, C.K.Y.: Feasibility of using ultrahigh-volume limestone-calcined clay blend to develop sustainable medium-strength Engineered Cementitious Composites (ECC). J. Clean. Prod. 262, 121343 (2020) 11. He, S., Qiu, J., Li, J., Yang, E.-H.: Strain hardening ultra-high performance concrete (SHUHPC) incorporating CNF-coated polyethylene fibers. Cem. Concr. Res. 98, 50–60 (2017) 12. Chen, Y., Yu, J., Leung, C.K.Y.: Use of high strength strain-hardening cementitious composites for flexural repair of concrete structures with significant steel corrosion. Constr. Build. Mater. 167, 325–337 (2018) 13. Ranade, R., Li, V.C., Stults, M.D., Heard, W.F., Rushing, T.S.: Composite properties of high-strength, high-ductility concrete. ACI Mater. J. 110(4), 413–422 (2013) 14. Chen, Y., Yu, J., Younas, H., Leung, C.K.Y.: Experimental and numerical investigation on bond between steel rebar and high-strength Strain-Hardening Cementitious Composite (SHCC) under direct tension. Cem. Concr. Comp. 112, 103666 (2020) 15. Kamal, A., Kunieda, M., Ueda, N., Nakamura, H.: Evaluation of crack opening performance of a repair material with strain hardening behavior. Cem. Concr. Comp. 30(10), 863–871 (2008) 16. Curosu, I., Liebscher, M., Mechtcherine, V., Bellmann, C., Michel, S.: Tensile behavior of high-strength strain-hardening cement-based composites (HS-SHCC) made with highperformance polyethylene, aramid and PBO fibers. Cem. Concr. Res. 98, 71–81 (2017) 17. van Zijl, G.P.A.G., Slowik, V.: A framework for durability design with strain-hardening cement-based composites (SHCC): state-of-the-art report of the RILEM technical committee 240-FDS. In: RILEM State-of-the-Art Reports. RILEM, Netherlands (2017)

Tensile and Compressive Performance of High-Strength ECC

1041

18. Li, V.C., Horii, H., Kabele, P., Kanda, T., Lim, Y.M.: Repair and retrofit with engineered cementitious composites. Eng. Fract. Mech. 65(2–3), 317–334 (2000) 19. Huang, B.-T., Li, Q.-H., Xu, S.-L., Liu, W., Wang, H.-T.: Fatigue deformation behavior and fiber failure mechanism of ultra-high toughness cementitious composites in compression. Mater. Des. 157, 457–468 (2018) 20. Lu, C., Yu, J., Leung, C.K.Y.: Tensile performance and impact resistance of strain hardening cementitious composites (SHCC) with recycled fibers. Constr. Build. Mater. 171, 566–576 (2018) 21. Huang, B.-T., Li, Q.-H., Xu, S.-L., Zhou, B.-M.: Tensile fatigue behavior of fiber-reinforced cementitious material with high ductility: experimental study and novel P-S-N model. Constr. Build. Mater. 178, 349–359 (2018) 22. Mechtcherine, V.: Novel cement-based composites for the strengthening and repair of concrete structures. Constr. Build. Mater. 41, 365–373 (2013) 23. Huang, B.-T., Li, Q.-H., Xu, S.-L., Zhou, B.-M.: Frequency effect on the compressive fatigue behavior of ultrahigh toughness cementitious composites: experimental study and probabilistic analysis. J. Struct. Eng. 143(8) (2017) 24. Yu, J., Chen, Y., Leung, C.K.Y.: Mechanical performance of Strain-Hardening Cementitious Composites (SHCC) with hybrid polyvinyl alcohol and steel fibers. Compos. Struct. 226, 111198 (2019) 25. Huang, B.-T., Li, Q.-H., Xu, S.-L., Zhang, L.: Static and fatigue performance of reinforced concrete beam strengthened with strain-hardening fiber-reinforced cementitious composite. Eng. Struct. 199, 109576 (2019) 26. Mechtcherine, V., Silva, F.d.A., Butler, M., Zhu, D., Mobasher, B., Gao, S.-L., Mäder, E.: Behaviour of strain-hardening cement-based composites under high strain rates. J. Adv. Concr. Technol. 9(1), 51–62 (2011) 27. Zhang, D., Yu, J., Wu, H., Jaworska, B., Ellis, B., Li, V.C.: Discontinuous micro-fibers as intrinsic ductile reinforcement for Engineered Cementitious Composites (ECC). Compos. Part B-Eng. 184, 107741 (2020) 28. Wu, H.-L., Yu, J., Zhang, D., Zheng, J.-X., Li, V.C.: Effect of morphological parameters of natural sand on mechanical properties of engineered cementitious composites. Cem. Concr. Comp. 100, 108–119 (2019) 29. ASTM. Standard Test Method for Compressive Strength of Hydraulic Cement Mortars. In: C109/C109M (ASTM International, West Conshohocken) (2013) 30. JSCE. Recommendations for design and construction of high performance fiber reinforced cement composites with multiple fine cracks (HPFRCC). Japan Society of Civil Engineers, Tokyo, Japan (2008)

Effect of Fiber Content Variation in Plastic Hinge Region of Reinforced UHPC Flexural Members Mandeep Pokhrel1(&), Yi Shao2, Sarah Billington2, and Matthew J. Bandelt1 1

2

Department of Civil and Environmental Engineering, New Jersey Institute of Technology, Newark, USA [email protected] Department of Civil and Environmental Engineering, Stanford University, Stanford, USA

Abstract. Ultra-high performance concrete (UHPC) is used for the construction of resilient structures that can sustain dynamic loadings such as blast, impact, and earthquake loadings, among others. In structural components subjected to such loading, it is essential to ensure the formation of a ductile plastic hinge mechanism for suitable load transfer mechanisms and global stability of the structure. Experimental research is needed to understand the formation of plastic hinges in UHPC materials and the impact of plastic hinges on the rotation capacity of reinforced UHPC structural components. The study presented herein aims to understand the spread of plasticity and formation of plastic hinge regions in reinforced UHPC flexural members. Two reinforced UHPC beams with variation in fiber volume fraction (i.e., Vf = 1% and 2%) were subjected to monotonic loading. The test results demonstrated that the reinforcement plasticity length increased by 26% with a decrease in fiber volume fraction from 2% to 1%. The plastic hinge region of specimens with 2% fiber content had crack localization within the maximum moment region, whereas the specimen with 1% fiber content had a more uniformly distributed localized crack pattern. Further, analytical models and a recently proposed equivalent plastic hinge length equation were used to predict and compare the flexural strength and rotation values at various damage states. Keywords: Plastic hinge length
Fiber content
Reinforced UHPC
Ultimate rotation capacity

1 Introduction and Background Ultra-high performance concrete (UHPC) is an advanced cement-based composite material designed with optimal particle packing density, such that, it possesses extremely high compressive strength (>120 MPa without heat treatment) and enhanced durability properties [1]. When combined with short discontinuous fibers, UHPC materials have high tensile strength (>5 MPa), tensile fracture toughness, and ductile strain-hardening behavior under uniaxial tension tests [1]. The mechanical properties of © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 1042–1055, 2021. https://doi.org/10.1007/978-3-030-58482-5_92

Effect of Fiber Content Variation in Plastic Hinge Region

1043

UHPC have led researchers and engineers to perform a large number of proof-ofconcept investigations under extreme loading conditions such as blast, impact, earthquake, and fire [2]. In high seismic zones, researchers are especially interested in the applicability of UHPC in plastic hinge regions of structural components undergoing large inelastic deformations [3]. Various classes of high performance fiber reinforced cementitious composites (HPFRCCs) have already been effectively used in plastic hinge region of structural components such as coupling beams, columns, and bridge piers in recent years [4]. The use of UHPC in the plastic hinge regions of structural components can enhance the load carrying capacity, ductility, and energy absorption capacity because the mechanical properties of UHPC can prevent premature failure associated with damage in plastic hinge regions, as is typically observed in structural components made with conventional concrete (e.g., spalling of cover, buckling of rebar, shear cracks, etc.). Structural components can be engineered to improve the damage tolerance of structures by using UHPC in plastic hinge region, while using conventional concrete in the remaining portions of the component [3]. Such an approach can minimize the high cost associated with UHPC while reducing the overall life cycle cost (i.e., maintenance and repair cost) of the structure. To optimize the initial construction cost, the fiber volume fraction used in a UHPC material can be reduced; however, the influence of such a reduction on structural ductility and performance of plastic hinge regions is not well understood. A previous numerical study conducted using a wide range of HPFRCC materials indicated that the tensile strength and ductility of the matrix can significantly alter the amount of damage and the length of the plastic hinge region in reinforced HPFRCC flexural members [5, 6]. Therefore, the use of low fiber content in regions undergoing large displacement reversals may result in an undesired failure mechanism (e.g., shear cracking) without the formation of a ductile plastic hinge mechanism. Other experimental and numerical studies have shown that the flexural behaviour of reinforced HPFRCCs (including UHPC) in terms of crack progression, reinforcement plasticity, and failure mechanism is significantly different than conventional reinforced concrete structural components [7–10]. Specifically, the failure mode of flexural members is found to be predominantly through the fracture of longitudinal reinforcement rather than compression crushing of an HPFRCC matrix. This is due to a crack localization phenomenon observed in reinforced HPFRCC structural components, wherein the plastic damage concentrates in the vicinity of a single or few flexural cracks. Several bond experiments with lap splice beam specimens have shown that higher bond strength of HPFRCC matrix restraints the formation of splitting cracks which leads to such a phenomenon [11–13]. Further, tension stiffening experiments of reinforced HPFRCC prisms have shown that there is localized strain hardening in longitudinal reinforcement at such localized cracks [14, 15]. Localized hardening of steel reinforcement can provide a strengthening mechanism at the critical section of reinforced HPFRCC flexural members, until the member loses its load-carrying capacity by reinforcement fracture. Shao and Billington [16] recently conducted an experimental study with reinforced UHPC beams consisting of two different reinforcement ratios (q = 0.96% and 2.10%). The study showed that there can be two different failure paths in reinforced UHPC

1044

M. Pokhrel et al.

beams depending on the amount of longitudinal reinforcement used. The use of low reinforcement ratio led to failure after crack localization failure path in which there were three major damage states: yielding, crack localization, and rebar fracture. The use of a high reinforcement ratio led to failure after gradual strain hardening failure path in which the intermediate damage state changed from crack localization to compression crushing (or softening). Compared to specimens that fail after crack localization, specimens that fail after gradual strain hardening show higher ductility and more failure warnings. The flexural failure paths and various damage states are found to be predominantly dependent on longitudinal reinforcement and matrix property based on similar experimental studies carried out on other HPFRCC materials [10, 17]. Although many experimental studies have investigated the flexural behavior of reinforced UHPC beams, their specimens mostly fail after crack localization [16, 17]. There is limited experimental study in the literature dedicated to investigate the reinforcement plasticity distribution, curvature distribution, UHPC surface strain variation, and damage propagation in plastic hinge region of UHPC beams that fail after gradual strain hardening. Further, the influence of fiber content variation on various parameters, as mentioned above, is important to understand the component flexural behaviour, especially the maximum load carrying capacity, deformation capacity, and structural stability of components constructed with such material and cross-section property. Moreover, a mechanics-based analytical approach is adopted in the study, to predict strength and rotation capacity of the such specimens, which provides a simplified way for the practicing engineers to model and design such components in large structural systems. To that end, an experimental study consisting of two reinforced UHPC beams with 1% and 2% fiber volume fraction (Vf ) were tested under a four-point bending test setup. Based on recent experimental studies [16, 17], these two beams were designed to fail after gradual strain hardening by adopting a high reinforcing ratio of 2.10%. The flexural response, length of reinforcement yielding, inelastic curvature distribution, crack distribution, and strain variation within the maximum moment region were investigated. Analytical models and a recently proposed equivalent plastic hinge length equation for ductile concrete composites were used to predict flexural strengths and rotation values at various damage states.

2 Experimental Program 2.1

Materials, Mixture Proportions, and Mechanical Tests

Two types of mixture proportions were used in the experiment as listed in Table 1. The propriety pre-mix blend contained a mixture of cement, quartz, and silica fume. There were three types of admixtures used to improve the workability of UHPC during the casting period. Standard smooth steel fibers with a diameter of 0.2 mm and length of 13 mm were used in both mixtures. The naming convention of the beam specimens were based on the percentage of fiber volume fraction used in each specimen. Therefore, UHPC-1% denotes the specimen with a fiber volume fraction of 1% and UHPC2% indicates the specimen containing a fiber volume fraction of 2%. The materials

Effect of Fiber Content Variation in Plastic Hinge Region

1045

were mixed in a horizontal shear mixer and poured from one end of the beam mold until the mold was filled up to the full height. Table 1. Mixture proportions (per m3 ) Specimen UHPC-1% UHPC-2%

Premix Water blend [kg] [kg] 1939 194 1960 196

Fibers [% Vol.] 1.0 2.0

Admix. A [kg] 20 20

Admix. B [kg] 26 26

Admix. C [kg] 28 28

All UHPC specimens were moist cured and tested at 56 ± 3 days of casting. The representative results of mechanical tests have been tabulated in Table 2. Both compression and flexural tests of the two types of UHPC mixtures were conducted in accordance with ASTM C1856-17 [18]. Cylindrical specimens of diameter 75 mm and height 150 mm were prepared and tested to obtain compressive strength and modulus of elasticity. Four-point bending tests were performed on UHPC prisms with a cross section dimension of 75  75 mm and a length of 300 mm to obtain the equivalent bending stress versus displacement response as shown in Fig. 1(a).

Table 2. Mechanical properties Description

a

ft [MPa]

Gf a [MPamm] 6.4 6.7 –

0

fc [MPa]

E [GPa]

fy [MPa]

fu [MPa]

UHPC-1% 5.2 170 41.6 – – UHPC-2% 9.2 180 42.5 – – Longitudinal – – 190 470 780 Rebar (#6) Transverse – – – 200c 510c – Rebar (#3) a Obtained using inverse analysis; bExtrapolated value; cManufacturer listed value

eu [%]

ef [%]

– – 10

– – 18b

–

–

A series of inverse analyses were conducted using two-dimensional finite element simulations to estimate the tensile properties of the two UHPC mixtures. The inverse analysis scheme adapted in this study has been successfully implemented by other researchers in the past to characterize the tensile stress-strain curve without resorting to the direct tension test [9, 10, 17]. A total strain-based fixed-crack constitutive model was used with trial multilinear tensile stress-strain curve as an input to the numerical model. It can be observed from Fig. 1(a) that the simulated flexural response from inverse analysis closely approximates the experimental flexural behaviour of unreinforced UHPC prisms. The corresponding tensile strength (f t ) and tensile fracture energy (Gf ) of the mixtures are listed in Table 2, which can be used in lieu of tensile

1046

M. Pokhrel et al. 80

UHPC-1% [Experiment] Inverse Analysis [ft: 5.2 MPa] UHPC-2% [Experiment] Inverse Analysis [ft: 9.2 MPa]

20

70

Applied Moment[kNm] [kNm] Applied Moment

Equivalent Bending Stress [MPa]

25

15

10

5

50 40 30 UHPC-1% UHPC-2% Yield Softening Fracture

20 10

0

0

0

(a)

60

0.5

1

1.5

2

Mid-span Displacement [mm]

2.5

0

1

(b)

2

3

4

5

6

7

8

9

10 11

Δ∆/L /Lshear-span [%] shear-span [%]

Fig. 1. (a) Equivalent bending stress vs. mid-span displacement of unreinforced UHPC beams (b) Applied moment vs. drift response of two reinforced UHPC beams with different damage states

parameters obtained from direct tension test to investigate the flexural behavior and plastic hinge region of reinforced UHPC beams tested in this study. ASTM A615 Grade 60 steel with a diameter of 19 mm was used as longitudinal reinforcing bar in both UHPC-1% and UHPC-2% specimens. A uniaxial tension test was conducted using an extensometer of gauge length 50 mm to obtain characteristic tensile properties of the longitudinal reinforcement as listed in Table 2. Transverse reinforcement of Grade 60 steel with manufacturer listed yield strength of 510 MPa was used in both specimens. 2.2

Test Specimens, Setup and Instrumentation

Two reinforced UHPC beam specimens were tested using a four-point bending setup as shown in Fig. 2(a). A digital image correlation (DIC) system was used to assess variations in strain in the constant moment region of 200 mm between the two point loads. Since the DIC system could not be extended further due to laboratory constraints, a series of linear variable displacement transducers (LVDTs) were used to measure the vertical displacement along one side of the beam to a distance of 200 mm from the center line of the right point load towards the roller support. Figure 2 (b–c) shows the design of the beam with longitudinal reinforcement layout, transverse reinforcement layout, location of strain gauges, and cross-section details. The strain in the bottom longitudinal reinforcement was measured at five locations by attaching post-yield strain gauges (YEFLA-2-3LJC-F from Tokyo Measuring Instruments Lab) with a maximum measurement capacity of 10%. Two longitudinal reinforcing bars of diameter 19 mm were used in the top and bottom sides of the cross section resulting into a tensile longitudinal reinforcement ratio of 2.10%. Transverse reinforcement was provided at a spacing of one-half of the effective depth (i.e., d=2) with bars of diameter 9.5 mm. The specimens were subjected to a monotonic

Effect of Fiber Content Variation in Plastic Hinge Region

1047

Vertical Actuator (245 kN)

Reinforced UHPC Beam (thickness: 150 mm)

DIC Surface Preparation

Roller Support

Pin Support

LVDT @ Center

LVDTs @ 50 mm c/c 200 mm

800 mm

800 mm

(a) -19 mm Rebar

-9.5 mm Stirrups @ d/2

150 mm

A

180 mm

220 mm Strain Gauge at Center

A 100 mm

Strain Gauges @ 50 mm c/c

40 mm (c)

1800 mm (b)

Fig. 2. (a) Test setup of reinforced UHPC beams with location of DIC surface and LVDTs (b) Specimen design detail with location of strain gauges (c) cross section at A-A

loading at the rate of 0.097 mm/s until they lost their load carrying capacity by fracture of the tensile longitudinal reinforcement.

3 Result and Discussion 3.1

Moment-Drift Response

The applied moment versus drift response of the two specimens with various damage states is shown in Fig. 1 (b). Drift is expressed in percentage (%) and is calculated by normalizing the vertical displacement at mid-span by the shear-span length (D=Lshearspan ). The initial elastic response of both beams including stiffness, moment at yield and drift at yield are similar. The beams were assumed to yield when the strain in the tensile reinforcement at mid-span reached the yield strain (ey ) of 0.2772%. At yield, the moment and drift capacities are similar because of the use of the same reinforcement ratio (q = 2.10%) in both beams. The strength and drift capacities at yield are less sensitive to variations in tensile strength of the matrix (or fiber volume fraction) in specimens with higher reinforcement ratios compared to specimens with lower reinforcement ratios [5]. After yield, the flexural load carrying capacity increased in both specimens due to the combination of fiber bridging action and localized hardening of the tensile reinforcement. The post-yield stiffness of both beams were similar; however, the nominal moment

1048

M. Pokhrel et al.

(Mn ) capacity (i.e., peak moment capacity) of UHPC-2% (72.2 kNm) was found to be lower than UHPC-1% (75.9 kNm). It was anticipated that the UHPC-2% specimen, which has twice the fiber content and a higher tensile strength than UHPC-1% beam, would have a higher nominal moment capacity than UHPC-1% specimen. However, a higher rate of post-yield strain accumulation in the compression zone of UHPC-2% specimen was observed (Fig. 5 (a)). This led to earlier softening of the compression matrix and a lower flexural load carrying capacity in UHPC-2% specimen than in UHPC-1% specimen. The drift of UHPC-2% specimen at the nominal level is lower than the drift of UHPC-1% by 31% because of the rapid strain concentration in the compression zone of the UHPC-2% specimen. For example, at 2.2% drift level, the compression zone strain ðec Þ in UHPC-2% beam was found to be 35% higher than in UHPC-1% beams (i.e., ðec Þ UHPC-2% = 0.42% whereas ðec Þ UHPC-1% = 0.31% as shown in Fig. 5 (a)) as further discussed in Sect. 3.5. Both specimens were able to achieve large deformations without significantly losing load carrying capacity because the hardened bottom longitudinal reinforcement acted as the tensile component of the flexural couple before reaching the fracture point. This failure mechanism with gradual strain hardening of tensile reinforcement is mostly found in reinforced HPFRCC flexural members with high reinforcement ratios [17]. It is interesting to note that the variation of fiber content did not influence the value of ultimate drift

Reinforcement Tensile Strain [%]

4.0

4.0

Lpy = 290 mm

2% Drift

2% Drift

3.0

4% Drift

3.0

8% Drift

8% Drift

Yield Strain

Yield Strain

2.0

2.0

1.0

1.0

0.0

0.0

0 0.001

50 100 150 200 250 Distance from Mid-span [mm]

Lcl = 150 mm

300

(b)

0

0.001

1% Drift 2% Drift 4% Drift 8% Drift Collapse Drift

0.0008 Curvature [1/mm]

1% Drift

4% Drift

(a)

50 100 150 200 250 Distance from Mid-span [mm]

Lcl = 150 mm

0.0006

0.0004

0.0004

0.0002

0.0002

0

300

1% Drift 2% Drift 4% Drift 8% Drift Collapse Drift

0.0008

0.0006

0 0

(c)

Lpy = 230 mm

1% Drift

50

100

150

200

250

Distance from Mid-span [mm]

300

0

(d)

50

100

150

200

250

300

Distance from Mid-span [mm]

Fig. 3. Longitudinal reinforcement tensile strain vs. distance from mid-span at various drift levels for (a) UHPC-1% specimen and (b) UHPC-2% specimen Curvature vs. distance from midspan at various drift levels for (c) UHPC-1% and (d) UHPC-2% specimens

Effect of Fiber Content Variation in Plastic Hinge Region

1049



capacity ( D=Lshearspan UHPC-1% = 10.02% and D=Lshearspan UHPC-2% = 10.15%). The results indicates that the ultimate rotation or drift capacity of reinforced UHPC beams with high longitudinal reinforcement (i.e., q [ 2%) is not sensitive to variation in fiber volume fraction compared to the beams with low to moderate reinforcement ratio (i.e., 1%\q\2%). 3.2

Strain Distribution

Figure 3 (a–b) presents the variation of reinforcement tensile strain from mid-span to the right side of the specimens (up to 250 mm). At 1% drift, it can be observed that the UHPC-1% specimen had a reinforcement yielding length, Lpy , of 50 mm, whereas the longitudinal reinforcement yielded over a length of 90 mm in UHPC-2% specimen. However, Lpy in the UHPC-1% specimen increased at higher drift level compared to UHPC-2% specimen because of the opening of flexural cracks along a longer span length as shown in Fig. 4. The fiber bridging action and matrix tensile strength of UHPC-1% specimen is lower than the UHPC-2% specimen. The lower strength of UHPC-1% allowed cracks top easily open at high drift levels and plasticity distributed uniformly over a longer length of reinforcement. The length of reinforcement yielding remained constant at higher drifts in both specimens as the inelastic strain concentration mostly occurred near a dominant crack location as was observed in previous studies involving reinforced HPFRCC flexural members [9, 19]. At 8% drift level, the length of reinforcement yielding region in UHPC-1% specimen (Lpy ¼ 290 mm) was longer than UHPC-2% specimen (Lpy ¼ 230 mm). These results suggest that the plastic hinge length in beams with higher fiber content (or higher tensile strength and matrix toughness) are shorter than those with lower fiber content (or lower tensile strength and matrix toughness). This is also in agreement with the length and type of cracking pattern shown in Fig. 4 and further discussed in Sect. 3.4. 3.3

Curvature Distribution

Curvature distribution along the span of the specimens were calculated using the vertical displacements data obtained from the six LVDTs. Mathematically, curvature at a section can be approximated using the elastic deflection theory as shown in Eq. (1). /¼

dh dh hi þ 1  hi ffi ¼ ds dx distance between LVDTs

ð1Þ

Where hi and hi þ 1 are the angles at sections i and i þ 1: These angles can be computed using the vertical displacements obtained from LVDTs along the span of the beam. Due to the opening of localized cracks at higher drifts, the recorded vertical displacement data at some locations were estimated using a trendline. Figure 3 (c–d) shows curvature distribution from mid-span to the right side of the specimens at incremental drift levels. At a lower drift level (1% or 2%), the curvature is maximum below the point load, which is the assumed hinge location under four-point bending setup. However, at larger drift levels (4% or more) the curvature is larger at mid-span because of the opening of a major crack at mid-span, such that the section at mid-span deforms more than the section

1050

M. Pokhrel et al.

below the point load as seen from the crack pattern at in Fig. 4. The overall trend of the curvature is similar to the theoretical curvature under four-point bending test where the curvature is maximum near the mid-span and sharply decreasing away towards the support. The curvature localization region (Lcl ) was 150 mm in both the specimens indicating there was no substantial effect of fiber content variation in curvature distribution of reinforced UHPC flexural members with high reinforcement ratios. 3.4

Crack Pattern in Plastic Hinge Region

Figure 4 shows the crack pattern and location of reinforcement
fracture in UHPC specimens at impending collapse level drift (i.e., D=Lshearspan UHPC-1% = 10.02%
and D=Lshearspan UHPC-2% = 10.15%). Both specimens contained multiple fine distributed cracks without any flexural crack localization up to 1% drift level. However, after yielding of tensile reinforcement, flexural cracks slowly began to open as the fiber-bridging action declined in some of the cracks. Four major flexural cracks widened in both specimens, but the major cracks were evenly spaced in UHPC-1% specimen compared to UHPC-2% specimen as shown in Fig. 4. Major cracks widened in the region away from maximum moment in the UHPC-1% specimen because the cracks could open at lower flexural load due to a lower tensile strength and fracture energy of UHPC-1% matrix. In UHPC-2% specimen, major cracks were confined to the maximum moment region and damage was pre-dominantly localized in a single crack (i.e., crack number 3). As crack number 3 widened, rebar plastic strain in the vicinity of that crack concentrated at a higher rate in the UHPC-2% specimen as compared to the UHPC-1% specimen (Fig. 5 (b)) as further discussed in Sect. 3.5. The distance between the extreme minor and major cracks were found to be longer in UHPC-1% specimen (Lminorcracks = 1350 mm and Lmajorcracks = 420 mm) compared to UHPC-2% (Lminorcracks = 935 mm and Lmajorcracks = 240 mm) which can be attributed to the comparatively weaker matrix and fiber-bridging action in UHPC-1% specimen compared to UHPC-2% specimen. UHPC-1% 200 mm

1

(a)

2

3

4

Lmajor cracks= 420 mm Lminor cracks= 1350 mm

UHPC-2% 200 mm

1

(b)

2

3

4

Lmajor cracks= 240 mm Lminor cracks= 935 mm

Fig. 4. Crack pattern in (a) UHPC-1% and (b) UHPC-2% specimens at impending collapse

Effect of Fiber Content Variation in Plastic Hinge Region 2.5

5.0 UHPC-2% Mid-span Tensile Rebar Strain [%]

Mid-span Compression Strain [%]

UHPC-2% UHPC-1% 2.0

1.5

Softening

1.0

Yield

0.5

0.0

(a)

1051

UHPC-1% 4.0

3.0

Softening

2.0

Yield

1.0

0.0 0

1

2 3 4 ∆ /Lshear-span [%] ∆/L [%] shear-span

5

6

(b)

0

1

2

3

4

5

6

∆ /Lshear-span [%] ∆/L shear-span [%]

Fig. 5. (a) Mid-span compression strain vs. drift and (b) mid-span longitudinal tensile reinforcement strain vs. drift

3.5

Variation of Strain in Maximum Moment Region

Figure 5 shows the variation of compression and tension strain at mid-span with incremental drift levels. The slope of the lines indicates the rate of strain accumulation in the compression zone (Fig. 5 (a)) or in the tensile reinforcement (Fig. 5 (b)) at midspan. After yielding of the specimens, the rate of compression strain accumulation in the UHPC-2% specimen became marginally higher compared to the UHPC-1% specimen because the high tensile strength matrix attracts larger forces at smaller deformation level. Further, higher bond strength in UHPC-2% specimen restricted the formation of splitting cracks, causing early rebar hardening over a small deboned length. The hardening led to widening of a mid-span crack and rapid strain concentration in the tensile longitudinal reinforcement in UHPC-2% specimen at a lower drift level as shown in Fig. 5 (b). For example, the strain in the tensile reinforcement at 2% drift level in UHPC-2% specimen was 2.00% and that in UHPC-1% specimen was 1.16%. The effect of this rapid strain variation caused softening of
UHPC-2% beam at a lower drift level compared to UHPC-1% beam ( D=Lshearspan UHPC-1% = 3.2% and
D=Lshearspan UHPC-2% = 2.2%). After softening of the specimens, the compression strain accumulation rate in the UHPC-1% specimen increased compared to the UHPC2% specimen. The difference in compression strain at the same drift level decreased progressively at higher drift levels. The localized hardening strain in tensile reinforcement at mid-span of the UHPC-2% specimen was much higher compared to the UHPC-1% specimen at the same drift levels. For instance, the strain in tensile reinforcement at 4% drift level in UHPC-2% specimen was 4.72% and that in UHPC-1% specimen was 2.91%. It was anticipated that the UHPC-2% would fail earlier by fracture of reinforcement based on this trend. However, UHPC-2% had a similar deformation capacity as UHPC-1% as discussed in Sect. 3.1. Further investigation is required to understand the failure mechanism at higher strain levels using post-yield strain gauges of a larger strain capacity.

1052

3.6

M. Pokhrel et al.

Prediction of Flexural Strength and Rotation Capacity

The flexural behavior and failure mechanism of reinforced HPFRCC is significantly different than the conventional reinforced concrete as demonstrated by several experimental studies [7–10]. A recently proposed analytical model (Fig. 6) was used to predict flexural strength and rotation capacity at different damage states [10, 20]. The predictability of the analytical model was measured by comparison with the experimentally obtained values. The flexural strengths and curvatures at various damage states were computed assuming linear strain distribution and using cross section properties of the specimens. The beams were assumed to reach the yield level when the tensile reinforcement strain reached the yield value (i.e., ey Þ: The analytical model shown in Fig. 6(b) considers a tensile stress block contribution which is ignored in the flexural calculation of conventional concrete components. Elastic deflection theory was used to compute yield rotation using the yield curvature value as shown in Eq. (2). It can be observed from Table 3 that the prediction ratio of both the parameters are close to 1.00, which indicates that the analytical formulation can be successfully used to compute yield rotation and moment capacity of reinforced UHPC flexural members.

Fig. 6. Analytical model for section analysis (a) cross-section (b) stress distribution at yield level [20] (c) stress distribution at nominal level using modified Hognestad stress block [10] (d) simplified stress distribution at nominal and ultimate level using Whitney stress block [20].

Two analytical models were used to estimate the nominal moment capacity: one with modified Hognestad compression stress block (Fig. 6(c)) and the second with simplified Whitney compression stress block (Fig. 6(d)). Both models considered localized hardening of the reinforcement bar as observed in tension stiffening Table 3. Comparison of experimental and analytical results at yield level Specimen

Yield Rotation [rad] Exp. Ana. Ana./Exp. UHPC-1% 0.0098 0.0095 0.97 UHPC-2% 0.0095 0.0098 1.03

Yield Exp. 58 60

Moment [kNm] Ana. Ana./Exp. 53 0.92 60 1.00

Effect of Fiber Content Variation in Plastic Hinge Region

1053

experiments [14, 15]. The beams were assumed to the reach nominal level when the strain in the compression zone reached 3% [10] or tensile reinforcement strain reached ultimate value (eu Þ [20]. It can be observed from Table 4 that the nominal moment predictability using a modified Hognestad stress block is better compared to the use of simplified rectangular Whitney stress block; however, both give reasonable estimates of strength. Table 4. Comparison of experimental and analytical results at nominal and ultimate level. Specimen

Nominal Moment [kNm] Exp. Ana.a Ana.b Ana./Exp.a Ana./Exp.b UHPC-1% 76 75 80 0.98 1.06 UHPC-2% 72 75 81 1.03 1.12 a Figure 6(c) [10]; bFigure 6(d) [20]

Ultimate Rotation [rad] Exp. Ana. Ana./Exp. 0.1002 0.083 0.83 0.1015 0.060 0.59

The ultimate rotation capacity was computed using Eq. (2). 1 hu ¼ hy þ hp ¼ /y Ls þ ð/u  /y ÞLp 2

ð2Þ

Where Ls is the shear span length (mm), /y and /u are the section curvatures of the structural member at yield level (mm−1) and collapse level (mm−1), respectively. In the above equation, Lp is the equivalent plastic hinge length (mm), which was computed using a recently developed expression based on a range of HPFRCC materials as shown in Eq. (3) [20]. Lp ¼ 0:02Ls þ

0:24qfy ft

ð3Þ

Where q is the longitudinal reinforcement (%), fy is the yield stress (MPa), and ft is the tensile strength of UHPC mixture (MPa). The analytical framework underestimated the ultimate rotation capacity in both specimens (Table 4). The reason for this discrepancy is the underestimation of equivalent plastic hinge length values for reinforced UHPC specimens tested in this experiment. The expression shown in Eq. (3) was developed using reinforced HPFRCC beams with typical reinforcement ratio (i.e., 0:70%\q\1:90%) and a maximum tensile strength of 8 MPa. As such, the majority of the specimens followed the failure after crack localization path with damage localization in a dominant crack. In the current experiment, both the specimens followed failure after gradual strain hardening failure path due to the use of a high reinforcement ratio. The damage was uniformly distributed over longer length as seen from the crack patterns. Therefore, there is a need to further improve the plastic hinge length expression for highly reinforced UHPC beams (i.e., q [ 2:0%) using a more rigorous parametric investigation.

1054

M. Pokhrel et al.

4 Conclusions This study provides valuable insight about the formation of plastic hinges with variation in fiber volume fraction. The following conclusions can be drawn from this study: • The variation in fiber content does not impact ultimate rotation capacity in reinforced UHPC beams with high longitudinal reinforcement ratio (q > 2.0%). • The length of plasticity in the longitudinal reinforcement increases with a decrease in fiber volume fraction because of the formation of multiple distributed flexural cracks along the plastic hinge region. • Distribution of visible cracks in the specimens indicated that the damage is much more localized in specimens with higher fiber content (Lmajorcracks = 240 mm) than those with low fiber content (Lmajorcracks = 420 mm) because the UHPC matrix with high fiber content had higher tensile strength, bond strength, and fracture energy which restrained the formation of splitting cracks and prevented opening of additional flexural cracks. • A parametric study with a wider variation in fiber content at high reinforcement ratios is necessary to further improve a recently developed plastic hinge length expression such that ultimate rotation capacity can be computed with higher accuracy. Acknowledgements. The authors gratefully acknowledge the financial support from the John A. Reif, Jr., Department of Civil and Environmental Engineering at New Jersey Institute of Technology and John A. Blume Earthquake Engineering Center at Stanford University. The authors would also like to recognize King Packaged Materials Company, Boisbriand, QC for providing materials necessary for the tests.

References 1. Graybeal, B.A.: Material property characterization of ultra-high performance concrete. FHWA, Virginia, USA, Technical report, No. FHWA-HRT-06-103, pp. 1–176 (2006) 2. Yoo, D., Yoon, Y.: A review on structural behavior, design, and application of ultra-highperformance fiber-reinforced concrete. Int. J. Concrete Struct. Mater. 10(2), 125–142 (2016) 3. Ichikawa, S., Matsuzaki, H., Moustafa, A., Elgawady, M.A., Kawashima, K.: Seismicresistant bridge columns with ultra high-performance concrete segments. J. Bridge Eng. 21 (9), 04016049 (2016) 4. Rokugo, K., Kanda, T., Yokota, H., Sakata, N.: Applications and recommendations of high performance fiber reinforced cement composites with multiple fine cracking (HPFRCC) in Japan. Mater. Struct. 42(9), 1197–1208 (2009) 5. Pokhrel, M., Bandelt, M.J.: Material properties and structural characteristics influencing deformation capacity and plasticity in reinforced ductile cement-based composite structural components. Compos. Struct. 224, 111013 (2019) 6. Pokhrel, M., Bandelt, M.J.: Plastic hinge region and rotation capacity in reinforced HPFRCC flexural members at collapse level. In: Proceedings of the Seventh International Colloquium on Performance, Protection & Strengthening of Structures under Extreme Loading & Events, Whistler, Canada, pp. 1–15 (2019)

Effect of Fiber Content Variation in Plastic Hinge Region

1055

7. Yoo, D.Y., Yoon, Y.S.: Structural performance of ultra-high-performance concrete beams with different steel fibers. Eng. Struct. 102, 409–423 (2015) 8. Bandelt, M.J., Billington, S.L.: Impact of reinforcement ratio and loading type on the deformation capacity of high-performance fiber-reinforced cementitious composites reinforced with mild steel. J. Struct. Eng. 142(10), 04016084 (2016) 9. Bandelt, M.J., Billington, S.L.: Simulation of deformation capacity in reinforced highperformance fiber-reinforced cementitious composite flexural members. J. Struct. Eng. 144 (10), 04018188 (2018) 10. Shao, Y., Billington, S.L.: Flexural performance of steel-reinforced engineered cementitious composites with different reinforcing ratios and steel types. Constr. Build. Mater. 231, 117159 (2020) 11. Bandelt, M.J., Billington, S.L.: Bond behavior of steel reinforcement in high-performance fiber-reinforced cementitious composite flexural members. Mater. Struct. 49(1–2), 71–86 (2014) 12. Lee, J.: Bonding behavior of lap-spliced reinforcing bars embedded in ultra-high strength concrete with steel fibers. KSCE J. Civil Eng. 20(1), 273–281 (2016) 13. Dagenais, M.A., Massicotte, B.: Cyclic behavior of lap splices strengthened with ultra high performance fiber-reinforced concrete. J. Struct. Eng. 143(2), 04016163 (2017) 14. Moreno, D.M., Trono, W., Jen, G., Ostertag, C., Billington, S.L.: Cement & Concrete Composites Tension stiffening in reinforced high performance fiber reinforced cement-based composites. Cement Concr. Compos. 50, 36–46 (2014) 15. Hung, C., Lee, H., Nga, S.: Tension-stiffening effect in steel-reinforced UHPC composites: Constitutive model and effects of steel fibers, loading patterns, and rebar sizes. Compos. B Eng. 158, 269–278 (2019) 16. Shao, Y., Billington, S.L.: Utilizing full UHPC compressive strength in steel reinforced UHPC beams. In: Proceedings of Second International Interactive Symposium on Ultra-High Performance Concrete, Albany, New York, USA, pp. 1–9 (2019) 17. Shao, Y., Billington, S.L.: Predicting the two predominant flexural failure paths of longitudinally reinforced high-performance fiber-reinforced cementitious composite structural members. Eng. Struct. 199, 109581 (2019) 18. ASTM: Standard practice for fabricating and testing specimens of ultra-high performance concrete. ASTM C1856/C1586-17, pp. 1–4 (2017) 19. Pokhrel, M., Bandelt, M.J.: Simulation of reinforced HPFRCC deformation capacity under flexure- and shear-dominated stress states. In: Proceedings of Computational Modelling of Concrete and Concrete Structures, Bad Hofgastein, Austria, pp. 633–640 (2018) 20. Pokhrel, M., Bandelt, M.J.: Plastic hinge behavior and rotation capacity in reinforced ductile concrete flexural members. Eng. Struct. 200, 109699 (2019)

An Eco-Friendly UHPC for Structural Application: Tensile Mechanical Response Amin Abrishambaf(&), Mário Pimentel, and Sandra Nunes CONSTRUCT-LABEST, Faculty of Engineering (FEUP), University of Porto, Porto, Portugal [email protected]

Abstract. This paper presents and discusses experimental results on the tensile mechanical performance of a newly developed ultra-high performance cementitious material, UHPC, incorporating spent equilibrium catalyst (ECat), a waste generated by the oil refinery industry, as a supplementary cementitious material. The results are compared to a previously developed conventional UHPC. The influence of ECat on the heat of hydration in UHPC is evaluated by isothermal calorimeter under a constant temperature of 20 °C. To determine the evolution of the tensile behaviour with time, a series of uniaxial tensile tests are performed on the specimens at different ages, i.e. 1, 3, 7, 28 and 91 days after casting. Afterwards, the fibre to matrix interfacial bond properties were characterized by executing a series of single fibre pullout tests at the age of 28 days on the steel fibres embedded in UHPCs with 0°, 30° and 60° orientation angles. The results confirmed the adequate performance of the new developed UHPC for the structural application. Keywords: Ultra-high performance fibre reinforced cementitious material Tensile behaviour
Spent equilibrium catalyst
Fibre to matrix bond properties
Fibre orientation




1 Introduction Ultra-high performance fibre reinforced cementitious composites (UHPFRC) designate a family of advanced cementitious materials with enhanced matrix packing density, very low water/binder ratio (w/b < 0.2), and containing a significant amount of high strength short and discrete steel fibres. This combination provides superior performances in terms of compressive strength, energy absorption capacity, ductility and durability [1–4]. Cementitious matrix of UHPFRC, designated as UHPC, exhibits large shrinkage values, mainly in the first days after casting, in which, unlike the conventional concrete, a large portion corresponds to the autogenous shrinkage [5]. Spent Equilibrium Catalyst (ECat) is a waste material generated by the oil refinery industry. ECat was shown to act as an internal curing agent, and its incorporation in the matrix of UHPFRC allowed reducing the autogenous shrinkage of the composite. ECat has a high porous microstructure which provides a high specific surface area with a

© RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 1056–1067, 2021. https://doi.org/10.1007/978-3-030-58482-5_93

An Eco-Friendly UHPC for Structural Application

1057

high water absorption capacity. Therefore, during the first day of hydration, the absorbed water in the porous ECat particles is released in the UHPC matrix and provide a beneficial effect to mitigate autogenous shrinkage. As supplementary cementitious material, ECat, including the one generated in Portugal and used in this research, exhibits high pozzolanic activity. Chapelle test results revealed that 1540 mg of Ca (OH)2 is consumed per gram of ECat during the pozzolanic action which is close to SF pozzolanic activity, i.e. 1577 mg [6]. This paper presents the experimental results on the mechanical performance (mainly tensile behaviour) of the new developed UHPC, incorporating spent equilibrium catalyst (ECat), as a supplementary cementitious material. The results are compared to a conventional UHPC previously developed. The influence of ECat on the heat of hydration in UHPC is evaluated by isothermal calorimeter under a constant temperature of 20 °C. To determine the evolution of the tensile behaviour with time, a series of uniaxial tensile tests are performed on the specimens at different ages. Afterwards, the fibre to matrix interfacial bond properties was characterized by executing a series of single fibre pullout test on the fibres with different orientation angles.

2 Experimental Program 2.1

Materials and Mixtures

Cement CEM I 4.25 R, limestone filler, Spent Equilibrium Catalyst (ECat) and silica fume (SF) with a specific surface area of 3110, 2680, 2660 and 2200 m2/kg were used in the preparation of the mixtures. ECat is generated by Sines Refinery, Portugal and shows about 30% water absorption capacity by mass at 24 h and at the saturated surface-dry basis. Figure 1 shows SEM images of ECat particles. The ECat’s water absorption was considered and added to the total water used in the UHPC mixture. The chemical compositions of CEM, ECat and SF used in this research are shown in Table 1. Siliceous natural sands with the specific surface area of 2630 m2/kg and 0.3% absorption were adopted in the mixtures. To guarantee the self-compacting property, a superplasticizer based on polycarboxylate ethers has been used. The mix compositions are shown in Table 2. In this table, UHPC and ECat_UHPC designate the conventional and the new developed mixture incorporating ECat [6]. A low fibre volume fraction of 0.5% is used to decrease the brittleness of UHPC mixtures during mechanical tests. However, this fibre volume fraction is low enough not to affect the mechanical properties of plain UHPC. The steel fibres are high strength, short and smooth brass-coated with a length, lf, of 13 mm and a diameter, df, of 0.2 mm. Both mixtures exhibited a slump flow diameter of 280 and 282 mm, respectively, without any compaction energy; and close 28-day compressive strength between 140–150 MPa, without heat/pressure treatment.

1058

A. Abrishambaf et al.

(a)

(b)

Fig. 1. SEM image ECat’s particle morphology in secondary electron mode: (a) general view of particles (250) and (b) internal porosity (10000). Table 1. Chemical compositions of the mix constituents (% by mass). SiO2 Al2O3 Fe2O3 CaO MgO SO3 Na2O K2O CEM 19.80 5.08 3.16 61.31 1.82 2.90 0.15 0.58 ECat 40.30 54.45 0.45 0.06 0.15 0.00 0.43 0.02 SF >90

LoI 2.54 1.05 2%), UHPFRC specimens subjected to uniaxial tensile loads show a strain hardening behaviour after the formation of the first crack. As a result, the tensile strength of the composite is higher than that at first cracking [6]. Figure 1 illustrates a simplified response of strain-hardening Fiber Reinforced Concrete (FRC). This idealized modelling approach distinguishes the tensile behaviour into three parts: Part I (Elastic behaviour up to cracking strength cc); Part II (Strain hardening behaviour with multiple cracking); Part III (Softening behaviour).

Fig. 1. Idealized response of strain-hardening FRC in tension

There are already few national recommendations on the Design and Construction with UHPFRC: the French AFGC-SETRA [7] and the Japanese JSCE [8] both provide recommendations on how to perform uniaxial tensile tests on UHPFRC materials. The Fib Model Code 2010 [9], which serves as a basis for the future code for concrete structures, also provides important information on the mechanical properties of Fiber Reinforced Concrete (FRC) elements. The Italian CSLP (Consiglio Superiore dei Lavori Pubblici) recently published a guideline about qualification, technical assessment certification and acceptance control of FRC. It also specifies how to test the flexural and direct tensile strength of FRC. Further details on test methods can be found in [10]. UHPFRC material is under detailed exploration: many experts from all over the world presented their research results regarding its tensile characterization. Wille et al. [11] experimentally investigated the direct tensile behaviour of nine dog-bone different UHPFRC specimens (by varying fiber type and fiber volume fraction). The analysed results showed, as expected, a strong dependency on fiber volume fraction. In Yoo D. et al. [12] the effects of steel fiber type on the tensile performance of UHPFRC were investigated. Four different steel fiber types were used: S-straight, T-twisted, H-hooked, and HH-half-hooked. The order of effectiveness in enhancing the tensile performance of dog-bone specimens was S-fibers > T-fibers > HH-fiber > H-fiber. Direct tensile tests are challenging to perform, since it is difficult to achieve homogeneously distributed stress throughout the specimen cross section and to control a stable load versus displacement/crack opening response [13]. Furthermore, specimen’s geometry and dimension, together with the gripping methods used during the test, can significantly influence the test results. Different

1070

N. Generosi et al.

tensile test setups have been proposed by researcher to evaluate the direct tensile behaviour of cementitious materials [11, 14, 15]. The objective of this research work is to investigate the mechanical properties of UHPFRC reinforced with different amount of hooked steel fibers, from 0.6% up to 2.6% by volume. First, the material has been characterized by means of compression and flexural tests. Then, direct tensile tests have been performed on dog-bone shape specimens to fully characterize their tensile properties and to evaluate the effect of different amounts of hooked steel fibers both on the tensile strength, deformation capacity and softening or hardening behaviour.

2 Materials and Methods A Commercial Portland-limestone blended cement (CEM I 52.5 R), in compliance with EN-197/1 [16] was used. The Blaine fineness of cement was 0.48 m2/g and its specific gravity was 3.15 g/cm3. As aggregate two different types of quartz sand with particle size 0–0.6 mm and 0.6–1.0 mm were suitably combined. Silica fume was added at a dosage of 125 kg/m3. In addition, an acrylic-based water-reducing admixture (WRA) was added in powder at dosage of 1.1% by weight of cement with a water to cement ratio (w/c) of 0.20. Hooked steel fibers (Fig. 2), 30-mm long with aspect ratio equal to 80, were added at increasing dosages: 50, 100, 150 and 200 kg/m3. The corresponding fibers volume fractions are 0.6%, 1.25%, 1.9% and 2.6%. The mixture proportions are reported in Table 1.

Fig. 2. Hooked steel fibers (a) and dog-bone specimens manufacturing (b) Table 1. UHPFRC mixtures (kg/m3) Specimen CEM I 52.5R Water Sand 0/0.6 Sand 0.6/1 Silica fume WRA UHPC 1000 200 400 600 125 11 UHPFRC-0.6 UHPFRC-1.25 UHPFRC-1.9 UHPFRC-2.6

Fibers – 50 100 150 200

Characterization of Ultra High Performance Fiber Reinforced Concrete (UHPFRC)

1071

3 Experimental Results 3.1

Compression and Flexural Tests (4  4  16 cm)

Three prismatic specimens were prepared for each concrete mixture, 4  4  16 cm in size, by casting them in steel forms. After 2 days they were demoulded and kept up to 28 days in climatic chamber at 20 °C and 50% R.H. to test their behaviour in the worst curing conditions. They were tested in bending and then in compression, according to the procedure described in EN 1015-11 [17]. Even if this type of test method is specific for mortars and it is not suitable for concrete samples, it was adopted in this study to obtain preliminary information on the mechanical properties of the specimens by varying the fibers dosage. Results obtained for the different UHPFRC mixtures are listed in Table 2, while load-midspan displacement curves of flexural tests are reported in Fig. 3. Table 2. Experimental results of compression and flexural tests on specimens with dimensions of 40  40  160 mm3 Specimen

UHPC UHPFRC0.6 UHPFRC1.25 UHPFRC1.9 UHPFRC2.6

Average CoV(%) Average CoV(%) Average CoV(%) Average CoV(%) Average CoV(%)

Compression strength (MPa) 98.5 5 121.4 5 124.0 5 138.8 4 143.4 4

Increase of compression strength – +23.3% +25.8% +40.9% +45.6%

Flexural strength (MPa) 13.8 5 17.2 7 27.2 9 31.7 5 43.6 4

Increase of flexural strength – +24.8% +97.9% +130.4% +217.1%

Flexural toughness Uf (kNmm) 1.14 11 9.34 3 16.23 8 20.15 5 28.63 3

It can be observed that the reference UHPC cured in dry environment showed an average compression strength of 98.5 MPa (a quite low value for a mixture with a w/c of 0.20, due to the dry curing conditions). The addition of hooked steel fibers was able to gradually increase both the compression and the flexural strength of the concrete mixtures. Concerning the compressive strength, the reason for this result can be attributed to the confinement effect of the fiber reinforcement. The maximum increase was found for the mixture with 200 kg/m3 of steel fibers (equal to about 2.6% by volume): compressive strength raised of about 45% while flexural strength showed an increase higher than 200%. Flexural toughness has been calculated by integrating the area under the loaddisplacement curves of flexural tests, up to a midspan displacement equal to 2 mm. The addition of steel fibers allows to considerably increase the flexural toughness of UHPFRC specimens. This value raises by increasing the amount of fibers, up to a value of about 28 kNmm for specimens reinforced with 2.6% of fibers. With respect to the reference UHPC mixture without fibers, the UHPFRCs were able to absorb up to 25time higher energy amount with the highest dosage of fibers.

N. Generosi et al.

20 18 16 14 12 10 8 6 4 2 0

UHPFRC-0.6 - Flexural test

A B

Load (kN)

Load (kN)

UHPC - Flexural test

C

0

0.5

1

1.5

2

2.5

3

3.5

20 18 16 14 12 10 8 6 4 2 0

4

A B C

0

Midspan displacement (mm)

20 18 16 14 12 10 8 6 4 2 0

Load (kN)

B C

0.5

1

1.5

2

2.5

3

3.5

1

1.5

2

2.5

3

3.5

20 18 16 14 12 10 8 6 4 2 0

4

A B C

0

Midspan displacement (mm)

0.5

1

1.5

2

2.5

3

3.5

Midspan displacement (mm)

Load (kN)

UHPFRC-2.6 - Flexural test 20 18 16 14 12 10 8 6 4 2 0

A B C 0

0.5

4

UHPFRC-1.9 - Flexural test

A

0

0.5

Midspan displacement (mm)

UHPFRC-1.25 - Flexural test

Load (kN)

1072

1

1.5

2

2.5

3

3.5

4

Midspan displacement (mm)

Fig. 3. Load-displacement curves of flexural tests on UHPFRC specimens

4

Characterization of Ultra High Performance Fiber Reinforced Concrete (UHPFRC)

3.2

1073

Direct Tensile Tests (Dog Bone Specimens)

Dog-bone specimens with a representative cross section of about 30 mm  45 mm and with a total length of 330 mm were used to experimentally determine the tensile stress-strain responses of UHPFRC (Fig. 4). Specimens were cast in wood frameworks (Fig. 2), after 2 days they were demoulded and kept up to 28 days in climatic chamber at 20 °C and 50% R.H. to test their behaviour in the worst curing conditions. A steel frame realized with welded steel plates and four steel cylinders (Fig. 4) was used to grab the specimen and to transfer the tensile force without applying compression at the specimen ends. In this configuration rotations at the ends of the specimen are allowed. Tensile tests have been performed on a tensile testing machine with a load capacity of 50 kN in displacement control, with a loading rate of 0.5 mm/min. Digital Image Correlation (DIC) was used to measure displacements and deformations in the specimens. During the tests, pictures of the frontal surface of the specimens have been acquired by two digital cameras (model Pixelink® B371F) at 2 frame per second, collecting about 600 images. The cameras were equipped with a lens having a focal length of 16 mm and placed about 400 mm away from the specimen, in order to reduce the perspective errors due to eventual out-plane motions. The specimen was illuminated using an LED spotlight. A second camera, placed on the right side, was used to monitor possible motions out of the plane. Tensile strain, e, was measured on a free length, LDIC, equal to 100 mm as marked in red in Fig. 4b. Results of the tensile tests have been reported in Table 3 (the average value of 3 specimens for each mixture and relative Coefficient of Variation (%)), specimens after testing have been shown in Fig. 6 and the stress-strain curves have been shown in Fig. 5. Tensile strength was calculated by dividing the tensile load recorded by the testing machine by the specimen’s cross section, which has been calculated as the mean value of three sections at both ends and the centre of the specimen (kept in the central part of the specimen with total length of 80 mm). The first cracking stress rt was defined as the tensile stress in the specimen corresponding to the first crack formation. The formation of the first crack has been determined by visual inspection (with the help of DIC frames) or looking at the stress-strain curves, in correspondence with a stress decrease in the elastic phase. The average maximum tensile strength rmax and the corresponding deformation emax, for each group of specimens were also reported in Table 3.

1074

N. Generosi et al.

(a)

(b)

Fig. 4. Specimens dimension in mm (a), and direct tensile test setup (b) Table 3. Experimental results of direct tensile tests on dog-bone specimens Specimen UHPC

Average CoV(%) UHPFRC-0.6 Average CoV(%) UHPFRC-1.25 Average CoV(%) UHPFRC-1.9 Average CoV(%) UHPFRC-2.6 Average CoV(%)

rt (MPa) 5.65 17 3.74 17 4.90 5 3.93 14 6.14 14

et (%) 0.037 28 0.029 31 0.035 17 0.036 25 0.062 34

rmax (MPa) emax (%) 5.65 0.037 17 28 4.49 0.07 14 19 5.40 0.49 6 45 5.88 0.56 17 5 8.11 0.87 10 40

In uniaxial tensile tests, specimens without fibers (UHPC) showed a linear elastic behaviour up to failure, with the formation of a single pass-through crack (Fig. 5 and 6). Failure was always brittle after the formation of one unique crack. The addition of hooked steel fibers, even at the lowest dosage (0.6% by volume), is able to avoid brittle failure and it allows the specimen to undergo plastic deformation after the formation of the first crack. However, with fibers dosages of 50 and 100 kg/m3, the formation of the first crack is reached at lower tensile stresses with respect to the plain UHPC specimen. This is probably due to a bad dispersion and misalignment of the fibers within the cross-section, thus reducing the resistant section: it might have constituted a discontinuity, thus anticipating the formation of the first crack. Only by exceeding a critical threshold on the fibers dosage (about 200 kg/m3) the addition of steel fibers is able to increase the tensile strength at first cracking.

Characterization of Ultra High Performance Fiber Reinforced Concrete (UHPFRC)

UHPC

UHPFRC-0.6

10

10 A

8

A

8

6

B 6

σ (MPa)

σ (MPa)

B C

4

C

4

2

2

0

0 0

0.5

1

1.5

2

2.5

3

0

0.5

1

ε (%)

1.5

2

2.5

3

ε (%)

UHPFRC-1.25

UHPFRC-1.9

10

10

A

A 8

B

8

6

C

6

σ (MPa)

σ (MPa)

1075

4 2

B (failed) C

4 2

0

0 0

0.5

1

1.5

2

2.5

3

0

0.5

1

ε (%)

1.5

2

2.5

3

ε (%)

UHPFRC-2.6 10

σ (MPa)

8 6 4

A

2

B C

0 0

0.5

1

1.5

2

2.5

3

ε (%)

Fig. 5. Stress-strain curves of direct tensile tests on dog-bone UHPFRC specimens

1076

N. Generosi et al.

It can be noticed that the post-cracking behaviour changes by varying the amount of steel fibers. For lower dosages (50 and 100 kg/m3) a softening behaviour is observed while with higher dosages (150 and 200 kg/m3) the post-cracking behaviour changes and significant stress-hardening branches can be noticed in the stress-strain curves (Fig. 5). This change from softening to hardening behaviour in direct tensile tests, when a critical threshold in fibers dosage is exceeded, is confirmed by some studies present in literature [11, 18]. Failure modes observed in UHPFRC specimens are different depending on the amount of steel fibers, as shown in Fig. 6. UHPFRC specimens remain intact due to the presence of the steel fibers while those without fibers failed immediately after the formation of a unique pass-through crack. Specimens with lower fibers dosages (UHPFRC-0.6 and UHPFRC-1.25) showed the formation of a big crack which grows at increasing load. Only specimens with higher fibers dosages (UHPFRC-1.9 and UHPFRC-2.6) showed the formation of multiple cracks. However, once the maximum load is reached, only one of them continues to grow in the softening branch.

Fig. 6. Failure modes of dog-bone UHPFRC specimens under tensile loads

4 Conclusions The objective of this study was to evaluate the tensile properties of UHPFRC mixtures, reinforced with different amount of 30-mm long hooked steel fibers, varying from 0.6% up to 2.6% by volume. Direct tensile tests have been carried out on dog-bone shaped specimens, with total length of 330 mm and cross section of 45  30 mm2. Compression and flexural tests have been performed on 40  40  160 mm3 specimens. Experimental results allowed to compare bending and tensile properties for UHPFRC specimens with different fiber dosages.

Characterization of Ultra High Performance Fiber Reinforced Concrete (UHPFRC)

1077

The flowing concluding remarks can be drawn: • The addition of hooked steel fibers allowed to significantly increase the compression and flexural strength of UHPFRC mixtures (up to 45% and more than 200%, respectively, at the highest dosage of 2.6%). • Flexural toughness has been greatly raised by increasing the steel fiber volume: with respect to the reference UHPC mixture without fibers, UHPFRCs were able to absorb up to 25-time higher energy amount with the highest dosage of fibers. • The addition of steel fibers significantly modified the behaviour of UHPFRC specimens subject to uniaxial tension, by avoiding their brittle failure even at low dosages (0.6% by volume). • The post-cracking behaviour in direct tensile tests was strongly influenced by the amount of steel fibers, passing from softening to hardening if high amount of fibers was used (higher than 1.9% by volume). • This study confirms that the post-cracking behaviour in uniaxial tensile tests is different from the post-cracking behaviour in bending tests. Once cracked, the UHPFRC is still able to increase its flexural strength even at low fibers dosage while its tensile strength can be increased only when high fibers volume is used.

References 1. Yu, R., Spiesz, P., Brouwers, H.J.H.: Mix design and properties assessment of Ultra-High Performance Fibre Reinforced Concrete (UHPFRC). Cem. Concr. Res. 56, 29–39 (2014). https://doi.org/10.1016/j.cemconres.2013.11.002 2. Yoo, D.-Y., Yoon, Y.-S.: A review on structural behavior, design, and application of ultrahigh-performance fiber-reinforced concrete. Int. J. Concr. Struct. Mater. 10(2), 125–142 (2016) 3. Song, Q., Yu, R., Shui, Z., Wang, X., Rao, S., Lin, Z.: Optimization of fibre orientation and distribution for a sustainable Ultra-High Performance Fibre Reinforced Concrete (UHPFRC): Experiments and mechanism analysis. Constr. Build. Mater. 169, 8–19 (2018). https://doi.org/10.1016/j.conbuildmat.2018.02.130 4. Corinaldesi, V., Donnini, J., Nardinocchi, A.: The influence of calcium oxide addition on properties of fiber reinforced cement-based composites. J. Build. Eng. 4, 14–20 (2015) 5. Banthia, N., Nandakumar, N.: Crack growth resistance of hybrid fiber reinforced cement composites. Cem. Concr. Compos. 25, 3–9 (2003). https://doi.org/10.1016/S0958-9465(01) 00043-9 6. Fantilli, A.P., Mihashi, H., Vallini, P.: Multiple cracking and strain hardening in fiberreinforced concrete under uniaxial tension. Cem. Concr. Res. 39, 1217–1229 (2009) 7. Bffup, A.: Ultra High Performance Fiber-Reinforced Concretes: Interim Recommendations: Scientific and Technical Committee. Assoc Française Genie Civ (2002) 8. Japan Society of Civil Engineers. Recommendations for Design and Construction of High Performance Fiber Reinforced Cement Composites with Multiple Fine Cracks (HPFRCC). Concr. Eng. Ser. 82: Testing Method 6–10 (2008) 9. fib Model Code for Concrete Structures 2010 (2013). https://doi.org/10.1002/ 9783433604090 10. Delle, C., In, S., Armato, C.: Consiglio Superiore dei Lavori Pubblici Servizio Tecnico Centrale. Cemento 93, 62–00146 (2003)

1078

N. Generosi et al.

11. Wille, K., El-Tawil, S., Naaman, A.E.: Properties of strain hardening ultra high performance fiber reinforced concrete (UHP-FRC) under direct tensile loading. Cem. Concr. Compos. 48, 53–66 (2014). https://doi.org/10.1016/j.cemconcomp.2013.12.015 12. Yoo, D.Y., Kim, S., Kim, J.J., Chun, B.: An experimental study on pullout and tensile behavior of ultra-high-performance concrete reinforced with various steel fibers. Constr. Build. Mater. 206, 46–61 (2019). https://doi.org/10.1016/j.conbuildmat.2019.02.058 13. Wille, K., Kim, D.J., Naaman, A.E.: Strain-hardening UHP-FRC with low fiber contents. Mater. Struct. Constr. 44, 583–598 (2011). https://doi.org/10.1617/s11527-010-9650-4 14. Kamal, A., Kunieda, M., Ueda, N., Nakamura, H.: Evaluation of crack opening performance of a repair material with strain hardening behavior. Cem. Concr. Compos. 30, 863–871 (2008). https://doi.org/10.1016/j.cemconcomp.2008.08.003 15. Jun, P., Mechtcherine, V.: Behaviour of strain-hardening cement-based composites (SHCC) under monotonic and cyclic tensile loading: Part 1 - Experimental investigations. Cem. Concr. Compos. 32, 801–809 (2010). https://doi.org/10.1016/j.cemconcomp.2010.07.019 16. EN 197-1. Cement - Part 1: Composition, specifications and conformity criteria for common cements (2011) 17. EN 1015-11. Metodi di prova per malte per opere murarie Parte 11 : Determinazione della resistenza a flessione e a compressione della malta indurita (2007) 18. Naaman, A.E., Reinhardt, H.W.: Proposed classification of HPFRC composites based on their tensile response. Mater. Struct. Constr. 39, 547–555 (2006). https://doi.org/10.1617/ s11527-006-9103-2

Slip-Hardening Bond: A Key to the Success of Ultra High Performance FRC Composites Antoine E. Naaman(&) Department of Civil and Environmental Engineering, University of Michgan, Ann Arbor, USA [email protected]

Abstract. Bond is recognized as a fundamental causal parameter in the mechanics of fiber reinforced composites. Its paramount function is to transmit forces between fibers and matrix and vice-versa. This paper focuses on “sliphardening bond” which can be achieved in cement composites with some fibers, as compared to commonly observed slip-softening bond or constant bond. In particular: 1) it describes how bond is characterized from a pull-out test or a pull-through test on a single fiber; 2) it gives examples of fibers with bond stress versus slip exhibiting slip-hardening behavior; and 3) it clarifies how sliphardening response can be likely achieved. Slip-hardening bond is an extremely important characteristic which should be evaluated at the onset of design; it implies that, at the composite level, once a crack is formed in the matrix, the fibers bridging the crack provide increasing resistance to crack opening. This is likely to encourage multiple cracking in the composite and helps lead to composites with strain-hardening behavior in tension, as well as large composite strains at failure. Slip-hardening bond is considered critical for the further development and success of high performance and ultra-high performance fiber reinforced cement composites [1, 6]. Keywords: Concrete
Fiber
Fiber concrete
Bond
Pull-out
Pull-through
Slip-softening
Slip-hardening
UHPC composites

1 Introduction Bond is recognized as a fundamental causal parameter in the mechanics of fiber reinforced composites. Its paramount function is to transmit forces between fibers and matrix and vice-versa. The success of composite action, as described by the composite’s strength, modulus, fracture energy, and ductility, depends on bond. Bond is not a physical variable; unlike the fiber or the matrix it does not have a volume, but it is as important. Of the three key components of a composite (fiber, matrix, and bond), bond is the most difficult to define, understand, characterize, control, optimize, … and, in short, the most elusive [Chapter 13 in 1]. Bond is also the most critical differential property among fibers of same material (such as in the case of steel fibers). In fiber reinforced cementitious composites, bond is mostly evaluated through single fiber pull-out tests during which the applied load and the fiber slip are continuously measured leading to a pull-out load versus slip relationship, from which a bond © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 1079–1089, 2021. https://doi.org/10.1007/978-3-030-58482-5_95

1080

A. E. Naaman

stress versus slip curve is derived [2–5]. Such a curve can be considered a fundamental constitutive relation for bond and can vary in nature from brittle to ductile and similarly from slip-softening to slip-hardening behavior. Moreover, precise pull-out tests allow us to uncover various bond mechanisms, and then characterize, model separately, and integrate them in analytical models to optimize composite properties. This paper focuses on “slip-hardening bond” which can be achieved in fiber reinforced cement composites, as compared to commonly observed slip-softening bond or constant bond. In particular: 1) it describes how bond is identified from a pull-out test or a pull-through test on a single fiber; 2) it gives examples of fibers with bond stress versus slip exhibiting slip-hardening behavior; and 3) it clarifies how sliphardening can be possibly achieved. So far, the most effective way to achieve sliphardening wiht steel fibers is through mechanical deformations along the fibers. Sliphardening bond is an extremely important characteristics of some fibers; it implies that, at the composite level, once a crack is formed in the matrix, the fibers bridging the crack could be designed to provide increasing resistance to crack opening. This is likely to encourage multiple cracking in the composite and helps lead to composites with strain-hardening behavior in tension, as well as large composite strains and high energy absorption at failure. This author considers that slip-hardening bond is essential for furthering the development and success of strain-hardening high performance and ultrahigh performance (UHPC) fiber reinforced cement composites.

2 Basic Pull-Out and Pull-Through Tests for Characterization of Bond In order to better understand bond, it is helpful to understand typical experimental bond tests. Two are described in Fig. 1, namely the pull-out test and the pull-through test. Both lead to a pull-out load versus slip relationship from which various bond parameters can be extracted. In the pull-out test, the fiber embedded length (that is the length of contact between fiber and matrix) decreases with an increase in slip; while with the pull-through test, the embeded length remains constant. The pull-out load versus slip response recorded from a test is used to determine the maximum and average bond stress (over a given slip) to use in various mechanical models and in design. Generally, the maximum load obtained in a pull-out or pull-through test is used to determine the maximum bond strength. In the mechanics of materials, bond stress is essentially defined as a shear stress and thus the terms are interchangeably used in this paper, often referring to bond-shear stress. Figure 1 (right end) also shows a “push-through” test generally used with microfibers and fibers that are very delicate to be gripped effectively. The specimen for such test is obtained by slicing a prism containing one longitudinal fiber; the slice is generally very thin and the fiber is pushed through by an indenter leading to a load versus slip response from which a bond measure is derived. This test is not addressed here but mentioned for completeness. Figure 2 shows photos of pull-out and pull-through test set-ups.

Slip-Hardening Bond

P

P

Fiber

Penetration surface

Push-Through

Pull-Through

Pull-Out

S End

Fiber

Slip

Indenter Matrix

Matrix

Matrix

P

Le

Variable contact length during test

Le

1081

Constant contact length during test

Fiber

Fig. 1. Typical illustrations of tests to determine the bond or shear stress versus slip response of a fiber embedded in a cement matrix [Ref. 1].

(a)

(b )

(c)

Fig. 2. (a) Pull-out test set-up. (b) Photo of ongoing pull-out test. (c) Photo of ongoing pullthrough test [Refs. 1–4, 8].

3 Basic Calculations of Bond Stress or Shear Stress Given the results of a pull-out test which measures load, P, and slip, S (Figs. 1 and 2), the following relation is typically used to estimate a bond shear stress versus end slip response [1]: (

PðSÞ sðSÞ ¼ wðL for a fiber of any section e SÞ PðSÞ sðSÞ ¼ pdðL for a circular fiber e SÞ

ð1Þ

1082

A. E. Naaman

where S is the end slip, sðSÞ is the bond or shear stress at slip S, PðSÞ is the pull-out load at slip S, w is the fiber perimeter (p in other notation), d is the fiber diameter, and Le is the embedded length of the fiber. Similarly, for a pull-through test where the area of contact between fiber and matrix remains constant, the bond shear stress at any slip can be calculated from: (

PðSÞ sðSÞ ¼ wL for a fiber of any section e PðSÞ sðSÞ ¼ pdL for a circular fiber e

ð2Þ

In the pull-through test, the slip can be defined at the midsection or an average between the two ends; the embedded length can be small but is generally taken to represent the average surface condition of the fiber; for example, for an indented fiber, a slice with at least one or two indentations is needed. The bond-shear stress versus slip relationship provides an excellent means to compare different types of fibers. Another measure that also helps evaluate the efficiency of a typical fiber under pull-out in comparison to other fibers, is the tensile stress generated in the fiber by the pull-out load recorded; it is simply given as the pull-out load at a given slip divided by the cross-sectional area of a fiber. Thus: rðSÞ ¼

PðSÞ Af

ð3Þ

The higher the tensile stress generated the more efficient is the fiber; an optimum design would seek to maximize the stress without leading to failure of the fiber. Often in practice, the bond-shear strength (or maximum stress), smax , is calculated for the maximum pull-out load Pmax but assuming S ¼ 0 in Eq. (1), that is, assuming the initial embedment length, Le , even though the test is a pull-out test. This is because generally the maximum load occurs at a relatively small slip. Thus for all practical purposes one can write: smax 

Pmax wLe

ð4Þ

Many parameters affect the shape of the bond stress versus slip response of a fiber and include chemical surface treatment as well as mechanical surface indentation or deformation. Numerous examples are given in Chapter 13 of Ref. [1].

4 Typical Bond-Shear Stress Versus Slip Curves 4.1

Typical Idealized Curves

For convenience, the pull-through test is easier to understand at first (Figs. 1 and 2). Indeed, in a pull-through test, the bond contact surface between fiber and matrix remains constant during the test, making the pull-out load directly proportional to the bond-shear stress (Eq. 2). Thus the load axis, can be visualized at the bond-shear stress

Slip-Hardening Bond

1083

axis. In Fig. 3, the contact surface has a width Le, but could generally be considered an increment dx in an integration for a pull-out test. Theoretically, if the bond shear stress versus slip is to be considered a true constitutive property, then the slip should be the observed displacement of the fiber with respect to the matrix over a distance dx considered representative of the fiber surface.

Fig. 3. Typical bond shear stress versus slip curves illustrating the definition of some variables and terminology [Ref. 1].

Figure 3 illustrates the types of bond-shear stress versus slip curves observed with steel fibers. Generally the curves start with an ascending portion (almost vetical linear at the scale plotted) up to a point beyond wich a significant decrease in stiffness (slope of the curve) occurs followed by one of four cases: 1) brittle failure or sudden drop of bond stress to zero); 2) decaying bond (also called slip-softning) or slow gradual decrease of bond stress with increasing slip; 3) constant bond, a mostly hypothetical case where bond remains constant with slip (this is also a theoretical boundary condition between slip-softening and slip-hardening); and 4) slip-hardening bond, where the bond stress increases with slip up to a reasonable and useful level of slip. Most smooth steel fibers as well as carbon, glass, and kevlar fibers and many polymeric (PP, PE, Nylon, polyesther) generally show a decaying bond-shear stress versus slip response. Some steel fibers with mechanical deformation may show slip-hardening bond up to a different extensts of slip; for instance hooked-end fibers embedded in

1084

A. E. Naaman

normal strength concretes, show a slip hardening bond behavior up to a slip of about 1.5 mm (about equal to the hook length) while, everything esle being equal, an equivalent diameter twisted steel fiber can show a slip-hardening bond up to 10 mm. The difference is critical in the development of strain-hardening FRC composites (defined as having a post-cracking strength higher than their cracking strength) and in drastically improving their energy absorbtion capacity. Although the curves in Fig. 3 assume a pull-through test to illustrate the direct proportionality between the pull-through load and the bond-shear stress, similar curves can be observed when derived from pull-out tests results using Eq. (1) instead of Eq. (2). 4.2

Example of Pull-Through and Pull-Out Test Results for Bond with Steel Fibers

Two examples of pull-through load versus slip response curves for a smooth steel fiber are shown in Fig. 4 [2]. The test was carried out as per Fig. 2c. Since according to Eq. (2), the bond-shear stress is directily proportional to the load, the y axis could represent the bond-shear stress. It can be observed that in this example, after the initial peak point, the bond slightly deteriorates with increasing slip, thus leading to a slipsoftening or decaying bond up to a slip representing about 20% of the embedded length. Note that for a pull-out load of 10 lb, the corresponding bond-shear stress is about 318 psi (2.2 MPa); a much smaller value of bond-shear stress would be used in design to reflect an average response related to the expected slip.

Fig. 4. Example of pull-through test illustrating decaying bond with a smooth steel fiber and influence of having the fiber embedded in a plain matrix versus a SIFCON matrix [Ref. 2].

Given any experimental pull-out load versus end slip curve for a fiber (as per the test of Figs. 2a and 2b), a bond shear stress versus end slip curve can be obtained from

Slip-Hardening Bond

1085

Eq. (1), point by point, and may take on various forms similar to those described in Fig. 3. Figure 5a shows the actual pull-out load versus slip response obtained from a pullout test such as shown in Fig. 2a. Here, after an initial steep ascending response, the load drops suddenly and then decreases at a slow rate with increasing slip. Using Eq. (1), the bond-shear stress is first computed from the pull-out load, and the corresponding bond-shear stress versus slip is plotted in Fig. 5b [5]. It can be observed that, after the maximum point, the stress decreases suddenly then stabilizes up to relatively large slips about 50% of embedded length. The value of bond at the plateau level (1.2 MPa) is only 38% of the maximum bond stress (3.2 MPa). Such behavior justifies the often used assumption in design that bond is about constant (for example taken as 1.2 MPa), in which case the initial portion of the curve is ignored.

Fig. 5. (a). Typical pull-out load verus slip response from a test such as in Fig. 2a. (b). Derived bond-shear stress versus slip response for the curve of Fig. 5a [Refs. 2, 5].

Figures 4 describes two examples of typical decaying or slip-softening bond-shear stress versus slip curves (the bond-shear stress is proportional to the load). If we ignore the initial portion which occurs over a very small slip, Fig. 5b can be used as a typical example of a constant (1.2 MPa) bond-shear stress versus slip response. Examples of slip-hardening response are described next.

5 Examples of Experimental Slip-Hardening Bond Shear Stress Versus Slip Response Most generally, the bond shear stress versus slip relationship of smooth fibers is observed to be with decaying frictional response; however, in some cases, sliphardening can be observed (Figs. 6 and 7). Slip hardening is attributed to some mechanical action which is not intuitive with smooth fibers. For instance, in the case of smooth brass-coated steel fibers embedded in UHPC, the brass-coating, if scraped from the surface during pull-out, creates a mechanical blocking that may lead to a slip

1086

A. E. Naaman

hardening response; this is illustrated in Fig. 6 taken from [5]. Note the very high values of bond stress achieved with an UHPC of compressive strength exceeding 200 MPa. In a similar manner monofilament PP fibers can fibrillate along their surface during pull-out significantly improving their frictional resistance leading in some cases to a slip-hardening bond behavior, although the numerical values of bond are much much smaller.

Fig. 6. Slip-hardening bond-shear stress versus slip curves for brass coated steel fibers embedded in UHPC matrices with various amounts of fine sand; embedded length = 6.5 mm [Ref. 5].

Figure 7 illustrates several bond stress versus end slip curves obtained from pullout load versus slip tests of high strength twisted steel fibers [5–8]. The x axis represents the normalized value of slip divided by the embedded length. It can be observed that, following the initial ascending portion, the bond stress for the twisted fibers increases continuously with increasing slip, leading to very high bond stresses at large slips (up to 80% of embedded length), prior to complete pull-out. The curves can be described as “slip-hardening”. Slip-hardening characteristics have been also observed with some other deformed fibers such as hooked-ends steel fibers, but hardening was limited along a slip smaller than the length of the hook, prior to decaying, suggesting the important contribution of the mechanical component of bond. Note generally that while slip-hardening is not very common with smooth fibers, it is more likely to develop with mechanically deformed fibers at least over the range of slip over which the mechanical deformation provides the most anchorage and resistance. There is need to explore fully this type of behavior and carry out in depth research to uncover if other treatments, such as chemical treatment of the fiber surface or a change in fiber material, lead to a slip-hardening bond-shear stress versus slip response.

Slip-Hardening Bond

1087

Fig. 7. Slip-hardening bond shear stress versus end slip curves obtained from experimental pullout load versus slip curves with twisted triangular steel fibers Eq. (1) [Refs. 1, 6–9].

6 Concluding Remarks As evident from the above discussion, bond is a very complex subject. Some commendable successes have been achieved so far in understanding and modeling bond in cementitious composites, but more can be done. An important simplification used in design, is to assume a bond shear stress that is constant and independent of slip; this can be a very optimistic and unsafe guess-estimate. Indeed, it is not realistic to consider bond stress without associating a maximum slip value to it; both should be quantified for a particular application. Thus, the author strongly recommends for standard design, to select an average value of bond-shear stress (such as from Eq. (1 or 2)) over a predetermined slip based on expected or prescribed crack opening in the composite, and for a given level of performance (service or ultimate limit state, or in-between). A very safe lower bound value can be taken as the equivalent bond obtained from the pull-out work such as estimated from Eq. (13.8) in Ref. [1] and in Ref. [10]. Among the numerous methods to improve bond in steel fibers, the use of suitably designed twisted triangular or square shaped fibers offer so far the most advantages [6– 9]. Proper design implies that the fibers un-twist during pull-out or crack opening. Note that round fibers do not offer the benefits of mechanical ribs when twisted; and very flat rectangular fibers, if twisted, will form tunnel like sections that may not be penetrated by the cement matrix and are undesirable sites of stress concentration. Crimping, which induces a sinusoidal wave form, while effective in improving mechanical bond, leads to a significant reduction in the equivalent elastic modulus of the fiber. Hooked and paddled ends fibers offer an effective anchorage concentrated at their ends, but not distributed along the fiber, thus unable to help for bond at large slips; thir corresponding pull-work [11], while better than that of smooth fibers, can be significantly smaller than that of equivalent twisted steel fibers. To achieve strain-hardening behavior in tension, the post-cracking tensile strength of the composite should exceed its strength at first precolation cracking, and one can

1088

A. E. Naaman

set, for a given composite and fiber, a critical value of fiber volume fraction to insure such behavior [1]. The post-cracking tensile strength is tied to the product s *Vf *L/d [1]; practical limitations on both Vf and L/d (such as for instance for steel fibers, L/d < 100 and Vf < 3% in the premix process), leave the bond as the next most critical causal variable. In UHPC where the practical limits on Vf and L/d are about attained, slip-hardening bond becomes crucial not only performance-wise (post-cracking tensile strength) but also ductility-wise (particularly in improving the strain at peak tensile stress), and therefore cost-wise as for any other FRC composite. Future developments in new fibers for concrete should focus on improving fiber efficiency and achieving better bond characteristics, preferably with bond stress-slip hardening behavior. Fibers with bond properties characterized by a bond shear stress versus slip response that is either elastic-plastic or better (slip hardening) up to relatively large slips, enhance crack bridging resistance of a composite under larger crack openings and thus enhance structural performance and damage tolerance in all respects. Such fibers allow the practical development of high performance and ultra-high performance (UHPC) FRC composites capable of exhibiting strain-hardening behavior in tension at relatively low volume fractions of fibers. They are suitable in structural applications in combination with reinforcing bars or prestressing strands such as in blast and seismic resistant structures, as well as in stand-alone applications (with no other reinforcement) such as in thin sheet products for housing, claddings for buildings, precast products, shells, pipes, and the like. Acknowledgements. The author has carried out research on bond in fibers for concrete since 1970, as part of his Ph.D. thesis. Since then, the following former colleagues and students have helped in the evolution of ideas briefly described in this paper: S.P. Shah, U. Gokoz, K. Visalvanich, G. Nammur, K. Kosa, H. Najm, J. Alwan, R.E. Robertson, P. Guerrero, C. Sujivorakul, A. Waas, D.J. Kim, and K. Wille. Their collaboration is sincerely acknowledged. The author offers his apologies to all those who have specifically addressed in their research “slip-hardening” bond but are not cited here due to space limitations.

References 1. Naaman, A.E.: Fiber Reinforced Cement and Concrete Composites. Techno Press 3000, 765 p. (2018). http://www.technopress3000.com/books. ISBN 978-0-9674939-3-0; LCCN 2017916342 2. Naaman, A.E., Najm, H.: Bond-slip mechanisms of steel fibers in concrete. ACI Mater. J. 88 (2), 135–145 (1991) 3. Guerrero, P., Naaman, A.E.: Effect of mortar fineness and adhesive agents on pullout response of steel fibers. ACI Mater. J. 97(1), 12–20 (2000) 4. Sujivorakul, C., Naaman, A.E.: Evaluation of bond-slip behavior of twisted wire strand steel fibers embedded in cement matrix. In: Balaguru, P., Naaman, A.E., Weiss, W. (eds.) Proceedings of ACI Symposium on Concrete: Material Science to Applications, a Tribute to S.P. Shah, ACI SP-206, pp. 271–292, April 2002

Slip-Hardening Bond

1089

5. Wille, K., Naaman, A.E.: Bond-slip behavior of steel fibers embedded in ultra high performance concrete. In: Dresden, V.M., Kaliske, M. (eds.) Proceedings of 18 European Conference on Fracture and Damage of Advanced Fiber-Reinforced Cement-Based Materials, Contribution to ECF 18, pp. 99–111. Aedificatio Publishers, Freiburg, September 2010 6. Naaman, A.E.: Fibers with slip-hardening bond. In: Reinhardt, H.W., Naaman, A.E. (eds.) High Performance Fiber Reinforced Cement Composites - HPFRCC 3, RILEM Pro 6, pp. 371–385. RILEM Publisations S.A.R.L., Cachan, France, May 1999 7. Naaman, A.E.: Engineered steel fibers with optimal properties for reinforcement of cement composites. J. Adv. Con. Technol. 1(3), 241–252 (2003) 8. Naaman, A.E., Sujivorakul, C.: Pull-out mechanisms of twisted steel fibers embedded in concrete. In: Proceedings of International Conference on Applications of Shotcrete, Tasmania, Australia, 7 p., April 2001 9. Sujivorakul, C., Naaman, A.E.: Modeling bond components of deformed steel fibers in FRC composites. In: Naaman, A.E., Reinhardt, H.W. (eds.) High Performance Fiber Reinforced Cement Composites (HPFRCC-4), , pp. 35–48. RILEM Pub., Pro. 30, June 2003 10. Kim, D.J., El-Tawil, S., Naaman, A.E.: Loading rate effect on pull-out behavior of deformed steel fiber. ACI Mater. J. 105(6), 576–584 (2008) 11. Alwan, J., Naaman, A.E., Hansen, W.: Pull-out work of steel fibers from cementitious matrices - analytical investigation. J. Cem. Concr. Compos. 13(4), 247–255 (1991)

Testing of Thin UHPFRC Cantilever Stairs with Bolted Connections Ioan Sosa(&), Camelia Negrutiu, Bogdan Heghes, and Adel Todor Technical University of Cluj-Napoca, Cluj-Napoca, Romania [email protected]

Abstract. The purpose of this research is to analyze the suitability of UHPFRC for applications to precast cantilever stairs that can be easily connected to the main structure. Several cantilever stair elements were tested under a concentrated static load applied at the free end. The anchored end was connected to the testing frame with four bolts that provided a partial fixed end. Each stair had a tread, a riser and an end plate casted with four holes to accommodate the bolts. The riser and the tread had a thickness of 20 mm whereas the end plates were 20 mm and 30 mm thick. The length of the stairs was 600 mm. Only short steel fibers were provided as reinforcement (2.35 vol.%) resulting a 180 MPa compression strength and 25 MPa flexural strength. The failure of the elements occurred at concentrated load values of 4 to 7 kN depending on the thickness of the end plate. The equivalent static force was above the one resulting from standardized static loads for staircases. Keywords: Fibers Precast
Bolts


Ultra-high performance concrete
Cantilever
Stairs


1 Introduction Ultra High Performance Fiber Reinforced Concrete (UHPFRC) is one of most promising concretes for the construction industry. Its strong points such as ultra-high compressive strength (over 150 MPa), very good durability, ductility provided by fiber reinforcement lead to slender cross-sections and a significantly decrease of the concrete elements’ weight [1, 2]. The reduced maintenance costs and improved performances recommend it as a sustainable construction material [3]. However, uncertainties about UHPFCR behavior and the lack of standardized design rules limited the expected boost of UHPFRC use. The most significant and complete publications related to UHPFRC are referring to the few commercially available compositions (Ductal, CERACEM, CEMTECmultiscale): Setra&AFGC French recommendations (2002 and updated 2012) [2], JSCE recommendation (2006) [3], U.S. Federal Highway report in 2006- updated state-of-the art report (2013) [4]. Some large-scale structures are available worldwide but are limited mainly to footbridge applications. Some of the most impressive are in: France (Pont du Diable Footbridge, 2008), Germany (Gärtnerplatzbridge Kassel, 2007), South Korea (Seonyu Footbridge, 2002), Japan (Sakata Mirai Footbridge, 2002), Spain (Ovejas ravine in Alicante, 2013). One of the uncertainties regarding UHPFRC is the fiber orientation and its dependability of the casting procedure. As most of the researches use © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 1090–1099, 2021. https://doi.org/10.1007/978-3-030-58482-5_96

Testing of Thin UHPFRC Cantilever Stairs with Bolted Connections

1091

self-compacting fiber reinforced concrete, the fibers movement and rotation possibilities are far more pronounced that those in a less workable concretes. Two trends are observed for fibers orientation depending elements shape: beam like and slab like elements. In the beam-like elements the concrete can unrestrictedly flow in the longitudinal direction conducting to a funnel type flow dominated by shear and wall effects. In the slab-like elements the concrete can flow unrestricted in any horizontal direction namely a radial flow [5, 6]. However, efforts are made to use the preferential orientation of the fibers as a way to improve the flexural behavior of UHPFRC elements [7, 8].

2 Research Scope The scope of the research is to develop a series of easy to assemble precast UHPFRC cantilever elements, with different connections types, that would appeal to the construction industry. The first stage was the experimental testing and analysis of the UHPFRC stairs with no other reinforcement but fibers. The possible advantages of UPFRC stairs are the thin walls and the cantilever system with a great impact on the esthetics of a building. Moreover, using bolts to connect the stairs to the main structure greatly reduces the time involved in the construction. The advantage of the bolts is that they can be either fixed in concrete during casting or more conveniently post-installed using chemical anchors. This way, the support elements for the stairs in the form of concrete walls or beams can be cast without interruption. The research findings will be used to extend the research to other types of cantilever elements that require a small self-weight and fast assembling such as canopies and balconies. Moreover, other types of connections are planned to be tested including by post tensioning which will also increase the capacity.

3 Experimental Research 3.1

Cantilever Stairs Design

The cantilever stairs were cast without reinforcement except the short steel fibers that were included in the concrete composition. Each stair consisted of a tread, riser and an end plate with four holes (25 mm diameter) to accommodate the bolts for the connection. Two sets of stairs were researched: Set A and Set B (Fig. 1). Set A had an end plate with a thickness of 20 mm and Set B had an end plate of 30 mm. Both Sets, A and B, had 20 mm thick treads and the rise. Supplementary, Set B was casted with a 30 mm strengthening shoulder at the intersection of the tread with the end plate. The variation was designed in order to increase the bending capacity of the end plate. For each set, two stairs elements were casted.

1092

I. Sosa et al.

Fig. 1. Design of the stairs: Set A (left) and Set B (right)

3.2

Concrete Composition, Specimens, Casting and Curing

The UHPFRC composition was developed within the Technical University of Cluj research group on concretes and the concrete was extensively tested for mechanical behavior before [9]. The cement and the sands were locally produced, whereas the silica fume and steel fibers were imported from the EU. The components (Table 1) were: early strength CEM I 52,5 R produced by Holcim, dry silica fume powder with min 90% SiO2 (Microsilica Grade 940) produced by Elkem, 0.175 mm diameter/6 mm long straight fibers produced by Baumbach Metall, modified acrylic polymer superplasticizer Dynamon SR3 by Mapei. The locally used aggregates consisted of quartz sands divided into fine sand (0 to 0.3 mm) medium fine sand (0.3 to 0.63 mm), and coarse sand (0.63 to 1.2 mm). No quartz powder was added. The fiber volume was 2.35 vol.%. The concrete resulting workability was high as the concrete freely flowed within the formworks without the need for vibrations. Table 1. UHPFRC composition [kg/m3]. Binder Cement Silica fume Water Super. W + S/B Fibers Sands 740 156 150 60 0.23 185 1242

The formwork of the stairs was crafted out of wood which was brushed with release agent to minimize the water absorption by the wood. The wood formwork is not that best option for UHPFRC as any water loss can affect the properties but was chosen as it is easy to work with and has low costs. Two identical formworks were produced in order to cast the stairs from the same concrete batch. As the direction of casting can induce preferential orientation of the fibers, the decision was to pour from the riser part of the stairs (Fig. 2). Due to the concrete flow and formwork walls, a favorable, longitudinal direction orientation of the fibers was presumed for the riser and the tread, meaning a potential better flexural capacity.

Testing of Thin UHPFRC Cantilever Stairs with Bolted Connections

1093

However, the fiber orientation in the end plate would be unfavorable in terms of flexure as the fibers would align parallel with the presumed flexural cracks.

Fig. 2. Stair elements positioned for casting (left), flexural test specimens (right)

In order evaluate the fiber orientation and flexural behavior, two series of thin plate elements were casted, both with a thickness of 20 mm, identical with the stairs tread and riser. One of the plate series was poured in a similar way with the stair tread (horizontal) and one in horizontal position similar with the riser (vertical). Furthermore, two series of 40 mm x 40 mm x 160 mm prisms were cast horizontally and vertically to further assess the influence of fibers orientation on flexural behavior (Fig. 2). The vertical cast prims would emulate the end plate in terms of fiber orientation. To evaluate the concrete compression strength, 50 mm x 50 mm x 50 mm cubes were casted. The relatively small size of the specimens was chosen because of the small particle size of the aggregates (sands) and the thin cross section of the stair elements. After pouring the concrete settled in the formworks for 24 h, followed immediately after demolding by thermal treatment for 24 h. The parameters of the thermal treatment were: temperature of 90°C and relative humidity of 80%. 3.3

Testing Methodology

All the specimens and the stair element were tested approximately one week after the finish of the thermal treatment. The compressive strength was determined on 50x50x50mm cubes using an automated testing machine. The flexural behavior was tested using the three-point flexural test setup for the prisms and the thin plates (Fig. 3). The span was 140 mm for all flexural test specimens. The thin plates were cut out of the casted 800 mm long plates. Displacement and force were recorded by a computer via a data acquisition system. Due to the test setup, the displacement transducer was placed on the machine frame, thus the initial readings for mid-span displacement included the machine movement. Subsequently, those initial movement due to preload were removed from the diagrams. The stair element was connected to a rigid steel frame using four M20 bolts and 50 mm washer. The stair element was positioned upside down to facilitate the application of the load (Fig. 4). One concentrated load was applied at the free end of the stair

1094

I. Sosa et al.

Fig. 3. Flexural test setup of prims (left) and thin plates (right)

element using 100 mm x 100 mm steel plate, thus resulting a span of 550 mm for the cantilever. The displacement transducer was connected at the very end of the cantilever. Both the force sensor and displacement transducer were connected to the computer.

Fig. 4. Test setup of the cantilever stairs.

4 Results and Discussion The obtained compression and flexural strengths results are listed in Table 2. The average compressive strength was 186 N/mm2 and is more than the minimum required for the ultra-high-performance concretes (150 N/mm2). The flexural strength and behavior were tested on two different kinds of specimens, each casted in two positions in order to evaluate the preferential fiber orientation effect due to casting. Regarding the thin plates the effect of casting direction was minimal,

Testing of Thin UHPFRC Cantilever Stairs with Bolted Connections

1095

Table 2. Compression and flexural strength of specimens Strength test

Specimen type

Specimen size [mm] Compression Cubes 50  50  Flexure Plate vertical 20  95  Flexure Plate horiz. 20  95  Flexure Prism vertical 40  40  Flexure Prims horiz. 40  40 

Average value

50 140 140 140 140

[N/mm2] 186 23 25 19 24

both the flexural strengths and displacements being in the same range (Figs. 5 and 6). Thus, is presumed that the riser and the tread of the stair elements have similar distribution of the fibers and consequently will have comparable response in flexure. A more pronounced fiber orientation effect was observed for the prims. The vertically cast prims had an approximately 20% lower flexural strength compared with the horizontally cast prims or either of the thin plates. Therefore, the end plate, casted in a similar position with the vertically cast prims is presumed to have lower strengths than the riser or the tread.

35

Vertical cast thin plates V1 V3 V4 V5

30

fct,fl [N/mm2]

25 20 15 10 5 0 0

500 1000 1500 2000 2500 3000 3500 4000 4500 5000

Middle span displacement [μm] Fig. 5. Flexural stress – displacement of vertical cast thin plates

The cantilever stairs elements test results are listed in Figs. 7, 8, 9 in terms of force vs. displacement. One of the elements from Set B was damaged during demolding and was excluded from testing. Set B stairs with an end plate of 30 mm and a strengthening shoulder have an almost 85% larger maximum loading capacity compared with the set A with an end plate of 20 mm. Both sets displayed a semi-ductile behavior, with slow development of the crack openings. The prescriptions of EN 1991-1-1:2004 regarding the imposed loads for stairs require a minimum concentrated load of 2 kN for stairs applied in the most unfavorable position. In the case of the tested cantilever stairs the

1096

I. Sosa et al.

35

Horizontal cast thin plates H1 H2 H3 H4 H5

30

fct,fl [N/mm2]

25 20 15 10 5 0 0

500 1000 1500 2000 2500 3000 3500 4000 4500 5000

Middle span displacement [μm] Fig. 6. Flexural stress – displacement of horizontal cast thin plates

most unfavorable position is the free end. Considering a safety factor of 1.5 the required imposed load is 3 kN. The tested cantilever stairs exceeded the loading requirements of the EN 1991-1-1:2004: Set A with 30% and Set B with 145%.

5 Set A, Element 1

Force [kN]

4 3 2 1 0 0

2

4

6

8 10 12 14 16 18 20 22 24 26 28 30 Free end displacement [mm]

Fig. 7. Force – Displacement of Set A, Element 1 (20 mm thick end plate)

The thickening of the end place from 20 mm to 30 mm also changed the pattern of cracks (Fig. 10). Whereas for the Set A all the cracks developed within the end plate, the Set B element displayed failure cracks on the riser as well. Nevertheless, for both types of elements the concrete cracks initiated on the end plate, at the first row of bolts, indicating a semi-rigid connection.

Testing of Thin UHPFRC Cantilever Stairs with Bolted Connections

1097

5 Set A, Element 2

Force [kN]

4 3 2 1 0 0

2

4

6

8 10 12 14 16 18 20 22 24 26 28 Free end displacement [mm]

Fig. 8. Force – Displacement of Set A, Element 2 (20 mm thick end plate)

8 Set B, Element 1 7 6

Force [kN]

5 4 3 2 1

28

26

24

22

20

18

16

14

12

10

8

6

4

2

0

0 Free end displacement [mm] Fig. 9. Force – Displacement of Set B, Element 1 (30 mm thick end plate + shoulder)

1098

I. Sosa et al.

Fig. 10. Crack patterns after failure: Set A with 20 mm end plate (left) and Set B with 30 mm end plate (right)

5 Conclusions The present research focused on the potential use of UHPFRC cantilever stair elements, with fiber reinforcement only and a simple bolted connection. The experimental flexural tests on cantilever elements and concrete specimens led to the following conclusions: • The cantilever stairs exceeded with up to 145% the required imposed loads from EN 1991-1-1:2004 • The bolted steel connection offers only a fixed end to the stairs, the crack initiation and failure was observed in all cases in the end plate • Increasing the end plate thickness by 10 mm increased the loading capacity by 85% • Preferential fiber orientation due to casting and formwork walls can decrease the flexural strength by 20% • The research was limited to a relatively small span for the stairs and a single type of connection: larger openings and moment resting connections will be tested and evaluated together with a finite element analysis.

References 1. Naaman, A., Wille, K.: The path to UHP-FRC: Five Decades of Progress. In: Proceedings of Hipermat 2012 3rd International Symposium on UHPC and Nanotechnology for High Performance Construction Materials, Kassel, 7–9 March 2012, pp. 3-15 (2012). 2. Respledino, J., Toutlemonde, F., et al.: Ultra-high performance fiber-reinforced concretes. Interim Recommend., AFGC -Setra (2002) 3. Japan Society of Civil Engineers, ‘Recommendations for design and constructions of ultra high strength concrete structures’, Japan (2006) 4. Russell, H., Graybeal, B.: UHPC ‘A State-of-the-Art Report for the Bridge Community’, US Highway Admin.,Report No. FHWA-HRT-13–060, USA (2013)

Testing of Thin UHPFRC Cantilever Stairs with Bolted Connections

1099

5. Lataste, B., et al.: Assessment of fiber orientation in ultra high performance fiber reinforced concrete and its effect on flexural strength. Mater. Struct. 43(7), 1009–1023 (2010) 6. Bartoli, L., et al.: The production effect on the performance of panels cast with selfcompacting fiber reinforced concrete. In: 3rd International RILEM Conference on Strain Hardening Cementitious Composites, 03 – 05 November, RILEM, Dordrecht 2014. 7. Yu, R. et al.: ‚Sustainable development of ultra-high performance fiber reinforced concrete (UHPFRC): towards to an optimized concrete matrix and efficient fiber application. J. Clean. Prod. 162, 220–233 (2017) 8. Huang, H., et al.: Improvement effect of steel fiber orientation control on mechanical performance of UHPC. Constr. Build. Mater. 188, 709–72 (2018) 9. Magureanu, C., et al.: – ‘Mechanical properties and durability of ultra high performance concrete’. ACI Mater. J. 109(2), 177–184 (2012)

Studying of Processing-Structure-Properties Relation of Strain Hardening Cementitious Composites (SHCC) Zhenghao Li1(&), Jiajia Zhou1, Cong Lu2, and Christopher K. Y. Leung1

2

1 Department of Civil and Environmental Engineering, Clear Water Bay, Hong Kong SAR, People’s Republic of China [email protected] Department of Civil Engineering, Southeast University, Nanjing, China

Abstract. Strain hardening cementitious composites (SHCC) are a class of fiber reinforced materials exhibiting tensile strain hardening behavior up to strain of several percent, accompanied by the formation of fine multiple cracks with openings below 50 lm. To model the full stress-strain relation of SHCC (which governs ductility and energy absorption) and the crack width versus strain relation (which governs durability), the sequential formation of cracks needs to be analysed. The cracking process is related to the internal structure of SHCC, such as the fiber and flaw size distributions, which varies with material rheology and mixing sequence. As a first study of the processing-structure-property relation of SHCC, tensile specimens are prepared with mixes exhibiting different viscosities. To account for sequential cracking, the variation in SHCC ‘structure’ is represented by the size variation of equivalent spherical flaws inside the member according to the normal distribution, while fibres are assumed to be uniformly distributed. By fitting the measured tensile stress-strain curves for various mixes with a micromechanical model developed at HKUST, the effect of matrix viscosity on the flaw size distribution is determined. The results will provide insight on the micromechanics-based design of SHCC for various requirements on ductility and crack control. Keywords: Strain hardening composites control
Durability


Tensile performance
Crack width

1 Introduction Strain hardening cementitious composites (SHCC), also known as engineered cementitious composites (ECC), exhibit tensile strain hardening behavior up to several percent strain, accompanied by the formation of fine multiple cracks with opening below about 50 lm. In structural applications, these properties will translate into high deformation capacity and energy absorption as well as high durability, as fine cracks will not facilitate the penetration of water or other chemicals to induce concrete deterioration or corrosion of steel rebars. Based on understanding of micromechanics and fracture mechanics, the criteria for achieving SHCC behavior with short random © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 1100–1111, 2021. https://doi.org/10.1007/978-3-030-58482-5_97

Studying of Processing-Structure-Properties Relation of SHCC

1101

fibers was first proposed in [1] and further refined in [2–4]. Based on these criteria, SHCCs with different compositions have been successfully made by various researchers groups. In engineering applications, besides knowing the failure mode of the material (hardening or softening), it is very important to have the full stress-strain relation up to maximum stress, as this will reflect the tensile strain that can be reached and the total energy that can be absorbed. Moreover, this constitutive relation can be employed in finite element analysis to predict the behavior of structural components. The crack width versus strain is another relation with practical significance. Penetration of water and chemicals can be accelerated by the opening of cracks, but the effect will be minimal if the crack width is below about 50 lm. At the serviceability state, the maximum strain in a member can be calculated. If the corresponding crack width is sufficiently small, long-term durability can be assured. While the stress versus strain and crack width versus strain relations can be experimentally measured, it is highly desirable to have micromechanical models that can predict them based on properties of the fiber, matrix and fiber/matrix interface as well as fiber dimensions and volume fraction. SHCC can then be designed to achieve the performance requirements for various applications [5]. Such models have been developed recently at HKUST [6, 7]. The strain corresponding to a certain applied stress and the crack width at a particular strain are dependent on the crack spacing, which is governed by two factors. The first factor is the effectiveness of stress transfer from the fiber at the crack back to the matrix, which has been analysed in [6]. The second factor is the variation of matrix cracking strength in the material, which determines if the transferred stress at a certain section is sufficient for cracking to occur. In [6] and [7], different assumptions have been made on how the cracking strength varies, and fitting of test data was conducted to find the parameters governing the variation. From a scientific point of view, the strength in various sections should depend on the variation of fiber and flaw size distributions within the material, which are in turn governed by the rheological properties and the mixing sequence. The effect of Marsh cone flow time (which reflects the viscosity [8]), on SHCC behavior was revealed in [9] while the effect of mixing sequence was shown in [10]. To be able to predict SHCC behavior (so the proper trial mix can be designed), a link between processing parameters, internal structure (i.e., fiber and flaw distributions) and properties of SHCC need to be established. The goal of our research program is to study the effect of processing parameters on the fiber and flaw size distributions, and the effect of such distributions on the tensile stress-strain relation and crack width versus strain relation. As the first paper reporting our finding, only the effect of viscosity (i.e., flow cone time) is considered, with the mixing sequence kept unchanged. Also, the variation in SHCC ‘structure’ is represented by the size variation of equivalent spherical flaws within the member according to the normal distribution, while the fiber content is taken to be uniform for all sections. Based on the fitting of measured tensile behavior with the model in [7], the effect of viscosity on the flaw size distribution can be determined. This information should be useful for the design of SHCC. In the following sections, the experimental program will be presented with major test results. After a brief description

1102

Z. Li et al.

of the micromechanical modelling, fitting of the tensile curves will be described and the relevance of the results will be discussed.

2 Experimental Program 2.1

Material

Three mixtures with the same mix proportion (Table 1) but different superplasticizer content were designed in this experimental investigation. For the SHCC specimens, the main constituents included Ordinary Portland cement, Class I Type F fly ash, silica sand with the size distribution of 125–180 lm, polycarboxylate-type superplasticizer and water. PVA fibers were added into the matrix at a volume fraction of 2%. All mixtures are listed in Table 1. The PVA fibers are 12 mm in length and 39 lm in diameter, and the Young’s modulus and tensile strength are 40 GPa and 1600 MPa respectively. Table 1. Mixture Proportions (by Weight) of the tested SHCCs Cement SHCC_I 0.2 SHCC_II 0.2 SHCC_III 0.2

2.2

Fly ash Silica sand Water 0.8 0.2 0.22 0.8 0.2 0.22 0.8 0.2 0.22

Super-plasticizer 0.50% 0.40% 0.36%

PVA fiber (Vf) 2% 2% 2%

Testing of Fresh Properties

In this study, the Marsh cone test was carried out to evaluate the fluidity of the fresh ECC mortar. A metallic Marsh cone with the dimensions recommened in [9] was employed (Fig. 1). The cone was fully filled with ECC mortar, and the time taken for all of the mortar to flow through the cone was measured and denoted as the flow time. The higher the flow time, the lower is the fluidity of the mortar. 2.3

Mixing, Casting and Curing

All mixtures were prepared using a 12 HL mixer, following the same mixing sequence, speed, and time. Solid ingredients, including cement, fly ash, and silica sand, were first mixed at low speed for 2 min. Water and super-plasticizer were then added into the dry mixture and mixed at high speed for 2 min. PVA fibers were added and mixed at high speed for 2 min. The mixtures were then cast into steel molds, and external vibration was applied to improve the compaction. After casting and curing the SHCC specimens in steel molds for 24 h, all the specimens were demolded and cured in the standard curing environment at 95% relative humidity and 23 ± 2 °C. When the curing period reached 28 days, the specimens were removed from the curing room until testing. Five samples were prepared for each mix.

Studying of Processing-Structure-Properties Relation of SHCC

2.4

1103

Tensile Test

The uniaxial tensile test was carried out with the dumbbell SHCC specimens at the age of 28 days. Figure 2 illustrates the dimension of the specimen. Before testing, four aluminum plates were glued on the both ends of each specimen to facilitate gripping. Tests were carried out using a 250 kN capacity MTS machine under a displacement control rate of 0.005 mm/s. Two external linear variable displacement transducers (LVDTs) were attached to the middle part of the specimen with a gage length of 80 mm to measure the deformation of the specimen during the test. The tensile stress– strain curve of each specimen was recorded.

120 mm

20 mm

Fig. 1. Marsh cone flow time test [9].

30

60

25 mm

300 mm

13

330

85

40

80

40

85

Fig. 2. Dumbbell specimen for tensile testing.

3 Experimental Results and Discussion The Marsh cone flow times of the fresh mortar and the tensile properties of the hardened SHCC samples are summarized in Table 2. The flow time of the fresh mortar decreased with increasing dosage of the superplasticizer. Table 2. Flow times of the fresh mortars and tensile properties of the hardened SHCC samples SHCC_I

Flow time First cracking strength (MPa) Ultimate strength (MPa) Ultimate strain (%) SHCC_II Flow time First cracking strength (MPa) Ultimate strength (MPa) Ultimate strain (%) SHCC_III Flow time First cracking strength (MPa) Ultimate strength (MPa) Ultimate strain (%)

19 s 3.37 4.73 3.34 30 s 3.44 5.21 4.64 48 s 3.22 3.54 2.10

3.33 3.36 2.12 4.02 5.30 4.68 4.19 4.92 4.79 4.96 2.12 3.76 3.55 4.06 3.72 3.28 4.92 5.30 5.22 5.64 4.73 5.55 6.22 6.03 3.28 3.25 3.20 3.51 4.25 3.93 3.82 4.46 4.37 3.87 2.09 3.15

1104

Z. Li et al.

Figure 3 shows the tensile stress–strain curves of the three groups of SHCC samples. All specimens show tension-hardening and multiple cracking behavior. When the flow time was 19 s, large variation existed in both the ultimate strain and the ultimate strength. The tensile strain could vary from 2.12% to 4.96%. Large variation in tensile strain capacity was also found when the flow time was 48 s. However, when the flow time was 30 s, the SHCC samples exhibited more consistent tensile behavior as well as greatly improved tensile strain capacity, with an average value of 5.45%. Flow time = 30s

Flow time = 48s 4

4

4

3

3 2

3 2

0

1

2

3

4

5

0

2 1

1

1 0

Stress(MPa)

5

5

Stress(MPa)

Stress(MPa)

Flow time = 19s

0

Strain(%)

1

2

3

4

5

6

0

0

1

Strain(%)

2

3

4

Strain(%)

Fig. 3. Tensile stress–strain curves of the SHCC samples.

4 Simulation of Tensile Behavior To interpret the variation of tensile behavior from the perspective of microstructure, curve fitting method was adopted. In this section, the micromechanical model applied will be briefly introduced first, then the fitting parameter will be obtained by inversion method. The correlation between viscosity, microstructure parameter, and tensile behavior will be discussed. 4.1

Micromechanical Model

The model applied in this article for predicting the tensile behavior of SHCC was proposed by Lu [7], which considered the effect of fiber distribution, flaw distribution and stress transfer distance under different load stages. Key features of the model are described below. Fibres were assumed to be 3D randomly distributed in this model, which means the V number of fibers across each section is equal to the theoretical value N ¼ prf2 , the f

L

embedment length of short side is randomly distributed from 0 to 2f , and the probability density of fiber inclination angle h is sinh in the range of 0 to p/2 [11]. Fibers are divided into 5 different categories according to different statuses or stress levels, namely Two-way debonding, Pullout-debonding, Two-way Pullout, Ruptured, and Pullout, which have different contributions to the crack bridging force. By summing up the contribution of each fiber, the stress-cracking opening relation is calculated. The stress transfer to matrix is calculated in two parts. The first part is the pulley force, which is caused by snubbing effect as shown in Fig. 4. The pulley force can be derived by force equilibrium and occurs right at the crack plane. The second part is

Studying of Processing-Structure-Properties Relation of SHCC

1105

frictional force. Frictional force is caused by chemical bond and frictional bond of interface, so the contribution of fibers are different at different statuses. Similarly, by summing up all contributions from fibers with different embedded lengths and inclination angles, the stress field near a single crack can be obtained.

Fig. 4. Schematic diagram for snubbing effect at exit point [7].

To model the distribution of matrix strength, a set of initial flaws are introduced to the matrix. The flaws are assumed to be spherical, with their size following the normal distribution. The flaw location is randomly assigned within the member as suggested by Kabele [12]. Then the equivalent penny-shaped cracks are calculated according to the largest flaw on each section as shown in Fig. 5 [1], and the normalized cracking strength can be derived from   pffiffiffi e pK 4 pffiffiffi 1 r fc ¼ þ c c 2 c 3 2 g

ð1Þ

where g

 2  pf =2 1 þ e 4þf2

ð2Þ

~fc refers to cracking strength at the is governed by snubbing coefficient f. In Eq. (1), r   section normalized by r0  Vf s Lf =df =2, in which Vf is the fiber volume fraction, Lf and df is the length and diameter of fiber. K and c are normalized fracture toughness and flaw size, defined by  pffiffiffiffiffi d K  Ktip =r0 c0 =g~ 1~1

c ¼ ðc=c0 Þ2d

ð3Þ ð4Þ

1106

Z. Li et al.

Where  c0 

Lf EC 2Ktip

2

p

ð5Þ

16ð1  v2 Þ2     ~d  2s=ð1 þ gÞEf  Lf =df

ð6Þ

With proper modeling of fiber and flaw distributions as well as stress transfer analysis, simulation of multiple cracking process can be conducted. In the process of the simulation, stress field for each load step is calculated and compared with the cracking strength distribution. If the stress reaches the cracking strength, the matrix will crack at the section and stress field will be calculated again. The process ends when the load reaches peak bridging force. Knowing the cracking strength and post-cracking bridging stress vs crack opening relation of different sections, the cracking sequence and width of each crack can be obtained. The stress-strain curve can then be established.

Fig. 5. Schematic diagram for calculating flaw size [7].

4.2

Parameters of Flaw Size Distribution

To study the effect of microstructure on the macro tensile behavior of SHCC, inversion method was used to get the distribution parameter of flaws, specifically the mean size and its standard deviation. Based on previous research [13–15], a typical set of material parameter was adopted for the simulation, as listed in Table 3. Table 3. A typical set of material parameter. Fiber

Fiber volume fraction Fiber length lf (mm) Fiber diameter df (mm) Fiber elastic modulus Ef (GPa) Nominal fiber strength

2% 12 0.039 18 1060 (continued)

Studying of Processing-Structure-Properties Relation of SHCC

1107

Table 3. (continued) Elastic modulus Em (GPa) Fracture toughness Ktip (MPa
m0.5) Poisson’s ratio m Interface Frictional bond s0 (MPa) Debonding fracture strength Gd (J/m2) Snubbing coefficient f Fiber strength reduction factor f’ Slip hardening parameter b

Matrix

20 0.15 0.2 1.31 1.08 0.5 0.33 0.58

5

Stress-crack opening relation for a crack

3 2

0.4 0.2

1 0

19s 30s 48s

0.6

4

Density

Bridging Stress (MPa)

With the parameter in Table 3, the bridging stress – crack opening relation curve obtained by the model is shown in Fig. 6 and the peak stress is about 5.1 MPa at 93 µm crack opening. To simplify the model and improve calculation efficiency, the same bridging stress – crack opening curve is used for each section. In other words, the variation of number of fibers among different sections is not considered, and the average bridging stress – crack opening relation is employed. With this approach, the section with the smallest value of maximum crack bridging stress, which governs the ultimate strength of SHCC, cannot be properly simulated. To deal with this problem, the ultimate stress of ‘the weakest section’ is set according to the test data, i.e. the average ultimate stress of all the samples in a batch. When the stress reaches the ‘failure stress’, the simulation program will stop and the simulated strain-stress curve will be plotted. Three main characteristic parameters are most concerned in the simulation, namely the first cracking strength, the ultimate stress and the ultimate strain. The sample is cut into 200 sections and a total of 60 individual flaws are incorporated. To fit the different tensile curves of the SHCC specimens with different viscosity, the mean and standard deviation of the flaw size distribution were varied. Note that the distributions are truncated at zero as the crack size cannot be negative. The attained flaw size distribution parameters are listed in Table 4, and probability density curves of flaw sizes are displayed in Fig. 7. With these parameters, the theoretical simulation of tensile behavior coincides well with the experimental curve as plotted in Fig. 8.

0

100

200

300

400

500

600

Crack Opening ( m)

Fig. 6. Stress-cracking opening relation for a single crack.

0

2

3

4

5

6

Diameter of flaws(mm)

Fig. 7. Probability density curve of simulated flaw size.

1108

Z. Li et al. Table 4. Parameters of flaw size distribution. Marsh flow cone time 19 s 30 s 48 s Distribution of flaw size Mean value (mm) 2.6 2.9 3.5 Standard derivation (mm) 0.5 0.6 1.0 Other parameters Number of flaws 60 60 60 Number of sections 200 200 200

It should be mentioned that the fitting parameters, namely the mean value and standard derivation of flaw size, are not the same as the physical flaw size and its distribution. As mentioned in a previous section, many assumptions have been made in the model. For example, flaws are assumed to be spherical and are treated as pennyshaped flaws in analysing their effect on the cracking strength of individual sections. The cracking strength of a section is assumed to be governed by the largest flaw only so interaction among individual flaws is neglected. Moreover, many material parameters employed in the model (shown in Table 3) are not directly measured but taken from the literature. These assumptions may cause some inconsistencies with the real material. Nevertheless, there should still be close correlation between the simulated parameters and the real flaw distribution. Also, from the point of view of engineering application, if these ‘effective’ flaw distribution parameters can be obtained as a function of flow properties (e.g., a certain flow cone time), the effect of processing on SHCC behavior can be predicted with the existing model. Although not a precise description of the actual microstructure, the use of the ‘effective’ flaw size parameters can be a useful first step in studying the processing-structure-properties relation of SHCC. In the following, the effects of viscosity on flaw size parameters and SHCC behavior will be discussed. 4.3

Correlation of Flowability, Distribution Parameter and Tensile Behavior

Flowability of mortars affects the flaw size distribution within the specimens and their relation is reflected by the fitting parameters. From a general perspective, mortar with better flowability is more likely to be better compacted. It is because it is easier for entrapped air (especially large bubble) to escape, and hence densifying the matrix. This point is consistent with the simulating parameters in Table 4 where the mean value and standard deviation of flaw size both increase with the Marsh cone flow time. From Table 2, the first cracking strength decreases with increasing flow time. In general, first cracking strength of a specimen is dependent on the largest flaw size. As shown in Fig. 7, the mean value and variation of flaws size of SHCC_III (48 s) is the largest, thus there will be more large flaws so the first cracking strength is expected to be the lowest. Similarly, the first cracking strength of SHCC_I (19 s) is the highest and that of SHCC_II (30 s) is medium.

Studying of Processing-Structure-Properties Relation of SHCC

1109

Fig. 8. Stress-strain curve with simulating parameter.

The ultimate strength of SHCC_II (30 s) is the highest among the three matrix, and also shows less variation. During the tensile test, when the stress of a certain section reaches its cracking strength, a crack is formed and the stress will be carried by the fibers crossing the section. Thus the ultimate strength is governed by the weakest section, which has the least number of fibers. For SHCC_I (19 s), a better flowability means lower viscosity and lower shear force generated during mixing. In this case, some fibers may stay as clusters, leading to poorer fiber dispersion and reduction of the ultimate strength. If the viscosity is very high, as in SHCC-III (48 s), the fibers can disperse well during mixing but the ultimate strength is found to be reduced (as shown in Table 2). There are two plausible explanations. Firstly, as shown in Fig. 7, there are many large flaws in the mix with low flowability, together with a high variation of the flaw size. As a result, there will exist some sections with high porosity (associated with large flaws) and others with low porosity. If the fibers are uniformly distributed within the matrix, sections with high porosity will exhibit lower fiber content (as the matrix fraction is reduced) [9], so the ultimate strength will also be reduced. Secondly, low flowability may be associated with poorer wettability of the fibers, causing reduction in the interfacial fracture energy and friction, which governs the fiber bridging stress. These aspects will be experimentally investigated in our future work. The ultimate strain of SHCC_II (30 s) is significantly higher than the other cases. The total strain of SHCC is directly influenced by the number of cracks and the average crack width at failure. As the same bridging stress – crack opening relation was used for the three SHCCs, the average crack width is the same at the same stress level. The number of cracks is affected by the ultimate stress and the matrix strength. In the tensile test, the sections with the cracking strength lower than ultimate strength may crack. With more sections undergoing cracking, the higher is the ultimate strain. For SHCC_III (48 s), the large amount of large flaws weakened the matrix strength and reduced the ultimate strength to only about 4 MPa. Therefore, only a small number of sections can crack before failure, resulting in a lower ultimate strain. For SHCC_I (19 s), the ultimate strength is about 5 MPa which is similar to SHCC_II (30 s). However, the relative small flaw size makes the cracking strength of most sections higher than 5 MPa. With a smaller number of cracked sections, the ultimate strain is reduced. From the analysis above, SHCC_II (30 s) represents a mixing process closer to the optimal. The resulting flaw sizes are not too big to influence the fiber distribution or weaken the section too much, and the proper viscosity can disperse the fibers well so

1110

Z. Li et al.

that the ultimate strength among different sections are similar. Therefore, the ultimate stress and strain are less variable, showing the best robustness.

5 Conclusion This study investigated the correlation between the flowability of SHCC mortar, flaw distribution parameters and macroscopic tensile behavior. Based on the numerical model, flaw distribution parameters are attained by inversion method, which is consistent with the rheology analysis and tensile results. • The first cracking strength shows a negative correlation with Marsh cone flow time. Low flowability of SHCC mortar could lead to large flaws and reduce the first cracking strength. • High flowability mortars could not provide enough shear force when mixing and lead to poorer fiber dispersion. Low flowability mortars may introduce too many large flaws and induce lower content of fibers in the weakest section. In both cases, the ultimate strength is reduced. • Ultimate strain is governed by the region where the cracking strength is lower than the ultimate strength of the SHCC, which is governed by the weakest section. The number of cracks that can form in the member is limited by both high flowability (which results in dense matrix with high cracking strength for most of the sections) and low flowability (which results in low ultimate strength). • Based on the above, the highest ultimate strength and ultimate strain are obtained for an intermediate flowability, which corresponds to Marsh cone flow time of 30 s in our study. In this study, the size distribution of effective spherical flaws is investigated by inversion method. In future work, the variation of both the physical flaw size and fiber distribution among various cross sections will be experimentally studied to provide insights for the development of an improved model for the comprehensive tensile behavior of SHCC. Acknowledgements. The support of this research by the Hong Kong Research Grant Council through the General Research Fund (Project Number 16215018) is gratefully acknowledged.

References 1. Li, V.C., Leung, C.K.Y.: Steady state and multiple cracking of short random fiber composites. ASCE J. Eng. Mech. 188(11), 2246–2264 (1992) 2. Li, V.C.: From micromechanics to structural engineering–the design of cementitious composites for civil engineering applications. JSCE J. Struct. Mech. Earth. Eng. 10(2), 37– 48 (1993) 3. Leung, C.K.Y.: Design criteria for pseudo-ductile fiber composites. ASCE J. Eng. Mech. 122(1), 10–18 (1996) 4. Kanda, T., Li, V.C.: A new micromechanics design theory for pseudo strain hardening cementitious composite. ASCE J. Eng. Mech. 125(4), 373–381 (1999)

Studying of Processing-Structure-Properties Relation of SHCC

1111

5. Leung, C.K.Y.: Performance-based design of SHCC components – research and challenges. In: ‘Strain Hardening Cement-based Composites’, Proceedings of an International Conference, Dresden, pp. 429–440. Springer, Cham (2017) 6. Lu, C., Leung, C.K.Y.: A new model for the cracking process and tensile ductility of strain hardening cementitious composites (SHCC). Cem. Concr. Res. 79, 353–365 (2016) 7. Lu, C., Leung, C.K.Y., Li, V.C.: Numerical model on the stress field and multiple cracking behavior of engineered cementitious composites. Constr. Build. Mat. 133, 118–127 (2017) 8. Roussel, N., Le Roy, R.: The marsh cone: a test or a rheological apparatus? Cem. Concr. Res. 35(5), 823–830 (2005) 9. Li, M., Li, V.C.: Rheology, fiber dispersion and robust properties of engineered cementitious composites. Mater. Struct. 46(3), 405–420 (2013) 10. Zhou, J., Qian, S., Ye, G., Copuroglu, O., van Breugel, K., Li, V.C.: Improved fiber distribution and mechanical properties of engineered cementitious composites by adjusting the mixing sequence. Cem. Conc. Comp. 34, 342–348 (2012) 11. Aveston, J., Kelly, A.: Theory of multiple fracture of fibrous composites. J. Mater. Sci. 8, 352–362 (1973) 12. Kabele, P., Stemberk, M.: Stochastic model of multiple cracking process in fiber reinforced cementitious composites. In: Proceedings of the 11th International Conference on Fracture. Turin: CCI Centro Congressi Internazionale srl, Citeseer (2005) 13. Yang, E., Wang, S., Yang, Y., Li, V.C.: Fiber-bridging constitutive law of engineered cementitious composites. J. Adv. Concr. Technol. 6, 181–193 (2008) 14. Kanda, T., Li, V.C.: Effect of fiber strength and fiber-matrix interface on crack bridging in cement composites. J. Eng. Mech. 125, 290–299 (1999) 15. Lin, Z., Li, V.C.: Crack bridging in fiber reinforced cementitious composites with sliphardening interfaces. J. Mech. Phys. Solids 45, 763–787 (1997)

The Effect of Fiber Content on the Post-cracking Tensile Stiffness Capacity of R-UHPFRC M. Khorami1,2, Juan Navarro-Gregori1(&), and Pedro Serna1 1

Institute of Science and Concrete Technology, ICITECH, Universitat Politècnica de València, 46022 València, Spain [email protected] 2 Universidad UTE, Facultad de Arquitectura y Urbanismo, Calle Rumipamba s/n y Bourgeois, Quito, Ecuador

Abstract. Concrete cracking can be controlled by adding fibers to concrete, with the expected desirable behavior under serviceability conditions due to narrower close space cracks compared to similar concrete without fibers. Using fibers to produce Ultra-High Performance Fibre Reinforced Concrete (UHPFRC) has enhanced the post-cracking tensile capacity of composite material and increased the related energy absorption capacity for the cracked member. Accordingly, the amount and type of fiber in the matrix affect postcracking behavior. In this experimental study, specimens reinforced by conventional steel rebars with a constant cross-sectional dimension and reinforced steel ratio were tested. The tested variables were: 1) type and length of fibers; 2) fiber content. The Direct Tensile Test was conducted, and the tensile behaviour of specimens was obtained. The results showed that the increment in fiber content (80 kg/m3 to 160 kg/m3 in this research) or the combination of micro and macro steel fibers with the same content (80 kg/m3 for each fiber type) had no significant effect on the post-cracking stiffness capacity. Moreover, all the RUHPFRC specimens provided the full tension stiffening with the quasi same post-cracking stiffness capacity close to the bare bar axial stiffness. Keywords: Tension stiffening
Post-cracking tensile stiffness
Serviceability
UHPFRC

1 Introduction In the structural concrete design, it is commonly assumed that reinforcement carries all tensile force at the crack face [1]. Away from the crack face, due to the bond between the steel rebar and the surrounding concrete, tensile stresses are shared between the concrete and steel rebar. This contribution of concrete between cracks in tension is commonly termed tension stiffening. This phenomenon has an effect on member stiffness and is essential for determining serviceability deflection [2] and crack widths [3]. The tension stiffening response should be included in the analysis to predict member behaviour, under serviceability conditions.

© RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 1112–1123, 2021. https://doi.org/10.1007/978-3-030-58482-5_98

The Effect of Fiber Content

1113

The effects of cracking and reinforcement on member stiffness can be taken into account with effective axial member stiffness (EAÞeff . The effective modulus method is adopted by the fib Model Code 2010. It suggests an effective modulus for reinforce
b ) and then the axial member response can be predicted using P ¼ E
b :As :em : (As ment (E and em are the reinforcing bar area and the average member strain, respectively). The load-strain relation and the effective modulus method to predict the member response are shown in Fig. 1.

Fig. 1. Load-strain relationship and effective modulus method (MC2010)

The presence of the fiber in the concrete leads to enhanced post-cracking behaviour and, as a result, improves the overall tensile response of the tensile elements. This improvement is due to the combination of tension stiffening and strain hardening mechanism. Strain hardening refers to the bridging effect and transmission of tensile stresses by fibers across crack faces, and tension stiffening refers to the bond behaviour of reinforcement and concrete [4]. Many theoretical and experimental studies have been performed to evaluate the fiber content effect on post-cracking behaviour at the material consideration level for FRC or UHPFRC [5–9], while real structural elements are combined by the reinforcement steel bar. Thus studying post-cracking behaviour for these reinforcement elements is essential. In line with this, the present work focuses on the axial tensile stiffness of cracked tensile elements (herein called post-cracking tensile stiffness). The effect of fiber content was studied by employing three different doses and two types of fibers for UHPFRC and comparing to Ultra-High Performance Concrete (UHPC). The average tension stress-strain of tensile elements was obtained. Post-cracking tensile stiffness was calculated by considering the tensile behaviour curve slope in the elastoplastic region of behavior. Finally based on the experimental results, the cracking behaviour and influence of fibers are also reported.

1114

M. Khorami et al.

2 Experimental Program 2.1

Materials

Four different types of UHPFRC were used in this study given the difference in fiber content and the type of steel fibers. Two types of steel fibers were considered in this study. The first type was high-strength micro smooth steel fibers with a fine diameter (df = 0.2 mm) and relatively short fiber length (Lf = 13 ± 0.1 mm). The second type was macro hooked end fibers (length Lf = 30 ± 1 mm and diameter df = 0.375 mm). The geometry and parameters of the steel fibers are presented in Table 1. Table 1. Properties of steel fibers

The four types of UHPFRC were named as follows: • (C160): the UHPFRC with fiber type (SF1) and fiber content of Vf = 2% or 160 kg/m3. • (C80): the UHPFRC with fiber type (SF2) and fiber content of Vf = 1% or 80 kg/m3. • (C8080): the UHPFRC with fiber type (SF1&SF2) and fiber content of Vf = 1% for SF1, and Vf = 1% for SF2. • (C0): the UHPC without fibers. The average compressive strength of UHPFRC and UHPC was determined at the tensile elements’ testing age with four cube specimens (100  100  100 mm). The resulting values are found in Table 2. Table 2. UHPFRC and UHPC average compressive strength Concrete type fc ðMPaÞ C160 158.55 C80 151.42 C8080 142.59 C0 136.69

The Effect of Fiber Content

1115

The matrix composition of UHPFRC herein used in all the concretes was based on our research group’s previous experiences [10–14]. Table 3 provides the matrix composition of the employed UHPFRC and the UHPC matrix. Table 3. Composition of the matrix mixture by the weight ratio for UHPFRC and UHPC Cement (type I)

Medium sand 0.6–1.2 mm 0.70

1.00

2.2

Fine sand 0.5 mm 0.37

Silica flour U-S500 0.28

Silica fume (Elkem Microsilica, grade 940) 0.22

Superplasticizer Water

0.037

0.20

Specimen Geometry and Test Set-Up

The uniaxial tensile tie test was carried out using prismatic concrete specimens with a 100x100 mm square cross-section and 1000 ± 3 mm length. The tensile element was reinforced with a 12-mm central rebar (Es ¼ 200 GPa, fy ’ 550 MPa). Ten specimens were cast for this study, six of them made with UHPFRC C160 and C80 (three for each type), and two specimens for each C8080 and C0. The uniaxial tensile test was conducted using a hydraulic jack machine with a loading capacity of 200 kN under displacement control at a loading rate of 0.5 mm/min. The complete details and testing procedure are available in M. Khorami et al. [15]. Figures 2a and 2b show the test setup and the R-UHPFRC tensile element. a)

b)

Fig. 2. Uniaxial tensile test: (a) test set-up, (b) R-UHPFRC specimen

Eight displacement transducers (DTs) were installed on the surfaces of specimens (four of them on the right side and four on the left side of the specimen) to record element elongation during the test and to capture any undesired bending applied to the specimen due to unforeseen load eccentricities. Each DT measured the length variation between the fixing points placed with a 350 mm length from the center of specimens toward the ends (see Fig. 2b).

1116

M. Khorami et al.

3 Test Results and General Discussion 3.1

Tensile Behavior

The average stress-strain curves obtained for each tensile element type are shown in Figs. 3a–c. The tensile stress in reinforcement (rs ) was calculated by dividing the tensile load (N) by the reinforcement area (As ). The tensile elongation was measured when the test started and after curing and the storage time. The authors have studied the shrinkage effect on R-UHPFRC. UHPFRC shrinkage would shorten the member without reinforcement, while embedded reinforcement would restrain concrete shortening. This leads to a negative pre-strain (es;sh ) with compressive stress in the rebar, and initial tensile strain in the UHPFRC (ec;sh ). Hence the real origin of the bare steel rebar response was modified by moving the compression strain value caused by shrinkage. The experimental value obtained for the reinforcement strain due to UHPFRC shrinkage was approximately (es;sh ¼ 0:40%). The cracking stress level at the interaction point between the fit line over the uncracked and cracked responses was defined. The slope of the elastoplastic region of the tensile behavior represents the post-cracking tensile stiffness for the R-UHPFRC tensile elements and was herein evaluated. The values calculated over the average response of two or three specimens (red-colored curve) for each concrete are presented in Table 4 according to the criteria shown in Fig. 4.

Fig. 3. Average tensile response of tensile elements for four concrete types.

The Effect of Fiber Content

1117

According to the obtained results (Table 4), the post-cracking tensile stiffness increased by a low value when increasing the fiber content when using the hybrid fibers for TC8080. Moreover, these values came very close to the axial stiffness of the bare bar with a value of c ¼ Es ¼ 200 GPa. To better understand the influence of fiber content on post-cracking tensile stiffness and the difference between them, the stressstrain relations for four concrete types are presented together, as shown in Fig. 5. Table 4. Elastic tensile stiffness and post-cracking tensile stiffness Specimen ID

Stress in reinforcement at cracking rs;cr ðMPaÞ

Stress in UHPFRC at cracking rc;cr ðMPaÞ

Post-cracking tensile stiffness c ðGPaÞ

TC0 TC80 TC160 TC8080

112.00 455.00 563.00 591.00

1.28 5.20 6.44 6.76

N.A 205.83 225.43 237.59

Fig. 4. Criteria for calculating R-UHPFRC tensile response.

Fig. 5. Tensile response of tensile elements with four concrete types.

1118

M. Khorami et al.

As can be seen, R-UHPFRC tensile elements TC160, and TC8080 almost have the same tensile cracking stress. Moreover, the post-cracking tensile stiffness for all three types of R-UHPFRC tensile elements is in parallel to the bare bar response, while TC0 exhibits a very similar behavior to the tensile elements with conventional concrete (this finding is discussed in more detail in Sect. 4). By comparing the tensile response of TC80 with TC160, and TC8080, it can be concluded that if using a double fiber content is necessary (e.g. for the durability aspect of UHPFRC), employing the hybrid fiber for R-UHPFRC (80 kg/m3 for the micro smooth steel fibers, plus 80 kg/m3 for the macro hooked end steel fibers) will present better tensile behavior on the one hand and, on the other hand, the employed steel fiber weight will be the same for both cases (TC8080 and TC160). The cracking behavior aspect and the influence of fiber on cracking propagation and crack width are discussed in Sect. 5.

4 UHPFRC Contribution in Tension The tension stiffening response refers to the tension carried by the concrete between cracks due to the reinforcing bar’s bond behavior. This ability increases the element’s stiffness before reinforcement yields and can be used to predict tensile behaviour, multiple crack spacing and crack widths. Figure 6 is a qualitative representation of the tensile behavior of both R-UHPFRC and R-UHPC. The cracked R-UHPFRC tensile element exhibits constant contribution in tension, which refers to the parallel region of the behavior with the bare steel bar (called herein full tension stiffening), while the tension stiffening response for the UHPC tensile element after crack stabilization gradually decreases as the applied load increases, and the member response curve moves closer to the bare bar response. The bond factor parameter (b) accounts for the variation in the concrete average tensile stress between cracks, and is generally expressed as an average tensile stress/cracking stress ratio. The bond factor value vary from zero to one for no bonded reinforcement and fully bonded, respectively (0\b\1).

Fig. 6. Qualitative representation of the tensile behavior of the R-UHPFRC and R-UHPC elements, full tension stiffening concept.

The Effect of Fiber Content

1119

The R-UHPFRC tensile elements (TC80, TC160, and TC8080) exhibit the full tension stiffening with Beta values equaling one (see Fig. 5). This R-UHPFRC property is essential when the serviceability state is evaluated, especially for deflection control. Consequently, the high tension stiffening capacity of R-UHPFRC leads to good performance under serviceability conditions and affirms the use of this material for special structures in which durability or permeability of concrete elements is essential. Moreover, it is evidenced that the cracking stress of the specimen without fibers (TC0) is much lower compared to all the other specimens with fibers (TC80, TC160, and TC8080). For modeling the post-cracking tensile response of RC or FRC members, an empirical relation is needed to represent the gradual transition from the uncracked axial response to the fully cracked member (bare steel bar). An alternative approach to predict the post-cracking tensile response consists in using an effective axial stiffness (EAÞeff for the cracked member, which depends on the member’s strain level. According to the fib Model Code 2010, the average strain of the RC or FRC member (em ) is calculated by taking into account the tension stiffening effect, and can be calculated by the average reinforcement strain (eb ) minus the average concrete strain (Dec ): em ¼ eb  Dec

ð1Þ

As the (Dec ) is variable for the RC and FRC tensile members, the R-UHPFRC tensile elements provide a constant tension stiffening effect. Hence post-cracking tensile modeling and the deflection calculation may involve less complexity.

5 Cracking Behavior 5.1

Fiber Content Effect

The crack distribution along the entire length element was determined to evaluate the crack behavior of the tensile elements. The number of cracks was obtained at the end of the test when the average tensile strain reached 2%. Water was used to wet the surface so that micro cracks would be visible given the narrow width of cracks, which could not be seen by the naked eye, and the crack pattern was painted on the specimen surface. The number of cracks was recorded on each lateral surface at two surface edges over the red line (see Fig. 7) and the average number of cracks was calculated. The cracking measurement approach is shown in Fig. 7.

1120

M. Khorami et al.

Fig. 7. Cracking measurement approach of the R-UHPFRC ties elements

It is worth mentioning that the R-UHPC tensile element exhibited a completely different cracking behavior with the localized macro cracks and wide crack spacing, while the R-UHPFRC ties provided a distribution crack propagation along the entire tie length. The average crack width (wm ) was calculated by dividing the total tension elongation (et  l) to the total number of cracks (n) as follows: wm ¼ ðet  lÞ=n

ð2Þ

where (et ) is the average tensile strain at the time the number of cracks was measured (2%), and (l) is tie length, which is 1000 mm. The parameter (n) is the average number of the cracks. The obtained results are presented in Table 5, which reveal a clear influence of crack width on the fiber content. Employing a high-dose fiber content for concrete causes higher bond strength and shorter transfer length [16]. Thus crack spacing will narrow, and the number of cracks will increase. This phenomenon was observed when comparing the results for similar tensile elements with different UHPFRC types in fiber content terms. In serviceability behavior terms, the multiple-cracking with the distributed crack propagation of R-UHPFRC led to very thin cracks (micro cracks) at the high tension strain (em ¼ 2% in this study). The serviceability limit state can be controlled by applying limitations for the tensile stresses in reinforcement. These limitations are made to avoid inelastic strain, unacceptable cracking or deformation. Eurocode 2 indicates that the tensile stress in reinforcement cannot exceed 0.8 fyk , where fyk is the characteristic yield strength of reinforcement.

The Effect of Fiber Content

1121

Table 5. Average crack width at the tensile strain of 2% Specimen ID TC80-1 TC80-2 TC80-3 TC8080-1 TC8080-2 TC160-1 TC160-2 TC160-3 TC0-1 TC0-2

Total number of cracks 57 46 43 63 79 137 113 114 9 11

Average crack width Eq. (2) (mm) 0.035 0.043 0.047 0.032 0.025 0.015 0.018 0.018 0.220 0.181

Mean value (mm) 0.042

0.028 0.014

0.200

By comparing the mean crack width of TC160 to TC8080, it can be concluded that using micro fibers led to a more efficient behavior compared to hybrid concrete (TC8080), and both cases were better than TC80. In addition, the micro cracking process was better controlled in UHPFRC with 160 kg/m3 fiber content. From the serviceability perspective, the tensile elements with C160 exhibited better behavior than those with C8080 despite the steel fiber content being the same.

6 Conclusions In the present study, the impact of fiber content on the axial post-cracking tensile stiffness capacity was studied by employing three UHPFRCs with fiber content and fiber type variation, and one type of non fiber UHPC. The uniaxial tensile test was conducted for tensile elements. Based on the experimental results, the following conclusive remarks are drawn: • The reduction of the axial tensile stiffness for the R-UHPFRC specimens between the elastic region and micro crack stabilization region (the parallel zone of the behaviour with the bare bar response) rapidly happens. Moreover, during the applied tensile load, most micro cracks appear at the same time in this region of behaviour. • The tensile strength capacity in the cracking region depends on the fiber content of R-UHPFRC. However, the slope of the overall tensile behaviour of all the RUHPFRC tensile elements (referring to the curve slope: c) almost parallels the bare bar response. Consequently, the R-UHPFRC tensile elements provide the full tension stiffening effect. • Providing full tension stiffening for the R-UHPFRC members with a constant value for concrete contribution for the cracked member leads to a facility of the computing deflection of reinforced members with UHPFRC, while for reinforced RC or FRC members an empirical model with descending branch after cracking is needed.

1122

M. Khorami et al.

• Multiple cracking behaviour, and thus a large number of R-UHPFRC micro cracks, emphasizes the high potential of UHPFRC composite material for structures in which durability and permeability aspects are essential. From the serviceability point of view, R-UHPFRC with a 2% micro fibers content provides better behavior (crack width at a high rate of tensile strain) compared to the R-UHPFRC with hybrid fibers (1% of micro fibers, plus 1% of macro fibers), while both contain the same fiber content. Acknowledgments. This study forms a part of Project BIA2016-78460-C3-1-R supported by the State Research Agency of Spain.

References 1. Bischoff, P.H.: Reevaluation of deflection prediction for concrete beams reinforced with steel and fiber reinforced polymer bars. J. Struct. Eng. 131, 752–767 (2005) 2. Visintin, P., Oehlers, D., Muhamad, R., Wu, C.: Partial-interaction short term serviceability deflection of RC beams. Eng. Struct. 56, 993–1006 (2013) 3. Bischoff, P.H.: Tension stiffening and cracking of steel fiber-reinforced concrete. J. Mater. Civ. Eng. 15, 174–182 (2003) 4. Bernardi, P., Cerioni, R., Michelini, E.: Analysis of post-cracking stage in SFRC elements through a non-linear numerical approach. Eng. Fract. Mech. 108, 238–250 (2013) 5. Kooiman, A., Walraven, C.: Modelling the post-cracking behaviour of steel fibre reinforced concrete for structural design purposes. HERON 45(4), 2000 (2000) 6. Buratti, N., Mazzotti, C., Savoia, M.: Post-cracking behaviour of steel and macro-synthetic fibre-reinforced concretes. Constr. Build. Mater. 25, 2713–2722 (2011) 7. Abrishambaf, A., Barros, J.A., Cunha, V.M.: Relation between fibre distribution and postcracking behaviour in steel fibre reinforced self-compacting concrete panels. Cem. Concr. Res. 51, 57–66 (2013) 8. Pereira, E., Barros, J.A., Ribeiro, A.F., Camões, A.: Post-cracking behaviour of selfcompacting steel fibre reinforced concrete. In: 6th International RILEM Symposium on FibreReinforced Concretes (2004) 9. Zhou, B., Uchida, Y.: Relationship between fiber orientation/distribution and post-cracking behaviour in ultra-high-performance fiber-reinforced concrete (UHPFRC). Cement Concr. Compos. 83, 66–75 (2017) 10. López, J., Serna, P., Navarro-Gregori, J., Coll, H.: Comparison between inverse analysis procedure results and experimental measurements obtained from UHPFRC Four-Point Bending Tests. In: Proceedings of the 7th RILEM Workshop on High Performance Fiber Reinforced Cement Composites (HPFRCC7), pp. 185–192 (2015) 11. López, J.Á., Serna, P., Navarro-Gregori, J., Camacho, E.: An inverse analysis method based on deflection to curvature transformation to determine the tensile properties of UHPFRC. Mater. Struct. 48(11), 3703–3718 (2014) 12. López Martínez, J.A.: Characterisation of the tensile behaviour of UHPFRC by means of four-point bending tests. Ph.D. thesis (2017) 13. Mezquida-Alcaraz, E.J., Navarro-Gregori, J., Lopez, J.A., Serna-Ros, P.: Validation of a non-linear hinge model for tensile behavior of UHPFRC using a Finite Element Model. Comput. Concr. 23, 11–23 (2019)

The Effect of Fiber Content

1123

14. Mezquida-Alcaraz, E., Navarro-Gregori, J., Serna-Ros, P.: Numerical validation of a simplified inverse analysis method to characterize the tensile properties in strain-softening UHPFRC. In: IOP Conference Series: Materials Science and Engineering, p. 012006. IOP Publishing (2019) 15. Khorami, M., Navarro-Gregori, J., Serna, P., Navarro-Laguarda, M.: A testing method for studying the serviceability behavior of reinforced UHPFRC tensile ties. In: IOP Conference Series: Materials Science and Engineering, p. 012022. IOP Publishing (2019) 16. Harajli, M., Hamad, B., Karam, K.: Bond-slip response of reinforcing bars embedded in plain and fiber concrete. J. Mater. Civ. Eng. 14, 503–511 (2002)

Controlling Strength and Ductility of Strain-Hardening Cementitious Composites by Nano-Engineering Ousmane A. Hisseine(&) and Arezki T. Hamou Cement and Concrete Research Group, University of Sherbrooke, Sherbrooke, QC, Canada [email protected]

Abstract. Considering the hierarchical nature of cracking in cement composites, multi-scale reinforcement bears the potential to enhance the fracture performance of fibre-reinforced cementitious composites. This study shows how nanoscale cellulose filaments (CF) can be used as a novel tool for tailoring the properties of strain-hardening cementitious composites (SHCC) towards improved strength and ductility. SHCC with fly ash-to-cement ratio of 1.2 and incorporating CF at rates 0.03, 0.05 and 0.10% of cement mass were developed following the micromechanical principles for pseudo-ductile cement composites. Results indicate that the incorporation of CF in SHCC allows nanoengineering matrix and interface properties by increasing matrix elastic modulus and imparting a significant slip-hardening effect. Consequently, higher complementary energy and lower crack tip toughness were obtained, thereby leading to enhanced ductility as also validated by tensile and flexural tests. As such, the incorporation of CF enhanced composite tensile strength by up to 23% and increased the ultimate strain capacity in tension by up to 26% and the deflection capacity in flexure by up to 36%. Therefore, nano-engineering SHCC with CF yields multi-scale composites with higher ductility without necessarily increasing the volume fraction of PVA fibres while exhibiting higher strength without necessarily increasing the binder content. Keywords: Cellulose filaments (CF)
Engineered cementitious composites (ECC)
Nanocellulose
Nanoengineered concrete
Recycled glass powder (RGP)
Ground-glass pozzolans (GP)
Strain-hardening cementitious composites (SHCC)

1 Introduction Bearing in mind the increasing socioeconomic burden associated with the rehabilitation and reconstruction of aging concrete infrastructure, today’s concrete technology practitioners are challenged to develop concrete recipes demonstrating the highest performance, the least ecological footprint, and the maximum performance-to-investment ratio. Among concrete types bearing the potential to meet such stringent requirements is strainhardening cementitious composite (SHCC) also known as engineered cementitious composite (ECC). SHCC belongs to high-performance fibre-reinforced cementitious © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 1124–1136, 2021. https://doi.org/10.1007/978-3-030-58482-5_99

Controlling Strength and Ductility of Strain-Hardening Cementitious Composites

1125

composites (HPFRCC). The salient features of SHCC include its significant tensile strain capacity reaching up to 11%, the tight multiple cracking as well as the strain-hardening behaviour at low fibre content, commonly  2% [1–4]. While extended multiple cracking is a fundamental requirement in SHCC, crack localization often leads to a softening behaviour, thereby jeopardizing composite ductility. One way to foster SHCC ductility is to consider nano-engineering matrix and interface properties. Nanoreinforcement can favour a multi-scale crack resistance such that sequential multiple cracking and strain-hardening can be obtained. The survey of literature indicates that the incorporation of 0.08 wt% graphene oxide (GO) in SHCC enhanced the tensile strength by up to 38% and the flexural capacity by up to 81% [5]. Nanocellulose fibres were proven to enhance the fracture energy of UHPC by more than 50% [6]. SHCC with nanocarbonate whiskers incorporated at 0.5 vol. % led to 30% enhancement in compressive strength, 53% in ultimate tensile strength, 114% in the tensile strain capacity [7]. Considering the effect of nano-reinforcement on enhancing the ductility of SHCC, the hypothesis of this study is that the use nanoscale cellulose filaments (CF) can also enhance the ductility of SHCC. CF are nanoscale rod-like cellulosic particles with 30–400 nm diameter and 100– 2000 µm length and belong to nanocellulose materials (NCM) such as cellulose nanocrystals (CNC); microfibrillated cellulose (MFC; and nanofibrillated cellulose (NFC) [8, 9]. Owing to their intrinsic mechanical strength imparting inherent strength to plants, NCM are attracting substantial attention in versatile applications including cement and concrete composites [8–10]. Our former investigations on CF demonstrate enhancement in elastic modulus (18%) [8], flexural capacity (25%), and toughness (96%) [9] attributable to higher microstructure properties (increased degree of hydration of  12% and enhanced micromechanical properties of C-S-H gel matrix of 12–25%) [8]. The current study aims at leveraging the advantages of CF to enhance the ductility of SHCC. Research outcomes are expected to contribute towards the development of high-performance cement composites while contributing towards enhancing concrete ecoefficiency.

2 Experimental Program 2.1

Materials Properties

SHCC ingredients include type HS cement, type F- fly ash (FA), and quartz sand (QS) with a maximum particle size of 600 µm. The cement has a specific gravity (SG) of 3.18, Blaine fineness of 438 m2/kg, and mean particle diameter (d50) of 12 µm. The FA used in the study fulfils the requirements of CAN/CSA A3000 specifications and has an SG of 2.55, Blaine surface area of 363 m2/kg, and d50 of 17. The GP used herein is of high alkali content and has an SG of 2.51 and d50 of 27 µm. As for the QS, it has an SG of 2.70, maximum particle size (dmax) of 600 µm, and a d50 of 250 µm. Figure 1 provides the particle size distribution of granular materials used in this study. Polyvinyl-alcohol (PVA) fibres (with 38 µm diameter, 8 mm length, 40 GPa elastic modulus and 1400 MPa tensile strength were added at 2% per volume). The cellulose filaments (CF) shown in Fig. 4 have 30–400 nm diameter, 100–2,000 µm length, and a

1126

O. A. Hisseine and A. T. Hamou

surface area of more than 80 m2/g. Further information about CF characteristics and sustainability features can be found in our former works [8]. Figure 2 depicts a scanning electron microscope (SEM) image of a CF diluted aqueous suspension at a concentration of 0.10%.

Fig. 1. Particle-size distribution of SHCC ingredients.

2.2

Mixture Proportions

Four SHCC were considered in this investigation. This includes a reference SHCC mixture as well as three SHCC incorporating cellulose filaments (CF). The reference SHCC has FA/cement ratio of 1.2, QS to binder ratio of 0.35 and a water-to-binder (w/c) ratio of 0.28. CF was incorporated at dosages of 0, 0.03, 0.05, and 0.10% per mass of cement. Table 2 shows mixture proportioning. The amount of high-range water reducing admixture was adjusted in function of CF in order to achieve the target slumpflow of  300 mm in the suspended mortar and  250 mm in final SHCC. Resulting mixtures were tailored via the micromechanics design approach of SHCC proposed by Li et al. [11–13]. 2.3

Mixing Procedures and Specimen Preparations

A pan mixer of type Mortarman 360 was used to prepare the different batches using the following sequence: • All granular materials were first dry–mixed for 7 min prior to adding water and HRWRA. • A 90% of the HRWRA diluted into 95% of the mixing water was added to the mixer slowly during 0.5 min then mixing continued for 2.5 min. For mixtures incorporating CF, a CF-water suspension was first prepared from readily dispersed CF (colloidal suspensions with 1.2% CF solid content). Thereafter, 90% of HRWRA was diluted into 95% of CF-water suspension. The remaining 10% of HRWRA is used for final adjustment of mixture flowability. • The mixer was then stopped for 0.5 min to scrape its blades and edges then mixing continued for another 1 min. The consistency of the suspended mortar was then checked such that when a mini-slump flow diameter of 300 mm was obtained [14],

Controlling Strength and Ductility of Strain-Hardening Cementitious Composites

1127

Fig. 2. SEM micrograph of nanoscale cellulose filaments (CF). Figure adapted from [9] with permission from ASCE Journal of Materials in Civil Engineering.

only the remaining water is added (after adjustment of the amount of water contained in the unused 10% of HRWRA). Otherwise, a gradual amount of HRWRA is added (along with an adjusted amount of the remaining 5% of water) until the desired 300 mm mini-slump flow diameter is achieved. • PVA fibres were added slowly during 1 min. Mixing continued for another 3 min to allow PVA fibres to be evenly dispersed. • Finally, following 2 min of rest, concrete is remixed for 1 min then sampled for the different tests. Specimens for mechanical properties were covered with plastic sheets and kept in a room with relative humidity and temperature of approximately 50% and 23 °C, respectively then demoulded or 24 ± 1 h later. The specimens were then sealed inside plastic bags then transferred for storage in a fog room at 100% RH and 22 °C temperature until the age of testing. 2.4

Testing

Testing included two distinct series: the first was conducted on non-fibrous (suspended) mortar to collect the parameters necessary for micromechanical tailoring of SHCC mixtures. Those tests are the fracture toughness, the direct tension, and the single-fibre pull-out tests. • The elastic modulus (E) of the matrix was evaluated on 100  200 mm cylinders at 28 days as per ASTM C469 [15]. • The fracture toughness (Km) of the matrix was determined following a procedure adapted from ASTM E399 [16]. The test was conducted at 28 days on (100  100  400 mm) plain prisms of 25 mm deep, 3 mm central notch tested in three-point bending configuration in a displacement-controlled mode using a displacement rate of 0.05 mm/min. • The direct tensile strength was conducted on dog-bone shaped coupons (Fig. 3) in order to determine the tensile strength of the plain matrix (rfc). The test was conducted

1128

O. A. Hisseine and A. T. Hamou

under a displacement-controlled mode at a rate of 0.2 mm/min. The test set-up is equipped with a 24.4 mm wire extensometer attached at the middle of the sample.

Fig. 3. Uniaxial tension test: Sample dimensions (left) and test configuration (right).

• Single fibre pull-out test (SFPT) was conducted to evaluate fibre/matrix interface properties necessary for micromechanical tailoring. A 5-N capacity of load cell was used for this purpose. The test was conducted under a displacement-controlled mode at a rate of 0.2 mm/min (See Fig. 4 (a) for the test set-up and Fig. 4 (b) for the actual test configuration). Further details about SFPT as well as the determination of interface parameters [frictional bond (s0), chemical bond (Gd), and slip hardening coefficient (b)] are provided elsewhere [17].

Fig. 4. Single-fibre pull-out test: Schematic of test set-up (a), actual test configuration with a focus on a single fibre being pulled from the matrix (b). Figure adapted from Hisseine et al. [17] with permission from Elsevier.

The second test series was conducted on SHCC to assess the behaviour at the composite level as well as to validate the outcome of the micromechanical tailoring. Those tests include the compressive strength, the uniaxial tensile strength and the

Controlling Strength and Ductility of Strain-Hardening Cementitious Composites

1129

flexural capacity. The compressive strength was assessed on 50  50  50 mm cubes (ASTM C109-16) [18]. The uniaxial tensile strength was evaluated at 28 days following the same procedures and specimen dimensions described earlier for the case of the plain matrix. The flexural capacity was evaluated on 100  100  400 mm prisms (ASTM C78-18) [19] at 28 days. Flexural tests were performed under four–point bending configuration in a displacement-controlled mode using a displacement rate of 0.05 mm/min. Elaborated experimental details can be found elsewhere [17, 20, 21]

3 Results and Discussions 3.1

Effect of Nano-Modification with CF on Micromechanical Properties

Table 1 presents the results of the micromechanical investigation conducted in this study. Results show that the incorporation of nanoscale cellulose filaments (CF) influence the properties of SHCC matrix as well as fibre/matrix interface properties. For matrix properties, CF appears to moderately increases the elastic modulus (Em), the fracture toughness (Km), and the first cracking strength (rfc). The enhancement in the Em (around 15%) can affect the fracture behaviour of SHCC. Given the definition of crack tip toughness (Jtip) as a function of the elastic modulus and the fracture toughness (Jtip ¼ Km2 =Em ), it is perceivable that that the increase in Em can lower Jtip and allow fulfilling the energy criterion (Jtip  Jb0 ) for SHCC [24]. In terms of interface properties, the use of CF slightly attenuated the frictional bond s0, significantly reduced the chemical bond Gd, and imparted a characteristic slip-hardening effect b. CF influences interface properties by altering the fibre/matrix interfacial transition zone (ITZ). The observed reduction in frictional bond s0 in the presence of CF is attributable to the interference of the omnipresent CF between the matrix and PVA fibres (Fig. 5), thereby reducing the contact sites between the matrix and the PVA fibres. While high frictional bond s0 favours higher bridging stress, too high s0 can cause fibre premature rupture, thereby jeopardizing composite ductility. Likewise, the observed reduction in the chemical bond Gd in the presence of CF is ascribable to the attenuation in contact sites between the PVA fibres and active cations from the matrix such as Ca2+ known to influence the affinity of PVA fibres to adhering to the matrix [22, 23]. The characteristic slip-hardening behaviour observed in systems with CF, on the other hand, can sprout from the propensity of the flexible nanoscale CF to causing a jamming effect during PVA fibre pull-out, thereby leading to resisting higher pull-out load. 3.2

Effect of Nano-Modification with CF on Fibre Bridging Capacity

Figure 6 presents the fibre bridging stress-crack opening response (r–d) for the different SHCC developed in this study. Using the micromechanical principles of SHCC, the strain hardening performance indicators for the different SHCC were also obtained and are presented in Table 2. The results of Fig. 6 indicate that the incorporation of CF has no observable influence on the maximum bridging capacity (r0) in spite of the aforementioned positive effect of matric strength [elastic modulus (Em), fracture

1130

O. A. Hisseine and A. T. Hamou Table 1. Mixture proportions

Mixture name

Mixture Composition (kg/m3) Cement Fly ash Quartz sand (FA) (QS)

Water HRWRA (solid extract)

M M M M

597 597 597 597

366 366 366 366

0.00CF 0.03CF 0.05CF 0.10CF

717 717 717 717

460 460 460 460

3.251 4.014 4.125 4.589

CF (solid extract) – 0.175 0.292 0.584

Fig. 5. Effect of nanomodification with CF on matric and interface properties of SHCC: (a) omnipresence of CF on PVA fibre surface, (b) interference of CF on fibre/matrix interface, and (c) nan-reinforcing of matrix by CF. Figure adapted from Hisseine et al. [17] with permission from Elsevier.

toughness (Km), and first cracking strength (rfc)]. This can be a consequence of the slight reduction in frictional bond s0 in the presence of CF. Nonetheless, all systems with CF recorded higher complementary energy Jb0 . Consequently, the assessment of strain-hardening indicators presented in Table 2 demonstrates that the index Jb0 /Jtip is higher in all SHCC incorporating CF than in the reference. Likewise, the ratio r0/rfc is maintained above 1.45 as required for substantial strain-hardening behaviour. Kanda and Li [24] have suggested that additional to satisfying the strength and energy criteria required, SHCC should also meet two more conditions: Jb0 /Jtip > 3 and r0/rfc> 1.45. With both ratios enhanced in all SHCC incorporating CF, it is perceivable that nanomodifying SHCC with CF fosters composite ductility. 3.3

Effect of Nano-Modification with CF on Composite Performance

To validate the outcome of the micromechanical investigation, the performance of resulting SHCC was evaluated at the composite level in terms of compressive strength, uniaxial tensile strength and flexural capacity.

Controlling Strength and Ductility of Strain-Hardening Cementitious Composites

1131

Fig. 6. Computed fibre bridging stress versus crack opening relationship (r–d) for varying GP content.

3.3.1. Effect of CF on compressive strength Figure 7 presents the effect of CF on the evolution of compressive strength up to 91 days. Overall, the incorporation of CF leads to strength enhancement particularly from 7 days and onward. While it is not evident to make a clear distinction between the effect of different CF dosages, it is observable that all CF concentrations led to higher compressive strength compared to that of the reference SHCC (without CF). With better effects observed at later curing ages, strength enhancement of  15-20% were recorded. This can be ascribed to the mechanism underlying the effect of CF on cement composites whereby the hydrophilic and hygroscopic CF can contribute to enhancing composite strength by an internal curing effect [9] as well as by a nanoreinforcing effect allowing to bridge cracks at the scale of hydrates [8]. Assessment of autogenous shrinkage in low water-to-cement ratio systems in our former work indicate that CF can contribute by internal curing effect to control volumetric instability whereby a reduction in autogenous shrinkage of up to 31% were obtained [9]. On the other hand, microstructure assessment as well as nanoindentation studies revealed that CF strengthen cement composites by a twofold effect: (i) increased degree of hydration (15%) and (ii) higher micromechanical properties of C-S-H matrix (  12–25%) [8]. 3.3.2. Effect of CF on uniaxial tensile strength and on flexural capacity Figure 8 depicts the uniaxial tensile behaviour for the different SHCC considered herein. The figure indicates that nano-modification by CF has influenced the first cracking strength (rfc), the post-cracking strength (rpc), the multiple cracking and strain-hardening response as well as the ultimate tensile strain capacity (eu). The highest enhancement in the first cracking strength rfc was obtained at 0.10% CF where an increase of  20% was observed. On the other had, the three respective CF dosages enhanced the average post-cracking strength rpc from 3.34 MPa in the reference SHCC to 3.48, 3.65, and 4.10 MPa, corresponding to enhancements of 5, 10, and 23%, respectively. As for the ultimate tensile strain capacity (eu), the incorporating CF enhanced eu from 3.01% in the reference SHCC [Fig. 8 (a)] to 3.33, 3.60, and 3.78% in the systems with 0.03, 0.05, and 0.10% CF [(Fig. 8 (b), (c), (d)]. This corresponds to enhancements of 11, 20, and 26%, respectively. The enhancement in eu is a consequence of a more prominent multiple cracking response reflected by the intensity and

M0.00CF M0.03CF M0.05CF M0.10CF

Mixture name

22.50 25.74 25.50 24.69

   

0.48 0.21 0.19 0.32

0.68 0.70 0.71 0.71

   

0.05 0.08 0.04 0.07

Matrix parameters Fracture Elastic toughness, modulus, Km Em pffiffiffiffi (GPa) (MPa. m)

2.59 2.90 2.90 2.92

    0.11 0.12 0.08 0.04

First crack uniaxial strength, rfc (MPa) 2.95 2.94 2.82 2.80

    0.52 0.56 0.66 0.78

2.65 0.27 0.25 0.21

    0.14 0.10 0.12 0.06

Interface parameters Frictional Chemical bond, bond, s0 Gd (MPa) (J/m2)

SHCC pseudo-ductility Maximum Complebridging mentary stress, energy, r0 Jb; (MPa) (J/m2) 0.003 4.82 64 1.24  0.39 5.03 67 1.51  0.20 5.15 66 1.31  0.34 5.18 67

Slip hardening coefficient, b

Table 2. Results of micromechanical investigation and strain-hardening indicators for SHCC

20.5 19.0 19.8 20.4

3.12 3.50 3.34 3.28

performance Jb; Crack Jtip tip toughness, Jtip (J/m2)

1.86 1.73 1.78 1.77

r0 rfc

1132 O. A. Hisseine and A. T. Hamou

Compressive strength, fc (MPa)

Controlling Strength and Ductility of Strain-Hardening Cementitious Composites

90 75 60

1133

M 0.0%CF M 0.03%CF M 0.05%CF M 0.10%CF

45 30 15 0 1

7

28 56 Age (days)

91

Fig. 7. Compressive strength development

frequency of peaks in the uniaxial tensile behaviour. This is more pronounced in the SHCC with 0.10% CF [Fig. 8 (d)]. The results of uniaxial tensile response are further confirmed by the flexural response shown in Fig. 9. The figure shows that the incorporation of CF influenced particularly the deflection capacity du (identified by the mid-span deflection corresponding to the maximum flexural load). As such, compared to the reference SHCC [Fig. 9 (a)] recording an average du of 3.91 mm, the incorporation of CF at 0.03, 0.05, and 0.10% [Fig. 9 (b), (c), (d), respectively] resulted in average du of 4.51, 4.92, and 5.31 mm. This corresponds to 15, 26 and 36% higher du. The improved deflection capacity demonstrates the effectiveness of CF in imparting higher ductility. These results can be linked to the micromechanical investigation where nano-modified SHCC showed higher complementary energy (Jb; ). The enhanced strain-hardening response and extended deflection capability in systems with CF can also be linked to the characteristic slip-hardening behaviour (b) imparted by CF as discussed earlier. The slip-hardening response imparted by CF was demonstrated elsewhere to allow the matrix to withstand pull-out loads even higher than the load at which the initial fibre/matrix bond deteriorated. As such, further energy is consumed to complete fibre pull-out. This can be reflected by the observed enhancement in the multiple cracking response as well as in the deflection capacity and in the overall ductility [17, 21].

1134

O. A. Hisseine and A. T. Hamou

Fig. 8. Uniaxial tensile strength

Fig. 9. Flexural capacity

Controlling Strength and Ductility of Strain-Hardening Cementitious Composites

1135

4 Summary and Conclusions Nano-engineered strain-hardening cementitious composites (SHCC) have been developed in this study following micromechanical tailoring. Cellulose filaments (CF) were incorporated at dosages of 0, 0.03, 0.05, and 0.10% in SHCC with fly ash to cement ratio of 1.2 and a water-to-binder ratio of 0.28. Results demonstrate that the incorporation of CF enhances both strength and ductility characteristics of SHCC. Specific findings are as follows: • The incorporation of CF influenced the mechanical properties of the plain matrix as well as the micromechanical properties at fibre/matrix interface. SHCC with CF recorded higher elastic modulus, a reduced chemical bond energy, and a characteristic slip-hardening effect. • The effect of CF on elastic modulus was positively reflected by an improved complementary energy and a reduced crack tip toughness. Consequently, higher strain-hardening performance was obtained. • Experimental validation by tensile and flexural tests proved that the incorporation of CF enhanced composite tensile strength by up to 23% and increased the ultimate strain capacity in tension by up to 26% and the deflection capacity in flexure by up to 36%. Finally, results support the effectiveness of nano-engineering SHCC with CF to foster composite strength and ductility without necessarily increasing the volume fraction of PVA fibres or the binder content. This provides a twofold benefit in terms of utilization of the most abundant and renewable natural polymer on the planet to enhance the performance of cement and concrete composites as well as to foster their ecoefficiency. Acknowledgements. This project is jointly supported by a Cooperative Research and Development (CRD) grant from the Natural Sciences and Engineering Research Council of Canada (NSERC), Canada Vanier Graduate Scholarship (CGS) program award no: 360284, Kruger Biomaterials Inc. (QC, Canada), and Euclid Chemicals. The authors are grateful to the financial support from all these partners.

References 1. Li, V.C., Leung, C.K.Y.: Steady state and multiple cracking of short random fibre composites. ASCE J. Eng. Mech. 188(11), 2246–2264 (1992) 2. Guan, X., Li, Y., Liu, T., Zhang, C., Li, H., Ou, J.: An economical ultra-high ductile engineered cementitious composite with large amount of coarse river sand. Constr. Build. Mater. 201, 461–472 (2019) 3. Yu, K.Q., Yu, J.T., Dai, J.G., Lu, Z.D., Shah, S.P.: Development of ultra-high performance engineered cementitious composites using polyethylene (PE) fibres. Constr. Build. Mater. 158, 217–227 (2018) 4. Ding, Y., Yu, J., Yu, K.Q., Xu, S.: Basic mechanical properties of ultra-high ductility cementitious composites: From 40 MPa to 120 MPa. Compos. Struct. 185, 634–645 (2018) 5. Meng, W., Khayat, K.H.: Mechanical properties of ultra-high-performance concrete enhanced with graphite nanoplatelets and carbon nanofibers. Compos. B Eng. 107, 113– 122 (2016)

1136

O. A. Hisseine and A. T. Hamou

6. Peters, S.J., Rushing, T.S., Landis, E.N., Cummins, T.K.: Nanocellulose and microcellulose fibres for concrete. Transport. Res. Rec. 2142, 25–28 (2010) 7. Ma, H., Cai, J., Lin, Z., Qian, S., Li, V.C.: CaCO3 whisker modified Engineered Cementitious Composite with local ingredients. Constr. Build. Mater. 151, 1–8 (2017) 8. Hisseine, O.A., Wilson, W., Sorelli, L., Tolnai, B., Tagnit-Hamou, A.: Nanocellulose for improved concrete performance: a macro-to-micro investigation for disclosing the effects of cellulose filaments on strength of cement systems. Constr. Build. Mater. 206, 84–96 (2019) 9. Hisseine, O.A., Omran, A.F., Tagnit-Hamou, A.: Influence of cellulose filaments on cement pastes and concrete. J. Mater. Civ. Eng. 30(6), 04018109 (2018) 10. Hisseine, O.A., Basic, N., Omran, A.F., Tagnit-Hamou, A.: Feasibility of using cellulose filaments as a viscosity modifying agent in self-consolidating concrete. Cem. Concr. Compos. 94, 327–340 (2018) 11. Li, V.C.: Engineered Cementitious Composites (ECC) – Tailored Composites Through Micromechanical Modeling, Fiber Reinforced Concrete: Present and the Future, pp. 64–97. Canadian Soc. Civil Eng., Montreal (1998) 12. Li, V.C., Wang, S., Wu, C.: Tensile strain-hardening behaviour of polyvinyl alcohol engineered cementitious composite. ACI Mater. J. 98(6), 483–492 (2001) 13. Li, V.C.: On engineered cementitious composites (ECC) – a review of the material and its application. J. Adv. Concr. Technol. 1, 215–230 (2003) 14. Kong, H., Bike, S.G., Li, V.C.: Development of a self-consolidating engineered cementitious composite employing electrosteric dispersion/stabilization. Cem. Concr. Compos. 25(3), 301–309 (2003) 15. ASTM C469/ C469M-14, Standard Test Method for Static Modulus of Elasticity and Poisson’s Ratio of Concrete in Compression, ASTM International, West Conshohocken, PA, 2014 www.astm.org 16. ASTM E399-12, Standard test method for linear-elastic plane-strain fracture toughness KIc of metallic materials, ASTM International, West Conshohocken, PA, 2012. www.astm.org 17. Hisseine, O.A., Tagnit-Hamou, A.: Characterization and nano-engineering the interface properties of PVA fibres in strain-hardening cementitious composites incorporating highvolume ground-glass pozzolans, Constr. Build. Mater. 234, 117213 (2020) 18. ASTM C109/ C109M-16a, Standard Test Method for Compressive Strength of Hydraulic Cement Mortars (Using 2-in. or [50-mm] Cube Specimens), ASTM International, West Conshohocken, PA, 2016 www.astm.org 19. ASTM C78/ C78M-18, Standard Test Method for Flexural Strength of Concrete (Using Simple Beam with Third-Point Loading), ASTM International, West Conshohocken, PA, 2018 www.astm.org 20. Hisseine, O.A., Tagnit-Hamou, A.: Development of ecological strain-hardening cementitious composites incorporating high-volume ground-glass pozzolans. Constr. Build. Mater. 238, 117740 (2020) 21. Hisseine, O.A., Tagnit-Hamou, A.: Nanocellulose for the development of ecological nanoengineered strain-hardening cementitious composites incorporating high-volume ground-glass pozzolans, article under review by Cement and Concrete Composites (2020) 22. Rodger, S.A., Brooks, S.A., Sinclair, W., Groves, G.W., Double, D.D.: High strength cement pastes. J. Mater. Sci. 20, 2853–2860 (1985) 23. Gulgun, M.A., Kriven, W.M., Tan, L.S., McHugh, A.J.: Evolution of mechano-chemistry and microstructure of a calcium aluminate-polymer composite: Part I—Mixing Time Effects. J. Mater. Res. 10(7), 1746–1755 (1995) 24. Kanda, T., Li, V.C.: Multiple cracking sequence and saturation in fibre reinforced cementitious composite, Concr. Res. Technol., JCI 9 (2) (1998) 19–33

Comprehensive Characterization of UHPFRC Mixes for Seismic and Durability Rehabilitation of Bridge Piers C. Sevigny-Vallières1(&), P. Marchand2, B. Terrade2, N. Roy1, F. Toutlemonde2, and A. Tagnit-Hamou1

2

1 Université de Sherbrooke, Sherbrooke, Québec, Canada [email protected] Materials and Structures Department, Université Gustave Eiffel, Marne-la-Vallée, France

Abstract. Towards design of UHPFRC jacketing for rehabilitation of bridge piers, one UHPFRC mix has been extensively characterized in order to provide the necessary material characteristics for such a design verification process. This mix is an innovative environmental-friendly UHPFRC mix, comprising recycled glass powder as cement and quartz powder replacement. The mechanical characterization of the mix comprises compressive strength and its evolution, Young’s modulus at early age and mature state, development of autogenous and total shrinkage, and tensile behaviour identified by flexural testing both on standard prismatic specimens as well as on thin plates representative of the UHPFRC jacketing layer cast around the existing concrete pier. Development of this effort has been undertaken in a joint project within ECOMAT international laboratory with Université de Sherbrooke (Canada) and the Materials and Structures Department of Université Gustave Eiffel (France). Keywords: UHPFRC
Glass powder
Early age properties properties
Tensile characterization
Bridge piers


Mechanical

1 Introduction Both in Canada and France, the bottom parts of a large number of concrete bridge piers along motorways, especially the central piers of common overpasses, suffer degradation due to chloride ingress, because this part is generally directly exposed to splash of deicing salts and salty polluted water caused by the traffic. Moreover, a significant part of these piers date back to more than 30 years ago, and the level of seismicity considered for their design as well as the corresponding detailing provisions do not meet present requirements for structures that should not endanger the operation of major corridors in case of an important seismic event. UHPFRC jacketing, possibly combined with additional transverse reinforcement, in substitution to existing damaged regular concrete cover, has appeared as a promising solution, which may prove technically and economically efficient. This option can offer a better durability to the existing column, the UHPFRC having a low porosity, and it can ensure confinement by the presence of the fibers in the matrix. © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 1137–1148, 2021. https://doi.org/10.1007/978-3-030-58482-5_100

1138

C. Sevigny-Vallières et al.

The obtained structural ductility of UHPFRC jacketed piers may, however, highly depend on the pier-axial reinforcement ratio, on axial loading rate as compared to the global capacity, and on detailing of the UHPFRC jacket implementation with respect to the existing pier and footing critical cross-sections (variation of the concrete area, end of lap bars, etc.). In addition, the structural behaviour of the jacket is affected by the quantity and orientation of the fibers in the matrix. The influence of casting conditions and fiber orientation on the tensile strength is of great importance for the design of the reinforcing jacket. Indeed, fibers oriented parallel to the longitudinal axis of the column allow a gain in the bending strength of the column and fibers oriented perpendicularly allow a gain in confinement. For design purposes, it is also important to know the shrinkage of the concrete in the first hours after setting. Restrained deformations, due to the presence of the existing column, could induce an initial confinement to the section and possibly lead to cracking of the matrix. Therefore, optimization of the combined seismic and durability rehabilitation must be investigated, with the calibration of a rational repair design methodology. Development of this effort has been undertaken in a joint project within ECOMAT international laboratory with Université de Sherbrooke (Canada) and the Materials and Structures Department of Université Gustave Eiffel (France). As a first step, one UHPFRC mix has been extensively characterized in order to provide the necessary material characteristics for such a design verification process. This mix is an innovative environmental-friendly UHPFRC mix, comprising recycled glass powder as cement and quartz powder replacement (UHPFRC-GP). This new concrete have been developed at the University of Sherbrooke for different applications [1] and for different replacement ratios [2]. The mechanical characterization of the mix comprises compressive strength and its evolution, Young’s modulus at early age and mature state, development of autogenous and total shrinkage, and tensile behavior identified by flexural testing both on standard prismatic specimens as well as on thin plates representative of the UHPFRC jacketing layer cast around the existing concrete pier. This research project has extended the knowledge of the structural properties as well as the early age properties of UHPFRC-GP. Indeed, in France and in Europe, different testing standards, applying specifically to UHPFRC, are available, which is not the case in Canada. It is therefore of great interest to combine North American and European expertise to perform a complete characterization of UHPFRC-GP. The range of answers of UHPFRC to these standards is also given in this study. The paper will focus on the results of this comprehensive characterization protocol, which will enable validating the repair design methodology based on the structural tests to be detailed and realized in a next step of the research program.

2 Test Description For the compression tests, the french standard NF-EN 12390-3 [3] was used. For the evaluation of the Young’s modulus and the Poisson’s ratio in the long term, a test protocol close to the standard NF EN 12390-13 [4] was followed, to which are added the specifications of the standard NF P18-470 [5] and NF P18-710 [6] for the UHPFRC. The strain measurement device J2P [7] was used to measure the Young’s

Comprehensive Characterization of UHPFRC Mixes

1139

modulus and Poisson’s ratio in a compression test. In order to determine the properties of UHPFRC at early age, the BT JADE [8] test and the FreshCon [9, 10] test were performed. The BTJASPE test [9, 11], developed at IFSTTAR, allows to follow the evolution of the Young’s modulus of concrete at early age, from the first hours after the concrete is poured. The temperature variations of the specimen are limited by a constant water circulation around the mould, regulated by a thermocouple installed in the center of the concrete mass. The dynamic Young’s modulus can be measured using the FreshCon test [9, 10] developed at the University of Stuttgart. The velocity of the compressional (P) and shear (S) waves is measured in order to quantify the stiffening of the cementitious matrix in the different directions. The BT JASPE and FreshCon tests were performed on a single sample and the J2P test was performed on 3 samples per test period. According to NF P18-427 [12], shrinkage must be measured at 2, 7, 28, 91 and 365 days. However, this technique does not provide information on the first hours of setting of the concrete. The BT JADE test [8], developed at LCPC (IFSTTAR), allows a continuous measurement of autogenous shrinkage as soon as the test specimens are poured. This measurement is particularly important for concrete with a low water to binder ratio, where autogenous shrinkage is particularly important and can cause cracking when deformation is restrained [8]. Sensors are used to measure and control the temperature with the specimen immersed in a thermostatically controlled bath. The deformations due to autogenous shrinkage only can be isolated in this way. Thus, the BT JADE test was carried out in conjunction with conventional shrinkage measurements, allowing information on the shrinkage at different phases of concrete maturation to be obtained. The autogenous shrinkage was measured by the BT JADE test [8], during the first two weeks of concrete curing. Total shrinkage was measured on prismatic specimens stored in an climate chamber at 20°C and at a relative humidity of 50%. Shrinkage measurements by the BT JADE device were performed on a single specimen and conventional shrinkage measurements were performed on 3 specimens. In order to characterize the tensile behaviour of UHPFRC, bending tests were performed as recommended by NF P18-470 [5]. Compared to the direct tensile tests, the bending tests present a simpler execution and allow a better accounting of the real fabrication conditions [13, 15]. Indeed, the uniaxial tensile test on a “dogbone” shaped specimen requires the use of adapted press jaws, which must be compatible with the shape of the specimen [16]. Furthermore, the reduced width of the center of the specimen causes a preferential orientation of the fibers due to a wall effect, which overestimates the tensile capacity of the UHPFRC [17]. For these reasons, the bending test was used. In order to obtain the stress-strain law, or stress-crack-opening law, a point-by-point inverse analysis, based on the equilibrium of moments and forces of a section in bending, was performed on each of the results obtained [5, 13–15]. In order to obtain the identity card of the UHPFRC under study, standard NF P 18-470 [5] suggests to perform 3-point bending tests on notched prisms to obtain the law of postpeak tensile behaviour. The notch allows the localization of the crack and an exact measurement of the resistance to crack opening. The 4-point bending tests give the elastic and plastic behaviour of prisms and thin plates. The prisms tested had a crosssection of 70 mm by 70 mm and a length of 280 mm. They were cast in inclined open metal moulds to favour a preferential orientation of the fibers. The study of the

1140

C. Sevigny-Vallières et al.

behaviour of the thin plates makes it possible to represent the projected application, i.e. the rehabilitation of bridge piers with UHPFRC-GP. For the fabrication of the reinforcement jacket, the concrete will be poured vertically. It is therefore interesting to study the strength of the UHPFRC-GP horizontally (providing a confinement) and vertically (providing bending strength). Two plate orientations were studied, vertical and horizontal, as shown in Fig. 1. The thin plates had a cross-section of 104 mm by 30 mm and a length of 600 mm. They were cast in closed moulds with a small inclination to allow flow along the panels.

Fig. 1. Casting direction of the plates

Since the stress distribution is different in notched (3PBT) and un-notched (4PBT) prisms, the inverse analyses performed are also different [5, 15]. As shown in Fig. 2, the stress distribution of the notched prism is linear in the uncracked area and depends on the elastic curvature. The stress distribution is non-linear in the cracked area and depends on the crack opening. Thus, by computing the forces equilibrium on the bending section and using the kinematic relationship developed by Casanova [18] to connect the curvature of the uncracked and the cracked part, it is possible to obtain the relationship between the equivalent tensile stress and the crack opening using the inverse analysis [5, 19].

Fig. 2. Strain and stress distribution in a notched prism

Comprehensive Characterization of UHPFRC Mixes

1141

4-point bending tests on an unnoticed prism are used to determine the yield strength under tension and, if the tensile behaviour is strain-hardening, an inverse analysis can be performed to obtain a post-cracking behaviour law (stress-strain). The stress varies linearly with the elastic curvature until the yield strength is reached. The relationship between stress and strain beyond the yield point is obtained by considering the overall moment-curvature relationship. Thus, by computing the forces equilibrium on the bending section, the equivalent tensile stress vs strain can be obtained by using the inverse analysis.

3 Material The ultra-high performance fiber-reinforced concrete mix studied in this test program has a water-to-binder ratio of 0.18 and contains 2% metal fibers. The quantity of cement is reduced by 25% and replaced by glass powder and 100% of the quartz powder is replaced by glass powder. The glass powder is considered as a supplementary cementitious material in Canadian standard (CSA, A3000).

4 Experimental Results 4.1

Compressive Strength, Young’s Modulus, Shear Modulus and Poisson’s Ratio

Table 1 shows the average compressive strength (fcm), Young’s modulus (Ecm) and Poisson’s ratio (m) obtained from the J2P measuring device [7] in pure compression tests. In comparison, the UHPFRCs commonly used in France have an average 28-day compressive strength between 150 and 200 MPa, a Young’s modulus between 45 and 65 GPa and a Poisson’s ratio of 0.2 [5, 6]. However, these values are in the same range at 90 days. Table 1. Evolution of long term properties of UHPFRC-GP Time (days) Fcm (MPa) 3 72 7 85 28 140 90 169

Ecm (GPa) 38.5 42.3 46.4 47.6

m 0.184 0.196 0.192 0.183

The evolution of the modulus of elasticity (E) of UHPFRC-GP, shown in Fig. 3, was measured with FreshCon [9, 10], BT JASPE [9, 11] and J2P [7] tests. The red curve represents the dynamic modulus of elasticity obtained for a reference UHPFRC mix. The dynamic Young’s modulus at early age, measured with the FreshCon test, reaches a value of 32.9 GPa at 7 days. The static Young’s modulus at early age, measured with the BT JASPE test and the J2P test, reaches a value of 37.6 GPa and

1142

C. Sevigny-Vallières et al.

42.3 GPa respectively at 7 days. In comparison, the UHPFRCs have a dynamic Young’s modulus between 40 and 45 GPa (17.7% higher) and a static Young’s modulus between 45 and 65 GPa (6 to 16% higher) [5, 6].

Fig. 3. Evolution of elastic modulus of UHPFRC-GP

The dynamic shear modulus (Gdyn) of the UHPFRC-GP, obtained in the FreshCon test [9, 10], is 15.9 GPa, compared to a value of 20 to 25 GPa for the UHPFRCs [5, 6], which is 20% lower in the short term. The expected value at 7 days, considering the Young’s modulus and Poisson’s ratio obtained with the J2P test is 20.4 GPa. The FreshCon test [9, 10] also determined that the initial setting of UHPFRC with glass powder begins approximately 16 h after the addition of the liquids and that the stiffness of the material stabilizes after 30 h. The initial setting is slightly delayed compared to the reference concrete, beginning 10 h after the addition of the liquids. 4.2

Shrinkage

The autogenous shrinkage of UHPFRC-GP and the total measured shrinkage are shown in Fig. 4. Following the initiation of concrete setting, the shrinkage increases rapidly and reaches 350 lm/m in 20 h. The measured shrinkage stabilizes after 50 days, reaching a maximum value of 700 lm/m. In comparison, UHPFRC commonly used in France have a shrinkage between 550 and 800 lm/m [15].

Comprehensive Characterization of UHPFRC Mixes

1143

Fig. 4. Autogenous and total shrinkage of UHPFRC-GP

4.3

Tensile Behaviour

The stress-strain relationship resulting from the inverse analysis on the 4-point bending tests on six un-notched prisms is shown in Fig. 5 and the mean curve is shown in Fig. 6. The post-cracking behaviour versus crack opening, derived from the inverse analysis of the 3-point bending tests on notched prisms, is shown in Fig. 7 and the mean distribution is shown in Fig. 8. Finally, thin plates were also tested in 4-point bending tests to obtain the stress-strain response. The results of the horizontal thin plates are shown in Fig. 9 and the mean curve is shown in Fig. 10. The results of the vertical thin plates are shown in Fig. 11 and the mean curve is shown in Fig. 12. Those stress-strain relationships were obtained between 29 days and 39 days. It should be noted that in Fig. 9 and Fig. 11, the H5, V1, V2 and V3 plates showed a crack location outside the instrumented area. The average strength obtained for UHPFRCs commonly used in France is between 7 and 12 MPa [6, 20].

Fig. 5. Stress-strain relationship obtained by inverse analysis (4PBT)

Fig. 6. Average stress-strain relationship obtained by inverse analysis (4PBT)

1144

C. Sevigny-Vallières et al.

Fig. 7. Post-pic behaviour obtained by inverse analysis (3PBT)

Fig. 8. Average post-pic behaviour obtained by inverse analysis (3PBT)

Fig. 9. Stress-strain relationship obtained by inverse analysis (4PBT-Horizontal thin plates)

Fig. 10. Average stress-strain relationship obtained by inverse analysis (4PBT-Horizontal thin plates)

Fig. 11. Stress-strain relationship obtained by inverse analysis (4PBT-Vertical thin plates)

Fig. 12. Average stress-strain relationship obtained by inverse analysis (4PBTVertical thin plates)

Comprehensive Characterization of UHPFRC Mixes

1145

The waves observed in the curves resulting from inverse analysis come from the successive cracks opening which create successive decrease and increase in the tensile stress. It is also resulting from the inverse analysis which tends to amplify these waves (inverse analysis is a kind of derivation of raw curves). These waves should not be considered to determine design curves.

5 Discussion The 28-day compressive strength of UHPFRC-GP is 6% lower than that of the reference UHPFRCs. However, at 90 days, the strength continues to increase and is in the same range as the common UHPFRC [6]. Young’s modulus is 5 to 20% lower than the short-term (7 days) reference values. In the long term, the modulus is in the same range as the reference values. It can thus be concluded that the resistance gain occurs later. It is well known that cement replacement by pozzolanic material such as glass powder results in delay in hydration which reduce the mechanical performances at early age but not at long term (56 days and more). Indeed, the strength of UHPFRC-GP increases considerably between 28 and 90 days. The shrinkage measured for UHPFRC-GP is in the same range of values as the UHPFRC. The tensile strength measured on prisms and thin plates is in the same range of values as the UHPFRC. In Figs. 5 to 8, it can be observed that the UHPFRC-GP prisms tested have a slightly strain-hardening behaviour. Figure 13, adapted from Mousa et al. [21], shows the idealized behaviour expected for a strain-hardening UHPFRC. The curves used for the design need to be smoothed, as presented in Fig. 14. In this figure, the different phases of the constitutive law are easily identifiable, i.e. the linear zone, corresponding to the elastic behaviour, and the constant zone, corresponding to the fine multi-cracking. The increase in stress corresponds to the opening of the cracks. The maximum stress reached corresponds to the localization of a crack and the pull-out of the fibers [13, 14]. Figure 7 and Fig. 8 show the last phases of the constitutive law as a function of crack opening. Figures 9 to 12 show that the tensile strength of horizontal plates is higher than that of vertical plates and that a softening behaviour is observed, due to the rapid localization of the crack. Thus, it can be concluded that the fibers have an orientation following the flow of the fluid to the base of the mould, a preferential horizontal orientation being obtained, as shown in Fig. 15. For the projected application, a preferential orientation of the fibers perpendicular to the column axis is therefore to be expected, which is favourable for a column confinement application. Nevertheless, a post-cracking tensile stress higher in the case of thin plates would have been expected compared to prisms. This is not the case here. The way of introducing the material into mould would certainly need to be improved.

1146

C. Sevigny-Vallières et al.

Fig. 13. Idealized expected tensile behavior of UHPFRC

Fig. 14. Data smoothing of 4PBT on prism for design

Fig. 15. Direction of flow and fiber orientation

6 Conclusion Based on the experimental results presented above, the following conclusions can be drawn: • Replacing a proportion of the cement and all of the quartz powder with glass powder makes it possible to obtain properties that are comparable (compressive and tensile strength, Young’s modulus, shear modulus, etc.) to those of the UHPFRCs commonly used in France, with better environmental impact • The gain in strength of UHPFRC-GP is delayed compared to the UHPFRCs commonly used in France at 28 days, with comparable strength being reached at 90 days. • The prisms tested in bending showed a hardening behaviour, whereas the thin plates rather showed a softening behaviour due to the rapid localization of the crack. • For the planned application, a preferential orientation of the fibers perpendicular to the column axis is therefore to be expected. The durability tests are in progress.

Comprehensive Characterization of UHPFRC Mixes

1147

Acknowledgement. This research project was financially supported by Chaire de recherche SAQ, Université de Sherbrooke (grant), LIA-Écomat and Université Gustave Eiffel. The authors would like to express their gratitude to Aghiles Begriche, PhD student at Université de Sherbrooke and to the technical personnel of Université Gustave Eiffel for their help and support.

References 1. Tagnit-Hamou, A., Soliman, N., Omran, A.: «Green Ultra-High Performance Glass Concrete». In: Proceedings of the First International Interactive Symposium on UHPC, Des Moines, Iowa, USA, (2016)https://doi.org/10.21838/uhpc.2016.35 2. Soliman, N.A., Tagnit-Hamou, A.: Development of ultra-high-performance concrete using glass powder – towards ecofriendly concrete. Constr. Build. Mater. 125, 600–612 (2016). https://doi.org/10.1016/j.conbuildmat.2016.08.073 3. AFNOR, «NF EN 12390–3 - Essais pour le béton durci- Partie 3- Résistance à la compression des éprouvettes» (2019) 4. AFNOR, «NF EN 12390–13- Essais pour le béton durci- Partie 13: Détermination du module sécant d’élasticitié et du module de compression» (2014) 5. AFNOR, «NF P18–470- Bétons fibrés à Ultra Hautes Performances- Spécifications, performance, production et conformité» (2016) 6. AFNOR, «NF P18–710-Complément national à l’Eurocode 2- Calcul des structures en béton: règles spécifiques pour les Bétons Fibrés à Ultra-Hautes Performances (BFUP)» (2016) 7. Boulay, C., et al.: «Un extensomètre à béton éliminant l’influence des déformations transversales sur la mesure des déformations longitudinales». Matér. Constr. 14, 35–38 (1981) 8. Boulay, C.: «Développement d’un dispositif de mesure du retrait endogène d’un béton au jeune âge», présenté à Huitième édition des journées scientifiques du Regroupement francophone pour la recherche et la formation sur le béton (RF)2B, Montréal, Canada, 2007, [En ligne]. Disponible sur: https://scholar.google.com/scholar?hl=fr&as_sdt=0%2C5&q= Boulay+C.+2007.+D%C3%A9veloppement+d%27un+dispositif+de+mesure+du+retrait+ endog%C3%A8ne+d%27un+b%C3%A9ton+au+jeune+%C3%A2ge%2C+Huiti%C3% A8me+%C3%A9dition+des+Journ%C3%A9es+scientifiques+du+Regroupement+ francophone+pour+la+recherche+et+la+formation+sur+le+b%C3%A9ton%2C+Montr% C3%A9al%2C+Canada.&btnG= 9. Boulay, C., et al.: «Monitoring elastic properties of concrete since very early age by means of cyclic loadings, ultrasonic measurements, natural resonant frequency of composite beam (emm-arm) and with smart aggregates» (2013) 10. Carette, J., Staquet, S.: Monitoring the setting process of mortars by ultrasonic P and S-wave transmission velocity measurement. Constr. Build. Mater. 94, 196–208 (2015). https://doi. org/10.1016/j.conbuildmat.2015.06.054 11. Boulay, C., et al.: How to monitor the modulus of elasticity of concrete, automatically since the earliest age? Mater. Struct. 47(1–2), 141–155 (2014). https://doi.org/10.1617/s11527013-0051-3 12. AFNOR, «NF P18–427- Détermination des variations dimensionnelles entre deux faces opposées d’éprouvettes de béton durci» (1996) 13. Baby, F., Graybeal, B., Marchand, P., Toutlemonde, F.: UHPFRC tensile behavior characterization: inverse analysis of four-point bending test results. Mater. Struct. 46(8), 1337 (2013)

1148

C. Sevigny-Vallières et al.

14. Baby, F., Graybeal, B., Marchand, P., Toutlemonde, F.: Proposed flexural test method and associated inverse analysis for ultra-high-performance fiber-reinforced concrete. ACI Mater. J. 109(5), 545–555 (2012) 15. Resplendino, J., Marchand, P.: Bétons fibrés à ultra-hautes performances. Recommandations/ Ultra-high performance fiber-reinforced concretes. Recommendations. AFGC, Paris, France, Revised edition (2013). https://doi.org/10.1002/9781118557839.ch47 16. di Prisco, M., Ferrara, L., Lamperti, M.: Double edge wedge splitting (DEWS): an indirect tension test to identify post-cracking behaviour of fibre reinforced cementitious composites. Mater. Struct. 46(11), 1893 (2013) 17. Delsol, S., Charron, J.-P.: Numerical modeling of UHPFRC mechanical behavior based on fibre orientation. In: Proceedings of the RILEM-fib-AFGC International Symposium on Ultra-High Performance Fibre-Reinforced Concrete, UHPFRC. p. 10 (2013) 18. Casanova, P.: «Bétons renforcés de fibres métalliques: du matériau à la structure. Etude expérimentale et analyse du comportement de poutres soumises à la flexion et à l’effort tranchant», École nationale des ponts et chaussées, Paris, France (1996) 19. Marchand, P.: «Caractérisation et utilisation des BFUP pour des applications structurelles» , PhD Thesis, Université Paris-Est, (2019) 20. Herrera, A.: «Fonctionnement des jonctions âmes-membrures en Béton Fibrés à UltraHautes Performances (BFUP)», PhDThesis, Université Paris-Est (2017) 21. Mousa, M., Cuenca, E., Ferrara, L., Roy, N., Tagnit-Hamou, A.: Tensile characterization of an “eco-friendly” uhpfrc with waste glass powder and glass sand. Strain-Hardening Cem.Based Compos. 15, 238–248 (2018). https://doi.org/10.1007/978-94-024-1194-2_28

Evaluation of the Splitting Tensile Strength of Ultra-High Performance Concrete An Hoang Le(&) NTT Hi-Tech Institute, Nguyen Tat Thanh University, Ho Chi Minh City, Vietnam [email protected]

Abstract. The splitting tensile strength of ultra-high performance concrete (UHPC) is much larger than that of normal concrete. It was found that the studies on UHPC has mainly focused on the direct tensile strength or flexural strength, while there has been insufficient work to evaluate the splitting strength characteristics of UHPC. Therefore, this study is aimed at presenting experimental and statistical evaluation of the splitting tensile strength of UHPC. The splitting tests were conducted on cylindrical specimens of 100
200 mm size. UHPC was designed to achieve a nominal compressive strength of 200 MPa at the age of 28 days. Macro steel fibers were used to reinforce the UHPC by volumetric percentages of 0, 1, and 2%. The effect of fiber volume on the splitting tensile strength was investigated by the test results. The values of the splitting tensile strength of UHPC with and without fibers in some previous studies were collected together with this study and subsequently verified with the predictions of the splitting tensile strength obtained from the existing models. The appropriateness of these existing models was clarified. Finally, based on the regression analysis on the collected test results, a simplified equation was proposed to estimate the splitting tensile strength of UHPC having compressive strength varying between 120 and 200 MPa. Keywords: Ultra-high performance concrete  Splitting tensile strength Fibers  Compressive strength  Splitting tests



1 Introduction Ultra-high performance concrete (UHPC) is considered as a composite material that comprises fine powder such as Portland cement, silica fume, fine sand. In addition to a very low water-to-binder ratio, a super plasticizer is required to be used in the UHPC mixture to ensure the very high workability [2, 17]. Due to a very dense matrix, UHPC is characterized by a very high compressive strength exceeding 150 MPa, possibly attaining 250 MPa and an outstanding durability [2–4]. The inclusion of internal fiber reinforcement ensures the ductile behavior of UHPC under tension and significantly increases the tensile, flexural and shear strength under different types of structural actions. The superior properties of UHPC can generate many structural advantages, thereby leading to a great attention to many researchers throughout the world. Along with the compressive strength, the tensile strength is one of the most important aspects © RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 1149–1160, 2021. https://doi.org/10.1007/978-3-030-58482-5_101

1150

A. H. Le

in the design of structures made of UHPC, especially when UHPC reinforced by fibers (UHPFRC) is applied. Generally, there are three test methods to determine the tensile strength of concrete either directly or indirectly: (1) Direct tensile strength test; (2) Splitting strength test; (3) Flexural tensile strength test [7]. The direct tensile test (DTT) has some impediments with its performance such as the slippage or the alignment between gripping apparatus and concrete specimens, the stress concentration at the gripping devices [15]. The results from the flexural tensile test (FTT) are usually sensitive to the geometry and size of concrete specimen, loading application and strain rate. Both DTT and FTT give the tensile behavior of concrete, however, these two test methods require special test equipment and comparatively high cost to conduct [7]. On the other hand, the splitting test (ST), which is an indirect tensile strength test of concrete, was commonly adopted by many standards (ASTM C496, AS 1012.10 – Standards Australia 2000) and easier to carry out than DTT and FTT [7, 15, 16]. In terms of UHPC and UHPFRC, most previous studies have focused on DTT and FTT in order to comprehensively provide the effect of fibers on the pre-peak and postpeak response of tensile stress - strain behavior. Few studies have been conducted in the literature to investigate the splitting strength of UHPC and UHPFRC. Graybeal (2006) [17] suggested some modifications of the splitting test in ASTM C496 to capture the post-cracking behaviors of UHPC cylinders. El-Helou et al. (2014) [13] reported the splitting test results of UHPC cylinders using 0%, 2% and 4% steel fibers, and compared with the direct tensile test. El-Din et al. (2016) [12] investigated the influence of fiber volume and aspect ratio on the splitting strength of UHPC and developed an equation to predict the splitting strength of UHPC. Shafieifar et al. (2017) [23] compared the splitting strength between UHPC and normal strength concrete (NSC) cylinders. Bae et al. (2017) [7] studied the indirect tensile strength tests including splitting test and flexural test of ultra high strength concrete (UHSC) with and without steel fibers to show the effect of compressive strength and fiber content on the splitting strength. The authors also proposed equation for calculating the splitting strength considering the concrete compressive strength and fiber content simultaneously. Goaiz et al. (2018) [15] presented a comparison among the splitting test, the double punch test and the direct tensile test of notched prisms for determining the tensile strength of reactive powder concrete (RPC – a type of UHPC). Guler et al. (2018) [26] studied on splitting strength of concrete using steel, synthetic and hybrid fibers. Moreover, these authors evaluated the existing strength models for predicting the splitting tensile strength. It was found that the current information in the literature about the splitting tensile strength of UHPC and UHPFRC is limited. Therefore, the main aim of this paper is to examine the splitting tensile strength of UHPC and UHPFRC by splitting tests on concrete cylinders of 100
200 mm. The effect of steel fiber volumes of 0%, 1% and 2% on the splitting tensile strength was investigated. Based on the database of splitting test results obtained from this study and previous studies, this paper provide a statistical evaluation on some existing models for estimating the splitting strength of UHPC and UHPFRC. Finally, a simplified equation was derived from the regression analysis to predict the splitting tensile strength of UHPC and UHPFRC having compressive strength varying between 120 and 200 MPa.

Evaluation of the Splitting Tensile Strength

1151

2 General Specifications 2.1

Materials and Specimen Preparation

A total of nine UHPC and UHPFRC mixtures were prepared for this study. The recipe of M3Q, which was developed at University of Kassel - Germany, was adopted for producing UHPC and UHPFRC [2]. The details of the mix proportions are given in Table 1. It should be noted that the M3Q mix was designed to provide a very high selfcompacting characteristic and a compressive strength of concrete cylinder at about 200 MPa. Therefore, the necessity for compacting of concrete using external vibration was eliminated. This is favorable for the preparation of test specimens. The steel fibers (smooth and brass coated surface) with a diameter df of 0.175 mm and a length lf of 13 mm were added to the UHPC mix in volume fraction of 0%, 1% and 2% (UHPC, UHPFRC-SF1% and UHPFRC-SF2%). The mechanical properties of steel fibers are illustrated in Table 2. A mixer (Zyklos Gleichlaufmischer ZZ 150 HE) with maximum capacity of 170 L was used for mixing each concrete batch. The sequence for mixing UHPC and UHPFRC is presented in Table 2. When UHPC and UHPFRC mixtures were ready, for each batch of concrete, they were poured into six cylindrical moulds having a diameter of 100 mm and a height of 200 mm without any vibrations. For six concrete cylinders of each concrete batch, three cylinders were used for uniaxial compression test and three remaining cylinders were used for splitting tests. A total of fifty-four concrete cylinders were fabricated for testing. All test specimens were covered by plastic sheets immediately after casting in order to prevent moisture loss. The test specimens were demolded 48 h after casting and cured at ambient temperature until testing day. It should be noted that all concrete cylinders were cast and tested for 28-day compressive strength and splitting tensile strength. Figure 1 shows the slump flow of UHPFRCSF2% and the cylinders for casting and concrete cylinders after demolding.

a) Slump flow (UHPFRC-SF2%)

b) Cylinders for casting concrere

c) Concrete cylinders of 100x200 mm for testing

Fig. 1. Specimen preparation for splitting tests

1152

A. H. Le Table 1. Composition of UHPC and UHPFRC mixes

Mix composition Unit Water kg/m3 CEM I 52.5R HS-NA kg/m3 Silica fume kg/m3 Super plasticizer Sika Viscorete 2810 kg/m3 Ground Quartz W12 kg/m3 Quartz sand 0.125/0.5 kg/m3 Steel fibers kg/m3

2.2

UHPC 188.0 795.4 168.6 24.1 198.4 971.0 –

UHPFRC-SF1% 186.0 787.5 166.9 23.7 196.4 961.3 78.5

UHPFRC-SF2% 184.0 779.5 165.2 23.6 194.4 951.6 157.0

Testing Procedure

The flowability of the fresh concrete for each mix was checked immediately after mixing, using a mini-slump cone on a flow table in accordance with DIN EN 12350-8: 2010-12 [9]. The slump-flow test includes two governing parameters: the slump flow and the flow time t500. The minimum value of slump flow is higher than 800 mm, indicating that the concrete mixtures in this study were fully self-compacting even with the use of steel fibers 2% by volume (see Fig. 1a). The compressive strengths fc and elastic modulus Ec were determined from the compression tests on 3 cylindrical specimens of 100
200 mm for each batch of concrete in accordance with DIN EN 12390-3:2009-07 [10] and DIN 1045-1 [8], respectively. Prior to the compression tests, the two ends of each concrete cylinder were ground using a grinding wheel so that the two ends were parallel and the load was transferred uniformly to the cross section. All cylinders were tested uniformly force-controlled using a 4000 kN capacity compression machine. In addition, elastic modulus was measured using a compress meter installed at the mid-height of concrete cylinder measuring the average compressive strain. Splitting tensile strength tests were executed on three concrete cylinders of 100
200 mm for each batch of concrete according to DIN EN 12390-6:2009 [11] to determine the average splitting strength, as shown in Figure. The splitting test was performed by applying compressive line loads along two opposing lengths on the side of the concrete cylinder (see Fig. 2a). Two timber strips having the dimensions of 300 mm in length, 10 mm in width, and 5 mm in thickness were located between the loading plates and the specimen surfaces along the full length of the specimen as bearing strips (see Fig. 2a). A conventional compression testing machine at a loading rate of 0.01 mm/s was employed for the splitting test. The splitting tensile strength was calculated using the following equation: fspl ¼

2P pLD

ð1Þ

where fspl is the splitting strength (MPa), P is the maximum applied load (kN), L is the length of the concrete cylinder (mm), and D is the diameter of the concrete cylinder (mm). Figure 2 shows the splitting test setup.

Evaluation of the Splitting Tensile Strength

1153

Table 2. UHPC and UHPFRC mixing plan Process Time Material filling – Dry mixing (2 min) 0 min – 2 min Water and superplasticizer (3 min) 2 min – 5 min Break (2 min) 5 min – 7 min Mixing 7 min – 15 min Steel fiber filling 13 min Break (10 min) 15 min – 25 min Final mixing 25 min – 25.5 min

UHPC

UHPFRC-SF1% a) Splitting test setup UHPFRC-SF2% b) Failure modes Fig. 2. Splitting test on concrete cylinders 100
200 mm and failure modes

3 Test Results and Discussions Table 3 presents the average compressive strength (fc) and the average elastic modulus (Ec) derived from the compression tests on three concrete cylinders for each concrete cast. Also, the average splitting tensile strength fspl measured from the splitting test on three concrete cylinders for each concrete cast is tabulated in Table 3.

1154

A. H. Le Table 3. Results of compression tests and splitting tests

Series Mix

1

Cast Cast Cast Cast Cast Cast Cast Cast Cast

2

3

a)

1 2 3 4 5 6 7 8 9

Steel fiber volume Vf (%)

Average elastic Average Average splitting compressive strength modulus Ec (MPa) tensile strength fspl (MPa) fc (MPa)

0 1 2 0 1 2 0 1 2

178.9 195.5 188.2 198.0 195.5 187.8 190.4 195.6 192.4

The average splitting tensile strength corresponding with fiber volume

48370 49645 48421 46937 47881 48580 46186 48689 48557

b)

3.4 11.2 19.5 4.4 14.2 19.5 5.0 10.8 14.0

The splitting tensile strength of all concrete batches corresponding with fiber volume

Fig. 3. The effect of steel fiber volume on the splitting tensile strength

The typical failure modes of concrete cylinders in splitting tests were observed after testing and shown in Fig. 2(b). The concrete cylinders without steel fibers (UHPC) experienced one failure surface along the line of the loading strip. The concrete specimens without steel fibers failed in brittle manner with a sudden failure when all specimens were totally spitted into two parts. However, the concrete cylinders with steel fibers (UHPFRC-SF1% and UHPFRC-SF2%) showed a ductile failure mode. All specimens using steel fibers remained intact with a small crack which appeared along the middle line of the concrete section. The crack of specimens using higher volume of steel fibers (2%) was observed to be shorter than that of specimens using lower volume of steel fibers (1%). The ductile failure mode is attributed to the influence of steel fibers. Steel fibers provide the distributed stresses along the failure surface, thus preventing the complete splitting failure. Table 3 shows that presence of steel fibers significantly increased the splitting tensile strength as compared to that in the case of no fibers. The higher volume of steel fibers also resulted in higher splitting tensile strength. In this study, UHPC had the splitting tensile strength varying between 3.4 and 5.0 MPa, while the spitting tensile strength was found to be between 10.8 and 14.2 MPa for UHPFRC-SF1%, and between 14.0 and 19.5 MPa for UHPFRC-SF2%.

Evaluation of the Splitting Tensile Strength

1155

Figure 3 shows the effect of steel fiber volume on the splitting tensile strength. The average splitting tensile strength of UHPFRC-SF1% and UHPFRC-SF2% were increased by 205.74% and 348.20% as compared to UHPC. Similarly, the average splitting tensile strength of UHPFRC-SF2% was increased by 46.59% as compared to UHPFRC-SF1%. The highest splitting tensile strength was achieved by UHPFRC-SF2% (cast 3 and cast 6) which had the highest steel fiber volume of 2%. This observation indicates that the splitting tensile strength increases significantly with increasing steel fiber volume from 0% to 1% and from 0% to 2%, while the increase rate in the splitting tensile strength is lesser with steel fiber volume between 1% and 2%. It was observed from Fig. 3(b) that there was a large scatter of the splitting strength from batch to batch for each fiber volume. This is due to the different distribution and dispersion of fibers when casting specimens in vertical direction without vibration.

4 Evaluation of Existing Models Most of the international design codes introduce equations based on the compressive strength (fc) to predict the splitting tensile strength (fspl). However, these codes do not take into account the effect of fibers. Previous researchers attempted to propose models considering the effect of fibers including fiber volume (Vf), aspect ratio (lf/df), fiber type to estimate the splitting tensile strength.

Table 4. Equations of existing models Authors

Equations

Explanations

Wafa and Ashour (1992) [29]

fspl ¼ 0:58ðfc Þ0:5 þ 3:02Vf2

fc  100 MPa

Song and Hwang (2004) [27]

fspl ¼ 0:63ðfc Þ0:5 þ 3:01Vf  0:02Vf2
 l pffiffiffiffi fc fspl ¼ 0:6 þ 0:4Vf dff

fc  100 MPa

Musmar (2013) [31]

4 MPa  fc  120 MPa

Thomas and Ramaswamy (2007) [28] fspl ¼ 0:63ðfc Þ0:5 þ 0:288Vf lf f 0:5  0:02V 2 fc  85 MPa f df c h i pffiffiffiffi
El-Din et al. (2016) [12] 20 F ¼ dlf Vf bf , where bf is the bond fspl ¼ 0:076 fc þ 10 F 3  fc f

Narayanan and Darwish (1987) [19]

fspl ¼ 20

fc ffiffiffiffiffiffi q þ 0:7 þ l Vf df f

qffiffiffiffiffiffiffiffiffi l Vf dff

factor depending on the type of fibers 125 MPa  fc  155 MPa N/A

Arioglu et al. (2006) [5]

fspl ¼ 0:321fc0:66

4 MPa  fc  120 MPa

ACI 318-14 (ACI, 2014) [1]

fspl ¼ 0:321fc0:66

21 MPa  fc  69 MPa

SNZ 3101 (SNZ, 2006) [25]

fspl ¼ 0:36fc0:5   fspl ¼ 2:12ln 1 þ   fspl ¼ 2:12ln 1 þ

DIN 1045-1 (DIN, 2008) [8] Mode Code 2010 (MC2010, FIB 2013) [14] JSCE (JSCE, 2007) [18]

fspl ¼ 0:23fc2=3



fc 10 =0:9  fc 10

25 MPa  fc  100 MPa fc  67 MPa 50 MPa  fc  120 MPa 20 MPa  fc  80 MPa

1156

A. H. Le

It was found that the compressive strength in all existing models and design codes is limited up to around 120 MPa and usually applicable to normal or high strength concrete (NSC and HSC). Therefore, there is a need to evaluate the suitability of the existing models and design codes for UHPC and UHPFRC. In this study, UHPC and UHPFRC are defined as concrete having compressive strength of cylinders (100
200 mm) higher than 120 MPa. Table 4 shows the equations of seven existing models and five codes collected in the literature, which are used for the evaluation. The results of splitting tests on the 165 cylinders of 100
200 mm obtained from previous works and this study were used. Table 5 summarizes the content of database of collected splitting tests. The compressive strengths were in the range 124.0216.52 MPa, and the steel fiber volumes varied between 0 and 6%. Table 6 shows the results of statistical analysis of the predictions from existing models and design codes. The statistical measures include mean, standard deviation (SD), coefficient of variation (COV), and integral absolute error (IAE), which were calculated by the ratio of predicted splitting tensile strength (fspl,pre) to test result (fspl,test). It should be noted that the value of IAE is determined using the following equation in Bae et al. (2017):
IAE ¼

X

2 0:5

fspl;testi  fspl;prei P fspl;prei
100

ð2Þ

It was revealed from Table 6 that the models proposed by Wafa and Ashour (1992), Thomas and Ramaswamy (2007), Narayanan and Darwish (1987), Arioglu et al. (2006) underestimated the splitting tensile strength, however the mean value of fspl,pre/ fspl,test obtained from each model was close to unity. Furthermore, all design codes gave an remarkable underestimation with low mean value of fspl,pre/ fspl,test. The models suggested by Musmar (2013), and Song and Hwang (2004) slightly overestimated the splitting tensile strength, while the model of El-Din et al. (2016) gave an overestimation with high mean value of 1.19. In terms of statistical variables, all models and

Table 5. Collected database of previous splitting tests Authors

Fc (MPa)

Fspl (MPa)

Vf (%)

Lf/df (mm/mm)

Steel fiber type

Number of specimens

Wang et al. (2018) [30] Shafieifar et al. (2017) [23] Pansuk et al. (2017) [22] Nehdi et al. (2015) [20] El-Din et al. (2016) [12] Othman (2016) [21] Shin (2017) [24] Bae et al. (2017) [7] Current study

129–185

7.4–17.1

0–3

13/0.2

Straight

12

138

20.7

2

12.5/0.2

Straight

3

136.81–147.69

10.38–16.27

0–1.6

13/0.2

Straight

18

151–173

9.4–39.8

1–6

8/0.2;12/0.2; 16/0.2

Straight

30

124–154

11.5–19.1

0–3

30/1; 50/1

21

151.9–174.1 184.8 149.4–216.52 178.9–198.0

6.8–15.3 12.9 9.01–11.96 3.38–19.53

1–3 2 0.5–2 0–2

13//0.2 13/0.2 13/0.2 13/0.2

Hooked ended Straight Straight Straight Straight

27 3 24 27

Evaluation of the Splitting Tensile Strength

1157

Table 6. Results of statistical analysis Authors

Standard deviation Coefficient of variation Mean (COV) (fspl,pre/fspl,test) (SD)

IAE

Wafa and Ashour (1992) [29] Song and Hwang (2004) [27] Musmar (2013) [31] Thomas and Ramaswamy (2007) [28] El-Din et al. (2016) [12] Narayanan and Darwish (1987) [19] Arioglu et al. (2006) [5] ACI 318-14 (ACI, 2014) [1] SNZ 3101 (SNZ, 2006) [25] DIN 1045-1 (DIN, 2008) [8] Mode Code 2010 (MC2010, FIB 2013) [14] JSCE (JSCE, 2007) [18]

0.98 1.02 1.01 0.84

0.41 0.44 0.44 0.45

0.43 0.43 0.44 0.54

16.25 15.97 17.38 21.02

1.19 0.87

0.70 0.49

0.59 0.57

23.00 20.48

0.81 0.62 0.40 0.58 0.52

0.50 0.38 0.24 0.34 0.31

0.62 0.60 0.60 0.59 0.59

21.88 26.65 34.57 27.78 29.90

0.60

0.37

0.62

27.61

design codes exhibited relatively high values of SD and COV, especially in the case of design codes. It is mentioned that the design codes do not consider the effect of fibers on the splitting tensile strength, thereby resulting in the low accuracy of the prediction as compared with the test results. Likewise, in terms of IAE, an increasing tendency in IAE appeared with design codes, while the predictions from existing models had lower values of IAE. Among existing models, the model of Wafa and Ashour (1992) was the most suitable approach to predict the splitting tensile strength because this models gave a safe prediction with low values of SD, COV and IAE as compared with the remaining models. It is also recommended that the effect of fibers should be considered in the design codes to obtain more accurate prediction of the splitting strength for UHPFRC.

5 Proposed Equation for Predicting the Splitting Tensile Strength of UHPC and UHPFRC The collected database of splitting tests in Table 5 was used for the regression analysis. The approach of El-Din et al. (2016) [12] was adopted for the proposed equation of the splitting tensile strength. In El-Din et al. (2016) [12], the splitting tensile strength is a function of the compressive strength fc and a variable (F) which can reflect the effect of steel fiber comprehensively. This variable (F) is given as follows: F ¼ Vf


lf
bf df

ð3Þ

where Vf is the fiber volume, lf/df is the fiber aspect ratio, and bf is the bond factor (bf = 0.5 for fibers having circular section, bf = 0.75 for hooked or crimped fibers).

1158

A. H. Le

Fig. 4. Regression analysis on the collected database

 ffi Deriving from the regression analysis of the relation between the ratio fspl pffiffiffi fc and F in Fig. 4, the equation for predicting the splitting tensile strength of UHPC and UHPFRC can be proposed as below: fspl



pffiffiffiffi lf fc ¼ 0:94Vf

bf þ 0:67 df

ð4Þ

It is mentioned that the above proposed equation is applicable for UHPC and UHPFRC having compressive strength from 120 to 200 MPa

6 Conclusions Some conclusions and recommendations can be drawn from this study as follows: • In this experimental tests, UHPC had average splitting tensile strength between 3.4 and 5.0 MPa, while UHPFRC had average splitting tensile strength varying between 10.8 and 14.2 for 1% steel fibers by volume and between 14.0 and 19.5 MPa for 2% steel fibers by volume; • The inclusion of steel fibers significantly increases the splitting tensile strength. Higher volume of steel fibers results in higher splitting tensile strength; • The failure mode of concrete cylinders under splitting load becomes more ductile with the use of steel fibers; • All selected design codes underestimated the splitting tensile strength with low accuracy. The effect of steel fibers should be considered in these codes; • Among existing models, the model of Wafa and Ashour (1992) was found to be the best choice for predicting the splitting tensile strength; • A simplified equation for estimating the splitting tensile strength of UHPC and UHPFRC having compressive strength between 120 and 200 MPa was proposed with considering the effect of steel fibers. More test results of splitting strengths should be collected get better regression analysis. • Further studies on splitting tensile strength of UHPC and UHPFRC using various fiber types, fiber aspect ratios, fiber volumes should be conducted.

Evaluation of the Splitting Tensile Strength

1159

Acknowledgements. This research is funded by Vietnam National Foundation for Science and Technology Development (NAFOSTED) under grant number 107.01-2019.325.

References 1. ACI, Building code requirements for structural concrete (ACI 318–14) and commentary. In: ACI, 318-14 Farmington Hills, MI, USA (2014) 2. An, L.H., Fehling, E.: Influence of steel fiber content and type on the uniaxial tensile and compressive behavior of UHPC. Constr. Build. Mater. 153, 790–806 (2017) 3. An, L.H., Fehling, E., Binglin, L., Kien, T.D., Chau, V.N.: Experimental study on structural performance of UHPC and UHPFRC columns confined with steel tube. Eng. Struct. 187, 457–477 (2019) 4. An, L.H., Fehling, E., Kien, T.D., Chau, V.N.: Evaluation of axial strength in circular STCC columns using UHPC and UHPFRC. J. Constr. Steel Res. 153, 533–549 (2019) 5. Arioglu, N., Girgin, Z.C., Arioglu, E.: Evaluation of ratio between splitting tensile strength and compressive strength for concretes up to 120 MPa and its application in strength criterion. ACI Mater. J. 103(1), 18–24 (2006) 6. ASTM, C.: 496–90 Test method for splitting tensile strength of cylindrical concrete specimens. In: Annual Book of ASTM Standards ASTM International, West Conshohocken (1990) 7. Bae, B.L., Lee, M.S., Choi, H.K., Choi, C.S.: Indirect tensile strength of UHSC reinforced with steel fibres and its correlation with compressive strength. Mag. Concrete Res. 69(15), 772–786 (2017) 8. DIN (Deutsches Institut für Normung) DIN 1045–1, Plain, Reinforced and Prestressed Concrete Structures – Part 1: Design and Construction. Beuth Verlag, Berlin, Germany (2008) 9. DIN EN 12350–8:2010-12, Testing fresh concrete-Part 8: Self-compacting concrete-Slumpflow test. German version EN 12350-8, Beuth Verlag, Berlin (2010) 10. DIN EN 12390–3:2009-7, Testing hardened concrete-Part 3: Compressive strength of test specimens, German version EN 12390-3, Beuth Verlag, Berlin (2009) 11. DIN EN 12390–6, Testing Hardened concrete – Part 6: Tensile splitting strength of test specimens. Beuth Verlag, Berlin (2009) 12. El-Din, H.K.S., Heba, A.M., Khater, M.A.E., Sayed, A.: Effect of steel fibers on behavior of ultra high performance concrete. In: First International Interactive Symposium on UHPC, USA (2016) 13. El-Helou, R.G., Moen, C.D., Cusatis G.: Ultra-High performance fiber-reinforced concrete: extensive material characterization, model validation, and structural simulations. Presentation at ACI Fall 2014 Convention Washington, DC (2014) 14. Fib (Fédération Internationale du Béton), Model Code for Concrete Structures 2010. Fib. Lausanne, Switzerland (2013) 15. Goaiz, H.A., Farhan, N.A., Sheikh, M.N., Yu, T., Hadi, M.N.S.: Experimental evaluation of tensile strength test methods for steel fibre reinforced concrete. Mag. Concrete Res. 71(8), 1– 42 (2018) 16. Goaiz, H.A., Tao, Y., Hadi, M.S.N.: Quality evaluation tests for tensile strength of reactive powder concrete. J. Mater. Civ. Eng. 30(5), 04018070 (2018) 17. Graybeal, B.A.: Practical means for determination of the tensile behavior of ultra-high performance concrete. J. ASTM Int. 3(8), 1–9 (2006)

1160

A. H. Le

18. JSCE (Japan Society of Civil Engineers), Standard Specifications for Concrete Structures – Design, JSCE Guidelines for Concrete No. 15 Tokyo, Japan, (2007) 19. Narayanan, R., Darwish, I.Y.S.: Use of steel fibers as shear reinforcement. ACI Struct. J. 84(3), 216–227 (1987) 20. Nehdi, M.L., Abbas, S., Soliman, A.M.: Exploratory study of ultra-high performance fiber reinforced concrete tunnel lining segments with varying steel fiber lengths and dosages. Eng. Struct. 101, 733–742 (2015) 21. Othman, H.A.B.: Performance of Ultra-High Performance Fibre Reinforced Concrete Plates under Impact Loads. PhD dissertation (Ryerson University, 2016) 22. Pansuk, W., Thuc, N.N., Yasushiko, S., Uijl, J.A.D., Walraven, J.C.: Shear capacity of high performance fiber reinforced concrete I-beams. Constr. Build. Mater. 157, 182–193 (2017) 23. Shafieifar, M., Farzad, M., Azizinamini, A.: Experimental and numerical study on mechanical properties of Ultra High Performance Concrete (UHPC). Constr. Build. Mater. 156, 402–411 (2017) 24. Shin, J.: Ultra-High Performance Concrete (UHPC) Precast Segmental Bridges: Flexural Behaviour and Joint Design. PhD dissertation University of Hamburg (TUHH) (2017) 25. SNZ (Standards New Zealand) NZS 3101:2006, Concrete structures standard’, SNZ Wellington, New Zealand, (2006) 26. Soner, G., Demet, Y., Fuat, K., Ashraf, A.: Strength prediction models for steel, synthetic, and hybrid fiber reinforced concretes. Struct. Concrete 20(1), 1–18 (2018) 27. Song, P.S., Hwang, S.: Mechanical properties of high-strength steel fiber-reinforced concrete. Constr. Build. Mater. 18(9), 669–673 (2004) 28. Thomas, J., Ramaswamy, A.: Mechanical properties of steel fiber-reinforced concrete. J. Mater. Civ. Eng. 19(5), 385–392 (2007) 29. Wafa, F.F., Ashour, S.A.: Mechanical properties of high-strength fiber reinforced concrete. ACI Mater. J. 89(4), 449–455 (1992) 30. Wang, R., Gao, X., Li, Q., Yang, Y.: Influence of splitting load on transport properties of ultra-high performance concrete. Constr. Build. Mater. 171, 708–718 (2018) 31. Musmar, M.: Tensile strength of steel fiber reinforced concrete. Contemp. Eng. Sci. 6(5), 225–237 (2013)

Author Index

A Abbas, Ali A., 359, 596, 730, 920 Abellán-García, Joaquín, 570, 864 Abrishambaf, Amin, 1056 Agra, Ronney Rodrigues, 233 Al Marahla, Razan H., 301, 392 Alberti, Marcos G., 693 Al-Naimi, Hasanain K., 359, 730, 920 Alsabbagh, Ahmed, 111 Al-Tabbaa, Abir, 466 Altoubat, Salah, 883 Ambrosini, D., 536 Amin, Ali, 368 Andrade, Carmen, 417 Arangjelovski, Toni, 380 Arango-Campo, Samuel E., 864 Ayoubi, Mazen, 221

B Bakhshi, Mehdi, 621, 815 Bakhshi, Mohammad, 703 Baloch, Hassan, 3 Bandelt, Matthew J., 1042 Barros, Joaquim A. O., 49, 703 Bauwens, Thomas, 64 Bernard, E. Stefan, 37 Berrocal, Carlos G., 477 Biermann, Dirk, 801 Billington, Sarah, 1042 Blanco, A., 610 Blazy, Julia, 209 Bogart, Kurt, 140 Bokern, Jürgen, 402 Boshoff, William P., 199

Breitenbücher, Rolf, 176 Burk, Sarah, 1022 C Caballero-Jorna, Marta, 322 Camille, Christophe, 548, 717 Capuzzo, V. M. S., 99 Cardoso, M. G., 99 Carmona, Sergio, 253 Cavalaro, S. H. P., 610 Chan, Ricardo, 897 Charron, J.-P., 453 Chen, E., 477 Cheung, Andrés B., 75, 745 Chocca, Claudia, 938 Clarke, Todd, 140, 548, 717 Cleven, Simon, 24, 402 Codina, R., 536 Combrinck, Riaan, 199 Conforti, Antonio, 423 Constantinescu, Horia, 433 Corinaldesi, Valeria, 1068 Curosu, Iurie, 1022 D da Silva Oliveira, Kaio Cézar, 270 da Silva Ramos Barboza, Aline, 270 Dai, Jian-Guo, 1034 de Carvalho, Isadora Queiroz Freire, 270 de Figueiredo, Antonio Domingues, 233 de la Fuente, A., 610 de Melo Lameiras, Rodrigo, 279 de Morais Silva, Wandersson Bruno Alcides, 270 De Schutter, Geert, 64

© RILEM 2021 P. Serna et al. (Eds.): BEFIB 2020, RILEM Bookseries 30, pp. 1161–1164, 2021. https://doi.org/10.1007/978-3-030-58482-5

1162

Author Index

De Smedt, Maure, 151 De Wilder, Kristof, 151 Dehn, Frank, 779 Desmettre, C., 453 Destrée, Xavier, 841 di Prisco, Marco, 347 Dias, Gabriela Silva, 270 Dias, Salvador, 49 Dittel, Gozdem, 991 Dlouhý, L., 681 Donnini, Jacopo, 1068 Doostkami, Hesam, 489 dos Santos, Ana Carolina Parapinski, 279 Dreier, Julia, 801

H Hafezolghorani, Milad, 527 Hamou, Arezki T., 1124 Heek, Peter, 908 Heghes, Bogdan, 433, 1090 Hewage, Dayani Kahagala, 548, 717 Hisseine, Ousmane A., 1124 Huang, Bo-Tao, 1034 Hunger, Martin, 402

E Enfedaque, Alejandro, 693 Ercegovič, Rok, 290 Estephane, Pierre, 883

J Jain, Kranti, 791

F Fataar, Humaira, 199 Fernández, María E., 938 Fernández-Gómez, Jaime A., 570 Ferrara, Giuseppe, 983 Ferrara, Liberato, 779 Fiengo, F., 536 Figueredo, Diego, 873 Filho, Christiano Augusto Ferrario Várady, 270 Fiorio, B., 971 Formagini, Sidiclei, 75, 745 Freitas, Danilo José Pereira, 270 G Gallias, J. -L., 971 Galobardes, Isaac, 897 Gálvez, Jaime C., 693 Garbeth, Michael R., 757 García, Nicolás, 873 Garcia-Taengua, Emilio, 301, 392, 827 Gebhard, Lukas, 87 Generosi, Nicola, 1068 Genikomsou, Aikaterini S., 651 Gettu, Ravindra, 770 Gherman, Oana, 433 Giaccio, Graciela M., 189, 423, 536 Giraldo Soto, Alejandro, 163 Girardello, P., 852 Glynn, Brian, 757 Gries, Thomas, 949, 991 Grünewald, Steffen, 3, 64, 779 Guzlena, S., 262

I Ian Gilbert, R., 368 Isla, F., 536

K Karzad, Abdul Saboor, 883 Kaufmann, Walter, 87, 163, 368 Khorami, M., 1112 Kim, Sungwook, 123 Kimm, Magdalena, 949 Kirkland, Brendan, 140, 548, 717 Kitagawa, Hirokazu, 1003, 1012 Kopálová, M., 681 Krasnikovs, Andrejs, 841 Kunieda, Minoru, 445 L Lameiras, R. M., 99 Lauch, K.-S., 453 Le, An Hoang, 1149 Leblouba, Moussa, 883 Lee, Namkon, 123 Lehký, David, 527 Leite, João, 49 Lesage, Karel, 3 Leung, Christopher K. Y., 1034, 1100 Li, Jiabin, 333 Li, Jianchun, 661 Li, Zhenghao, 1100 Liebscher, Marco, 1022 Lipowczan, Martin, 527 Litina, Chrysoula, 466 Löfgren, Ingemar, 477 Look, Katharina, 908 Lu, Cong, 1100 Luccioni, B., 536 Lundgren, Karin, 477

Author Index M Maalej, Mohamed, 883 Magalhães, Margareth S., 313 Mahrenholtz, Christoph, 221 Maiworm, Bastian, 991 Makita, Tohru, 1003, 1012 Marchand, P., 1137 Mark, Peter, 176, 380, 908 Markić, Tomislav, 87 Markovski, Goran, 380 Martinelli, Enzo, 983 Mashiri, Fidelis, 548, 717 Mata-Falcón, Jaime, 87 Matthys, Stijn, 3 Mechtcherine, Viktor, 1022 Medina, Néstor Fabián Acosta, 279 Menna, Demewoz W., 651 Merta, Ildiko, 245 Mezquida-Alcaraz, Eduardo J., 489, 639 Mirza, Olivia, 140, 548, 717 Mobasher, Barzin, 815, 963 Molins, Climent, 253 Moy, Charles K. S., 897 N Naaman, Antoine E., 1079 Najari, A., 610 Nakov, Darko, 380 Nasri, Verya, 621, 815 Navarro-Gregori, Juan, 639, 670, 1112 Negi, Bichitra S., 791 Negrini, Alberto, 489 Negrutiu, Camelia, 433, 1090 Nell, Wilhelm, 221 Nemeth, Gergely, 245 Novák, Drahomír, 527 Nunes, Sandra, 209, 1056 Núñez-López, Andrés M., 570, 864 O O’Flaherty, Tomas, 558 Oliveira, T. T, 99 Ortiz-Navas, Francisco, 670 P Park, Gijoon, 123 Patel, Devansh, 815 Pepe, Marco, 983 Pereira Lima, Iva E., 584 Pereira, María E., 938 Petrone, Fernando, 938 Picazo, Álvaro, 693 Pimentel, Mário, 209, 1056 Pires, Ana R. L., 75, 745

1163 Pleesudjai, Chidchanok, 815 Plückelmann, Sven, 176 Pokhrel, Mandeep, 1042 Polanec, David, 290 Poletanovic, Bojan, 245 Polvere, Rafael R., 75, 745 Příbramský, V., 681 Pukl, Radomír, 527 R Ramezansefat, Honeyeh, 703 Ramos Barboza, Aline S., 584 Rezazadeh, Mohammadali, 703 Riva, P., 852 Rodríguez, Gemma, 938 Rodríguez, Iliana, 873 Roig-Flores, Marta, 322, 489 Rossi, Pierre, 515 Roy, N., 1137 S Sabir, Amna, 949 Saha, Suman, 49 Sakale, G., 262 Sanjuán, Miguel A., 417 Santana, F. B., 99 Sedran, Thierry, 515 Segura-Castillo, Luis, 873 Serafini, Ramoel, 233 Serna, Pedro, 322, 489, 639, 670, 1112 Sevigny-Vallières, C., 1137 Shahrbijari, Kamyar Bagherinejad, 49 Shao, Yi, 1042 Sirtoli, D., 852 Slama, A. -C., 971 Smarslik, Mario, 176 Sosa, Ioan, 433, 1090 Sousa, Carlos, 209 Soyemi, Olugbenga B., 596 Spyridis, Panagiotis, 801 Stephen, Stefie J., 770 Suksawang, Nakin, 12, 111 Šušteršič, Jakob, 290 Suwanpinij, Piyada, 949 T Tagnit-Hamou, A., 1137 Tailhan, Jean-Louis, 515 Talavera-Sánchez, Santiago, 670 Tang, Zixuan, 466 Tavares, Maria Elizabeth N., 313 Teixeira, Paulo José B., 313 Terrade, B., 1137 Todor, Adel, 1090

1164 Tolêdo Filho, Romildo D., 983 Torkaman, Javad, 133 Torres-Castellanos, Nancy, 570 Torrijos, María C., 189, 423, 536 Toutlemonde, F., 1137 Tran Thanh, Hai, 661 Tri, Le V., 445 Tsutsui, Masaki, 445 V Valente, Isabel B., 49, 703 van den Bos, A. A., 503 van der Aa, P. J., 503 Vandevyvere, Brecht, 333 Vandewalle, Lucie, 151, 333, 347 Verstrynge, Els, 151, 333 Vivas, Juan C., 189, 536 Vlasák, Oldrich, 24 Vrijdaghs, Rutger, 151, 347, 402 W Walter, Lars, 801 Wang, Jingquan, 963

Author Index Wangler, Michelle, 991 Watanabe, Yuji, 1003, 1012 Watts, Murray, 368 Willrich, Fábio Luiz, 279 Wilson, William., 558 Winterberg, Ralf, 757 Wolf, Sébastien, 24, 841 Wtaife, Salam, 111 Wu, Jia-Qi, 1034 Y Yanai, Shuji, 1003, 1012 Yao, Yiming, 963 Yohannes, Daniel, 12 Yu, Jing, 1034 Z Zajc, Andrej, 290 Zerbino, Raúl L., 189, 423, 536 Zhai, Mengchao, 963 Zhang, Y. X., 661 Zhou, Jiajia, 1100