Fiber-Reinforced Polymer: Reinforcement for Concrete Structures 9812384014, 9789812384010, 9789812704863

Citation preview

Reinforcement for

Concrete Structures VOLUME

1

zyx

This page intentionally left blank

Edited by

zyxw zy zy zyx zy Kimg Hwee TAN

National University of Singapore, Singapore

Singapore

8-1 0 July, 2003

Reinfo rc e me nt for Concrete Structures

Proceedings of the

VOLUME

Sixth International Symposium on FRP Reinforcement for Concrete Structures

1

(FRPRCS-6)

zyxw

r heWorld Scientific

NewJersey London Singapore Hong Kong

Preface Research on the application of fibre-reinforced polymer (FRP) as reinforcement for concrete structures appeared in as early as the 1960s. However, it was not until the late 1980s that such research has escalated, leading to field applications. The interest in non-metallic reinforcement was fuelled by the corrosion problem associated with steel reinforcement that surfaced around the world at that time, and the downturn of the aerospace industry, where fibre-reinforced polymers have been widely used due to its high specific strength and modulus, and other superior characteristics. I was fortunate to spend my sabbatical with Professor Naaman at the University of Michigan, USA, during the Fall and Winter of 1991 and with Professor Okamura at the University of Tokyo, Japan, during Spring and Summer of 1992. The former introduced to me this new material that has since fascinated many in the research community and construction industry. In Tokyo, in particular, I was overwhelmed by the mountains of research that were embarked by universities, public institutions and private companies on the development and application of FRP rods as reinforcement for concrete structures. There were round bars, flat bars, square bars, braided bars, sanded bars, strands, grids and links, and even three-dimensional reinforcement. Several applications in footbridges, foundation beams, tunnel linings, and floating structures suddenly mushroomed all over Japan and the rest of the world. That probably constituted the first era in the application of FRP reinforcement in concrete structures. The FRPRCS Symposia Series was initiated in 1993, and subsequently held every two years in the continents of America, Europe and Asia, on a rotational basis. The previous symposia were held in Vancouver, Canada (1993), Ghent, Belgium (1995), Sapporo, Japan (1997), Baltimore, USA (1999), and Cambridge, UK (2001). This year marks the 10th anniversary of the FRPRCS Symposia Series, and the Department of Civil Engineering at the National University of Singapore is honored to host the 6th International Symposium on FRP Reinforcement for Concrete Structures (FRPRCS-6) in Singapore.

zyx zyxw V

zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQ

VI

The planning and preparation works for FRPRCS-6 in effect began almost six years ago in 1997 when I was asked in Sapporo, to be the second Asian host for the FRPRCS International Symposium. At that time, there was still little awareness of the material known as FRP reinforcement in Asia outside Japan, and if any, the interests were centered mainly on externally bonded FRP systems rather than FRP reinforcing rods. The Kobe earthquake in 1995 has brought about rapidly increasing interests in the use of FRP systems in structural rehabilitation, and that marked the beginning of the second era in FRP applications in concrete structures. To promote awareness and interests in the development and application of FRP reinforcement in Singapore and the region, the FibreReinforced Society (Singapore) was formed in September 2002 and has since been a co-organizer of this Symposium The FRPRCS-6 International Symposium will signify the beginning of the third era, in which one could witness global interests in FRP reinforcement, as well as the use of FRP reinforcements as structural shapes, and in masonry and steel structures. This set of proceedings contains a total of 140 papers from 26 countries, in two volumes. Each technical paper had been reviewed and selected for presentation by at least two members of the International Scientific Committee, to whom I would like to express my gratitude. Volume 1 of the proceedings contains four invited keynote papers and 63 technical papers dealing with: (i) FRP Materials and Properties; (ii) Bond Behaviour; (iii) Externally Bonded Reinforcement (EBR) for Flexure, Shear and Confinement; and (iv) FRP Structural Shapes. The topics covered in Volume 2 are: (v) Durability and Maintenance; (vi) Sustained and Fatigue Loads; (vii) Prestressed FRP Reinforcement and Tendons; (viii) Structural Strengthening; (ix) Applications in Masonry and Steel Structures; (x) Field Applications and Case Studies; and (xi) Codes and Standards. Seventy-three papers are included in Volume 2. The FRPRCS-6 International Symposium also witnessed the formation of the International Steering Committee, which comprises the chairmen of the current and previous FRPRCS Symposia. The main purpose of this Committee is to chart the future directions for the Symposia Series. It has appointed a three-man taskforce to determine the Best Paper (Research), Best Paper (Application) and Honorable Mention Awards, which were first introduced at FRPRCS-6. The three gentlemen in the taskforce were Professor C.W. Dolan from USA, Professor F.S.

zyx zyxwvu zyx vii

Rostdsy from Germany, and Professor H. Okamura from Japan. All of them are well known in the areas of FRP reinforcement and structural concrete. The organization of the Symposium would not have been possible without the generous contributions from the sponsors, who are degussaMBT (S) Pte Ltd, Fyfe Asia Pte Ltd, Mapei Far East R e Ltd, Sika (S) Pte Ltd, Lee Foundation (Singapore) and Defence Science & Technology Agency. I would also like to express my sincere thanks to the American Concrete Institute, USA, Institution of Engineers, Singapore, Japan Concrete Institute, Japan, and The Concrete Society, UK, for supporting the event. Last, but not least, I would like to acknowledge the help of my colleagues, in particular, Balendra, Mansur and Maalej, and the Secretariat, comprising Christine, Siti and Sarimah, who have devoted many hours in getting the Symposium organized.

Kiang Hwee Tan Singapore July 2003

zyxwv

FRPRCS-6 Organizing Committees

National University of Singapore Organizing Committee Chairman : K.H. Tan Members : T. Balendra, M.A. Mansur, M. Maalej Secretariat : C.S. Tan, Siti Rohani, Sarimah

International Steering Committee S.H. Rizkalla, USA L. Taerwe, Belgium K.H. Tan, Singapore T. Uomoto, Japan

C.J. Burgoyne, UK C.W. Dolan, USA A. Nanni, USA H. Okamura, Japan

International Scientific Committee

K.H. Tan, Singapore (Chairman) K.W. Neale, Canada L.C. Bank, USA K. Pilakoutas, UK B. Benmokrane, Canada S.H. Rizkalla, USA C.J. Burgoyne, UK J. Sim, Korea E. Cosenza, Italy R.N. Swamy, UK C.W. Dolan, USA L. Taerwe, Belgium G.B. Guimaraes, Brazil J.G. Teng, China M. Harajli, Lebanon R. Tepfers, Sweden P. Hamelin, France T. Ueda, Japan L. Hollaway, UK T. Uornoto, Japan G. Manfredi, Italy P. Waldron, UK K. Maruyama, Japan Z. Wu, Japan U. Meier, Switzerland Q.R. Yue, China A.E. Naaman, USA A. Nanni, USA

viii

zy zyxw Contents

I

I

VOLUME 1

KEYNOTE PAPERS

FRP Reinforcements in Structural Concrete: Assessment, Progress, and Prospects A.E. Naaman

3

Progress and Prospects of FRP Reinforcements: Survey of Expert Opinions A.E. Naaman

25

Durability Design of GFRP Rods for Concrete Reinforcement T. Uomoto

37

New Types of Continuous Fiber Reinforcements for Concrete Members T. Ueda

51

FRPMATERIALS AND PROPERTIES ~~

zyx

Performance of ThermoplasticFiber Reinforced Polymer Rebars A.B. Mehrabi, C.A. Ligozio, A.F. Elremaily and D.R. Vanderpool

79

Experimental Study on Poisson’s Ratio for FRP Tendons M. Tanaka, M. Khin, T. Harada and K. Venkataramana

89

Stress-Strain Model for FRP-Confined Concrete for Design Applications L. Lam and J.G. Teng

99

ix

zyxwv zyxw

X

Accelerated Techniques to Predict the Stress-Rupture Behaviour of Aramid Fibres (Best Paper - Research) K.G.N.C. Alwis and C.J. Burgoyne

111

BOND BEHAVIOUR Bond Characteristics of Various FRP Strengthening Techniques S.H. Rizkalla and T. Hassan

123

Bond Strength between Fiber-Reinforced Polymer Laminates and Concrete T. Kanakubo, T. Furuta and H. Fukuyama

133

Local Bond Stress-Slip Relations for FRP Sheets-Concrete Interfaces (Best Paper - Research) J.G. Dai and T. Ueda

143

-

Bilinear Stress Slip Bond Model: Theoretical Background and Significance T. Ulaga, T. Vogel and U. Meier

153

Non Linear Bond-Slip Law for FRP-Concrete Interface M. Savoia, B. Ferracuti and C. Mazzoti

163

Experimental Analysis of Interface between CFRP and Concrete using Cylindrical Specimens A.C. Dos Santos, T.N. Bittencourt and R. Gettu

173

FRP Adhesion in Uncracked and Cracked Concrete Zones G. Monti, M. Renzelli and P. Luciani

183

Neural Network Prediction of Plate End Debonding in FRPPlated RC Beams S.T. Smith, J.G. Teng and M. Lu

193

Bond Behaviour of CFRP Strips Glued into Slits M. Blaschko

205

XI

EXTERNALLY BONDED REINFORCEMENT FOR FLEXURE Load Capacity of Concrete Beams Strengthened with External FRP Sheets Z.J. Wu and J.M. Davies

217

Reinforcing Effects of CFRP and AFRP Sheets with Respect to Flexural Behaviour of RC Beams 0. Joh, Z. Wang and H. Ibe

227

Flexural Behaviour of RC Beams Externally Reinforced with Carbon Fiber Sheets Y. Takahashi and Y. Sat0

237

Strength and Failure Mechanism of RC T-Beams Strengthened with CFRP Plates K. Lee and R. Al-Mahaidi

247

Effect of Beam Size on Interfacial Shear Stresses and Failure Mode of FRP-Bonded Beams K.S. Leong and M. Maalej

257

Debonding Failure of RC Structural Members Strengthened with FRP Laminates G. Camata, E. Spacone and V. Saouma

267

Effect of End Wrapping on Peeling Behaviour of FRP-Strengthened Beams P. Pornpongsaroj and A. Pimanmas

277

An Experimental Study on Debond-Control of AFRP for Flexurally Strengthened RC Beams S. Sawada, N. Kishi, H. Mikami and Y. Kurihashi

287

Tests on RC T-Beams Strengthened in Flexure with a Glued and Bolted CFRP Laminate A. Nurchi, S. Matthys, L. Taerwe, M. Scarpa and J. Janssens

297

zyxwv

xii

zyxwv zyxwv

Parametric Studies of RC Beams Strengthened in Flexure with Externally Bonded FRP S . Limkatanyu, H. Thomsen, E. Spacone and G. Camata

307

Concrete Cover Failure or Tooth Type Failure in RC Beams Strengthened with FRP Laminates M.M. Lopez and A.E. Naaman

3 17

Influence of Material Properties of FRPs on Strength of Flexural Strengthened RC Beams G.F. Zhang, N. Kishi and H. Mikami

327

Ductility of Reinforced Concrete Beams Strengthened with CFRP Strips and Fabric M. Valcuende, J. Benlloch and C.J. Parra

337

A Review of Ductility Determination of FRP Strengthened Flexural RC Elements D.B. Tann, P. Davies and R. Delpak

347

A Semi-EmpiricalApproach for the Prediction of Deflections of FRP Strengthened RC Slabs D.B. Tann

357

Crack Widths in RC Beams Externally Bonded with CFRP Sheets Y. Zhang, H. Toutanji and P. Balaguru

367

Numerical Simulations for Strengthened Structures with Hybrid Fiber Sheets H. Niu and Z. Wu

377

Fibre-Section FE of FRP-Strengthened RC Beam in Flexure, Shear and Confinement G. Monti and M. Barbato

387

Interaction between Internal Bars and External FRP Reinforcement in RC Members G. Zehetmaier and K. Zilch

397

xiii

Strengthening of RC Two-way Slabs with Composite Materials 0. Limam, G. Foret and A. Ehrlacher

407

Evaluation of Externally Bonded CFRP Systems for the Strengthening of RC Slabs K.Y. Tan, J.G. Tumialan and A. Nanni

417

Flexural Strengthening of Two-way Slabs Using FRPs H. Marzouk, U.A. Ebead and K.W. Neale

427

Tensile Properties of Concrete in FRP Strengthened Two-way Slabs H. Marzouk, U.A. Ebead and K.W. Neale

437

EXTERNALLY BONDED REINFORCEMENT FOR SHEAR Shear Critical RC Beams Strengthened with CFRP Straps G. Kesse and J. M. Lees

447

Effective Shear Strengthening of Concrete Beams using FRP Sheets with Bonded Anchorage B.B. Adhikary, H. Mutsuyoshi and M. Ashraf

457

zyxw zyxw

Behaviour of Concrete Structures Strengthened in Shear with CFRP A. Carolin and B. Taljsten

467

Strengthening of RC T Beams in Shear with Carbon Sheet Laminates (CFRP) G.S. Melo, A.S. Aratijo and Y. Nagato

477

Strength Analysis of Sheared Beams Retrofitted with Strengthening Materials 2.H. Xiong and M.N.S. Hadi

487

Shear Performance with Externally Bonded Carbon Fibre Fabrics A. Li, C. Diagana, Y. Delmas and B. Gedalia

497

XIV

Evaluation of Shear Capacity of RC Columns Strengthened by Continuous Fiber T. Furuta, T. Kanakubo and H. Fukuyama Shear Design Equations for FRP RC Beams M. Guadagnini, K. Pilakoutas and P. Waldron

507

zyx 517

Strengtheningof Corrosion-DamagedRC Columns with FRP S.N. Bousias, T.C. Triantafillou, M.N. Fardis, L.A. Spathis and B. O’Regan

527

Shear Strengthening of Concrete Bridge Decks using FRP Bar P. Valerio and T.J. Ibell

539

EXTERNALLY BONDED REINFORCEMENTFOR CONFINEMENT

zy zyxwvu

Stress-Strain Relationship for FRP-Confined Concrete Cylinders G. Wu, Z. Lu and Z. Wu

55 1

Stress-Strain Relationship for FRP-Confined Concrete Prisms G . Wu, 2. Wu and Z. Lu

561

Concrete Cylinders Confined by CFRP Sheets Subjected to Cyclic Axial Compressive Load T. Rousakis, C.S. You, L. De Lorenzis, V. Tamuis and R. Tepfers

57 1

Concrete Cylinders Confined by Prestressed CFRP Filament Winding under Axial Compressive Load T. Rousakis, C.S. You, L. De Lorenzis, V. Tamuis and R. Tepfers

581

Concrete Confined with Fiber Reinforced Cement Based Thin Sheet Composites H.C. Wu and J. Teng

59 1

xv

zyxwv zyx

Hoop Rupture Strains of FRP Jackets in FRP-Confined Concrete L. Lam and J.G. Teng

60 1

Externally Confined High Strength Concrete Columns under Eccentric Loading J. Li, M. Moulsdale and M.S.N. Hadi

613

Creep Performance of CFRP Confined Concrete Cylinders M. Thkriault, M.-A. Pelletier, K. Khayat and G. Al Chami

623

Development/Splice Strength of Steel Bars in Concrete Confined with CFRP Sheets M.H. Harajli and B.S. Hamad

633

Lateral Prestressing of RC Columns with FRP Jackets A.A. Mortazavi, K. Pilakoutas and M.A. Ciupala

643

Confinement of RC Rectangular Columns Using GFRP A. Prota. G. Manfredi and E. Cosenza

653

Behaviour of RC Columns Retrofitted by Fibre Reinforced Polymers under Cyclic Loads H. Shaheen, T. Rakib, Y. Hashem, I. Shaaban and A. Abdelrahman

663

Photogrammetrically Measured Deformations of FRP Wrapped Low Strength Concrete A. Ilki, V. Koc, B. Ergun, M.O. Altan and N. Kumbasar

673

FRPSTRUCTURAL SHAPES Rectangular FRP Tubes Filled with Concrete for Beam and Column Applications A.Z. Fam. D.A. Schnerch and S.H. Rizkalla

685

Flexural Behaviour of GFRP-Polymer Concrete Hybrid Structural Systems M.C.S. Ribeiro, A.J.M. Ferreira and A.T. Marques

695 zyxwvut

xvi

A New Concept for an FRP Panelized Rapid Deployment Shelter N.M. Bradford and R. Sen

705

Experimental Investigation of Pultruded FRP Section Combined with Concrete Slab A. Biddah

7 15

zyxw zyxwvu

1-

~

zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCB

VOLUME TWO

DURABILITY AND MAINTENANCE ~~~

~~

~

Research on Strength and Durability of GFRP Rods for Prestressed Concrete Tendons M. Sugiyama and T. Uomoto

727

Durability of Concrete Beams Reinforced with GFRP Bars under Different Environmental and Loading Conditions K. Laoubi, E.F. El-Salakawy, B. Benmokrane and M. Pigeon

737

Environmental Effects on RC Beams Strengthened with Near Surface Mounted FRP Rods F. Micelli, A. La Tegola and J.J. Myers

749

Synergistic Hydrothermal Effects on Durability of E-Glass Vinylester Composites W. Chu and V.M. Karbhari

759

Durability of GFRP Composites under Tropical Climate Y.S. Liew and K.H. Tan

769

Effects of Different Long-Term Climatic Conditions on FRP Durability P. Labossihre, K.W. Neale and I. Nishizaki

779

xvii

Durability of Aramid and Carbon FRP PC Beams under Natural and Accelerated Exposure H. Nakai, H. Sakai, T. Nishimura and T. Uomoto

785

Effects of Wet Environment on CFRP-Confined Concrete Cylinders F. Micelli, L. De Lorenzis, and A. La Tegola

795

Alkali Aggregate Reactive Mortar Cylinders Partly Restrained by External CFRP Fabric B.J. Wigum

805

ASR Expansion Reduction and Ductility Improvement by CFRP Sheet Wrapping A. Hattori, S. Yamamoto, T. Miyagawa and Y. Kubo

815

Durability of GFRP Rebars in Concrete Beams under Sustained Loads at Severe Environments T.H. Almusallam, Y.A. Al-Salloum, S.H. Alsayed and A.M. Alhozaimy

823

Influence of Sustained Stress on the Durability of GFRP Bars Embedded in Concrete V. Dejke, 0. Poupard, L.O. Nilsson, R. Tepfers and A. Air-Mokhrar

833

A Maintenance Strategy for FRP Strengthening Systems P. Desiderio

843

zyx

SUSTAINED AND FATIGUE LOADS Viability of using CFRP Laminates to Repair RC Beams Corroded under Sustained Loads T. El Maaddawy and K. Soudki

855

Fatigue Bond of Carbon Fiber Sheets and Concrete in RC Slabs Strengthened by CFRP A. Kobayashi, S. Matsui and M. Kishimoto

865

xviii

Fatigue Performance of RC Beams Strengthened with CF Sheets Bonded by Inorganic Matrix H. Toutanji, Y. Deng and M. Jia

875

Fatigue Performance of RC Beams Strengthened with Externally Prestressed PBO Fiber Sheets Z . Wu, K. Iwashita, T. Ishikawa, K. Hayashi, N. Hanamori, T. Higuchi, A. Ikeda, T. Takeda, S . Murakami and T. Ichiryu

885

Prestressed CFRP Sheets for Strengthening Reinforced Concrete Structures in Fatigue R. El-Hacha, R.G. Wight, P.J. Heffernan and M.A. E r h

895

Fatigue Behaviour of Bridge Deck Specimen Strengthened with Carbon Fiber Polymer Composites J. Sim and H.S. Oh

905

Static and Fatigue Tests on Precracked RC Beams Strengthened with CFRP Sheets Z.Y. Wu, J.L. Clement, J.-L. Tailhan, C. Boulay and P. Fakhri

913

Fatigue Investigation of Concrete Bridge Deck Slab Reinforced with GFRP and Steel Strap A.H. Memon, A.A. Mufti and B. Bakht

923

zyxw zyxw

PRESTRESSED FRP REINFORCEMENT AND TENDONS Fatigue of High Strength Concrete Beams Pretensioned with CFRP Tendons B.B. Agyei, J.M. Lees and G.P. Terrasi

935

Transverse Confinement of Deck Slabs by Concrete Straps V. Banthia, A.A. Mufti, D. Svecova and B. Bakht

945

Design of Anchorage Zones for FRP-Prestressed Concrete T.J. Ibell, L. Gale and M.C. Choi

955

xix

A Simple Continuous System of Shear Reinforcement with Polyacetal Fiber R. Tuladhar, Y. Utsunomiya, Y. Sat0 and T.Ueda

965

Analytical Modeling of Splitting Bond Failure for NSM FRP Reinforcement in Concrete L. De Lorenzis and A. La Tegola

975

Strengthening of RC Beams with External FRP Tendons: Tendon Stress at Ultimate R.A. Tjandra and K.H. Tan

985

Comparative Analysis on Stress Calculation Methods for External FRP Cables L. An, T. Yamamoto, A. Hattori and T. Migayama

995

zyxwvu

Moment Redistribution in Continuous Monolithic and Segmental Concrete Beams Prestressed with External Aramid Tendons A.F. Araujo and G.B. Guimarles

1003

Experimental Investigation on the Ductility of Beams Prestressed with FRP M.M. Morais and C.J. Burgoyne

1013

Time-Dependent Flexural Crack Width Prediction of Concrete Beams Prestressed with CFRP tendons P.X.W. Zou and S.T. Smith

1023

STRUCTURAL STRENGTHENING Multiscale Reinforcement Concept for Employment of Carbon Fiber Woven Mesh K. Yamada, S. Ishiyama, H. Mihashi and K. Kirikoshi

1037

Woven Composite Fabric to Strengthen Structurally Deficient RC Beams H.Y. Leung, R.V. Balendran and T. Maqsood

1047

xx

zyxwvu

Calibration of Partial Safety Coefficients for FRP Strengthening G. Monti and S. Santini

1057

Comparison between FRP Rebar, FRP Grid and Steel Rebar Reinforced Concrete Beams M. Ozel, L.C. Bank, D. Arora, 0. Gonenc, D. Gremel, B. Nelson and D. McMonigal

1067

Concrete Beams Strengthened with Pre-Stressed Near Surface Mounted Reinforcement H. Nordin and B. Taljsten

1077

Strengthening of One-way RC Slabs with Openings using CFRP Systems H.D. Zhao and K.H. Tan

1087

Experimental Results of One-way RC Slabs with Openings Strengthened with CFRP Composites P. Casadei. T.J. Ibell and A. Nanni

1097

Seismic Behaviour of Reinforced Concrete Beam-Column Joint Strengthened with GFRP Y. Ouyang, X.L. Gu, Y.H. Huang and Z.Z. Qian

1107

FRP Seismic Strengthening of Columns in Frames M.A. Ciupala, K. Pilakoutas and N. Taranu

1117

Retrofitting of Shear Walls Designed to BS8110 for Seismic Loads using FRP K.H. Kong, K.H. Tan and T. Balendra

1127

Strengthening of Interior Slab-Column Connections with CFRP Strips K. Soudki, T. Van Zwol and R. Sherping

1137

Effectivenessof FRP Plate Strengtheningon Curved Soffits A.D. Porter, S.R. Denton, A. Nanni and T.J. Ibell

1147 zyxwvuts

xxi

Strengthening Performance of FRP Sheets Bonded to Concrete Tunnel Linings Z. Wu, W. He, J. Yin, Y. Kojima and T. Asakura

1157

Strengthening of Concrete Structures in Torsion with FRP B. Taljsten

1167

FE Modelling of FRP-Repaired RC Plane Stress Elements N. Khomwan and S.J. Foster

1177

APPLICATIONSIN MASONRY AND STEEL STRUCTURES Blast Resistance of Prototype In-Built Masonry Walls Strengthened with FRP Systems (Honourable Mention) M.K.H. Patoary and K.H. Tan

1189

Retrofit Techniques using Polymers and FRPs for Preventing Injurious Wall Debris J.E. Crawford and K.B. Morrill

1199

Experimental Behaviour of Masonry Panels Strengthened with FRP Sheets G. Marcari, G. Manfredi and M. Pecce

1209

Flexural Strengthening of URM Walls with FRP Systems V. Turco, N. Galati, J.G. Tumialan and A. Nanni

1219

zyxw

Shear Strengthening of URM Clay Walls with FRP Systems S. Grando, M.R. Valluzzi, J. G. Tumialan and A. Nanni

1229

Effect of FRP Mesh Reinforcement on Shear Capacity and Deformability of Masonry Walls S . Russo, R. Gottardo and D. Codato

1239

Strengthening of Masonry Structures under Compressive Loads by using FRP Strips M.R. Valluzzi, D. Tinazzi and C. Modena

1249

xxii

Seismic Behaviour of Masonry Structural Walls Strengthened with CFRP Plates X.L. Gu, Y. Ouyang, W.P. Zhang and F.F. Ye

1259

Advanced Composite Materials for the Repair of Steel Structures A.H. Al-Saidy, T.J. Wipf and F.W. Klaiber

1269

zyxwv

FIELD APPLICATIONS AND CASE STUDIES

Construction and Evaluation of Full-Scale CFRP Prestressed Concrete DT-Girder N.F. Grace and G.A. Sayed

1281

Flexural Behaviour of Bridge Deck Slabs Reinforced with FRP Composite Bars E.F. El-Salakawy, C. Kassem and B. Benmokrane

129I

Details and Specifications for a Bridge Deck with FRP Framework, Grid and Rebar L.C Bank, M.G. Oliva, J.S. Russell, D.A. Dieter, J.S. Dietsche, R.A. Hill, B. Gallagher, J.W. Carter, S. Woods and A.H. Anderson

1301

Construction, Testing and Monitoring of FRP RC Bridges in North America B. Benmokrane, E.F. El-Salakawy, G. DesgagnC and T. Lackey

1311

Strengthening of Concrete Structures with Prestressed and Gradually Anchored CFRP Strips (Best Paper - Application) I. Stoecklin and U. Meier

1321

Strengthening of Concrete Bridges with Carbon Cables and Strips T. Keller

1331

New Corrosion-Free Concrete Bridge Barriers Reinforced with GFRP Composite Bars E.F. El-Salakawy, R. Masmoudi, B. Benmokrane, F. Brikre and G. DesgagnC

1341

zyxw

zyxw xxiii

Strengthening of Steel Silos with Post-Tensioned CFRP Laminates L. De Lorenzis, F. Micelli and A. La Tegola

1351

Seismic Performance Improvement of the Bell Tower in Serra S. Quirico by Composites E. Cosenza, I. Iervolino and E. Guglielmo

1361

Strengthening with CFRP under Simulated Live Loads A. Hejll, A. Carolin and B. Taljsten

1371

Composite Structural Systems - From Characterization to Field Implementation V.M. Karbhari, H. Guan and L. Zhao

1381

Optimal Cost Design for Beams Prestressed with FRP Tendons I. Balafas and C.J. Burgoyne

1391

FRP in Civil Engineering in China: Research and Applications L.P. Ye, P. Feng, K. Zhang, L. Lin, W.H. Hong, Q.R. Yue, N. Zhang and T. Yang

1401

CODES AND STANDARDS Design Concepts of the New Swiss Code on Externally Bonded Reinforcement T. Vogel and T. Ulaga

1415

zyxwv

Design Guideline for CFRP Strengthening of Concrete Structures B. Taljsten

1425

Design Practice of Framed Building Structure Based on AIJ Design Guideline 2002 K. Kobayashi, H. Fukuyama, T. Fujisaki, S. Fukai and T. Kanakubo

1435

XXlV

zyxwvutsrqpon

Evaluations of Continuous Fiber Reinforced RC Members based on AIJ Design Guildeline 2002 K. Nakano. Y. Matsuzaki., T. Kaku and K. Masuo

1445

Design Procedure of NSM FRP Reinforcement for Strengthening of RC Beams L. De Lorenzis and A. Nanni

1455

~~

zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA

Keynote Papers

This page intentionally left blank

zyxwvu

FRPRCS-6, Singapore, 8-10 July 2003 Edited by Gang Hwee Tan QWorld Scientific Publishing Company

zyxw zyx

FRP REINFORCEMENTS IN STRUCTURAL CONCRETE: ASSESSMENT, PROGRESS AND PROSPECTS A.E. NAAMAN Department of Civil and Environmental Engineering University of Michigan, Ann Arbor, 48109-2125, USA

Key technical issues regarding the use of FRP reinforcements are reviewed, and assessment of the state of progress made. They include ductility in bending, shear resistance, dowel resistance, and a brief discussion related to fire, heat and durability. It is observed that technical problems can all be resolved, but each at a significant increase in cost. This adds to the already high cost of FRP reinforcements in comparison to steel, discouraging their use except for very special applications. However, in thin concrete products and laminated cementitious composites, FRP reinforcements in the form of meshes, textiles or fabrics are not only competitive on a technical basis but also on a cost basis. Recommendations for use of FRP reinforcements in cost-effective applications are made.

INTRODUCTION

zyxw

The applicability of Fiber Reinforced Polymer (FRP) reinforcements to concrete structures as a substitute for steel bars or prestressing tendons has been actively studied in numerous research laboratories and professional organizations around the world 33. FRP reinforcements offer a number of advantages such as corrosion resistance, non-magnetic properties, high tensile strength, lightweight and ease of handling. However, they generally have a linear elastic response in tension up to failure (described as a brittle failure) and a relatively poor transverse or shear resistance. They also have poor resistance to fire and when exposed to high temperatures. They loose significant strength upon bending, and they are sensitive to stress-rupture effects. Moreover, their cost, whether considered per unit weight or on the basis of force carrying capacity, is high in comparison to conventional steel reinforcing bars or prestressing tendons. From a structural engineering viewpoint, the most serious problems with FRP reinforcements are the lack of plastic behavior and the very low shear strength in the transverse direction. Such characteristics may lead to premature tendon rupture, particularly when combined effects are present,

'-*,

4 FRPRCS-6: Keynote Paper

such as at shear-cracking planes in reinforced concrete beams where dowel action exists. The dowel action reduces residual tensile and shear resistance in the tendon. Solutions and limitations of use have been offered and continuous improvements are expected in the future. The unit cost of FRP reinforcements is expected to decrease significantly with increased market share and demand. However, even today, there are applications where FRP reinforcements are cost effective and justifiable. Such cases include the use of bonded FRP sheets or plates in repair and strengthening of concrete structures, and the use of FRP meshes or textiles or fabrics in thin cement products. The cost of repair and rehabilitation of a structure is always, in relative terms, substantially higher than the cost of the initial structure. Repair generally requires a relatively small volume of repair materials but a relatively high commitment in labor. Moreover the cost of labor in developed countries is so high that the cost of material becomes secondary. Thus the highest the performance and durability of the repair material is, the more cost-effective is the repair. This implies that material cost is not really an issue in repair and that the fact that FRP repair materials are costly is not a constraining drawback. This paper provides a summary of key results and assessment of several research studies carried out by the author and his students at the University of Michigan on the use of FRP reinforcements in reinforced, prestressed and partially prestressed concrete members, and in laminated cementitious composites’2-28. It also provides some information on what the author believes are the most interesting and cost-effective applications of FRP reinforcements for today’s market conditions. Results of an opinion survey of a number of experts on “assessment of progress and prospects of FRP reinforcements,” are reported in a parallel paper and should be used to complement the information described here.

zyxw

STRUCTURAL DUCTILITY Structural ductility is of main concern in concrete beams reinforced or prestressed with FRP reinforcements due to FRP materials’ linear elastic behavior up to rupture without yielding. Unless ductility requirements are satisfied, FRP materials cannot be used reliably in structural engineering applications. Extensive experimental and analytical studies were carried out on structural ductility of concrete beams prestressed or partially prestressed with FRP tendons. Their main objective was to evaluate the ductility of

zyx

FRP Reinforcements in Structural Concrete 5

these beams by various measures, and suggest ways to improve such ductility for structural applications. Details are given in Refs. [9,12,13,14,16]. TC9 TC6a TC6

zyxwvuts zyxwvutsr zyxwvuts zyxwv

TCO ROC

RFC WCa- 1

RFCa-2 TS12

TS9

TS6

0.0

1.0

2.0

Ductility Index

3.0

4.0

5.0

’

p

Figure 1 Comparison of ductility index for beams with FRP and steel tendons

l6

Conventional definitions of the ductility index which are based on yielding of the reinforcement, are inappropriate for the evaluation of structural ductility in concrete beams reinforced or prestressed with FRP tendons. A new definition of ductility index was proposed. The new definition is expressed in terms of the ratio of the total energy to the elastic energy at the failure-state of a beam. It is applicable to beams with steel as well as brittle FRP reinforcements, thus providing a common basis for comparison. Results confirmed without the shadow of a doubt, that, everything else being equal, beams with FRP tendons tend to have lower ductility indices than beams with steel strands. This is illustrated in Fig. 1 where the ductility index is calculated on the basis of energy consideration from the following equation:

where Elotalis the total energy consumed to failure and Eelaslrris the elastic energy recovered at failure. For an elastic perfectly plastic material, Eq. (1) leads to a ductility index equal to the ratio of ultimate deflection to the

6 FRPRCS-6: Keynote Paper

zy zy zyxw

deflection at yield, that is, same as the conventional ductility index. Generally, higher ductility can be achieved by using proper design parameters or by considering several improvement methods such as fiber reinforcement or confinement. These methods are described next.

Approaches to Improve Structural Ductility

Structural ductility can be improved in several ways the most obvious being by the use of ductile materials. However, both FRP tendons and concrete are rather brittle and structural ductility must be achieved by other means. The approaches considered include the ideas described next 9.14916221.

Confinement b y fiber reinforcement. Increasing the compression strain capacity of the concrete by adding discontinuous fibers to the concrete, the fibers being steel or polymeric, should have similar effect as lateral confinement. First, fibers substantially influence the stress strain curve of concrete in compression, leading to a much more ductile behavior; this can be used to increase the non-recoverable part of the deflection at ultimate, leading to a substantial improvement in ductility index. Second, the use of fibers in appropriate amounts would easily increase the fracture energy of concrete by one to two orders of magnitude and, therefore, would balance the effects of elastic energy released by the FRP tendons should their failure occur. For applications requiring non-magnetic properties, polymeric fibers can be used. Jeong and Naaman 9 ~ 1 2 3 1have 6 shown that the use of fibers in otherwise over-reinforced (brittle) concrete beams led to ductility indices ranging from 2.9 to 5.45 and energy ratios from 3.7 to 9.2 (the control beam had values of 1). Note also that fibers can be used selectively in the structure, such as in the compressive zone only, or where a failure mechanism (producing hinges) is designed to occur. Their addition to concrete has a number of other benefits such as improving the shear capacity of the concrete matrix, decreasing crack widths, improving the inelastic bond between the reinforcement and concrete, and holding the concrete cover against spalling.

Confinement b y spirals and/or stirrups. Increasing the strain capacity of the concrete at failure by providing lateral confinement through spirals or stirrups made from steel or FRP reinforcements would increase the spread of plasticity in the compression zone of the concrete and lead to improved ductility. Using spirals with FRP is far more effective than bending them sharply into rectangular or square stirrups. Layered tendon and effective prestress design. Here it is suggested to place the prestressing reinforcement in layers and design the effective prestress in

zyx

FRP Reinforcements in Structural Concrete 7

each layer so as to provide a step like progressive failure with increasing deflections. Partial PrestressinR or hybrid combination o f reinforcements. Using partially prestressed concrete where prestressed FRP tendons are combined with conventional steel reinforcing bars or specially designed low strengthhigh ductility FRP bars allows sufficient flexibility to achieve increased ductility. Unbonded tendons. From an analytical viewpoint, the use of unbonded tendons, internal or external, is very attractive since the stress in the tendons does not reach ultimate prior to the failure of the concrete. This allows maximum ductility use from the concrete side while the risk of failure of the reinforcement is minimal. However, the use of unbonded FRP tendons implies the use of perfect anchorages that can sustain fatigue loading; moreover, external tendons can be very vulnerable to vandalism, and should they fail, they release an enormous amount of elastic energy that can be devastating. The transition failure from bonded to unbonded may provide an attractive solution as described next. Controlled bond failure. Design the interface between the FRP reinforcement and the concrete matrix so as to trigger a bond failure when the stress in the tendons reaches a threshold level, thus moving from a bonded tendon configuration to an unbonded tendon configuration. Technologically, this approach should not provide any difficulty. ODtimizina sectional ductility throunh ProQer reinforcement. Designing the section and proportioning the reinforcement in order to take advantage of the full strain capacity of concrete simultaneously with that of the reinforcement, is an essential design objective. Everything else being equal, sectional ductility can be improved by properly proportioning and placing the reinforcement in the section, and by selecting the effective prestress in the tendons. In particular, in order to better utilize the low strain capacity of the FRP reinforcement, it seems appropriate to design the section so as to achieve a neutral axis at ultimate as low as possible within the section. This somehow imply a section that is close to being over-reinforced.

zyxwvu

Assessment Summary on Ductility

Concrete structures using FRP reinforcements can be made sufficiently ductile to meet most design requirements. However, this will only come at a significant increase in cost.

8 FRPRCS-6: Keynote Paper

zyxwvu

BEHAVIOR IN SHEAR The majority of research on coricrete structures using FRP reinforcements has been on members that are not shear critical. Unlike flexural behavior, shear resisting behavior is quite complex by itself even in ordinary reinforced or prestressed concrete members. Furthermore, the experimentally derived prediction equations for the shear capacity of prestressed concrete members using steel tendons has not yet been proven applicable when FRP tendons are used. This is because the mechanical characteristics of FRP reinforcement, such as no yielding behavior, low shear or transverse strength, and low elastic modulus, are significantly different from those of steel tendons.

Dowel Action The longitudinal reinforcement, which is designed primarily to resist flexural tension, is often required to carry a shear force by dowel action across a diagonal tension crack. If the crack opens (rotates) slightly, a shear displacement will result from the rotation of a beam about the crack tip and the shear slip due to the shear force along the crack face. To resist differential shear displacement between the crack faces, the bars or tendons develop dowel shear forces. This counteraction of the bars or tendons to displacement is called "dowel action" (Fig. 2).

P

Dowel shear

zyx Shear displacement

Figure 2 Dowel action at shear cracks in a concrete beam In a diagonally cracked prestressed concrete beam, dowel action leads to a dowel bending moment and a shear force in the tendon itself, in addition to the tensile force due to the effective prestressing force and the applied load. As the bending moment and the shear force due to dowel action increase

zy

FRP Reinforcements in Structural Concrete 9

with loading, bending and shear stresses initiate simultaneously in the FRP tendon and become larger. Under these combined tensile and shear stresses, the tendon may fail prematurely, that is before reaching its unidirectional tensile strength. Generally the available tensile strength of FRP reinforcements decreases as their shear stress increases. Thus, dowel action reduces the allowable tensile stress in the tendon beyond that already caused by the effective prestressing force and applied load. Park and Naaman 18,25926327328carried out an experimental investigation of the shear behavior of concrete beams prestressed with CFRP tendons. They observed a mode of failure, not encountered with steel reinforcements, described as shear tendon rupture failure which is due to tendon rupture by dowel shear at the shear-cracking plane (Fig. 3). It is attributed to the brittle behavior and low transverse resistance of FRP tendons. This mode of failure is unique to FRP tendons and was not previously observed when steel tendons are used, because of the steel’s high transverse resistance and yielding characteristics. Shear tendon rupture failure may result in a serious reduction in load carrying capacity and ductility. Typical failure of a beam prestressed with FRP tendons that failed by shear tendon rupture is shown in Fig. 4.

zyxwv zyxwvu zyxw zyxw I

,

Horizontal cracking and spalling of concrete cover

(4

(b)

(4 (4 Figure 3 Typical failure modes observed in shear test of beams with steel or FRP tendons: (a) shear-tendonrupture failure; (b) shear-tensionfailure; (c) shearcompression failure; (d) flexural-tensionfailure

zyxwvutsr

10 FRPRCS-6: Keynote Paper

zyxw

Figure 4 Typical shear tendon rupture failure and crack pattern of beams prestressed with CFRP tendons

70

60

, :,

, ,

, ,

p,

, ,

C8 (CS)

(STR)

fO, , , ‘,2, , , f4, , , ‘,6, , STR: Shear-tendon rupture ST: Shear-tension failure CS: Compression-strut failure

, ,

300

250

zyxwvut zyxwvuts np7

50

S5(ST) C7(STR)

(ST)

/--A10

200

za 40

.Y,

150

30

bl

z

1; :, zyxwvutsr [,

Deflection (mm)

5B

bl

100

20

50

10

0 0

0.1

0.2

0.3 0.4 Deflection (in)

0.5

0.6

0 0.7

Figure 5 Observed load-deflection curves for short beams that failed primarily in shear. Beam S5 is prestressed with steel tendons. All other beams are prestressed with carbon FRP tendons. Beam CS3 had no stirrups. The prestressing index varied from 0.23 for beam C9 to 032 for beam C5. The effective prestress was about 0.42&, except for beam C6 with&, = 0 (from Park and Naaman)

zy

FRP Reinforcements in Structural Concrete 11

Conclusions from Study of Shear Following are some N~~~~~ 18.25,26,27,28.

conclusions drawn from the studies by Park and

zy zyxw

The shear-tendon rupture failure is a unique mode of failure characteristic of concrete beams prestressed with FRP tendons. It is due to tendon rupture by dowel shear at the shear-cracking plane. This premature failure is attributed to the poor resistance of FRP tendons in the transverse direction and their brittle behavior. Such a failure will result in lesser beam shear resistance and lesser ductility (Fig. 5). The shear-tendon rupture failure occurred at the flexural-shear-cracking plane in beams with FRP tendons, even when the effective prestress ratio was low (about 40%) and the required amount of steel stirrups were provided according to the ACI code. The ultimate shear displacement and crack width of prestressed beams which failed by shear-tendon rupture were respectively about one third and one half those of similar beams with steel tendons. Adding steel fibers is a possible way to improve the shear resistance of concrete beams prestressed with FRP tendons by avoiding or delaying shear-tendon rupture failure.

DOWEL BEHAVIOR OF TENSIONED FRP TENDONS Background and Significance It is generally agreed that the resistance in shear of steel bars or prestressing tendons is not less that 40% of their tensile yield resistance, and yielding in shear is the prevailing mode of failure under transverse loading. Thus, dowel resistance at shear cracks is generally assumed non critical when steel bars or tendons are used, because they yield under load. This is not the case for FRP tendons which do not undergo yielding. Moreover their observed shear resistance seem to be test dependent. Their failure due to transverse shear by dowel effect along cracked planes may lead to premature failure of structural members. Such behavior must be understood and accounted for in design. An experimental and analytical investigation of dowel action was carried out to study the dowel behavior of CFRP tendons subjected to combined tensile and shear forces (Fig. 6). Details are given in Refs. [26,27]. Key results are summarized next.

zyxwvutsrq

12 FRPRCS-6: Keynote Paper zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLK

1

F C

C

zyxwvu +Tj)

Section A-A

Section B-B

Section C-C

Note:( ) in mm

Figure 6 Typical cross section and loading arrangement of dowel test specimen [from Park and Naaman 26]

zyx zyxwvu zyxwv zyxwvu FRP Reinforcements in Structural Concrete 13

Displacement (mm)

2

0

8

6

4

C " " " " ' " " " " ' 1

1

zyxwvuts zy 30

zyxwvutsr zyxw ! ii ; 6

h v)

.-a

20

5 0 4 k! 42 2

k!

42

._ ...._..CFRP Tendon

$ 2

- 10

Steel Tendon Reinforcing Bar

I

0

0

I

0.05

I

,

,

I

I

,

I

, I ,

I

I

,

I I , , , I

0.1 0.15 0.2 Displacement (in)

I

0.25

,

2

$

1

I

I

0 0.3

Figure 7 Shear force versus displacement curves of dowel tests with steel and FRP tendons

Main Conclusions from Dowel Action Tests The ultimate dowel shear resistance of FRP tendons was 2 to 2.5 times smaller than that of steel tendons and their shear displacement at ultimate was about six times smaller (Fig. 7). Roughly the transverse shear resistance of FRP tendons varied from about 7% to 20% of their tensile strength (in comparison to 40% for steel). Dowel specimens with non-prestressed FRP tendons showed some ductility under increasing load due to crushing and cracking of the concrete surrounding the tendons. The ultimate dowel shear force of FRP tendons subjected to tensile and shear forces decreased elliptically as the tensile force increased. Theoretically, this failure behavior satisfies the maximum work theory, commonly referred to as the Tsai-Hill criterion, which is best described by an interaction curve. The ultimate dowel shear displacement of the FRP tendons subjected to tensile and shear forces decreased linearly with the increase of the tension ratio. Everything else being same, adding stirrups or adding fibers to the concrete, or using higher strength concrete lead to an increase in

14 FRPRCS-6: Keynote Paper

zyxwv zy zy

ultimate dowel resistance and a decrease in the corresponding shear displacement at failure.

5 ) The dowel response of specimens with shear planes inclined at 45O was more ductile than that with vertical shear planes, due to crushing and cracking of the concrete surrounding the tendons. However their ultimate shear force was only about 57% of that of specimens with vertical shear planes. Assessment Summary on Shear

Concrete structures using FRP reinforcements can be designed for shear; a safety factor higher that with steel reinforcement may be needed. As with the case of ductility, this will undoubtedly increase the cost. Clearly, from a structural design viewpoint, unbonded external tendons which are subjected neither to shear nor to dowel effects would be best.

HEAT AND FIRE RESISTANCE Few studies were devoted to evaluating the heat resistance, behavior under fire, or behavior after fire exposure, of concrete structures with FRP reinforcements 11,29330v31,34. It is fair to say that the concrete here protects the FRP reinforcement, and that the resin of the FRP reinforcement is the weakest link under high temperature exposure or fire. There is evidence that if the temperature is kept 20°C to 30°C below the glass transition temperature of the resin, the reinforcement remains fully effective 35. For practical purposes, this means a temperature below about 40°C. However for temperatures in the range of 60°C to 90°C deterioration due to creep may be significant depending on the type of resin matrix 35. There is also legitimate concern about the integrity of a structure following a fire, and the potential release of toxic fumes during a fire. A great deal of research is needed in applications where fire is an important design criterion. Although FRP reinforcements are not corrosion sensitive like steel, they do have other durability problems when subjected to various environmental conditions 32. Information is still very much needed in this area in order to provide assurance to the user, without the shadow of a doubt, that FRP reinforcements are indeed durable in the long term under the same environmental conditions for which steel reinforcements are exposed. Developing FRP reinforcements with improved heat resistance and stability for various environmental conditions will invariably lead to an increase in material cost.

zy z

FRP Reinforcements in Structural Concrete 15

DO COST COMPETITIVE APPLICATIONS EXIST AT THIS TIME? The technical drawbacks of FRP reinforcements (brittleness and low shear resistance) can be technically accommodated in a concrete structure but each at a significant increase in cost. However, these drawbacks which are quite critical for conventional concrete structures, seem to be less critical for applications in thin concrete products and thin laminated cementitious composites such as ferrocement 24. This is because ductility in laminated composites is guaranteed by the arrangement of the reinforcement system, and because meshes do no fail abruptly but rather through some progressive fracturing of their different wires or fibers15217322; also in thin sheets, vertical shear is not critical but interlaminar shear could be. Advanced fiber reinforced polymeric (FRP) meshes (textiles, fabrics, grids), may offer enormous advantages in spite of their initial high cost. This is because, unlike steel wire meshes, they can be tailored to exact requirements (i.e. fiber denier or diameter, mesh opening, etc..) at little extra cost, and they offer better corrosion resistance than ordinary steel reinforcements in thin products where the concrete cover to the reinforcement is small. Moreover they can be delivered in virtually any length, they are lightweight and they can be easily shaped to requirements. Thin reinforced concrete products such as cement sheets, ferrocement elements, glass fiber reinforced cement cladding and the like, generally utilize a high percentage of reinforcement (2% to 8% by volume) in comparison with conventional reinforced concrete (less than 2.5%); moreover the steel reinforcement in these products which consists of welded wire fabric, steel wire mesh, expanded metal mesh, or discontinuous fibers is two to ten times more expensive, on a unit weight basis, than conventional steel reinforcing bars. Fabricating a smaller diameter steel wire from a steel rod drives the cost of steel meshes very high. The smaller the diameter is, the higher the cost. The cost of the mesh is based on unit weight while the mesh mechanical efficiency is based on volume fraction in the composite. Since the unit weight of steel ranges from 5 to 8 times that of FRP materials, and since the composite properties are based on volume fraction of mesh reinforcement, cost comparison based on equal performance may favor FRP meshes. Also, the reinforcement content in thin cement products such as ferrocement is high in comparison to conventional reinforced concrete; as the cost of steel meshes is up to six times that of reinforcing bars, ferrocement (and generally thin concrete and laminated cementitious products) are ideally suited as an immediate market application for fiber reinforced polymeric (FRP)

16 FRPRCS-6: Keynote Paper

zyxwvu z zyxwv zyxwvu

reinforcements. Moreover, it is likely that future developments and applications will make FRP meshes (grids, textile, fabrics) increasingly cost competitive, especially when life-cycle cost analysis is considered, and advanced reinforcing configurations such as 3D meshes and mats become available.

FRP REINFORCEMENTS IN LAMINATED CEMENTITIOUS COMPOSITES AND HYBRID COMPOSITES When FRP meshes (or textiles, or fabrics) are used in thin cement products, the composite is called a laminated cementitious composites. A number of studies on the properties of such composites can be found in Refs. [ 15,17,19,20,22,24]. The FRP meshes used for demonstration in these studies included Kevlar (aramid fibers), Spectra (highly oriented polyethylene fibers), Carbon, PVA (poly-vinyl-alcohol), and polypropylene. The term “hybrid composite” implies a combination of continuous meshes such as carbon or steel, with discontinuous fibers such as PVA or steel. The fibers are generally premixed with the mortar matrix.

zyxwv zyxwvutsrqp

Typical Results

mm

._ u)

R

6 u)

2

zrn S

5

S

m” 0

‘F. u)

m

iil

zyxwvuts 0

5000

5

10

15

20

25

30

4000

25

3000

20

2

I

15

2000

10

5

0

0.2

0.4

0.6

0.8

1

0

Midspan Deflection, in.

Figure 8 Typical bending response of hybrid composite plate with PVA meshes and PVA fibers

zz

zyxw zyxwv FRP Reinforcements in Structural Concrete I7

zyxwvut zyxwv 5

0

10

mm

15

20

25

7500

e

50

L f 6000

Y

5 0

40

30

f

5 F

z '-

0

m

3- 3000 V

m n

e

No Fibers

.-F 4500

z

cc

V

20

w Y c

w

Kevlar Mesh Vr = 1.15%

-,-f 1500

z-m Y

0

- 10

c

0

I-

3

3 0.

u

w

Lu

0

0

0.2

0

0.4

0.6

1

0.8

Deflection (in.)

mm 0 3

X

7500

Y

5 I

-

10

, ,

I

.

15

20

,,

I

I

30

25

, , , ,

, ,

HYBRID COMPOSITE

v1 VI

PVAl: L ii 6 mm; d = 0.014 mm PVA2: L = 132 mm; d = 0.2 mm

6000

50

--

40

.-F

=

4500

30

0

e

m

=3

%

v

-

3000

20

1500

10

w Y

C

2

.-f

3

u

w

o

0 0

0.2

0.4

0.6

0.8

1

1.2

Deflection (in.)

Figure 9 Typical stress deflection curves of LCC plates: a) reinforced with two layers of FRP mesh (no fibers), and b) same with 1% PVA fibers by volume

I8 FRPRCS-6: Keynote Paper

zyxwv zyxw zyxwv

Typical results of laminated cementitious composites (LCC or ferrocement) bending tests using FRP meshes are shown in Figs. 8 and 9. Instead of plotting load versus deflection curves, the equivalent elastic flexural stress 2

for a rectangular plate ( oe = M I bh ), assuming uncracked section, was plotted versus the deflection. The equivalent elastic flexural stress accommodates automatically different geometric properties and test conditions, providing a good basis for comparison. All specimens described 0 ( 0 . 5 ~ 3 ~ 1in); 2 they were tested in in Figs. 8 and 9 were 1 2 . 7 ~ 7 5 ~ 3 0mm bending under third point loading with a span of 225 mm (9 in). Figure 8 shows a typical response of hybrid laminated cementitious composite plates reinforced with 2 layers of PVA meshes and PVA fibers; it also illustrates in this case the additive contribution of the fibers and the mesh. Figure 9 compares the stress-deflection response in bending of composite plates containing only two layers of either carbon, or Kevlar or Spectra mesh. Since the two layers of each mesh correspond to different volume fractions of reinforcement, some scaling should be used in comparing the effects of different mesh materials. However, Fig. 9a should be compared with Fig. 9b where the bending response of similar specimens, now containing in addition 1% fibers by volume, is shown. It can be observed that the addition of fibers is beneficial in many ways; it led to significant improvement in bending strength (more than 50%), deflection at maximum load, and toughness (measured as the area under the loaddeflection curve). Fibers also led to increased cracking strength, finer crack widths and smaller crack spacing.

Conclusions from Sudy of Laminated Cementitious Composites

1. Laminated cementitious composites using high performance FRP reinforcements (Kevlar, Spectra, and carbon), in the form of meshes, textiles or fabrics, exhibit excellent strength and ductility properties. A modulus of rupture close to 27 MPa was obtained with only 1.15% total volume fraction of Kevlar mesh. The addition of 1% by volume of discontinuous PVA fibers increased the modulus of rupture to about 39 MPa. Spectra and Kevlar type meshes led to similar results. 2. Fiber reinforced plastic meshes do enhance the multiple cracking process in laminated cementitious composites to the same extent known in ferrocement reinforced with steel meshes. Good multiple cracking developed with Kevlar and Spectra meshes, whether discontinuous fibers were added or not to the mix. More than sufficient ductility was also achieved.

zyxw

FRP Reinforcements in Structural Concrete 19

3. The addition of 1% PVA fibers by volume to specimens reinforced FRP meshes, led, on the average, to an increase in first cracking strength from about 5.3 MPa to about 8 MPa. 5 . When only 2 layers of Kevlar and Spectra meshes were used, failure was by interlaminar shear delamination. The addition of discontinuous fibers to the matrix led to increases in post-cracking strength of up to 50% and changed the failure mode from delamination to vertical shear.

The use of hybrid composites defined as composites reinforced with a combination of continuous meshes and discontinuous fibers offers clear advantages. They are ideally suited for applications in thin concrete products. In these composites, at least two layers of mesh are placed near the outer surfaces of the composite, and intermediate layers of mesh are replaced by discontinuous fibers. The fibers may be premixed with the mortar matrix or used in a mat structure. The combined utilization of fibers and meshes provides synergistic advantages due to each reinforcing system, in which the mesh contributes most to the load resistance and the fibers, while contributing their share of resistance, mostly contribute to crack control, toughness and ductility. Laminated cement based composites such as described above promise to become viable competitors to any existing fiber reinforced cement based thin sheet products as well as ferrocement products. Their efficiency in terms of modulus of rupture and ductility is unmatched at this time by any other cement composite including conventional RC and PC.

Assessment Summary on Applications in Thin Concrete Products

z

Thin cement based products using FRP reinforcements (meshes, textiles, fabrics) are not only competitive performance-wise, but also cost-wise. The use of FRP reinforcements in such applications should be pursued aggressively and is the best way to establish a credible record for the future.

GENERAL CONCLUSIONS AND RECOMMENDATIONS ON APPLICABILITY OF FRP REINFORCEMENTS From the above assessment, a number of recommendations can be inferred, and while they could be argued, they seem to be safe at time of this writing. Given the poor ductility and shear resistance of FRP reinforcements, it seems that the most effective way to use them, while avoiding excessively costly solutions, is under the following conditions:

20 FRPRCS-6: Keynote Paper

For FRP bars or tendons: 0

0

0

Prestressed members in order to take advantage of their strength; Prestressed, preferably unbonded construction, in order to make sure that their tensile strength is not attained at member failure, thus insuring some ductility; Prestressed, preferably unbonded, preferably external tendons in order to avoid dowel shear failure; Unless special requirements are desired and justify a significant increase in cost, it can be stated that, at time of this writing, FRP reinforcements are not cost-effective in conventional reinforced concrete structures.

For FRP sheets: For repair and strengthening, for confinement, or for protection from aggressive environments. (This application has not been discussed in this paper but is evident from the current state of progress and utilization worldwide).

zyxw

For FRP meshes or textiles or fabrics:

Definitely in thin concrete products and cement sheet applications and in combination with fibers; they exhibit superior performance while being cost effective.

ACKNOWLEDGMENTS The research work of the author has been funded in the past by several grants from the US National Science Foundation and by the University of Michigan. Their support is gratefully acknowledged.

REFERENCES Hundreds of references are available for information on the topics discussed in this paper. Given space availability, the author lists below only a very small number of references, among which the proceedings of previous FRPRC symposia and the main research studies used to formulate the above opinion assessment.

zyx zy

FRP Reinforcements in Structural Concrete 21 1. ACI Committee 440, Guide for the Design and Construction of Concrete Reinforced with FRP Bars, Manual of Concrete Practice, American Concrete Institute, Farmington Hills, 2001. 2. ACI Committee 440, Guide for the Design and Construction of Externally Bonded FRP Systems for Strengthening Concrete Structures, ACI 440.2R-02, American Concrete Institute, Farmington Hills, 2002. 3. ACT, Proceedings International Symposium on FRP Reinforcements in Concrete Structures, A. Nanni, and C.W. Dolan, Editors, ACI SP-138, American Concrete Institute, Farmington Hills, 1993. 4. ACI, Proceedings 4th International Symposium on FRP Reinforcements in Reinforced Concrete Structures, FRPRCS 4, ACI October 1999. 5. Benmokrane, B., and El-Salakawy, E.F., (Editors), “Durability of FiberReinforced Polymer (FRP) Composites for Construction,” Proceedings of 2”d International Conference, May 2002, Univeristy of Sherbrooke, Canada, 715 pages. 6. Burgoyne, C., (Editor), “FRP Reinforcements in Concrete Structures FRPRCS,” Proceedings, University of Cambridge, U.K., 200 1. 7. Japan Concrete Institute (JCI), Technical Report on Continuous Fiber Reinforced Concrete, TC 952: Committee on Continuous Fiber Reinforced Concrete, Tokyo, Japan, 1998. 8. Japan Society of Civil Engineers, JSCE Research Subcommittee on Continuous Fiber Reinforcing Materials, “Application of Continuous Fiber Reinforcing Materials to Concrete Structures,” International Concrete Library, No. 19, June 1992 9. Jeong, S.M., Naaman, A.E., and Tan, K.H., “Investigation of Beams Partially Prestressed with Carbon Fiber Reinforced Composite Tendons,” Proceedings of the FIP X I I Congress, Washington, May 1994, pp. B56-B61. 10. Harris, H.G., Somboonsong, W., and KO, F.K., “New Ductile Hybrid FRP Reinforcing Bar for Concrete Structures,” Journal of Composites for Construction, Vol. 2, No. 1, ASCE, Reston, Virginia, Feb. 1998, pp. 28-37. 11. Katz, A., Berman, N., and Bank, L., “Effect of Cyclic Loading and Elevated Temperature on the Bond Properties of FRP Rebars,” International Conference on the Durability of Fiber Reinforced Polymer for Construction, Sherbrooke, Canada, 1998, pp. 403-413. 12. Naaman, A.E., and Jeong, S.M., ”Considerations of Structural Ductility with External Tendons,” Proceedings, Workshop on Behavior of External Prestressing in Structures, St. Remy, France, June 1993.

z

zyxwvutsr

22 FRPRCS-6: Keynote Paper

13. Naaman, A.E., Tan, K.H., Jeong, S.M., and Alkhairi, F., "Partially Prestressed Beams with Carbon Fiber Composite Strands: Preliminary Tests Evaluation," Proceedings of ACI International Symposium on FRP Reinforcements in Concrete Structures, American Concrete Institute, ACI SP-138, 1993, pp. 441-464. 14. Naaman, A.E, "Ductility Implications of Prestressed and Partially Prestressed Concrete Structures Using Fiber Reinforced Plastic Reinforcements," FIP Symposium 93, Modern Prestressing Techniques and their Applications, Kyoto, Japan, October 1993, pp. 757-766. 15. Naaman, A.E., and Al-Shannag, J. "Ferrocement with Fiber Reinforced Plastic Meshes: Preliminary Investigation" Proceedings of the Fifth International Symposium on Ferrocement, Manchester, England, September 1994. P. Nedwell and N.R. Swamy, Editors, E. and FN Spon, London. 16. Naaman, A.E. and Jeong, S.M., "Structural Ductility of Beams Prestressed with FRP Tendons." Proceedings 2nd International Symposium on Non-Metallic (FRP) Reinforcement f o r Concrete Structures, L. Taerwe, Editor, Ghent, Belgium, August 1995; RILEM Proceedings 29, E & FN Spon, London, pp. 379-386. 17. Naaman, A.E., and Guerrero, P., "Bending Behavior of Thin Cement Composites Reinforced with FRP Meshes," Proceedings of First International Conference on Fiber Composites in Infrastructures, ICCI 96, Edited by H. Saadatmanesh and M. Ehsani, University of Arizona, Tucson, January 1996, pp. 178-189. 18. Naaman, A.E., and Park, S.Y., "Shear behavior of concrete beams prestressed with CFRP tendons: Preliminary test evaluation," Proceedings of International Conference, FRPRCS-3, Sapporo, Japan, October 1997. 19. Naaman, A.E., "High performance fiber reinforced cement composites: distinctive attributes for repair and rehabilitation," Proceedings of International Conference on Structural Failure ICSF-5, National University of Singapore, Nov. 1997; 11 pages. 20. Naaman, A.E., "Ferrocement: a High Performance Cementitious Third International Composite Laminate," Proceedings of the Conference on Analytical Models and New Concepts in Mechanics of Concrete Structures, Wroclaw, Poland, June 1999, pp. 199-206. 21. Naaman, A.E., "FRP Reinforcements in Concrete Structures: Design Issues, Potential Solutions, Realistic Applicability," Proceedings of the Second Middle East Symposium on Structural Composites for Infrastructure Applications, A.H. Hosni, I. Mahfouz and S. Sarkarni, Editors, April 1999, pp. 99-118.

zyxwv zyxwv

zy zyxwv zyxw FRP Reinforcements in Structural Concrete 23

22. Naaman, A.E., and Chandrangsu, K., “Bending Behavior of Laminated Cementitious Composites Reinforced with FRP Meshes,” ACI Symposium on High Perfonnance Fiber-Reinforced Concrete Thin Sheet Products, Edited by A. Peled, S.P. Shah and N. Banthia, American Concrete Institute, Farmington Hills, ACI SP 190, 2000, pp. 97-1 16. 23. Naaman, A.E., Park, S.Y., Lopez, M.M., and Till, R., Parameters Influencing the Flexural Response of RC Beams Strengthened using CFRP Sheets, Proceedings of FRPRC 5, C. Burgoyne, Editor, University of Cambridge, U.K., July 2001. pp. 117, 125. 24. Naaman, A.E., “Ferrocement and Laminated Cementitious Composites,” Techno Press 3000, ISBN 0-9674030-0-0, Ann Arbor, Michigan, 2000, 372 pages. 25. Park, S.Y., and Naaman, A.E., “Failure Criteria for CFRP Tendons Subjected to Tensile and Shear Forces,” Proceedings of 2nd International Conference on Composites in Infrastructure (ICCI 98), H. Saadatmanesh and M.R. Ehsani, Editors, January 1998, University of Arizona, Vol. 2, pp. 188-202. 26. Park, S.Y., and Naaman, A.E., “Dowel Behavior of Tensioned FRP Tendons,“ ACI Structural Journal, Vol 96, No. 5, September-October 1999, pp. 700-806. 27. Park, S.Y., and Naaman, A.E., “Shear Behavior of Concrete Beams Prestressed with FRP Tendons,” PCI Journal, Vol. 44, No. 1, Jan.-Feb. 1999, pp 74-85. 28. Park, S.Y., Naaman, A.E., Lopez, M.M., and Till, R.D., “Shear Strengthening of R.C. Beams Using Glued CFRP Laminates,” Proceedings FRP Composites in Civil Engineering, Vol. 1, J.-G. Teng, Editor, Elsevier Science Ltd., Dec. 2001, pp. 669-676. 29. Plecknik, J., et al., “Temperature Effects on Epoxy Adhesives,” Journal of Structural Engineering, Vol. 106, No. 1, ASCE, Reston, Virginia, 1986, pp. 99-1 13. 30. Saafi, M., “FRP Composites of Construction and the Fire Issue: Preliminary Laboratory Results,” 2nd International Conference on Durability of Fiber Reinforced Polymer (FRP) f o r Construction, Montreal, Canada, 2002. 31. Saafi, M., “Design of FRP Reinforced Concrete Structures under Fire Conditions,” CICE International Conference on FRP Composites in Civil Engineering, Kowloon, Hong Kong, December 2001.

24 FRPRCS-6: Keynote Paper

zyxwv zyxwv

32. Sen, R., Shahawy, M., Mullins, G., and Spain, J. (1999). “Durability of CFRPKoncrete Epoxy Bond in a Marine Environment,” ACZ Structural Journal, Vol. 96, No. 6, pp. 906-914. 33. Taerwe, L., (Editor), “Non-Metallic (FRP) Reinforcement for Concrete Structures,” 2nd International Symposium, Ghent, Belgium, RILEM Proceedings 29, E & FN Spon, 1995, London. 34. Tanano, H., et al., “Tensile Properties at High Temperatures of Continuous Fiber Bars and Deflections of Continuous Fiber Reinforced Concrete Beams under High Temperature Loading,” in Non-Metallic (FRP) Reinforcement for Concrete Structures, Japan Concrete Institute, Vol. 2, 1997, pp. 43-50. 35. Zureick, A. and L. Kahn (2001) “Rehabilitation of Reinforced Concrete Structures Using, Fiber-Reinforced Polymer Composites,“ ASM Handbook, Volume 21: Composites, pp. 906-913.

zyxwvu

FRPRCS-6, Singapore, 8-1 0 July 2003 Edited by Gang Hwee Tan QWorld Scientific Publishing Company

zyxwv zyxwv zyxwvu

PROGRESS AND PROSPECTS OF FRP REINFORCEMENTS: SURVEY OF EXPERT OPINIONS A.E. NAAMAN

Department of Civil and Environmental Engineering University of Michigan, Ann Arbor, 48109-2125, USA

This paper summarizes the opinion survey of a number of international experts in the field of FRP reinforcements carried out in response to the following questions: (1) Given the past 15 years of research, evaluation, and applications, what is your assessment of the current state of progress? Are we in an impasse? ( 2 ) Given the technical concerns posed during the initial development phase (ductility, poor transverse resistance, durability, stress-rupture, prestressing hardware, fire resistance, cost etc.), was any "leap-frog progress" made? If yes, what were the main elements? If not, why not? What are the main catalysts or obstacles? (3) What are the prospects for FRP reinforcements in near future applications in civil infrastructure systems? Do you expect major increases in applications? Give possible examples?

INTRODUCTION The author carried out a limited survey of opinions on the use of FRP reinforcements in structural concrete, reinforced and/or prestressed. Specifically the survey excluded the use of externally bonded sheets for repair and strengthening; it addressed only the use of FRP reinforcements as replacement of steel reinforcing bars or prestressing tendons in new structures. The opinion of a number of international experts was sought. Their responses are summarized below. It is important to realize that each bullet represents an opinion from a different person or group of persons. Some editing was used for uniformity. Very similar opinions were not repeated. However, care was taken to keep the information as clear and as close as possible to its original form. Not all persons contacted responded to the survey. Those who responded are listed in acknowledgments. The experts in the survey were selected to represent different continents and viewpoints. Special effort was made to preserve anonymity of the respondants. The opinions received cover the entire spectrum, from what could be described as optimistic to rather conservative, careful, risk averse, and pessimistic. It is hoped that the reader will have the opportunity to examine the information gathered and use it to better formulate hisher own opinion.

26 FRPRCS-6: Keynote Paper

SURVEY RESULTS

zyxwvu

zyxwvu zy

Question 1: Given the past 15 years of research, evaluation and applications, what is your assessment of the current state of progress? Are we in an impasse? Current State of Progress

Examples of opinions received include:

Since ACI 440.1R-01 came out one year ago, the number of commercial application is increased in the RC arena, while some work - especially in R&D - needs to be undertaken in PC type of applications. Major areas of interest regard bridge decks reinforcement, slabs on ground, and tunnel boring machine applications.

Considerable progress has been made in developing FRP rebars (glass and carbon fiber) for reinforcing slabs (considering no viable bar existed in the mid 80s). Less progress has been made on beams where shear capacity and FRP stirrup design remain a problem. No progress on FRP reinforced axial members (columns) was made. Perhaps there is no need for that. ACI design guide exists. Since the application of performance-based design is not ripe in structural engineering, the use of FRP reinforcements did not progress significantly. Moreover, a number of research issues on FRP applications need to be resolved to meet various technical and societal requirements. Some progress has been made in improving the durability of FRP rebars (particularly with glass fiber reinforcement). More work is needed on development of higher temperature and moisture resistant resin systems. Progress has been made on developing material specifications for FRP rebars. The use of FRP grids as reinforcement has been unfortunately neglected in the US - - and perhaps has the greatest potential for success. Considerable progress has been made in developing FRP tendons with carbon fibers. While there are some applications in beams, the use of FRF' post-tensioning in slabs appears to have the best potential, especially in bridge decks where craclung is a perennial problem.

Progress and Prospects of FRP Reinforcements 27

The state of the work is quite good and getting better. It is primarily economic limits that keep the technology in check. FRP repair is going great even with incomplete science because the economics work.

zyxwv z zyxw zyx

There has been significant progress at the research level. During the past decade, we have witnessed exponential growth in research and field demonstrations of FRP composites in civil engineering applications. In the United States, the growth was fueled by financial support from the National Science Foundation, the National Institute of Standards and Technology, the Federal Highway Administration, the National Cooperative Highway Research Program, the Department of Defense Advanced Research Projects Administration-Technology Reinvestment Project (DARPA-TRP), and state funding agencies. Since the late 199Os, deployment of FRP composites in highway bridges has increased considerably due to funding through the Transportation Equity Act-21 / Innovative Bridge Construction Program. 0

At least one researcher at almost every university in the US is involved in research on FRP materials; a field application of FRP materials in bridges has taken place in practically every state in the US.

Besides external prestressing, the application of FRP reinforcement (rods, tendons, grids) have been rather limited to special cases where the low unit weight, "non-corrosive" nature (in comparison to steel reinforcement), and non-magnetic properties, are being capitalized on. These include use in shaft walls so that the penetration and introduction of tunneling shield machines could be easily carried out, guideways for magnetic levitation systems, special facilities housing medical and sensitive equipment for which steel reinforcement would pose interference, and in seafront structures where the aggressive environment is adverse for steel reinforcement.

Generally spealung the fundamental properties of RC and PC structural members using FRP reinforcements are clear. The advantages of FRPs in comparison to steel rebars are also clear as well as their disadvantages such as cost, difficulty of handling in PC applications, and difficulties related to bent stirrups. While FRP reinforcements are used in special structures such as buildings for high-energy accelerators, there is concern about their heat resistance and their performance under fire.

zyx

The progress in FRP rods and tendons have in the last 5 to 7 years, definitely slowed down compared to FRP sheets. It can simply be

28 FRPRCS-6: Keynote Paper

observed from the number of publications in each field. We have to think of how to overcome the technical deficiencies of FRF’ materials to make them more attractive. However, as a first step, if the various codes for FW-reinforced concrete could be simplified and unified for general applications, and if the cost of F W reinforcements brought down, we may be able to see more applications.

To some degree, there has been progress. I think that the introduction of design guides will help in the short-term, but I also feel that the design guides will need to be substantially rewritten in coming years to ensure that efficient use is made of FRP materials, rather then mere “use”, to offset the undoubtedly higher short-term costs. FRP should not be seen as a direct replacement for steel, but rather as a material in its own right. Demonstration projects are still required, but there needs to be one BIG one. An analogy here could be Ironbridge, built in 1779 in Shropshire, UK, as a demonstration by the local iron mill that iron could be used to build bridges. Had they built a mere 20 ft span, no-one would have flinched, but they built an impressively large bridge, and the use of iron started. Equally, in the construction of the Millennium Dome in London in 1999, site welding was used on the main towers, something that academics would (at the time) usually discourage from attempting. This large-scale demonstration has helped to bring site welding back into the limelight. Therefore, the use of FRF’ reinforcements for concrete requires one large bridge or building, which could be seen as a focal point for its more widespread use. How such a project would arise is a much more difficult question.

zyxw zyxw

There have been literally hundreds of research studies, many practical guidelines (JCI, ACI, FIB, etc ...) accompanied by a lot of “fanfare” to encourage the use of FRP reinforcements, but the main issues remain same: ductility, shear resistance, bend-ability, fire resistance and most importantly cost. As long as the alternative “steel” provides better competitive solution in each of the above, it is not clear when and where FRP reinforcements will become equally competitive, except perhaps in structures built in space where weight is paramount.

A serious problem for FRP applications is acceptability at the professional-consultancy level. Many traditional steel-reinforced or prestressed concrete researchers and designers look at FRP-reinforced concrete researchers, consultants, or experts, as having “jumped on the bandwagon,” with little or no real grasp of some fundamental issues in structural engineering, but with the goal of attracting research funding

Progress and Prospects of FRP Reinforcements 29

zyxwvu zyxwvutsr

easily, or selling the material. They look at FRP research as slightly inferior, and untrustworthy to some extent. This might be a real obstacle to the use of FRP materials in reputable large consultancies. Are we at an Impasse?

We are not at an impasse. Progress is substantial. Innovation takes time. While there have been some inappropriate uses of FRP prototype materials there are also many applications that show a convincing trend toward future acceptance for specific applications.

In a way we are at an impasse. With extensive research baclung, and guidelines available, where are the projects and where are the plans for future large-scale projects? Compare to the beginning of steel, or reinforced concrete; and that was more than a century ago. Today the pace should be exponential. But where are the planned applications? I would rather be at the beginning of reinforced or prestressed concrete. We are indeed in an impasse despite the numerous, sometimes ambitious, test programs worldwide. The major obstacles are the high material cost and some particular technical problems which can be summarized as follows: a) Reinforced concrete Most applications seem to focus on GFRP (glass FRP). However, how to explain that we put glass in an alkaline hostile environment as a solution to the corrosion problem of steel? This is a contradictory situation. Even AR-glass is subject to long-term deterioration. GFRP is used because of cost reasons, compared to CFRP (carbon FRP) and AFRP (aramid FRP). However, the design of flexural members is governed by serviceability conditions which implies the need to provide a sufficient cross section of reinforcement in ordcr to increase flexural stiffness and reduce crack width. This additionally increases the cost. FRP stirrups are a serious practical obstacle in practice: 1) They need to be pre-shaped and cannot be bent at the site 2) There is significant strength reduction in the bends. Suggesting to use longitudinal FRP reinforcement and steel stirrups is inconsistent. b) Prestressed concrete Stress-rupture limits the efficient use of GFRP in prestressing. When CFRP and AFRP are used in RC applications, these high-strength

zyxw

30 FRPRCS-6: Keynote Paper

materials are not fully utilized. Hence there is benefit in using them in prestressed concrete. However, here arises the problem of reliable, cost effective anchorages. For pre-tensioning, requirements are less severe than for post-tensioning. In the first case individual wires, strands or strips need to be anchored temporarily at a smaller capacity than for big cables. Some of the special anchorages developed so far are complicated to use and not applicable to efficient day-to-day practice. This is true for pretensioning as well as post-tensioning, where in comparison, large steel cables are commonly tensioned. Another problem with CFRP tendons is the inherent brittleness which requires severe precautions during handling and tensioning. Regarding AFRP, problems arise with high relaxation (up to 20% in wet conditions), sensitivity to moisture, and high transverse thermal expansion. Relatively limited progress has been made.

zyx

Yes indeed we are at an impasse. Given the tremendous effort (technical and promotional) that introduced FRP reinforcements to the professional community (through research and research centers, symposia, educational materials, practical guidelines, technical journals and newsletters) today's applications of the technology should be widespread; but it is not. While we understand better the technical issues, the solutions so far provided are not sufficiently attractive to make FRP reinforcements competitive in both the short and long-term. Unless the cost of the most promising FRP materials (that are carbon based composites) diminishes significantly, we can practically do with stainless steel reinforcements all that can be done with FRP reinforcements, except for making them as lightweight. And with steel, we have a long-term proven experience.

zyx

Question 2: Given the technical concerns posed during the initial development phase (ductility, poor transverse resistance, durability, stressrupture, prestressing hardware, fire resistance, cost etc...) was any "leapfrog progress" made? Zfyes, what were the main elements? If not, why not? What are the main catalysts or obstacles? Examples of opinions received include: The issues of ductility, crack control, cost etc. is hard to deal with in FRP-reinforced concrete. I believe that the whole concept of "hiding"

zyxwv z zyxw

zyxw zyxw zyxw Progress and Prospects of FRP Reinforcements 31

FRP reinforcements inside concrete is problematic, until long-term performance has been demonstrated. 0

Many technical issues related to the use of Flip composites have been addressed in the ACI 440 guidelines. Still, more needs to be learned, and more experience in the field is needed before unresolved concerns can be treated with the familiarity used in steel reinforcement design.

In prestressed concrete, the issue is much more that of economics. The cost of carbon FRP strand is still 3 to 6 times that of steel strand, and steel works very well with little problems. The few bridges built with FRP reinforcements (for demonstration or trial) were so inefficient that they did not make a breakthrough.

There was significant progress. In prestressed concrete (including external prestressing) the use of FRP reinforcements enhances durability under severe environmental conditions. New hardware for prestressing was developed. 0

0

No leap-frog progress was achieved so far. The advantages of FRP reinforcements were not demonstrated convincingly. Cost should be correlated with the level of performance for FRP applications; however, their long-term performance level cannot be evaluated yet. Progress will depend on the assessment of life-cycle cost. Progress with FRP reinforcements should be characterized as evolutionary (not leap-frog), which is not bad. I do not see any major obstacles to continue the evolutionary development. The largest obstacle appears to be the low glass transition temperature of the polymer resin systems currently in use (around 100 "C) and also the poor fire resistance of the materials (however, this issue is most often not addressed by structural engineers directly, even with conventional materials.) The main catalyst for continuing progress is the search for more durable and predictable construction materials that are exposed to severe environmental conditions, therefore malung applications in highway structures much more attractive than in conventional building structures.

32 FRPRCS-6: Keynote Paper

zy zyxw

Research in the area of FRP reinforcement has been extensive over the past 15 years and has advanced the state of the art considerably. A testament to that are the codes and/or proposed guidelines that have been published in Japan, Canada, Europe, and the US. However, we still have a long way to go. The primary impediment to the deployment of FRP reinforcements is "cost". FRP reinforcements have been used in a number of bridges, and the main reason was because these were either experimental, or demonstration, or showcase projects. It would have been very difficult, even impossible, to justify their use based on cost. With the exception of limited special applications, such as in nonmagnetic or highly corrosive environments, the use of FRP reinforcements has been limited. 0

Progress has been made over the years on the issues mentioned, but it is not clear if they could be considered "leap-frog progress." Some of the durability tests, for example, may not be useful for design purposes. Many engineers are still unfamiliar with FRP rods and tendons. Further education of the industry and the profession on FRP materials is definitely needed.

0

There has been leap-frog progress in terms of ductility, by loolung to the concrete, rather than the reinforcement. In the other areas listed, some progress has been made (certainly in durability and fire resistance), but it cannot be considered "leap-frog.''

0

The next true leap-frog advancement must be in terms of how we use FRP materials. There is need to write a guide (at the international level and for an international audience) which addresses issues about how BEST to use FW materials in concrete. It could be spearheaded for instance by some of the people responding to this survey and started following FWRC6. It would show designers the versatility of FRP reinforcements and this would be beneficial in terms of providing a more credible, widespread confidence in FRF'. I don't see any major break-through for RC and PC applications, which happened during the last decade, despite the very extensive research efforts.

zyxwv

Progress and Prospects of FRP Reinforcements 33

Question 3: What are the prospects for FRP reinforcements in near future applications in civil infrastructure systems? Do you expect major increases in applications? Give possible examples? Examples of opinions received include:

zyxw

I do not expect major increases in applications.

Major increases in FRP applications are expected in the future, primarily due to the increased knowledge and confidence that engineers and designers will develop for this new technology. With the exclusion of columns reinforcement, where FRP reinforcements should not be used, every other structural component (slabs, beams, bridge decks) could be designed using FRP composites as main reinforcement. In highway applications (particularly in bridge decks) the future appears to be promising both for FRP rebars (glass or carbon) and FRP tendons due to the high cost of bridge deck maintenance. Otherwise only specialized applications that require exceptional durability (wastewater etc.) and electromagnetic transparency (MRIs) appear to have any rationale.

Repair and rehabilitation will remain the top contender application. The field of precast building systems may offer an entirely new technology.

FRP reinforcements will see growth in structures where the primary design considerations are corrosion, magnetic interference, and other special considerations. Durability, damage control, and long life are requirements for environmental sustainability. Examples of applications include long lasting skeletons (structural subsystem) of buildings where the remaining subsystems are changeable with time. The elasticity of FRP reinforcements leads to reduced residual deflections in seismic structures. Their non-magnetic properties will remain a key advantage in certain applications. FRP reinforcements will be mostly used in special structures, or in structures under severe environmental conditions (such as coastal bridges exposed to chlorides). They will be used if the client demands it, or if a policy is set by an agency to use them. However, their relatively high cost will limit their use. Moreover, construction

34 FRPRCS-6: Keynote Paper

companies do not care about using FRP reinforcements since they do not reap the benefits; while manufacturers of FRPs do. Further applications of FRP reinforcements in civil infrastructure systems are possible. This requires the combined efforts of researchers, educators, code writers and professional organizations. In the research arena, in particular, we need to think beyond the usual boundaries. For example: a) Is it not practical to replace partially corroded steel reinforcement with FRP reinforcement or add FRP reinforcement to existing steel reinforcement in concrete structures? In this manner, we achieve a "hybrid reinforcement", and by doing so, could we not improve the ductility of the member? (b) Can FRP rods and tendons be made sufficiently flexible so that they could be bent by hand or simple mechanical tools to form stirrups and the like on site? Can we develop a special thermoplastic that would allow such bending to be done, for example, using a "hot-press"? (c) Can we utilize FRP reinforcements as "smart materials" for structural monitoring?

zyxw

I do not know of a single near-future application of any sort, although I do know of a very recent one ...(that did not go as desired) .... I think this boils down once again to lack of confidence in the product, for whatever myriad of reasons particular designers have.

I do not foresee a major increase in the use of FRP until its use is first shown to be cost-effective in the SHORT term. It does us no good to point out the long-term cost savings (with no performance records to prove it). Clients are not interested in this. We need to cut down the initial cost, which is done by using the material in novel ways. One immediate issue I can think of is bends in FRP bars. There is absolutely no logical reason why we have bends. They cost a fortune to manufacture, they are unalterable and they are not needed if a helix is placed around the bar in the anchorage location; this anchors the bar just as well, but it is not mentioned in FRP design guides. Nor is any serious attempt made to mention ductility injection into concrete through helical confinement or short fibers, which would lead to redistribution of moments being possible and full use of compression reinforcement. The rational use of FRP is a research priority, in my opinion, and through such rational use we will create a product, which is CHEAPER than steel-reinforced concrete, even in the short term. At that stage, expect an explosion in use.

I predict almost no progress in practical applications in the near future except if the material cost could be substantially reduced, which seems

Progress and Prospects of FRP Reinforcements 35

z

unlikely. Applications will remain limited to small-scale demonstration projects. Exceptions are concrete elements in very aggressive chemical environment and requirements for electromagnetic neutrality (rooms for specific electronic equipment).

zyxw

For reinforcement of QUEEN CONCRETE, in the form of wires, reinforcing bars or prestressing strands, STEEL, conventional or stainless, is still KJNG of the hill.

0

CONCLUSIONS It is clear from the above questions and answers that in spite of the extensive research so far carried out worldwide on the use of FRP reinforcements in concrete structures, there is a relative “malaise” about their use and future success. It is also clear that at time of this writing, a number of international experts do not see “leap-frog’’ advances and exponential progress in applications. Similarly to many physical phenomena described by an inverted S curve, there has been a rapid initial development in FRP reinforcements, followed by a slow steady progress and, at time of this writing, we seem to be in that steady progress phase; however, it is not clear if what lies ahead is going to be continuing slow steady progress, or exponentially increasing progress, or fall in disgrace and forgetfulness phase. The curves of Fig. 1 illustrates our current position. Predicting the future at this time is more challenging than a “wait and see” approach.

v)

Y

v1

zyxwvuz

zyxwvutsrqpo

8 a

. 2 a v)

b

Figure 1 Schematic illustration of progress for use of FRP reinforcements in concrete structures

36 FRPRCS-6: Keynote Paper

zyxwvu

ACKNOWLEDGMENTS A number of experts were asked to respond to the questions of the survey described in this paper. Those who responded on time are listed next. The author is grateful for their cooperation and their candid answers. They are: Lawrence Bank, University of Wisconsin, Madison, USA; Charles Dolan, University of Wyoming, Laramie, USA; Hiroshi Fukuyama, Building Research Institute, Tsukuba, Japan; Issam E. Harik, University of Kentucky, Lexington, USA; Tim bell, University of Bath, UK; Ayman Mosallam, California State University, Fullerton, USA; firoshi Mutsuyoshi, Saitama University, Japan; Antoine E. Naaman, University of Michigan, Ann Arbor, USA; Antonio Nanni, University of Missouri, Rolla, USA; Ferdinand. Rostazy, Technical University of Braunschweig, Germany; Luc Taerwe, University of Ghent, Belgium; Kiang Hwee Tan, National University of Singapore, Singapore; Thanasis Triantafilou, University of Patras, Greece.

zyxwvutsr

FRPRCS-6, Singapore, 8-10 July 2003 Edited by Gang Hwee Tan OWorld Scientific Publishing Company

zyxw zyxwv

DURABILITY DESIGN OF GFRP RODS FOR CONCRETE REINFORCEMENT T. UOMOTO International Center for Urban Safety Engineering, Institute of Industrial Science, University of Tokyo 4-6-IKomaba, Meguro, Tokyo, Japan

To deal with the corrosion of reinforcing steel in concrete, FRP has been used throughout the world. They do not corrode even in chloride environments by sea water and deicing salt. Considering the durability of the material, FRP will become a major reinforcing material for concrete in highly corrosive environment. One of the problems of FRP is that some of the FRP rods and sheets deteriorate due to other causes such as alkali attack, acid attack, ultraviolet ray attack, etc. Among them, alkali attack to glass fibers and GFRP is the largest problem. It is difficult to apply the material as internal reinforcement of concrete. To deal with the problem, many attempts are performed. This paper explains how to deal with the problem to produce high alkali resistant GFRP using durability design.

INTRODUCTION

zyxw zy

Concrete structures throughout the world have been deteriorated severely due to chloride induced steel corrosion. To deal with the problem, many attempts were performed, such as to use galvanized steel bars, epoxy-coated bars, catholic protection, etc. Among these attempts, FRP was evaluated as one of the best methods to deal with the problem because FRP does not corrode even in chloride environment. As mentioned by JSCE research committee I), many researches were performed mostly in Japan, North America and Europe to utilize FRP as concrete reinforcement since 1980's. In Japan, a large amount of FRP has been applied to reinforced concrete structures not only to new structures but also to existing structures using recommendations by JSCE *, 3), etc. The types of FRP commonly being used are rods embedded in concrete for new structures and sheets applied to the surfaces of existing structures.

38 FRPRCS-6: Keynote Paper

Concrete structures are normally used for more than 50 to 100 years, and the reinforcements must be also durable enough to reinforce the concrete for the same period of time. Although FRP does not corrode in chloride environment, we have already clarified that FRP deteriorates in other environments such as high concentration of alkali and acid, ultra-violet rays from sunlight, etc. To deal with the problem, care must be taken how to use FRP materials as reinforcements for concrete structures4). One method is to use high durable material such as CFRP as concrete reinforcement. Another method is to change the properties of the existing FRP so that it may not deteriorate easily in these environments. Considering these conditions, this paper is written to explain briefly through our works in IIS, the cause of FRP deterioration and basic concept to deal with these problems. In this paper, explanation is given on FRP rods using carbon fibers, Aramid fibers and glass fibers.

zyx

MAIN CAUSES OF FRP DETERIORATION

FRP is a composite material, composed of millions of fibers and resin. The diameters of fibers are in the range of 6 (carbon fibers) to 15 microns (Aramid fibers and glass fibers). As shown in Figure 1, when tensile load is applied to FRP, fibers carry load and resin transfers stress to the neighboring fibers. The resin can also protect fibers from ingress of harmful ions from their environment. In this paper, carbon fiber reinforced plastics, Aramid fiber reinforced plastics and glass fiber reinforced plastics are abbreviated as CFRP, AFRP and GFRP. Deterioration of both fibers and resin, and also the transition zone between fibers and resin govern the durability of FRP. This makes the deterioration mechanism of FRP complicated compared to steel. As most of the mechanical properties are governed by fibers*), if the fibers are not deteriorated, FRP can resist against load in most cases. But when resin is attacked and deteriorated, the fibers fall off from the surface and FRP reduces strength (See Figure 2).

zy

Durability Design of GFRP Rods 39

zyxwvu

Figure 2 Deteriaration of GFRP due to alkali attack

zy zyxwv zyxwvu

40 FRPRCS-6: Keynote Paper

Considering the properties and usage of FRP, important items on deterioration to be considered are listed below. Items 1) to 3) are for reinforcements embedded in concrete (rods), and items 4) to 6) are for surface reinforcements (mainly sheets). 1) Static fatigue fracture 2) Fatigue fracture6") 3) Alkali resistance7) 4) Acidic resistan~e'~) 5 ) Ultra-violet ray resistance") 6) Freeze-thaw resistance STRENGTH OF FIBERS AND FRP AFTER DETERIORATION

Table 4 is obtained from the previous research works of IIS, University of Tokyo12. 14, 15) using the material mentioned in Tables 1, 2 and 3. Again, the first 3 items are related to reinforcements embedded in concrete, and the remaining 3 items are related to external reinforcement applied to the surface of existing concrete structures. The items, such as static fatigue, cyclic fatigue and alkali resistance are the 3 important items to be considered in case of FRP rods. As shown in the table, CFRP rods have little problem in the aspect of durability. In the case of GFRP, rods reduce their strength tremendously in all the cases. This is caused mainly by the deterioration of glass fibers. In the case of AFRP, although alkali resistance and cyclic fatigue properties are good, care must be taken on static fatigue strength. This is caused by the creep rupture of Aramid fibers. The items, such as acidic resistance, ultra-violet ray resistance and freeze-thaw resistance are the 3 important items to be considered in case of external reinforcement, such as out cables or sheet reinforcements. As shown in the table, carbon fibers and glass fibers have no special problem on durability except ultra-violet ray resistance. The resin deteriorates from the surface under ultra-violet rays and this causes the deterioration of CFRP and GFRP. In any of the cases, measures must be taken to prevent resin from the deterioration or to replace FRP before deterioration exceeds a certain limit. In case of Aramid fibers, the fiber itself deteriorates by ultraviolet rays and high concentrated acid, especially in case of Kevler 49. To deal with the problem, measures must be taken to prevent the deterioration from sunlight and acidic water or to replace FRP before deterioration exceeds a certain limit.

zyx zy

Durability Design of GFRP Rods 41

Table 1 Properties of fibers (monofilament ) f

ble 1 Properties o ble 1 Properties o ble ble 11 Properties Properties oo

Ta

Ta

Ta Ta

Table 2 Properties of resins for FRP

I

I

I Tensile Strength

Type Average S.D.

I

I I

Ripoxy-RSO2 84.9 1.15

zyx I

Ripoxy-H6001 67.00 4.96

Table Table 1 Properties 1 Properties of fibers of fibers (monofilament (monofilament )f )f .

I

Elastic Modulus (MPa) Maximum Strain

(%) Notes

I

Average S.D. c . of v. Average S.D. c . of v.

3110 37 0.0 1 5.22 0.1 1 0.02 AFRP,GFRP

I

4000 39 0.01 1.95 0.27 0.14 CFRP

Table 1 Properties of fibers (monofilament ) f GFRP CFRP Type AFRP Tensile 1340 1690 1690 Average Strength Table 1 Properties of fibers (monofilament ) c . of v. of fibers (monofilament 0.0836 Table 1 Properties )f Elastic Average 45.70 52.1 1 135.30 Modulus 1.946 0.476 Table 1 Properties of fibers (monofilament ) f 23.63 c . of v. of fibers (monofilament 0.0175 Table 1 Properties )f

f

42 FRPRCS-6: Keynote Paper

Table

Durability of fibers and FRP (Strength ratio)14) I Fiber Carbon Glass Notes Aramid 1

Static Fatigue Cyclic Fatigue Alkali Resistance

95%

Acidic Resistance

100%

Ultra-Violet Ray lesistance Freeze-Thaw kesistance

100%

~~

zyxwv zyxw zyxwv 92%

60-85?’0*

45%

15%

NaOH, 40 OC, 1OOOhr 100% HCI, 40 OC, 120 days 81% 0.2MJ/m2/hr, 1 OOOhr ~

:RP Rod

Static Fatigue

CFRP 91%

Cyclic Fatigue

85%

Alkali Resistance

100%

AFRP 46%

GFRP 30%

70%

23%

98%

29%

Notes 20°C, 100 years (Cal.) lOOMpa Amp., 2 million cycles NaOH, 120days, 40 OC

Acidic Resistance 69% 77% 90% 3 years exposure Ultra-Violet Ray Lesistance 100% Freeze-Thaw Table 1100% Properties of fibers (monofilament ) f Lesistance Table 1 Properties of fibers (monofilament ) f (Note) * In case of Technola is 85% and Kevlar 49 is 60%.

From these results, the following can be concluded as written in the previous paper by the author “Durability of FRP as Reinforcement for Concrete Structures” (ACMBS-3, 2000) 14), excluding the problems of fire and surface defects due to knives, etc: 1) We do not have to consider much about the durability of carbon fibers and CFRP used as internal and/or external reinforcement, except the deterioration of resins caused by ultra-violet rays. 2) In case of Aramid fiber and AFRE’, they have good durability properties except static fatigue, ultra-violet rays and acidic attack. When used as internal reinforcement, care has to be taken on static fatigue properties. Limitation of tensile stress is needed according to the duration time. When used as external reinforcement, not only the sustained load but also deterioration due to ultra-violet rays and acidic environment must be considered.

zy zyx

Durability Design of GFRP Rods 43

3) Glass fiber and GFRP have poor durability except acidic resistance and freeze-thaw resistance. GFRP is not recommendable for internal reinforcements. When glass fiber or GFRP is used as external reinforcements, care must be taken of the deterioration due to sustained load, fatigue load, alkali resistance and ultra-violet rays. DEVELOPMENT OF NEW GFRP TO INCREASE DURABILITY As explained above, it is difficult to use glass fibers of GFRP as internal reinforcement of concrete structures. To deal with the problem the following ideas may be used: 1) Develop new alkali-resistant glass fibers to reduce the effect of alkali attack from concrete. 2) Develop new GFRP to reduce the attack of alkali from surface of FRP. In the case of l), I hope the producers of glass fibers would challenge the work. Up till now I do not hear any good news on success in producing these fibers. In the case of 2), several attempts have been made in Japan, obtaining some good results. Here in this paper, two cases on development of new GFRP are explained.

zyxwvu zyx zyx

Development of Hybrid AGFRP Rod

As shown in Table 5 and Figure 3, new Hybrid AGFRP rods are developed". They are composed of both E-glass fibers and Aramid (Technora) fibers. Aramid fibers are placed at the surface portion of the rods and glass fibers at the center. As shown in the table, the volume content of fibers are fixed to about 66%, changing the amount of Aramid fibers from 19% to 30% and glass fibers from 48% to 36%.

zyxwvu zyx

ble 1 Properties o ble 1 Properties o ble 1 Properties oI I ble 1 Properties I I I oI

Ta Ta

Tabolume fraction of E-Glass fiber

Vf (%)

1Total volume fraction of fibers

IThickness of outer layer (Aramid)

Ta ITensile strength of rod

I

(mm)

(Nlmm')

36.3

66.8

Vf (%)

I

0.79

14.18

I 1 I

42.4

65.8

0.59

13.83

I I I

48.4

I

67.5 0.46

13.49

I

44 FRPRCS-6: Keynote Paper

Figure 3 AGFRP rods with Aramid fibers and glass fibers (Nishimura T.)")

zyxw

Figure 4 shows the retaining ratios of tensile strength after immersion in alkaline solution (NaOH:lmol/l, 40degrees) for up to 120 days. As can be seen from the figure, original GFRP has lost their strength to 30% at the age of 90 days, but these hybrid AGFRP rods possess more than 80% of the initial tensile strength at the age of 120 days. Figure 5 shows that AGFRP-7 (Aramid fiber content:23.4%, glass fiber content:42.4%) has improved resistance against alkaline solution drastically compared to other FRP rods. This is caused by the resistance of the rod against alkali at the surfaces. Measured result of Na shown in Figure 6 gives the information that Na ion does not penetrate into the rod compared to GFRP rods.

zyxwvut zyxwvuts zy 1

0. 9

0. 3

0

50 100 Immersion time (days)

150

Figure 4 Retention ratio of tensile strength of AGFRP (Nishimura T.)")

zyx zy z

Durability Design of GFRP Rods 45

12 1

02 0 0

50

zyxw

100 I n m ersbn tin e (days)

150

Figure 5 Retention ratio of tensile strength of FRP (Nishimura T.)") zyxwvutsrqponmlkjihgfedcbaZYXWVU

Figure 6 Na distribution within AGFRP rod after immersion to alkaline solution (Nishimura T.)

46 FRPRCS-6: Keynote Paper

New GFRP rod with Surface Treatment

zyxwvu

Figure 7 shows the deteriorated portion of glass fibers in GFRP when immersed in alkaline solution". The photograph shows that the glass fibers are deteriorated from the interfaces between fibers and resin. To obtain the distribution of Na ions in the cross section, EPMA was used to the same specimen. Figure 8 shows the results obtained by Katsuki'. As shown in the figure, high concentration of Na is observed at the boundary between fibers and resin. These two results show that GFRP is attacked by the Na ions from the boundary of fibers. From these observations, one good method to improve the resistance against alkali is to form good interface between fibers and resin so that alkali can not penetrate into GFRP from outside. To prove this idea, some tests have been done to improve the gap between fibers and resin. Figure 9 shows the validity of the method. As shown in the figure, GFRP has improved resistance against alkali to high extent. Although other properties are still under tests, this method may become a good solution to obtain durable GFRP rods and sheets to reinforce concrete structures as in the case of CFRP and AFRP.

Figure 7 SEM photograph showing deterioration of glass fibers in GFRP rod (Katsuki F.)')

zyx

Durability Design of GFRP Rods 47

zyxwvuts

Figure 8 Distribution of Na ions in GFRP by EPMA (Katsuki F.)') 1 0.9

0.4

zyxwvu

0.3

0

zyxwvu 60

30

90

120

Immersion time (days)

Figure 9 Retention ratio of tensile strength of FRP (Sugiyama M . ) ' ~ )

CONCLUSIONS The following conclusions can be drawn: (1) Tensile strength of FRF' rods can be used to evaluate the durability of FRP rods in different conditions.

48 FRPRCS -6: Keynote Paper

zyxwvu

(2) Durability of carbon fibers and CFRP used as internal andor external reinforcement is good, except the deterioration of resins caused by ultraviolet rays. (3) In the case of Aramid fiber and AFRP, they have good durability properties except static fatigue, ultra-violet rays and acidic attack. When used as internal reinforcement, care has to be taken of the static fatigue properties. Limitation of tensile stress is needed according to the duration time. When used as external reinforcement, not only the sustained load but also deterioration due to ultra-violet rays and acidic environment must be considered. (4) Commercially available glass fiber and GFRP have poor durability except acidic resistance and freeze-thaw resistance. They are not recommendable for internal reinforcements. When glass fiber or GFRP is used as external reinforcements, care must be taken to the deterioration due to sustained load, fatigue load, alkali resistance and ultra-violet rays. (5) Durable GFRP against alkali solution can be obtained by changing their composition, such as to combine glass fibers with Aramid fibers when producing FRP. A new AGFRP has improved the durability to a very high extent. (6) Interface between glass fiber and resin governs the resistance of GFRP against alkali. The improvement of the interface is effective to increase the alkali resistance of GFRP to high extent.

zyxwv

ACKNOWLEDGEMENTS

The author would like to thank Mr. Tsugio Nishimura (IIS, University of Tokyo), Dr. Futoshi Katsuki (Shibaura Institute of Technology) and Mr. Matoyoshi Sugiyama (Nippon Electric Glass Co. Ltd) for granting the permission to use their data in this paper. REFERENCES

JSCE Research Committee on CFRM( 1993), “State-of-the-Art Report on Continuous Fiber Reinforcing Materials”, Concrete Engineering Series 3, JSCE 2. JSCE Research Committee on CFRM( 1997), “Recommendation for Design and Construction of JSCE Concrete Structures Using 1.

zyx

Durability Design of GFRP Rods 49

Continuous Fiber Reinforcing Materials”, Concrete Engineering Series 23, JSCE

3.

4.

5.

6.

7.

zyx zyxwvut JSCE Research Committee on Upgrading of Concrete Structures with Use of CFS(200 l), “Recommendations for Upgrading of Concrete Structures with Use of Continuous Fiber Sheets”, Concrete Engineering Series 4 1, JSCE

Kobayashi, K., Uomoto, T. and Cho, R (1988), “ Prestressed Concrete Structure using FRP Tendons (in Japanese), Prestressed Concrete, V01.30, NO.5, ~ ~ 1 9 - 2 6

Hodhod, H.A.G.A.A.(1992): “Employment of Constituents Properties in Evaluation and Interpretation of FRP Rods Mechanical Behaviour”, Doctoral Thesis, Univeristy of Tokyo

Uomoto, T. et a1 (1995) “Fatigue Strength of FRP Rods for Concrete Reinforcement”, Building for the 2 1’‘ Century, Edited by. Y.C.Loo, EASEC, pp. 1659-1664

Uomoto, T. and Katsuki, F (1996), “Deterioration Mechanism of Glass Fiber Reinforced Concrete and Prediction of Strength Reduction”, Integrated Design and Environmental Issues in Concrete Technology, Edited by K.Sakai, FN & Spon, pp.137-146

8.

Katsuki, F. (1996), ”Evaluation of Alkali Resistance of FRP Rods for Concrete Reinforcement with Different Types of Fibers” (in Japanese), Doctoral Thesis, University of Tokyo

9.

Uomoto, T. and Ohga, H. (1996), “Performance of Fiber Reinforced Plastics for Concrete Reinforcement”, Advanced Composite Materials in Bridges and Structures, pp. 125-132

10. Uomoto, T., Nishimura , T., Kato, Y.: Development of new AGFRP tendon with high alkali resistance, Seisan Kenkyu, Vo1.48, No.9, pp.457-460. 1996 11. Yamaguchi, T. (1998), “Study on deterioration of FRP Rods for Concrete Reinforcement on Ultra-Violet Rays and Creep Rupture” (in Japanese), Doctoral Thesis, University of Tokyo 12. Uomoto, T., et al. (1998), “Strength and Durability of FRP Rods for Prestressed Concrete Tendons” (in Japanese), Report of the Institute of Industrial Science, University of Tokyo, Vo1.39, No.2, No.244 13. Nishimura T., et a1 (1999), “Temperature Effect on Fiber Strength in Different solutions”, JCI General Meeting, V0.2 1, No.2, pp.288-293

zyxwvutsr zyxwvuts

SO FRPRCS -6: Keynote Paper

14. Uomoto, T(2000): Durability of FRP as Reinforcement for concrete structures, ACMBS-3 15. Uomoto, T (2001): Durability considerations for FRP reinforcements, FRPRCS-5, pp. 17-32, Thomas Telford, London

16. Sugiyama, M and Uomoto T.: Development of new GFRP rods with high alkali resistance, Seisan Kenkyu, 2003 (in press)

zyxwvu zyxwvuts zyxwvu

FRPRCS-6, Singapore, 8-10 July 2003 Edited by Kiang Hwee Tan QWorld Scientific Publishing Company

zyxw

NEW TYPES OF CONTINUOUS FIBER REINFORCEMENTS FOR CONCRETE MEMBERS T. UEDA Division of Structural and Geotechnical Engineering, Hokkaido University Sapporo 060-8628, Japan

In this paper continuous fibers for concrete reinforcement with rather unique mechanical and/or other features are introduced. Carbon and Polyacetal Continuous Fiber Flexible Reinforcement, which are flexible enough to wind freely by hand may provide a solution to overcome a weak point of typical continuous fiber reinforcements and at the same time ease congestion of reinforcement in concrete members in highly seismic regions. Polyacetal fiber also has another feature that is a high fracturing strain of 69%. Despite its low stiffness that can be compensated by large amount, Polyacetal fiber gives a better ultimate deformation due to its high fracturing strain. Another item introduced in this paper is continuous fiber reinforcement externally bonded with a very soft resin. Previous studies mostly focus on properties of fibers not those of adhesive resin. Soft adhesive layer provides a better solution to enhance the delamination strength together with a ductile failure manner.

INTRODUCTION Continuous fiber reinforcement (CFR) provides us with a new option of internal and external concrete reinforcements. Unlike steel long dominated as the only concrete reinforcement in the past, CFR is non-corrosive. It has a very high strength to weight ratio hence reduces the amount of reinforcement. It is easy to handle during construction as cutting CFR requires only a simple cutter. CFR, however, shows some vulnerability such as low fracturing strain and no plastic deformation. Easy to cut is a good feature but at the same time a weak point. In order to utilize concrete strength and deformability fully its reinforcement should have a fracturing strain greater than 6 % (preferably 10 %). This is especially true for concrete members requiring high deformability which can be found in highly aseismic members. Elastic deformation without plastic deformation (or yielding) is not necessarily a weak point if the material has a high fracturing strain. The yielding, however, shows warning much clearly before failure.

52 FRPRCS -6: Keynote Paper

External bonding is a typical method for retrofitting concrete members and CFR is sometime used as a reinforcing material. External bonding creates another type of failure mode, that is delamination of externally bonded reinforcement. This failure mode does not happen in ordinary concrete members with only internal reinforcement. To provide a full anchorage (or development length) may not be economical or, in some cases, practical due to the space limit. Therefore, delamination is unavoidable in many cases. In this paper two new types of continuous fiber reinforcement and a resin for use as concrete reinforcement are introduced. Their new features are believed to help overcome some of the weak points found in typical CFR. The first one is “continuous fiber flexible reinforcement” or CFFR. CFR cannot be bent after the impregnated resin gets hardened due to the small plastic deformability of both fiber and resin. CFFR is introduced to overcome this problem. Two types of CFFR will be introduced in this paper -- with post-impregnation of resin and without resin. The second one is a continuous fiber whose fracturing strain is much greater than those of a typical CFR and is called Polyacetal Fiber (PAF). PAF does not require resin since the strengths of the bundled fiber at both straight and bent portions are not reduced. PAF in fact is the material for CFFR without resin. The resin to be introduced here is soft adhesive resin, which is used to enhance delamination strength. Particularly a very soft resin with Young’s modulus of 1 MPa is introduced.

zyxwvu zyxw

CONTINUOUS FIBER FLEXIBLE REINFORCEMENT (CFFR) Carbon Continuous Fiber Flexible Reinforcement (CCFFR)

The concept of continuous fiber flexible reinforcement was introduced in the late 1990s by a joint team of Nippon Steel Composites and Hokkaido University’, Carbon Continuous Fiber Flexible Reinforcement (CCFFR) is a bundle of thousands of carbon fibers inserted in a transparent PolyvinylChloride (PVC) tube that is injected by Vinyl-Ester type low viscosity high flowable resin (see Figure 1). Before the injection CCFFR is flexible enough to be bent, wound and placed by hand as you wish (see Figure 2). The original intention of CCFFR was to eliminate one of the disadvantages with typical continuous fiber reinforcement (CFR) such as carbon, aramid and glass CFR, that is the fact that you cannot bend CFR as you do for steel reinforcement. Since CCFFR can be bent by hand, it becomes even easier to handle at construction site. It can ease the difficulties encountered with

’.

z

zyxw z

New Types of Continuous Fiber Reinforcements 53

Continuous Carbon Fiber

+

zyx 1

Vinyl Ester type low viscosity high flowable resin

Figure 1. Carbon Continuous Fiber Flexible Reinforcement4

J

congested reinforcements. Although the elimination of the resin injection should provide us better constructability, it is less likely the case due to the fact that the tensile strength of bundled carbon fibers is significantly less than that of the fiber. PVC tube serves as electric isolator for carbon fiber since direct contact between carbon fiber and steel reinforcement causes steel to corrode more easily. In order to prove the advantage of CCFFR a series of experiments on reinforced concrete columns confined with CCFFR as the intermediate lateral tie was conducted3. Five column specimens were tested under the combination of flexure and shear. The size of all the columns was 350x350~1050mm and that of footing to which it was monolithically attached was 900x900x800 mm. All the columns contain the same amount of longitudinal reinforcement i.e., 8 D25 bars (deformed bar with a diameter of 25 mm); 4 each on both sides. The minimum amount of steel tie reinforcement of 9 D10 stirrups with a center-to-center spacing of 190 mm was provided. Besides the two reference columns S 1 and S2, three columns, S3 and S4 containing an additional 0.1 % carbon fiber and S5 containing an additional 0.2 % carbon fiber by volume were tested. The pattern of winding of CCFFR around the main bar is shown in Figure 2. The material properties of steel reinforcing bars and CCFFR are shown in Table 1. The ultimate strain of CCFFR is 15600 p and the thickness of the plastic tube is 1 mm. The concrete strengths and loading sequences are given in Table 2. Pressure injection of resin took 10 minutes after winding CCFFR. When the resin hardened 24 hours after the injection, the end portion of

zyx zyxw

zyxwv zyxwvu zyxwvut zyxwvutsrq zyxwvuts

54 FRPRCS -6: Keynote Paper

st

ln a 0 m \c

E

0 4-

0

II U S

m

P 0

44-

. m

0

0

5l (D

k rv) L

m

m E E v) (v

P

*

N

zyx

Figure 2. Arrangement of CCFFR4

CCFFR was passed through a steel pipe and fixed to the pipe with an expansive material. It was then allowed to set for 24 hours. The steel pipes were anchored with washers and nuts at the column top surface after the concrete was placed. The observed performance of each specimen is presented in the form of applied force versus column tip deformation in Figures 3 through 5 . The cyclic and reversed-cyclic hysteretic load-deformation relationship of the

New Types of Continuous Fiber Reinforcements 55

zyxwvut zyx zy zyxwvu zyxwvutsr zyxwvu

Table 1. Material Properties in Specimens Strengthened by CCFFR CrossYoung's Yield Yield sectional modulus strength strain area (mm2) (GPa) (MPa) 04 D25 506.7 173 367 2121 D10 71.3 183 364 1891 CCFFR 16.89') 265 1) Cross-sectional area of a bundle of carbon fibers only

Reinforcement type

Ultimate strength (MPa) 551 357 69.8 kN

Table 2. Compressive Strength of Concrete and Loading Sequence for Specimens Strengthened by CCFFR Specimen S1 (Ref) S2 (Ref) s3 s4 Average fc ' ('Pa) 44.7 40.0 40.3 35.1 Sequence of & T& & &? Loading') 1) & One-sided cyclic loading; ?& Reversed cyclic loading

s5

35.4

&?

reference columns S1 and S2 suggest that the columns experienced rapid strength degradation due to insufficient confinement from the tie reinforcement. Specimen S2 even suffered a huge damage from splitting along the longitudinal bar to the loading point. To such columns that lack ductility, 0.1 % volume fraction of carbon fiber as CFFR was added. These columns are S3 and S4. Column S3, companion of column S1, did not only counteract the degrading strength but also enhanced it by 7.5 % before failure by rupture of CCFFR at 11% lateral drift, which means that the ultimate deformation was also enhanced. Column S4, companion of column S2, also counteracted the highly decayed strength in S2 with deformation enhancement. The strength increment was 9.4 %, while the ultimate drift was 6.3 %. This demonstrates that addition of CCFFR could effectively confine the core concrete with an introduction of its ductile failure rather than quick extension and widening of diagonal crack leading to shear failure. Column S5 contained 0.2 % volume fraction of carbon fiber as CFFR and was the companion of S2 and S4. It showed even further enhancement of shear strength and ultimate deformation. The strength increment and the ultimate drift was 23.1 and 8.2 % respectively. The superior performance of S5 over the companions indicates the greater confinement efficiency provided by the increased amount of CCFFR and the its appropriate

zyxwvuts zyxwv

56 FRPRCS -6: Keynote Paper zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONML Localized Rotation

300

zyxwvutsrqponmlkjihgfedc

\

, zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA

0

1

2

3

zyxwvuts 4

5

6

7

8

9

10

I1

Effective Lateral Drift (%)

Figure 3. Load-Deformation Curves of Reference Specimens without CCFFR (Specimens S1 and S3 placed horizontally in the photo)4

winding pattern as well as the efficiency of the clamping system at its extremities. The CCFFR provided the intended lateral confinement to the volumetric dilation of core concrete without any premature rupture of CCFFR. All the column specimens with CCFFR showed the rupture of CCFFR at bend which controlled the ultimate deformation. This fact implies that predictions of CCFFR strength at bend and strain development in CCFFR are necessary. An experimental investigation of bent-portion of CCFFR (see Figure 6) shows clearly that the bent strength ratio (ratio of bent portion strength to straight portion strength) has a tendency to decrease with an increase in angle of winding (0 in Figure 7 ) as shown in Figure 8 where d is CCFFR diameter and L is CCFFR length in concrete before bend. As

zyxwv

New Types of Continuous Fiber Reinforcements 57 zyxwvuts

s2 zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHG s4zyxwvutsrqponmlkjihg

zyxwvuts I -sz

200 300

-300

zyxwvutsr zyxwvuts

-11

-9

-7

-5

-3 -1 1 3 5 Effective Lsteral Drift (%)

7

9

11

Figure 4. Load-Deformation Curves of Specimens with CFFR (Specimens S2 and S4 placed horizontally in the photo)4

other CFR bar reinforcement, the bent portion strength decreases as the bent radius decreases (which means here the decrease in diameter of main reinforcement around which CCFFR was wound). Numerical analysis with nonlinear finite element method is one of the methods to predict mechanical behaviors such as strain development in reinforcement. The author’s group at Hokkaido University recently developed a three-dimensional nonlinear finite element program for both concrete and steel-concrete composite members’. In order to apply this program for analysis of members with CCFFR, the constitutive model for force transfer versus slip relationship at the bent portion should be implemented. It is considered that the force transfer mechanism is characterized by the plastic tube compressive deformation and the resin

zyxwvutsr zyxw

58 FRPRCS -6: Keynote Paper zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONML

s2

s5 zyxwvutsrqponmlkjihgfedc

zyxw

1 -300 -11

-9

-7

-5 -3 - I 1 3 5 EffectiveLateraI Drift ( I % )

7

9

11

Figure 5. Load-Deformation Curves of specimens with Different Amount of CFFR (Specimens S2 and S 5 placed horizontally in the photo)4

failure in compression. The winding angle affects the force transfer-slip relationship. Based on those observations the empirical model was proposed4.

zy z zyxwv

New Types of Continuous Fiber Reinforcements 59

4

a End Clamp Tube for CCFFR b Continuous Arrangement of CCFFR c Main Bar 25 mm diameter d Steel Stirrup 10 mm diameter e Prestressing rod to fix Base Plate to the Strong Floor.

Displacement-control pull from Hydraulic actuator

Top-threaded 300mm long steel tube containing highly expansive material to confine CFFR

300

20mm diameter Prestressing rods to fix specimens with the base plate n

,

zyxwv zy Specimen

350

1

Base Plate

600x600~25

e l I 1

Laboratory Strong Floor I

1

(All dimensions in mm) Figure 6. Element tension test for bent-portion of CCFFR4

60 FRPRCS -6: Keynote Paper

zyxwvu Angle of winding

Main Bar

zyx

zyxwvut zyxwv zyxwvu zyxwvut Figure 7. Angle of winding CCFFR4

1.2

.-0 c,

a

p:

c c,

MI C

!!

;j c,

:

m

1

0.8

0.6

F

0.4 0.2

-5

5

15

25

35

45

Angle Deg.

Figure 8. Bent Strength Decrease with Winding Angle4

Polyacetal Continuous Fiber Flexible Reinforcement (PCFFR) Polyacetal fiber (PAF) has been used as ground reinforcement, however is rather new for concrete reinforcement6’’. It has a low stiffness but a high fracturing strain in comparison with a typical continuous fiber such as carbon, aramid and glass as shown in Table 3. Other features of PAF are that practically no strength reduction exists at bent portion or in the case without impregnating resin. These features make PAF a good material for CFFR. Unlike CCFFR, PAF does not require impregnating resin or attention to the bent-portion strength. In order to prove the applicability of PAF as CFFR (PCFFR), a series of tests on reinforced concrete columns with steel and PAF tie reinforcement

zyxwv zyx

New Types of Continuous Fiber Reinforcements 61

Table 3. Properties of Fiber for CFR

Carbon (PAN)

Carbon (Pitch)

Aramid

High strength type High elasticity type Ordinary type High strength and elasticity type High strength type High elasticity type

Fracturing strain

Density

PA)

(g/mm3)

2600-4500

200-240

1.3-1.8

1.7-1.9

2000-2800

350-450

0.4-0.8

1.8-1.9

780-1000

38-40

2.1-2.5

1.6-1.7

3000-3500

400-800

0.4-1.5

1.9-2.1

2800

130

2.3

1.45

3110

77

4.4

1.39

35003600

74-75

4.8

2.6

zyxwvu zy zyxw

E-glass Glass

Young’s modulus (GPa)

Tensile strength (MPa)

Fiber

Alkali 8oo 70-76 2-3 2.27 resistance 3500 Polyacetal 1730 20’’ 6-9 1.45 Polyester 219 1.04 21 1) Polyacetal fiber shows material nonlinearity. The initial stiffness is 40 GPa.

was conducted under reversed cyclic loading. Figure 9 and Photo 1 show the arrangement of PCFFR, which was wound around the longitudinal reinforcement. The details and test results of each specimen are given in Table 4. The test results indicate that PCFFR increases the ultimate deformation. Comparison of the load-deformation curve between specimen P2-2 with PAF and steel tie reinforcement and specimen P2-1 with only steel tie reinforcement is shown in Figure 10. The tie reinforcement ratios of both specimens are similar as shown in Table 4. Figure 10 clearly indicates that the ultimate deformation of the specimen with PCFFR is significantly greater than that of the specimen with only steel tie reinforcement. The reason may be that PCFFR was placed within the core concrete. It should be mentioned that the greater deformation was possible because of the absence of PAF fracture even at the ultimate deformation, which was observed in a comparable specimen S4 with CCFFR (see Table

4).

zyx z

zyxwvuts zyxwvut

62 FRPRCS -6:Keynote Paper

Outer PAF

Strain gage

Inner PAF

Back

Front

View from front face

Front

Back

View from back face

zyx

Figure 9. Arrangement of PCFFR (in specimen S l )

Although no fracture of PAF was observed in all the specimens, prediction of strain in PCFFR is important. This is because the tensile fracture of PCFFR is still possible and because the contribution as shear reinforcement of PCFFR can be quantified by the tensile force carried by PCFFR at ultimate. The measurement of PCFFR strain in details shows that

New Types of Continuous Fiber Reinforcements 63

early strain development in PCFFR indicates better efficiency than steel tie reinforcement and is caused by direct contact between PCFFR and the longitudinal reinforcement. Gap likely to exist between steel tie reinforcement and longitudinal reinforcement cannot make the steel tie as efficient as PCFFR. Based on the experimental observation a simple formula to predict the PCFFR strain as a function of maximum deformation was proposed by assuming a constant ratio 0.61 of average steel tie reinforcement strain to PCFFR one4.

Photo 1. Arrangement of PCFFR

zyxwvu zyxwvutsr zyxw zy zy zyxw

64 FRPRCS -6: Keynote Paper

Table 4. Details and Test Results of Reinforced Columns with PCFFR and CCFFR

4

fc PY P m all iu (MPa) @N) (mm) @N) (mm) (SJS,) P1-1 2.7 0.51 23.7 176.6 P1-2 185.5 10.28 211.2 33.8 3.4 2.7 0.51 0.52’’ 29.3 4.2 P1-3 2.7 0.51 0.79’’ 32.3 174.9 10.28 196.3 43.7 P2-1 2.04 0.68 27.4 139.2 8.02 158.2 30.4 3.8 P2-2 2.04 0.17 0.58’) 29.8 128.1 7.96 159.8 65.2 8.2 s4 2.0 0.21 0.10” 35.1 213.4 13.54 255.0 57.3 4.2 1) PCFFR for specimens P1-2, P1-3 and P2-2 and CCFFR for specimen S4 2) ps: longitudinal reinforcement ratio, pw:steel tie reinforcement ratio, pcf: continuous fiber reinforcement ratio as tie reinforcement,f’,: concrete compressive strength, Py: yielding load, 8,: yielding deformation, Pmm: maximum load, dU: ultimate deformation, and p: ductility ratio

Specimen

Ps

Pw

PA) PA)

Pcf

(%)

zyxwv zyxw Deformation (mm)

~

Figure 10. Load-Deformation Curves of Columns with PCFFR and Steel Tie Reinforcement

zyxw z

New Types of Continuous Fiber Reinforcements 65

CONTINUOUS FIBER WITH HIGH FRACTURING STRAIN Here “high fracturing strain” is stressed in comparison with “high strength” and “high stiffness”, which are often cited as a good feature of reinforcing material. High strength and/or stiffness reduce the needed amount of material. However the high cost of material, which is likely for the material with high strength or stiffness, reduces the attractiveness of the material. On the other hand, lower strength andor stiffness can be compensated by providing more material. In ordinary reinforced concrete members concrete crushing rather than fracturing of steel reinforcement is the cause of member failure, This is mainly due to the high fracturing strain of the steel reinforcement. On the contrary fracturing of continuous fiber reinforcement is usually the cause of member failure in concrete member with typical CFR because of its rather low fracturing strain (0.4 to 5 % as shown in Table 3). In order to utilize the concrete strength fully a higher fracturing strain is necessary. Unlike strength or stiffness, adding the material cannot compensate for the low

zyxwvz Deformation (mm)

Figure 1 1 . Load-Deformation Curve of Column with PAF Sheet Jacketing Failing in Flexure

zyxwvutsr

66 FRPRCS -6: Keynote Paper

zyxwv zyx

fracturing strain. The only solution is to use material with a high fracturing strain. Enhancement of Ultimate Deformation

The fracturing strain of Polyacetal Fiber is 6 to 9 %, which are 2 to 6 times of those in carbon, aramid and glass fibers as shown in Table 3. PAF can be used as both external and internal reinforcement to utilize its high fracturing strain. The internal reinforcement of PAF was already introduced as PCFFR in the previous section. PAF a s external reinforcement is in a sheet form for jacketing and bonding. A series of tests on reinforced columns with and without PAF sheet was conducted7. Three specimens PJ1, PJ2 and PJ3 that were identical except for PAF sheet ratio were prepared. PAF sheet ratios were 0, 0.146 and 0.291 % for PJ1, PJ2 and PJ3 respectively, while ratio of steel tie reinforcement was 0.151 %.

Photo 2. PAF Sheet for Jacketing at Ultimate Deformation without its Fracture

New Types of Continuous Fiber Reinforcements 67

Figure 11 indicates that the ultimate deformation of the specimens with PAF sheet is greater than that of the specimen without PAF sheet. In most of the specimens PAF sheet did not fracture at the ultimate deformation (see Photo 2), therefore the load-deformation curves show very ductile behavior even after the ultimate deformation. This phenomenon is not seen in specimens with carbon and aramid fiber sheet jacketing in which the sheet fracture causes a sudden drop of the load carrying capacity.

z z zyxwvu

Enhancement of Shear Strength

Another series of tests on the enhancement of shear capacity of columns by PAF sheet jacketing was conducted7. Three specimens T1, T2 and T3 identical except for the PAF sheet ratios that are 0.077, 0.153 and 0.29 1 % respectively were prepared. Two reference specimens P and C whose steel tie reinforcement ratio is 0.151 %, same as those in specimens TI, T2 and T3, were prepared, P without PAF sheet jacketing while C with carbon fiber sheet instead of PAF sheet. The carbon fiber sheet ratio is 0.017 %. Comparisons were made between specimens T1, T2 and T3 and specimen P; and between specimen T2 and specimen C. Because of its high fracturing strain the PAF sheet in the experimental specimens did not fracture when the shear capacity was reached. Specimens TI and T2 gave higher shear capacities than specimen C with carbon fiber sheet whose stiffness as

250 200

3 150 TI

8

-I

100

zyxwvu

50

0

zyxwvu 0

20

40

60

80

I00

Deformation(mm)

Figure 12. Load-Deformation Curve of Column with PAF Sheet Jacketing Failing in Shear

zyxwvut

68 FRPRCS -6: Keynote Paper

reinforcement (defined as product of reinforcement ratio and Young’s modulus) was similar to or greater than those of PAF sheet’ as shown in Figure 12. The carbon fiber sheet fractured at the peak load causing shear failure. While specimens T1, T2 and T3 show very ductile manner even after the peak load, specimens C and P show rather brittle manner that is the nature of shear failure.

zyxwvu zy z

Prevention of Total Collapse under VerticalLoads

Another type of fiber, a polyester fiber with even higher fracturing strain of around 20 % was applied to jacketing method’. The material constants are given in Table 3. Generally fiber materials with high fracturing strain indicate low Young’s modulus. Polyacetal and polyester fibers are no exception. The Young’s modulus of the former and the latter is 10 to 20 % and 1 % of that of carbon fiber respectively. The experimental study on polyester fiber sheet jacketing’ indicates that if the jacketing material does not fracture, at least the total loss of load carrying capacity for vertical load in columns can be prevented despite its extremely low stiffness. CONTINUOUS FIBER REINFORCEMENT BONDED WITH SOFT RESIN

For externally bonded continuous fiber reinforcement three types of resin are necessary; namely resin for primer, resin for adhesive and resin for impregnation. Recent studies found that usage of soft adhesive resin can improve the delamination capacity in both pullout bond test and beam test” 10

Bond Behavior Bond properties of CFR, such as average bond strength and local bond-slip relationship, were investigated by pullout test as shown in Figure 13. It is known that a higher stiffness of CFR gives a higher pullout force of the CFR externally bonded to concrete when delamination happens. Recent experimental results’ indicate another interesting fact that using adhesives with low shear stiffness, which is introduced by either increasing the thickness or decreasing of the elasticity of adhesives, can also improve the ultimate load transfer ability of CFR-concrete interfaces as shown in Figure 14. Decreasing the shear stiffness of adhesives reduces the interfacial strain distribution gradient as well as increases the effective bond length

zyxwvu zy zyx

New Types of Continuous Fiber Reinforcements 69 Load cell

t

7

/

zy zyxwvu zyxwvu

Steel plates attached to the both sides of CFR

4

I

CFR ‘

K

Bolts for fixing the concrete block on the base Vinylon tape

Enhancing anchorage bolts

Prestressed bolts

Hinges

‘”

Concrete block with four pre-set tubes

Steel basement for fconcrete block

ig floor

zyxw

Figure 13. Setup for Pullout Bond Test

significantly (see Figure 15). Unlike increasing the CFR stiffness, decreasing the shear stiffness of adhesives leads to a lower interfacial maximum bond stress and more ductile bond-slip behavior, although both ways increase the ultimate load transfer capacity. It should be noted that effects of adhesive resin are significantly different from those of impregnating resin. A lower stiffness of impregnating resin for CFR does not give a higher pullout capacity, instead it may cause fiber to fracture more easily.

70 FRPRCS -6: Keynote Paper zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONM &C F R P - 2 5 3 G P mm

+C F R P - 5 0

CFRP-759GPamm +AFRP-63lGPamm

%

6GPamm GFRP-262GPamm

zyxw

20 -.+---FPR fracture I 10 --zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFED 0

1

3

2

4

Ea Pa)

(a) Effects of adhesive’s elastic modulus

&CEXP-25

,--.60 50

-

-

z 4 030- :

20 10 -

3GPamm

+C F R P - 7 5 9 G P a m m

zyxwvuts +i zyxw zyxwvuts

0 -

-.

_--”

-

-1 I

*FRP

fki-

I

1

(b) Effects of adhesive’s thickness

Figure 14. Effects of Shear Stiffness of Adhesive on Pullout Bond Force

zyxw zyxwvut zyxwv

New Types of Continuous Fiber Reinforcements 71 zyxwvut

+ CFRP-253GPamm ,/ G F R P - 8

7GPamm + AFRP-18 5GPamm

!i

O

O

CFRP-50 6GPamm CFRP-756GPamm GFRP-262GPamm 0 GFRP-436GPamm AFRP-319GPamm AFRP-63 7GPamm

.

m

2

4

'Et/taGPa/mm8 ) zyxwvutsrqponm

Figure 15. Effects of Adhesive's Shear Stiffness on Effective Bond Length

Member Behavior

In beams with externally bonded CFR sheet or plate as flexure reinforcement, failure mode caused by delamination of CFR is often observed. Provision of better bonding characteristics of the interface can improve the performance of a strengthened member. In a previous experiment", a new adhesive resin with low Young's modulus and high deformability was applied between primer and FRP reinforcement. The CFR is carbon fiber sheet (CFS) with impregnating resin whose stiffness is 2.0 GPa. CFS used in this study is unidirectional strengthening zyxwvu

Strain gage on steel Strain gage on sheet >I zyxwvutsrqponmlkjihgfedcbaZYX 0

II> e +e

By considering the velocity discontinuities with the layer mechanisms, we get the other two sufficient conditions for collapse.

zyxwvu

414 FRPRCS-6: Externally Bonded Reinforcement for Flexure

Interface Mechanism

In the case of interface mechanism, velocity discontinuities is considered in interfaces. -3

U = 0 , t 2 = 0 and c l = O i n a .

zyxw zyxwvu

A sufficient condition for collapse is:

+2L2.rk: +-e2+e3 (41L)~:: e' +e2

IQI 2 ?[41Lr;;' e' +e2

e Z f e 3 2 2.3 +2 L 72c 1

(6)

e' +e2

Mixed Mechanisms

zyx zyxw

In the case of mixed mechanisms, the velocity discontinuities are considered in one layer and one interface. The mixed mechanism case concerning layer 1 and interface (2,3) are explored. It is assumed that the rate of generalised shearing strain between layers 1 and 2 is null.

e ' + e 2 2~

IQI'T-cosa + 2Yo"I

It

1+

e2 +e3

--I

-I

s i n a + N 1 2cosa+N12sina-

L

+ N:2t cosa 1-

*-yo

I-Yo

e2 +e3

(2LI -(I -yo) L) T y +-

el +e2

2

L

2.3

72c

L

(7)

1

When considering velocity discontinuity with a mixed mechanism, three other similar conditions sufficient for collapse are obtained. MODELLING AND COMPARISON WITH TEST RESULTS The tension zone in concrete below the neutral axis is neglected. An approximated method is used to calculate the depth of the neutral axis. Failure can occur with kinematic field discontinuity in layer 1 (compressive concrete ), layer 2 (steel) or layer 3 (CFRP strips). Failure can occur with

zy

RC Two-way Slabs with Composite Materials 415

zyxw

kinematic field discontinuity in interface (1,2), with separation of the external strengthening membrane (CFRP strips) from the concrete. Failure can occur also with kinematic field discontinuity in interface (2,3) called “peeling off failure”; where the whole thickness of cover concrete is removed. By considering a = 45”, eight possible collapse mechanisms and eight sufficient condition for failure are obtained. 2,3

2,3

zyxw

It was noted that z,c = zZc is the concrete shear stress strength, and is about 2.5 MPa. Also,

1,2

T , ~ = r,zmis

the shear stress strength at the

1.2

interface 1,2 in the x-direction, and zZc = r2z, is the shear stress strength

zyxwvu

at the interface 1,2 in the y-direction. The value r, is the section strength rate with CFRP strips in the x-direction and r2is the section strength rate with CFRP strips in the y-direction. For the present RC strengthened slabs, 5 = r2 = 25%, and z, is the shear stress strength at the bonding interface between concrete and CFRP strips, and is about 2.5 MPa. The calculated results of maximum capacity at each mechanism are given in table 1.

ble 1 Properties

Ta

Layer mechanisms

Mixed mechanisms

Interface mechanism

Mechanisms

Ultimate loads (KN)

I and 2 2 and 3 I and3 I and (2,3) 3 and (1.2) 2 and (1,2) 2 and (2,3) (1,2) and (2,3)

538 255 344 439 I86 123 341 3 74

zyxw

According to the tests, failure occurs with strips debonding and the ultimate load capacity is about 120 kN. The present model predicts failure with a mixed mechanism with CFRP strips debonding and steel yielding, and gives ultimate load capacity as equal to about 123 kN. A good agreement between theoretical and experimental results is therefore found. CONCLUSION

Results of the experimental study indicate that externally bonded CFRP plates can be efficiently used to strengthen two-way RC slabs. Limit analysis approach can predict correctly the ultimate load capacity of CFRP bonded RC slabs. This analysis is validated by comparison with test results.

416 FRPRCS-6: Externally Bonded Reinforcement f o r Flexure

zyxwv zy z zyx

The model presents eight possible collapse mechanisms including three layer mechanism, four layer mechanism and one interface mechanism. It also gives simple sufficient conditions and the ultimate load capacity for each collapse mechanism.

REFERENCES 1.

Meir U., “Bridge repair with high performance composite material.” Mater Technique, 1987;4: 125-8.

2. Teng J.G., Lam L., Chan W., Wang J., “Retrofitting of deficient RC cantilever slabs using GFRP strips.”, J. Comp. Constr. L 2000; 4(2): p. 75-84 3. Garden H.N., Quantrill R.J., Hollaway L.C., Thorne A.M., Parke G.A.R., “An experimental study of the anchorage length of carbon fibre composite plate used to strengthen reinforced concrete beams ”, Construction and building materials, 12(1998), pp 203-219. 4. Rabinovitch O., Forstig, Y., (( Strengtheneing of RC slabs with circular composite patches a high-order approach D, composite structures, p225238, ~ 0 1 5 5 , 2 0 0 2 5. Famiyesin, O.O.R., Hossain K.M.A., Chia Y.H., Slade P.A., (( Numerical and analytical predictions of the limit load of rectangular two way slabs D, 2001, Computes and Structures 6. Limam O., For&t G., Ehrlacher A., “RC beams strengthened with composite material: a limit analysis Approach and Experimental Study”, composite structures, 59 (2003) 467-472. 7. Philippe M., Naciri T. Ehrlacher A., “A tri-particle model of sandwich panels”, Composite Science and Technology, 1999, p. 1 195-1206. 8. Johansen, K.W., “Yield Line Theory”, Cement and concrete Association, London, 1962 9. Salengon J., (( Calcul a la rupture et analyse limite D, Presses de 1’E.N.P. C.

z

zyxwvu zyxwv zyxwvuts

FRPRCS-6, Singapore, 8-10 July 2003 Edited by Kiang Hwee Tan QWorld Scientific Publishing Company

EVALUATION OF EXTERNALLY BONDED CFRP SYSTEMS FOR THE STRENGTHENING OF RC SLABS K. Y. TAN, G. TUMIALAN AND A. NANNI Department of Civil Engineering University of Missouri, Rolla, USA Rolla, MO65409-0710

The use of carbon fiber-reinforced polymer (CFRP) composites as externally bonded reinforcement (EBR) for the repair and strengthening of deficient structures has been taking place since the late 1980’s. Continuous efforts in material development and research activities, with strong links to engineering practice, give this application more and more interest worldwide. This paper presents an experimental study on flexural strengthening of reinforced concrete (RC) slabs with different CFRP systems, using different EBR techniques. All the slabs were tested to failure under simply supported conditions. CFRP EBR increased the flexural strength and reduced the deflections and crack widths of the strengthened slabs. Two modes of failure, delamination and rupture of the CFRP reinforcement were observed

INTRODUCTION The wide acceptance and attractiveness of the externally bonded reinforcement (EBR) technique using epoxy-bonded plates can be attributed to the development of strong structural adhesives. The development of high strength-to-weight ratio, ease of fabrication and bonding and excellent resistance to electrochemical corrosion of fiber reinforced polymer (FRP) composites has given this technique even more acceptance worldwide. Over the last years, various types of FRP systems and EBR techniques have been developed and extended the possibilities of FRP EBR. In this experimental program, three different commercial products, which include pultruded laminate plates, fiber laminate sheets and pultruded laminate bars, were used to strengthen the RC slabs by four installation techniques. The techniques used were cold cured adhesive bonding, prestressing, wet lay-up and near surface mounted (NSM). All the strengthened slabs were tested to failure under simply supported conditions, subjected to a 6-point concentrated static loading system. A control slab was used as a baseline to compare the strengthened slabs.

418 FRPRCS-6: Exfernally Bonded Reinforcementfor Flexure

zyx

zyxw zyx '

EXPERIMENTAL PROGRAM Slab Details

A total of five slabs (1000 x 220 x 6300 mm) (39.4 x 8.6 x 248 in.) were fabricated and cured under normal laboratory conditions. All the slabs were reinforced in the longitudinal direction with two $10 mm (#3) and four $13 mm (#4) deformed steel bars, and in the transverse direction with $10 mm (#3) steel bars, spaced at 200 mm (7.9 in.) center-to-center. The minimum clear cover for the slabs was 30 mm (1.2 in.). (See Fig. 1.)

zyxwv

' L m / I 6.30

0.22

zy zyxwvu

LONGITUDINAL STEEL DISTRIBUTION

-1

''0°

1

f c = 27.6 MPa (4000psi) fy=413.7MPa(60ksi) Lenght = 6.3m(20.8ft) Dimensions in meters

T-.LD.I-2m,.nj 022

0.03

CROSS SECTION A-A

Figure 1 . Typical Cross Section of Slab

Three different CFRP systems and four different EBR techniques were used for the strengthening of the RC slabs. Table 1 summarizes the test matrix. The procedures for the installation of each system are described in the following section. All the strengthened slabs were tested after a curing period, under normal laboratory conditions, of 7 days after applying the adhesive.

zyxwvutsr zyx zy Table 1: Test Matrix

~

EBR Techniaues

CFRP Systems

Control

N/A

N/A

A

Cold cured adhesive bonding

2 strips of laminate plates

Prestressing

Zstrips of laminate plates

Manual wet lay-up

I ply of Fiber laminate sheets 8 strips of laminate bars

Slab

B C D

Near surface mounted

zyxwvu

CFRP Systemsfor the Strengthening of RC Slabs 419

Material Properties

zyx zy zyx

a)

Concrete. The compressive strengths of concrete for the test specimens

b)

Steel reinforcement. An average yield stress of 413.7 MPa (60 ksi) and

are presented in Table 2.

E-modulus of 200 GPa (29000 ksi) were obtained from tensile tests. c) CFRP systems. Table 3 shows the lower boundary mechanical properties of the CFRP systems provided by the manufacturer. No independent tests were performed to characterize the material. d) Adhesive. The epoxy gel adhesive and saturant were both two-part systems. After mixing, the epoxy gel adhesive had a paste-like consistency while the saturant had a liquid form. Table 4 shows the adhesive properties provided by the manufacturer. Table 2: Compressive Strength of Concrete at 28 days

zyxwvu zy

Slab

Compressive Strengths, MPa (psi)

Control

30.2 (4380)

A, B

33.8 (4900)

C, D

42.4 (6140)

Table 3: Material Properties of the CFRP Systems

CFRP System

Cross section, Af mm2 (in’)

Ef kN/mm2 (Mi)

Pultruted laminate 120 (0.186) 164 (23) plate Fiber laminate 240 (34.8) 117 (0.181) sheet* Pultruded laminate 112 (0.174) 164 (23) bar * Based on dryfiber cross-sectional area.

Tensile strength, N/mm2 (Ksi)

Ultimate strain, %

2500 (360)

1.6

3800 (550)

1.55

2900 (420)

1.8

&

Table 4: Adhesive Properties

ble 1 Properties ble 1 Properties

Ta Ta

Epoxy gel Saturant

69 (10 )

69 (10)

96.5 (14) 82.8 (12)

4.1 (600) 3.4 (500)

2 2

420 FRPRCS-6: Externally Bonded Reinforcement f o r Flexure

zyxw

INSTALLATION PROCEDURES

All the concrete surfaces were sandblasted and cleaned to ensure good bonding before strengthening. The adhesive was mixed in the specified ratio until a uniform and complete mixture was observed.

Slab A: Cold cured adhesive bonding Laminate Plate The epoxy gel was uniformly spread on the areas where the CFRP plates were to be placed. The CFRP plates were cut into the designed length and pressed into the wet epoxy gel using a hard roller. Air trapped between layers was rolled out before the epoxy gel sets. The thickness of the epoxy gel was approximately 1.5 mm (0.06 in.). (See Fig. 2a.)

Slab B: Prestressed Laminate Plate The installation of prestressing CFRP system started with the preparation of the moveable anchorage. This consisted of gluing one end of the CFRP plate between two steel plates, held in place by means of screws. After the moveable anchorage was cured for 24 hours, a fixed anchorage was installed and the CFRP plate was glued between the steel plate and the concrete surface. The steel plate was fastened to the concrete surface by means of an insert. The fixed anchorage was cured for another 24 hours before the CFRP plate could be stressed. While waiting for the fixed anchorage to cure, another fixed anchorage was attached, on the other end of the slab, to the concrete surface by means of an insert. Once the two fixed anchors were installed, the system was ready for stressing with a hydraulic jack. During the prestressing process, an epoxy gel was spread uniformly on all areas where the CFRP plate has contact. The thickness of the epoxy gel was approximately 1.5 mm (0.06 in.). Trapped air was released by rolling. Each laminate plate was stressed to an initial elongation of OS%, which represented 33% of the ultimate strain. After the epoxy gel cured, the moveable anchor was removed while the fixed anchors remained in place. (See Fig. 2b.)

zy zyxwvu

Slab C:Manual Wet Lay-up Laminate Sheet

An adequate layer of saturant was spread uniformly on all areas where the CFRP laminate sheet was to be placed. A single ply of CFRP laminate sheet was cut to design lengths and pressed down with a “bubble roller”. A second layer of saturant was reapplied to complete impregnation prior to cure. (See Fig. 2c.)

zyxw

zyxw zy

CFRP Systemsfor the Strengthening of RC Slabs 421

Slab D: Near Surface Mounted Laminate Tapes

Eight grooves approximately 3 mm (1/8 in.) wide and 15 mm (5/8 in.) deep, 12.6 cm (5.0 in.) center-to-center, were cut into the substrate of slab. The grooves were vacuum cleaned and then filled with saturant. The laminate bars were cut to design lengths and lightly pressed into the grooves. The grooves were refilled after part of the saturant was absorbed by the microcavities of the concrete. (See Fig. 2d)

(a) CFRF' Plates

(b) Prestressed CFRP Plates

zyxwvu

(c) CFRP Sheet

(d) NSM CFRP Tapes Figure 2: Test Specimens

TEST SETUP AND TEST PROCEDURE Two heavy-duty pin rollers were used to support the slab on a span of 6.0 m (20 ft.). The distance between each point load was 1.2 m (4.0 ft.). (See Fig. 3). A total of 5 linear variable differential transducers (LVDT' s) were placed at 1.5 m (5 ft.) from each other, starting from the supporting edge,

422 FRPRCS-6: Externally Bonded Reinforcement for Flexure

z

for displacement monitoring. Strain gages were used to measure strains on the CFRP systems and concrete. An 89kN (20 kip) capacity load cell was used to measure the applied loads. All the data from the electronic devices were recorded by a data acquisition system at a frequency of 1Hz. PI

l-l

150

120

zyxw zyxwvutsrq zyxwv zyx

PI

PI

+ ++ 1 50

I 50

I

I

1 50

41 2 0 4--120 &120 +120

Figure 3: Test Setup (dimensions in meter)

zyxwvu

TESTRESULTS Mode of Failure

A measure of the efficiency of the different CFRP EBR can be obtained by considering the modes of failure and the failure loads of the slabs. Fig. 4 illustrates the failure modes of the strengthened slabs and Table 5 summarizes the experimental results. The Control Slab exhibited a typical under-reinforced flexural failure. The test was discontinued after the steel yielded before the concrete crushed at a load of 5.60 kN (1.26 kips) due to excessively large cracks at the tension zone. This slab was used as a baseline to compare the remaining slabs. Slab A had a failure caused by delamination at a load of 13.7 kN (3.08 kips). As a reference, the theoretical failure load based on laminate rupture for Slab A was computed as 20.3 kN (4.56 kips), which is 48% larger than the experimental load. The introduction of initial prestressing provided Slab B with the ability to resist high loads prior to cracking. The cracks that developed in Slab B were fewer and finer as compared to Slab A. At a load of 20.7 kN (4.67 kips), sudden slippage took place at the fixed anchorage. The failure load for Slab B was close to the theoretical ultimate load, 22.7 kN (5.10 kips). In Slab C, a portion of the CFRP laminate sheet ruptured at a load of 21.3 kN (4.78 kips). After a brief time, the slab failed completely. No sign of debonding was observed at both ends of the slab. The experiment failure

CFRP Systems for the Strengthening of RC Slabs 423

zyx

load was 37% lower that the theoretical ultimate load, 29.2 kN (6.57 kips). This difference led to the conclusion that FRP rupture may have been caused by stresses concentration at the crack edges. Slab D reached the expected flexural capacity. The NSM laminate bars at the constant moment region ruptured at a load of 24.1 kN (5.43 kips). The test results positively proved that a good and uniform bond existed between the NSM laminate bars and the concrete.

(a) Slab A

(b) Slab B

(c) Slab C

(d) Slab D

Figure 4: Failure modes of the strengthened slabs.

zyx

The load vs. deflection curves for all the slabs are shown in Fig.5. The Control Slab started to yield after a load of 4.4 kN (1 .O kips) and continued to deform thereafter. All strengthened slabs responsed approximately linear before the concrete crack and with stiffnesses of about 84% greater than the Control Slab. At a load of 13.7 kN (3.08 kips), Slab A failed suddenly and exhibited low ductility. Due to the prestressing in Slab B, lower deflections were observed at the same level of loads as compared to Slab A. The plateau formation after 17.8 kN (4.0 kips) was interpreted as the beginning of delamination and slippage at the fixed anchorages.

zyx zy zyxwvu zyxw

424 FRPRCS-6: Externally Bonded Reinforcement for Flexure

The CFRP laminate sheet had an axial stiffness, E& of approximately 1.5 times larger than the other CFRP systems. The influence of the Efifin the stiffness of the strengthened slab was clearly observed in Slab C. Slab C had a stiffness similar to Slab B before the steel yielded. The CFRP laminate bar have a similar Eflfwith CFRP laminate plate, hence, Slab D and Slab A had a similar stiffness and behavior after initial cracking.

zyxw

Table 5: Test Result

Slab

Max. applied load, P kN (Kip)

E//EJ, (%)

Failure Increment, moment *, (%) kN-m (lc-ji)

Normalized Increment * *

(1) Control

(2)

(3) 5.6 (1.26)

(4) 46.8 (34.5)

(5)

__

(6)

--

A B

15.3 15.3

13.7 (3.08) 20.8 (4.65)

76.3 (56.3) 102.2 (75.4)

63 118

4.1 7.7

--

C 21.8 21.3 (4.78) 104.0 (76.6) 122 D 14.2 24.1 (5.43) 114.6 (84.5) 145 *Include slab self-weight. ** Column 5 divided by column 2

0

Midspan Deflection (in) 6 8 1 0

2

4 I

1

,

25

I

//

5.6 10.2

2

!I finnn ----

15000

4000

10

Control Slab SlabA SlabB SIabC

e)

2000

+ 5

zyxwvu zy -.--&--

0

7

o,o

-

-ii

10 l5 20 Midspan Deflection (crn)

25

Figure 5: Load vs. Deflection Curves

~looo

30

zyxw zyx zy zyxw

CFRP Systemsfor the Strengthening of RC Slabs 425

zyxw

The load vs. strain curves of Slab A, B, C and D are presented in Fig. 6. The following is observed: a) The concrete strains for all the specimens were less than 0.003. b) The CFRP laminate plate at Slab A delaminated when the strain at the mid span reached 0.56%, which was 35 % of the ultimate strain. c) The strains along the prestressed CFRP plate at Slab B started with an initial elongation of 0.5%. Strain gages close to the anchorage showed a dramatic increase at 18.5 kN (4.15 kips), indicating that delamination had reached the edge of the fixed anchorages. d) Slab C had strains patterns similar to Slab A. c) The CFRP bars in Slab D were uniformly stressed along the slab until they reached the ultimate strain.

Smm(l0.6)

(d) Slab D

Figure 6 : Load vs. Strain Curves

CONCLUSIONS

The following conclusions can be drawn from this experimental program: a) Significant increases in flexural capacity ranging from 63% to 145% were registered in all strengthened slabs as compared to the Control Slab.

.

426 FRPRCS-6: Externally Bonded Reinforcement f o r Flexure

b) During the test, it was observed that the CFRP EBR delayed the presence of the first visible cracks and reduced the deflection. c) The slab strengthened with CFRP plates failed due to delamination starting from the constant moment region and propagated towards the cutoff points. d) Prestressing of the CFRP laminate plate had a positive influence on the behavior of strengthened RC slab. The load capacity was substantially increased and the deflection and crack formations were substantially reduced. e) Premature failure at fiber laminate sheet was due to high stresses concentration at crack locations . f ) The slab strengthened with NSM bars exhibited a behavior such that CFRP reinforcement was fully utilized.

zyxw zyxw zyxw zyxwv zyxwv zy

ACKNOWLEDGMENTS

The authors would like to acknowledge the support of the S&P Clever Reinforcement Company. REFERENCES

.

1. Taerwe L., Matthys S., Pilakoutas K. and Guadagnini M., “European Activities on the Use of FRP Reinforcement, fib Task Group 9.3 and ConFibreCrete Network”, 5Ih International Conference on FRP Reinforced Concrete Structures (FRPRCS-5), Cambridge, UK, July 1618, 2001, Vol. 1, pp. 3-1 5 . 2. Triantafillou T., Matthys S. and Taerwe L., “Design of Concrete Members Strengthened with Externally Bonded FRP Reinforcement” 5Ih International Conference on FRP Reinforced Concrete Structures (FRPRCS-5), Cambridge, UK, July 16-18,2001, Vol. 1, pp. 157-166. 3. American Concrete Institute International, “Guide for The Design and Construction of Externally Bonded FRP Systems for Strengthening Concrete Structures ”,ACI-440.2R-02, Farmington Hill, Michigan. 4. S&P Clever Reinforcement Company, “Guide Line for S&P FRP Systems”, Brunnen, Switzerland, June 2000.

zyxwvu zyxwvu zyxwvu

FRPRCS-6, Singapore, 8-10 July 2003 Edited by Gang Hwee Tan @World Scientific Publishing Company

zyxw zyxwvu zy

FLEXURAL STRENGTHENING OF TWO-WAY SLABS USING FRPs H. MARZOUK Faculty of Engineering, Memorial University of Newfoundland 300 Prince Philip Dr., St. John’s, Newfoundland, Canada, A I B 3x5

U.A. EBEAD AND K.W. NEALE Department of Civil Engineering, University of Sherbrooke 2500 boulevard de I’UniversitC, Sherbrooke, QuCbec, Canada, J1K 2R1 Strengthening of two-way slabs using Fibre Reinforced Polymer (FRP) materials is presented. The behaviour of two-way slabs strengthened in flexure is discussed. Carbon FRP strips and glass FRP laminates can be used to increase the flexural capacity of two-way slabs to an average of 36% over that of the reference (un-strengthened) specimen. An increase of the initial stiffness was achieved for flexural specimens; however, an apparent decrease in the overall ductility was evident. A finite element analysis of the flexural-strengthened two-way slabs is also discussed. An incremental elastic-plastic concrete model is implemented. In compression, the concrete model is elastic until a yield point is reached after which irrecoverable plastic strain exists. Pre-cracking and post-cracking behaviours of concrete are considered in the study with special emphasis on the impact of the FRP materials on the concrete fracture energy and hence on the concrete tension stiffening. A full bond is assumed between concrete and the steel and FRP materials. The comparison between the finite element analysis results and the experimental results shows a good agreement.

INTRODUCTION Increasing attention has been placed to the applications of advanced composite materials especially glass fiber reinforced polymer (GFRP) laminates and carbon fiber reinforced polymer (CFRP) strips in the structural engineering field. There is a wide range of recent, current, and potential applications of these materials that cover both new and existing structures. The flexural capacity of concrete beams can be increased by bonding FRP sheets, strips or laminates to the tension side.’,* In addition, shear resistance of beams can be increased by using GFRP laminates

428 FRPRCS-6: Externally Bonded Reinforcement for Flexure

wrapped over three sides of beams at locations of high shear s t r e ~ s e s A .~ large number of research works have dealt with the de-bonding of FRP sheets to concrete beams.4 Some mechanical and finite element models have been developed to provide design guidelines and to investigate theoretically possible modes of failure of FRP strengthened beams based on experimental data.5x6Limited research work has been conducted on the strengthening of reinforced concrete slabs especially two-way slabs using FRP materials. Some research works dealt with the strengthening of one-way slabs using FRP materials in which slabs were treated in a very similar way to beams.7

zyxwvu

EXPERIMENTAL PROGRAM Materials

One m3 of concrete contains 1160 kg of gravel, 690 kg of sand, 350 kg of cement, and 175 liters of water. The compressive strengths of concrete at the day of the test ranged from 30 MPa to 35 MPa. The steel reinforcement bars were CSA grade 400 deformed bars. Unidirectional GFRP laminates and CFRP strips were used as strengthening materials. CFRP strips have a thickness equal to 1.2 mdlayer, tensile strength of 2800 MPa, and elastic modulus of 170 GPa. In addition, GFRP laminates have a thickness of 1 mdlayer, tensile strength of 600 MPa, and elastic modulus of 26.13 GPa. As per the manufacturer specifications, CFRP strips and GFRP laminates use different two-component epoxy adhesive resins. Test slabs and setup

zyx

The tested simply supported specimens were square with 1900-mm side length and 150-mm thickness as shown in Figure 1. Column stubs were square of 250-mm side dimension and were located at the slab center. Two un-strengthened specimens were used as reference specimens, Ref-P-0.35% and Ref-P-0.5% of reinforcement ratios of 0.35% and 0.5%, respectively. Specimens CFRP-F-0.35% and CFRP-F-0.5% and Specimens GFRP-F0.35% and GFRP-F-0.5% were strengthened using GFRP laminates and CFRP strips, respectively. A hydraulic actuator of 700 kN capacity facing the specimen was used to apply a uniform central load through the column stub. A load cell was used to measure the load using four calibrated electrical resistance strain gages fixed to the inner cylinder of the load cell. Linear Variable Displacement Transformers (LVDT’s) were built in the actuator to measure the central deflection of the slabs. The central loads

zyx zyx

Flexural Strengthening of Two-way Slabs 429

were applied using displacement control through a computerized function generator with a rate of 0.25 m d m i n . Strengthening and Loading Procedure

The concrete surface to be strengthened was roughened carefully using a vibrating hammer. Dust and bond inhibiting materials were removed carefully from the concrete and FRP surfaces. The epoxy resin was applied on the concrete and FRP surfaces before bonding. FRP materials were located at the tension side of the slab according to the configuration shown in Figure 1. Fifty percent of the ultimate load carrying capacity of the reference specimens was used as an initial loading for the specimens prior to strengthening. The applied loads were completely removed to represent a state of shoring two-way slabs in the field prior to strengthening. Afterwards, the specimens were removed from the loading frame for strengthening. After one week of curing, the specimens were relocated in the loading frame and were subjected to the central load until failure. K

'x

zy zyxw

..mAddinanal FRF'stnps layer

zyxwv Figure 1: Layout of the flexural-strengthening scheme

TEST RESULTS AND DISCUSSION Ultimate Load Carrying Capacity

Strengthening increases the load capacity of the slabs. Specimens CFRP-F0.35% and GFRP-F-0.35% showed an increase of 44.4% and 38%, respectively in the load capacity over that of the reference specimen, Ref-P-

zyxwv

430 FRPRCS-6: Externally Bonded Reinforcement f o r Flexure

0.35%. Moreover, Specimens CFRP-F-0.5% and GFRP-F-0.5% showed an increase of 36.4% and 25.8 %, respectively in the load capacity over that of the reference specimen, Ref-0.5% as shown in Table 1. Table 1: Experimental test results Title

Cracking load, KN 73 84 70 68 80 83

Ref-0.35% Ref-0.5% CFRP-F-0.35% GFRP-F-0.35YO CFRP-F-0.5% GFRP-F-0.5%

Defl. at cracking load, mm 7.00 6.25 7.25 7.69 6.03 6.35

Ultimate load, KN 250 330 361 345 450 415 ~

~~

Defl. at ultimate load, mm 42.0 1 35.57 18.08 27.72 21.03 26.71

Deformational Characteristics The average deflection at the ultimate load of the strengthened specimens was about 0.63 that of the corresponding reference specimens. In general, the strengthened specimens experienced smaller deformation compared to the corresponding reference specimens due to the impact of the brittleness of FRP materials on the overall behaviour of the slabs. Table 1 summarizes the deflection values at first crack load (at the un-strengthened stage) and at the ultimate load for the tested specimens. Figure 2 shows the loaddeflection relationship for the tested specimens.

f"oO

--C

0

5

10

15

zyxwv

zyxw CFRP F 05%

zyxwvuts zyxwvuts 25 30 Deflectionh r n l

20

YI

10

45

I

Figure 2: Load deflection relationship for some of the tested slabs

Failure characteristics The failure mode of the reference specimens was classified as flexuralductile. Flexural reinforcement yielded and the specimens showed relatively large deflection values before reaching the ultimate load. Figure 3 shows a typical failure mode for a specimen strengthened using CFRP for flexural

zyx

zy

Flexural Strengthening of Two-way Slabs 431

and shear strengthening, respectively. FRP materials contributed to increasing the capacity until the bond between the FRP material and concrete had failed. De-bonding cracks appeared at a late stage of the loading causing a separation of the FRP materials as shown in Figure 3. The specimens failed due to accelerated flexural failure following the FRP debonding. Neither CFRP strips nor GFRP laminates fractured.

Figure 3: A tested specimen after failure (CFRP-F-0.5%)

FINITE ELEMENT ANALYSIS Material Modelling

A plasticity-based concrete constitutive model was used in this study'. The model utilizes the classical aspects of the theory of plasticity. A complete representation of the model is defined by considering the following concepts: strain rate decomposition into elastic and inelastic strain rates; elasticity; yield; flow; and hardening. The concrete model in compression is elastic until the initial yield surface limit is reached as shown in Figure 4. The initial yield surface defines the elastic limit at which the linear-elastic constitutive relationships are valid. Further stresses of concrete cause an expansion of the initial yield surface so that new yield surfaces are developed. The constitutive concrete model addresses the tensile behaviour of concrete by considering several aspects. These aspects are cracking, shear modulus degradation, fracture energy, and tension stiffening. Cracking is considered the most significant factor of the concrete tensile behaviour. In the case of plain concrete, the fracture energy, G,. , is defined as the energy

zyx

432 FRPRCS-6: Externally Bonded Reinforcement f o r Flexure

zyxwv z zyxw zyxw zy

required to form a unit area of crack surface and is considered a material property based on the brittle fracture concept of Hillerb~rg.~ The fracture energy,G,, is estimated as the numerical integration of the function

between the tensile stress, o t ,and the “crack width” or displacement, uI, for the post-peak zone, i.e.:

zyxwvu

Gf = Jo,du

Compression frulure envelope (surface)

w,f:

i

D

z

‘ -Bt-axial

(1)

zyxwvutsrqp n c6mpresslon

Uni-axial

-

tension

ension

surface

Figure 4: Concrete in compression

Figure 5: Concrete in tension

In the case of reinforced or strengthened concrete, the calculations are made based on the assumption of a smeared crack approach. Cracks exist in reinforced concrete subjected to tensile stresses along with the steel reinforcement. Subsequently, interfacial shear stresses between the concrete and the reinforcement are transmitted to the concrete between cracks as tensile stresses. Hence, concrete bonded to the reinforcement is loaded with tensile stresses causing an increase of the overall stiffness. This phenomenon is referred to as tension stwening. The numerical integration of the o,- E, curve can be referred to as the fracture energy density, Wf.lo Concrete behaviour in tension is linear until the cracking stress is reached. The post-peak zone can be defined using broken lines as shown in Figure 5. It has been decided to define the tension stiffening of concrete by considering only two points on the post-peak zone of the of- E , relationship. The FW-concrete interaction is assumed similar to that of steel reinforcement-concrete interaction. Hence, F W materials are defined as smeared external reinforcement located at the tension side of the slab.

zyx zy zyxwvu zyxwvu zyxwv Flexural Strengthening of Two-way Slabs 433

The concrete crl - E , relationship, and hence the fracture energy density are calibrated. The calibration is based on the agreement of the FEA results and the available experimental results. The agreement is achieved after several FEA implementations for different values of the fracture energy density. This calibration is conducted with respect to the ultimate load carrying capacity of the slabs. For each implementation, the fracture energy density of concrete strengthened with FRP materials is calculated as follows:

zy zyxw &l,mSX

Wr=

zyxwvutsrqponm (2)

0

For the assumed tension stiffening model, the fracture energy density can be calculated as follows:

Wf=3El,may Material Properties

0,"

(3)

The modulus of elasticity of concrete, E,, is calculated as 26600 MPa. The equal biaxial strength of concrete is assumed 1.16 times that of the uniaxial strength of concrete.l 1 The yield stress of concrete is assumed 20 MPa. The tensile strength of concrete, or , is assumed 0.08 times the uniaxial strength of concrete that is equal to 2.8 MPa." The post-peak o,- E, relationship of strengthened or un-strengthened concrete is assumed linear descending to . 'The steel reinforcement is zero tensile stress at maximum strain

assumed to have a yield stress of 440 MPa and a modulus of elasticity of 210 GPa. The assumption of the full bond between FRP materials and concrete is inherited by the definition of these materials as smeared reinforcing layers located at the tension side of the concrete slabs.

Geometric Modelling One quarter of the slab is modelled due to the geometrical and loading symmetry using a 5 x 5 mesh. The general layout of the finite element model is shown in Figure 6 . Degenerated 8-node quadrilateral shear-flexible shell elements with six degrees of freedom at each node are used for modelling the slab. The degrees of freedom are three translations and three rotations. This permits the transverse shear deformation to be accounted for. Nine Simpson-type integration points are used along the thickness of each shell element. In addition, a reduced 2 x 2 Gaussian integration rule is used over the X - Y plane of the elements. Eight-node brick elements are used to

434 FRPRCS-6: Externally Bonded Reinforcement f o r Flexure

represent the column stub through which the load is applied. The brick element has three translational degrees of freedom per node in the X, Y , and Z directions. The discrepancy between the degrees of freedom of the column stub brick element and the panel shell elements is overcome using the Multi Point Constraints (MPC) technique. The MPC technique allows constraints to be imposed between different degrees of freedom in the model. Non-linear spring elements define the simply supported with corners free-to-lift boundary condition as in Figure 6 .

Steel Reinforcement and FRP Representation The slab reinforcement is treated as smeared unidirectional layers. These layers are embedded in concrete and located at the centerline of the actual reinforcing bars in the slabs. The layers are smeared with a constant thickness that is equal to the area of each reinforcing bar divided by the reinforcing bars spacing. CFRP strips and GFRP laminates are represented in a similar way to the rebars. FRP materials are treated as smeared unidirectional layers located at the tension surface of concrete. The definition of FRP materials as smeared reinforcement inherits the assumption of full bond with the concrete surface. In addition, the impact of steel reinforcement and FRP materials on the tensile properties of concrete is modelled through the suggested tension stiffening relationship for the FRP strengthened portions of the slab. For the un-strengthened portions of the slab, the tension stiffening model is used as was recommended by Marzouk and Chen.”

zyxwv zyxw

RESULTS OF THE FINITE ELEMENT ANALYSIS Load Carrying Capacity and Deformation

The assumptions of the fracture energy density, W, led to a good agreement between the FEA and the test results in terms of the load carrying capacity as shown in Table 2. The FEA underestimated the values of the deflections for either the strengthened and un-strengthened specimens. The FEA gives a stiffer deformational behaviour compared to the experimental results.

Flexural Strengthening of Two-way Slabs 435

zyxwv

Figure 6: Geometric model layout Table 2: FEA results Title

Pexw

PtheoAW

.P e x d Ptheo

zyxwvutsrqpon CFRP-0.5% GFRP-O.S?'o

450 415

424 416

1.06 1 .oo

S J MMARY AND CONCLUSIONS

The use of strengthening CFRP strips and GFRP laminates with the suggested dimensions were sufficient to achieve positive results for flexural-strengthening of slabs. The strengthened specimens using FRP strips or laminates showed an average gain in the load capacity of about 36% over that of the reference (un-strengthened) specimens. In addition, the strengthened specimens showed a stiffer behaviour than that of the reference specimens. However, a decrease in ductility and energy absorption was recorded due to the brittle nature of the strengthening of the FRP materials. For the suggested strengthening technique, de-bonding between FRP materials and concrete was the main cause of failure. Slabs failed soon after de-bonding occurred due to exceeding flexural capacity. None of the strengthening material type experienced rupture or failure. A finite element model was used to analyze strengthened two-way slabs. The finite element results are calibrated so that a good agreement

zyxw

zy zyxw zyxw

436 FRPRCS-6: Externally Bonded Reinforcement for Flexure

with the experimental results is achieved. The full bond between the steel reinforcement or FRP materials and concrete can be assumed in the analysis and lead to reasonably accurate results with low computational cost. REFERENCES

1. Chaallal, O., Nollet, M. and Perraton, D. “Strengthening of Reinforced Concrete Beams with Externally Bonded Reinforced Plastic Plates: Design Guidelines for Shear and Flexure”, Canadian Journal for Civil Engineering, 25, 1998, pp. 692-704. 2. Ritchie, P., Thomas, D., Lu, L. and Connelly, G., “External Reinforcement of Concrete Beams Using Fiber Reinforced Plastics”, ACI Structural Journal, 88(6), 1991, pp. 490-500. 3. Triantafillou, T. C., “Shear Strengthening of Reinforced Concrete Beams Using Epoxy-Bonded FRP Composites”, ACI Structural Journal, 95 (2), 1998, pp. 107-1 15. 4. Meier, U., Deuring, M., Meier, H. and Schwegler, G., “CFRP Bonded Sheets”, In Proceedings, Fiber-Reinforced-Plastic (FRP) Reinforcement for Concrete Structures: Properties and Applications, Duebendof, Switzerland, 1993, pp. 423-434. 5 . Nitereka, C. and Neale, K., “Analysis of Reinforced Concrete Beams Strengthened in Flexure with Composite Laminates”, Canadian Journal for Civil Engineering, 26, 1999, pp. 646-654. 6. Malek, M. A., Saadatmanesh, H. and Ehsani, M. R., “Prediction of Failure Load of R/C Beams Strengthened with FRP Plate Due to Stress Concentration at The Plate End”, ACI Structural Journal, 95 (l), 1998, pp. 142-152. 7. Kikukawa, K, Mutoh, K., Ohya, H., Ohyama, Y . and Tanaka, H., “Flexural Reinforcement of Concrete Floor Slabs by Carbon Fiber Textiles”, Composite Interfaces, 5 (5), 1998, pp. 469-478. 8. Hibbitt, K. and Sorensen., “ABAQUS Users Manual (Version 6.2.)”, Providence, R. I.: Hibbitt, Kalrsson and Sorensen Inc., 2001. 9. Hillerborg, A., “Numerical Methods to Simulate Softening and Fracture of Concrete”, Fracture Mechanics of Concrete, 1985, pp. 141- 170. 10. Marzouk, H. and Chen, Z., “Finite Element Analysis of High Strength Concrete Slabs”, ACI Structural Journal, 90(5), 1993, pp. 505-5 13. 11. Hussein, A. and Marzouk, H., “Behavior of High Strength Concrete under Biaxial Stresses.” ACI Structural Journal, 1998,97( l), pp. 27-36.

zy

zyxwvutsrq zyxwvu

FRPRCS-6, Singapore, 8-10 July 2003 Edited by Gang Hwee Tan @WorldScientificPublishing Company

TENSILE PROPERTIES OF CONCRETE IN FRP-STRENGTHENED TWO-WAY SLABS

zyxwvu zyxwv zyx

H. MARZOUK Faculty of Engineering, Memorial University of Newfoundland 300 Prince Philip Dr., St. John's, Newfoundland, Canada, A l B 3x5

U.A. EBEAD AND K.W. NEALE Department of Civil Engineering, University of Sherbrooke 2500 boulevard de I'Universite', Sherbrooke, Que'bec, Canada, J I K 2RI

Reinforced concrete behaviour in tension can significantly be changed due to strengthening. An overall increase in the post-peak stiffness based on the tensile stress-strain relationship is observed. A simplified bilinear model is introduced to define the behaviour of FRP-strengthened concrete in tension. An expression of the fracture energy density is introduced to define the area under the concrete tensile stress-strain relationship. The tensile stress-strain relationship of concrete is referred to as the tension-stiffening model. It is shown numerically that the ultimate load capacity of two-way slab specimens is sensitive to the fracture energy density. Hence, a distinction has to be made between the definitions of the tension-stiffening model of FRP-strengthened and unstrengthened concrete. This distinction is the focus of this paper.

INTRODUCTION

This paper presents a finite element analysis (FEA) aimed at investigating the effect of FRP strengthening on the tensile properties of concrete. The experimental results of the strengthened slabs are used to calibrate the finite element model based on the ultimate load carrying capacity as presented in reference 1. The calibration FEA study simulates six specimens tested experimentally'. The tested simply supported specimens are square with 1900-mm side length and 150-mm thickness as shown in Figure 1 . Column stubs are square of 250-mm side dimension and are located at the center of the slab. Two unstrengthened specimens are used as reference specimens, Ref-P-0.35% and Ref-P-0.5% of steel reinforcement ratios of 0.35% and 0.5%, respectively. Specimens CFRP-F-0.35% and CFRP-F-0.5% and specimens GFRP-F-0.35% and GFRP-F-0.5% are strengthened using GFRP

zyxw zyxw

438 FRPRCS-6: Externally Bonded Reinforcement f o r Flexure

laminates and CFRP strips, respectively. Figure 1 shows the strengthening configuration of the slabs. Details of the experimental program are described in reference 1. The FEA study is carried out using the generalpurpose finite element code ABAQUS.3

12-lOmm

A

zyxwvu zyxw 8-1Omm

I

300 1830

I

1830

I

I 4

I

Figure 1: A schematic representation of a strengthened specimen

FRP Strengthened Two-way Slabs 439

z

CONCRETE CONSTITUTIVE MODEL A plasticity-based concrete constitutive model is used in this study. Details of the concrete constitutive model are described in reference 2. The constitutive concrete model addresses the tensile behaviour of concrete by considering several aspects. These aspects are cracking, shear modulus degradation, fracture energy and tension-stiffening. In this study, an emphasis is placed on the appropriate tensile behaviour of concrete in two-way slabs strengthened with externally-bonded FRPs. Concrete bonded to the reinforcement or FRP materials is loaded with tensile stresses causing an increase of the overall stiffness. In addition, the distribution of cracks in the concrete is dependent on whether the concrete is plain or reinforcedstrengthened. The term tension-stiffening is introduced in the finite element analysis to consider the effect of the steel or FRP reinforcement on the concrete tensile behaviour. In the case of plain concrete, the fracture energy, Gr , is defined as the energy required to form a unit area of crack surface and is considered a material property based on the brittle fracture concept of Hillerb~rg.~ The fracture energy,Gf, is estimated as the numerical integration of the

zyx zyx z zyx zyxwvu zyxwvu

function between the tensile stress, cr,, and the “crack width” or displacement, uI, for the post-peak zone of the

0 ,-

u, relationship, that is.,

Gr = 10,dU

In finite element simulations that adopt the smeared crack approach like the one in this study, the tensile stress-strain relationship rather than stressdisplacement relationship is referred to when describing the concrete tensile behaviour. Some expressions have been developed to correlate the post-peak stress and strain for concrete in tension. Based on some experimental evidence on high strength concrete4, it was found that the post-peak relationship is referred to may be defined according to the following relationship: For E , & , ~

where:

zyxw zyxwv zyxw

zyxwvu z zyxwvutsrq

440 FRPRCS-6: Externally Bonded Reinforcement f o r Flexure

a =c30,"

where E, is the concrete tensile strain and E

(3) , is ~ the concrete tensile strain at

cr = 0,". In addition, the value of c3 is 0.3 1 for normal strength concrete'

and modified to 0.28 for high strength concrete4. Also, p is equal to 1.70 for normal strength concrete and 1.67 for high strength concrete. In this analysis, the effect of the reinforcement ratio as well as whether the concrete is strengthened or not is considered. The post-peak zone can be defined using line segments rather than a continuous relation~hip.~.~ A tabulated form for the values of the tensile stress,o, ,and the tensile strain, E , , can be used to define the tension-stiffening model. It has been decided to define the tension-stiffening of concrete by considering only two points on the post-peak zone of the 0,- E , relationship as shown in Figure 2. The numerical integration of the concrete tensile stress-strain ( 0, - E, )

zyxwv z

relationship represents the fracture energy density, Wr , and can be calculated as follows:6 6,IIl.X

zyxwvutsrqpon

W f =J o , ~ E ,

(4)

0

where

E,,,,

is the tensile strain when the tensile stress vanishes.

Denoting the maximum tensile stresso: and based on this approach, the fractur; energy density, W,can be calculated as follows:

With respect to the properties of the materials, the modulus of elasticity of concrete, E,, is calculated as 26600 MPa. The equal biaxial strength of concrete in compression is assumed 1.16 times that of the uniaxial compressive ~ t r e n g t hThe . ~ yield stress of concrete is assumed 20 MPa. The tensile strength of concrete, 0: , is assumed 0.08 times the uniaxial strength of concrete, that is equal to 2.8 MPa.4 The post-peak 0, - E, relationship of strengthened or unstrengthened concrete is assumed linear descending to zero tensile stress at the maximum strain E , , , ~ . The steel reinforcement is assumed to have a yield stress of 440 MPa and a modulus of elasticity of

zyxw zy

FRP Strengthened Two-Way Slabs 441

zy

210 GPa. The assumption of the full bond between FRP materials and concrete is implicit by defining these materials as smeared reinforcing layers located at the tension side of the concrete slabs. Details of this stage of the finite element analysis can be found in reference 1 including the geometric modeling and steel and FRP materials representation.

I

zyxwvutsrqpon zyxw Attempt 1 Attempt 2 Attempt 3 Attempt 4

Tensile strain, E t

zyx

Attempt 5

zyx

Figure 2: Tension-stiffening model

In this FEA study, the

0,

-

E,

relationship, and hence the fracture

energy density, W, are calibrated. Values of ,m are assumed upon which the fracture energy density, W, is calculated according to Equation 5 and altered accordingly. The calibration is based on the agreement of the FEA results and the available experimental results. Several attempts are implemented in the finite element code by altering the definition of the tension-stiffening model. Altering the definition of the tension-stiffening model is done by changing the values of&,,,, and hence the area under the bilinear relationship that is the fracture energy density, W,. This calibration is conducted with respect to the ultimate load carrying capacity of the slabs. Table 1 shows the FEA calibration implementation of the attempts for different values of E,,,, . This table shows a comparison between the ultimate capacity of the slabs based on the experimental testing, P,, and on the FEA, PFEA.

zyxw zyx ble 1 Properties 442 FRPRCS-6: Externally Bonded Reinforcement f o r Flexure

Table 1: The FEA calibration runs

Ta

ble 1 Properties ble 1 Properties ble 1 Properties ble 1 Properties ble 1 Properties ble 1 Properties ble 1 Properties

Ta

Ta

Ta

Ta

Ta

Ta

Ta

zyxw zyx

zyxwv zyxw zy FRP Strengthened Two-way Slabs 443

Tensile Stress-Strain Relationships

Figures 3 and 4 show the tensile stress-strain relationships for specimens of the calibration study at the center of the slab based on the FEA. It is clear that, due to the contribution of FRP strengthening materials, the post-peak behaviour of slabs is stiffened. The slope of the tensile stress-tensile strain is decreased in the post-peak zone indicating the contribution of the FRP strengthening materials in increasing the post-peak stiffness of concrete in tension. The stiffened post-peak tensile stress-strain relationship leads to higher values of fracture energy density, W, within a certain range of the strain. This complies with the initial assumptions of the fracture energy density shown in Table 1. 2.5

-

2

a

r_ 2

zyxwvutsr +GFRP-F-0.35%

1.5

g

2

1

c I!?!

0.5

a

0.001

0.002

0.003

0.004

0.005

Tensile strain

Figure 3: Tensile behaviour for GFRP strengthened slabs at the slab center 2.5 -

2

a

z3 1.5 5 .-s 1 Lo

C

I-

0.5 0 0

0.001

0.002

0.003

0.004

0.005

Tensile strain

Figure 4: Tensile behaviour for CFRP strengthened slabs at the slab center

444 FRPRCS-6: Externally Bonded Reinforcementfor Flexure

zyx

SUMMARY AND CONCLUSIONS In this paper, a finite element analysis is presented. A calibration study is conducted on a finite element model and is used to analyze FRPstrengthened two-way slabs. The finite element results are calibrated so that a good agreement with the experimental results is achieved. An FRP tension-stiffening model is recommended to predict the complete behaviour of concrete in tension. The recommended model describes the tensile behaviour of concrete slabs strengthened using FRP materials. The model takes a form of a bilinear relationship between the tensile stress and tensile strain. It was found that a distinction had to be made between the plain, reinforced and strengthened concrete tensile behaviour when defining the tension-stiffening model. FRP-strengthened concrete exhibits a stiffer postpeak response than conventional reinforced concrete. REFERENCES 1.

2. 3. 4. 5. 6. 7.

Ebead, U., Marzouk, H. and Neale, K.W., “Flexural strengthening of two-way slabs using FRPs”, 6’ International Symposium on FibreReinforced Polymer Reinforcement for Concrete Structures (FWRCS6), 2003. Hibbitt, K. and Sorensen., “ABAQUS users manual (Version 6.2)”, Providence, R. I.: Hibbitt, Kalrsson and Sorensen Inc., 2001. Hillerborg, A., “Numerical methods to simulate softening and fracture of concrete”, Fracture Mechanics of Concrete, 1985, pp. 141-170. Marzouk, H. and Chen, Z. M., “Fracture energy and tension properties of high strength concrete”, Journal of Materials in Civil Engineering, 1995,7(2), 108-1 16. Guo, 2. and Zhang, X., “Investigation of complete stress-deformation curves for concrete in tension”, ACI Materials Journal, 1987, 84(4), pp. 278-285. Marzouk, H. and Chen, Z., “Finite element analysis of high strength concrete slabs”, ACI Structural Journal, 90(5), 1993, pp. 505-5 13. Hussein, A. and Marzouk, H., “Behaviour of high strength concrete under biaxial stresses”, ACI Structural Journal, 1998,97( l), pp. 27-36.

zyxwvu

ExternallyBo Bonded Reinforcement for Shear Externally

z

This page intentionally left blank

zyxwvutsr zyxwv

FRPRCS-6, Singapore, 8-10 July 2003 Edited by Kiang Hwee Tan @World Scientific Publishing Company

zy

SHEAR CRITICAL REINFORCED CONCRETE BEAMS STRENGTHENED WITH CFRP STRAPS G. KESSE AND J. M. LEES Cambridge University Engineering Department, Trumpington Street, Cambridge, CB2 IPZ, United Kingdom

This paper reports on the strengthening of shear critical reinforced concrete beams with pre-stressed Carbon Fibre Reinforced Polymer (CFRP) straps. The straps can be used to enhance the shear capacity of beams and to change the mode of failure from shear to flexure. In the experimental programme, reinforced concrete cantilever beams were tested with or without straps for a particular shear span to depth ratio. The main parameters that were varied during the experiments were: the number of straps, the strap location, the level of pre-stress and the existing crack state. The experimental results are presented and discussed and conclusions drawn regarding the potential shear capacity enhancement, the stiffness improvement and the change in failure mode due to the presence of the straps.

INTRODUCTION

Throughout the world, an increasing number of reinforced concrete (RC) structures are being assessed as having inadequate shear capacity. The reasons for this include, design codes used in earlier years being less stringent than today’s standards and, in some cases deterioration of the internal steel reinforcement. Even for structures designed adequately, loads greater than the design capacity are being applied and thus the structures have been rendered unsafe. Such structures must be strengthened in order to serve their intended purpose. External Prestressed Carbon Fibre Reinforced Polymer (CFRP) straps provide a means of increasing the shear capacity of a concrete beam. The straps are closed loops formed from CFRP tape. The straps are installed at a specified spacing within the shear span and then prestressed. Experiments by researchers at the Swiss Federal Laboratories for Materials Testing and Research (EMPA)’ and the University of Cambridge’ have shown that the shear capacity of RC beams can be enhanced using this system and the mode of failure changed from shear to flexure. The straps do not corrode and hence offer a further advantage in aggressive environments. However, the material is brittle and thus all the parameters that govern the behaviour must

448 FRPRCS-6: Externally Bonded Reinforcement f o r Shear

be fully understood. The key parameters are the level of prestress, the number of straps required, the strap spacing and the number of tape layers in the strap. An additional factor is that, in previous studies, the straps have been installed before any external loading was applied. This is not necessarily representative of existing beams which might have sustained some damage prior to being strengthened or repaired. This paper reports on experimental work where the strap locations and the number of layers in the straps were varied. The purpose was to determine how these factors influence RC beams failing in shear and the resulting modes of failure. Work on pre-cracked or damaged beams will also be discussed.

zyxwvu zyxwv

EXPERIMENTS

Rectangular cantilever beams of dimensions 1200 mm x 105 mm x 280 mm were designed for this series of experiments. The support block had dimensions of 600 mm x 105 mm x 300 mm as shown in Figure 1. The cantilever shape was chosen in preference to a conventional simply supported beam because it presented a single shear span that furnishes less voluminous data while enabling the detailed monitoring of cracks and beam behaviour. In addition, the fabrication and testing procedures were greatly simplified.

100.0

applied load

Steel pad

General layout Figure 1. Beam layout with 2 straps

c105.0

Section

z

Shear Critical RC Beams with CFRP Straps 449

Figure 1 also shows the beam layout and reinforcement details. Electronic strain gauges were attached to the longitudinal tension steel and the internal shear links. The mix design and the properties of the reinforcement are given in Table 1. The beams were installed in the testing frame approximately 7 days after casting (see Figure 2a).

zyx 1 1

zyx

Table 1. Material DroDerties and reinforcement details

Reinforcement details

I

Dia. (mm) Bar 4 6

I

Location

Shear links: Beams 2-8 Shear links: Beam 1 Beam -tension and 12 compression Beam -tension 16 reinforcement Tape for straps

Youngs Modulus = 130000 N/mm’ Ultimate strain = 11000 micro strain Thickness = 0.16 mm.Width = 12 mm

Area of Steel(mm2)

Yield Stress (N/mm2)

25.12 @, 200c/c 6 1.56@,75c/c

395 400

226.22

500

402.17

500 Concrete

Max aggregate size 10 mm Concrete mix: 1:2:2

Forming and prestressing of the CFRP straps The prestressing material comes in the form of a tape consisting of unidirectional carbon fibres in a thermoplastic matrix. The thickness of the tape is approximately 0.16 mm with a width of 12 mm. To form a strap, the tape is wound around the beam web until the desired number of layers is obtained. The outermost two layers of the strap are welded together but the inner layers remain non-laminated. Strain gauges were attached to the outer layer of each strap. The prestressing procedure involves an arrangement as shown in Fig 2b. The strap is jacked upwards, and then metal shims inserted underneath the space created below the lower steel plate. When the required prestress has been attained, the jack is released and the steel pad then rests on the shims. EXPERIMENTAL RESULTS The loading was applied vertically upwards in increments of 1.5 kN. After each load increment, the crack pattern was marked and photographs taken.

450 FRPRCS-6: Externally Bonded Reinforcement f o r Shear

Figure 2a. Beam in testing frame

Figure 2b. Arrangement for prestressing

The experiments were arranged in three major stages. The first stage involved testing un-strengthened beams to establish the minimum and maximum beam capacities and failure modes. In the second stage, strengthened beams were tested. The key parameters under investigation were the number of layers of tape, the number of straps and the strap spacing. The final stage considered damaged beams where the beams were cracked before the straps were installed. A summary of the experimental programme, results and failure modes can be found in Table 2. For comparison purposes, the strap strengthening increment predicted using the 45" truss analogy* is also shown (*the approach assumes the use of passive ductile materials and may be inappropriate for prestressed CFRP straps).

zyxw

Table 2. Summary of test results

ble 1 Properties ble 1 Properties ble 1 Properties ble 1 Properties

Ta

Ta

Ta

Ta

Beam number

B 1-NS-NL B2-NS-NL B3-2s-10L B4-1S-5L B5-1s-5L-P B6-2s-10L-F B7-1 S-1 OL B8-2S-5L

ieam number-no. straps-nolayers-; P denotes a precracked beam. F-flexural failure; DT diagonal tension failure; S strap failul e.

zyx zy

Figure 3 shows the load versus displacement (measured directly on top of the loading point) curves for all the beams tested. Beams 1 and 2 were used to establish the maximum and the minimum beam capacities in terms of flexural failure and shear failure. Beam 1 was tested at d d = 4.5 and

z zyxwv zyxwv Shear Critical RC Beams with CFRP Straps 451

contained 6 mm shear links at 75 mm centers whilst the rest of the beams were all tested at d d = 3.0 with the shear reinforcement as shown in Figure 1 (4 mm links at 200 mm centers). Beam 1 failed in flexure with the yielding of the tension reinforcement before the crushing of concrete. The load displacement curve has been omitted from Figure 3 for ease of comparison. Beam 2 failed in shear. The outermost shear crack developed and propagated resulting in the failure of the beam by the splitting of the compression zone (Figure 4a). For the strengthened beams, the initial beam behaviour was similar to that of the un-strengthened beams in terms of the stiffness and crack pattern. For the beams with 1 strap, the strap was located in the middle of the shear span (345 mm from support block) whilst for beams with 2 straps, the straps were equally spaced within the shear span (230 mm apart). Beams 4 and 8 both failed in shear with the shear crack crossing the strap followed by the snapping of the strap. Beam 7 also failed in shear but the strap did not break since the shear crack formed between the strap and the support block leading to failure in the concrete. Beam 3 attained its full flexural capacity and crushing of concrete occurred at the corner between the beam and support. Snapping of the first strap followed flexural failure. In the case of the pre-cracked beams (B5 and B6), the beams were initially loaded to about 34 kN (i.e. 70% of unstrengthened beam’s shear ultimate capacity) and then unloaded before the straps were installed. The beams were then loaded until failure.

zyxwvu

zyx

120

100

-5 m

0 J

zyxwvutsrq zy 80

60

40 20

0

0

2

4

6

8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 Displacement (m m)

Figure. 3. Load versus displacement curves

452 FRPRCS-6: Externally Bonded Reinforcement for Shear

zyxw

DISCUSSION The strengthened beams were significantly stronger (55 to 100%) than the unstrengthened beam. The straps are not bonded to the concrete and act like a tie inducing transverse compression through the web of a beam. The strain in a strap leg is thus uniform but the prestress force and the stiffness of the straps will differ depending on the number of layers of tape used. Since all the straps were prestressed to 50% of their ultimate capacity, straps made of 5 layers of tape had an initial prestress of 12.5 kN whilst the 10 layer straps had an initial prestress of 25 kN. In addition, the stiffness of the 10 layer straps was twice that of the 5 layer straps.

zy z zyxwvu

Influence of the number of layers Beams with 1 strap

Beam 4 and beam 7 both had a single strap installed in the middle of the shear span and prestressed to 50% of the strap capacity. Figures 4b and 4c show the crack pattern when the applied load reached 65 kN. It could be observed that the outer shear crack for beam 7 could not develop as far as that of beam 4 and the stiffer 10-layer strap managed to reduce the progress of this crack. In contrast, the crack in beam 4 quickly progressed beyond the strap. The failure mode of beam 7 suggests that the 10-layer strap stopped the shear crack between the strap and the loading point from developing. On the other hand, it had a limited influence on the crack between the support and the strap and thus the beam failed due to this crack. For beam 4 the strap did not exert as much influence on the outer crack. Nevertheless, final failure occurred at a load of 81 kN which was higher than the shear capacity of the unstrengthened beam.

Beams with 2 straps Beam 3 and beam 8 had 2 straps installed from the beginning of the test with 10 and 5 layers respectively. Whilst beam 3 failed in flexure, beam 8 failed in shear. The crack progress can be observed from the traced crack pattern (Figures 4d and 4e). Up to 50 kN, the two beams behaved in the same manner and the crack pattern looked very similar. But when the crack crossed the straps, the straps had a strong influence on the crack progress. It could be observed that the cracks in beam 8 had progressed about twice the distance of that of beam 3.

Shear Critical RC Beams with CFRP Straps 453

(el

zyxwvu zyxwvuts I

zyxwvu zyxw 1

I '

I

"

Figure 4. Beam crack patterns at specified loads (a) Beam 2 at 48 kN (b) Beam 4 at 65 kN (c) Beam 7 at 65 kN (d) Beam 8 at 65 kN and (e) Beam 3 at 65 kN

454 FRPRCS-6: Externally Bonded Reinforcement f o r Shear

zy zy zyxwvu zyxwvu zyxwvu

The prestress force applied to the concrete and the stiffness of the straps thus control the rate of crack growth and also influence the mode of failure. However, when a particular strap retards crack progress, other cracks in unstrengthened regions remain free to grow and can lead to beam failure. The influence of the strap spacing will be discussed in the next section. As beam 3 failed in flexure it can also be concluded that the straps can effectively enhance the shear capacity and change the mode of failure.

Influence of number of straps /strap spacing

The 45" truss analogy would predict that a single strap would be ineffective. However, even beams with a single strap showed a capacity increase.

Beams with 5 layer straps

Beams 4 and 8 each had 5 layer straps but whereas beam 8 had two straps, beam 4 had only one. The load deformation curves (Figure 5 ) show both beams attaining almost the same load capacity and both failing in shear. Although the beams had a different number of straps, the difference in behaviour was not significant due to the nature of the failures.

zyxwvu zyxwvu

90 7 80 70 z 60 5 50 40 -I 30 20

3

I

10 0

I

I

I

I

I

I

I

I

I

1

zyxwvut 0

2

4

6 8 10 12 14 Displacement (m m)

16

18

20

Figure 5. Comparison of beams with 5-layer straps

Beam 4 had a single strap a distance 1.5d from the support and a crack developed between the support and the strap. The strap was also not able to stop the outer crack from developing and crossing through the strap. As a result, these two cracks lead to failure of the beam. On the other hand, beam 8 had two straps. The location of the first strap was such that it stopped the first crack from developing. The second strap

Shear Critical RC Beams with CFRP Straps 455

zy zyxwvu zyxwvu zyxwvut

also limited the propagation of the outer crack. However, the section in between the straps was unstrengthened and the crack that developed in this region easily then passed through the first strap to fail the beam. The result suggests that the importance of the strap spacing is connected to the stiffness of the straps. For the beams with the lower stiffness 5 layer straps, the final crack pattern and ultimate shear capacity were similar regardless of the strap spacing. Beams with I0 layer straps

Beams 3 and 7 both had 10-layer straps but the number of straps differed. The load deflection curves show beam 3 failing in flexure whilst beam 7 failed in shear (see Figure 6). The 2 straps of beam 3 managed to control the crack growth whilst the single strap of beam 7 did not restrain the propagation of the inner shear crack leading to the failure of the beam in shear. 110 100

90 80 70 60

50 40 30 20 10 0

I

0

zyxwvut 5

10

15 20 25 Displacement (m m )

30

35

40

Figure 6. Comparison of beams with 10-layer straps

Since the unstrengthened beam failed by the propagation of the outermost shear crack (see Figure 4a) and it is reasonable to assume that stopping this outer crack might be sufficient to prevent shear failure. However, the tests show that although this can be achieved, failure may develop in adjacent unstrengthened regions in later stages of testing. Hence, provided that the stiffnesses of the straps were adequate, two straps spaced at a distance d apart were more effective. Pre- cracked beams

From the load displacement curves (see Figure 3 ) the peak loads did not change significantly for the pre-cracked beams but the stiffnesses of these

zyxwvut

456 FRPRCS-6: Externally Bonded Reinforcement for Shear

beams were lower than those of beams with straps applied at the beginning of the test. This was expected as the pre-cracking induced some permanent damage before strapping. The ability of the pre-cracked beam to attain the same peak load could be a function of the load to which the beam was initially subjected. From strain readings taken from the internal steel links, the links were not carrying much load in the preloading stage and thus the beam had not sustained significant damage. The prestressed straps will also help close any existing cracks and thus it could be that the influence of existing damage is mitigated, The influence of existing cracks in a passive system may well be different. Tests are continuing where the pre-crack load will be higher than that used in the earlier tests.

CONCLUSIONS

zyxw

Based on the experiments carried out to date, all of the beams tested with straps, whether single or double and irrespective of the number of strap layers, had a shear capacity enhancement at least 55% higher than that of an equivalent unstrengthened beam. All of the beams with a single strap failed in shear whilst some of the beams with two straps attained their full flexural capacity. The pre-cracking of the beams before installing the straps seems to have no significant influence on the shear capacity gain. The crack path does not appear be influenced significantly by the straps however the crack growth and widths are reduced due to the presence of the straps. The mode of failure depends on the strap stiffness, the strap location and the crack path. Further tests are being carried out to separate the effect of the prestress force and the stiffness of the straps.

ACKNOWLEDGEMENTS

zyx z

The authors are grateful to EMPA for their support with this project.

REFERENCES

1. Lees, J.M., WinistBrfer A.U. and Meier U., “External Prestressed CFRP Straps for the Shear Enhancement of Concrete ”, ASCE, Journal of Compositesfor Construction, 6(4), Nov 2002. pp. 249-256 2. Chan, K.M.C., Prestressed Non-laminated Carbon Fibre Reinforced Plastic Straps, Fourth Year Project Report, Dept. of Engineering, University of Cambridge, UK, 1999/2000.

zyxwvu

FRPRCS-6, Singapore, 8-10 July 2003 Edited by Kiang Hwee Tan QWorld Scientific Publishing Company

zyxw zyxw

EFFECTIVE SHEAR STRENGTHENING OF CONCRETE BEAMS USING FRP SHEETS WITH BONDED ANCHORAGE

zyxwvu B. B. ADHIKARY Frank Lam &Associates, Austin, Texas, USA

H. MUTSUYOSHI Department of Civil and Env. Engineering, Saitama University, Saitama, Japan M. ASHRAF Engineering Associates (EA) Pvt. Ltd., Karachi, Pakistan This paper presents the results of an experimental study for shear strengthening of reinforced concrete beams using externally bonded FRP sheets. The study focused on effect of bonded anchorage of sheets in delaying or preventing sheet debonding. Three different models to estimate the contribution of FRP to the shear capacity (V’) of RC beams are discussed and two separate equations to calculate V’are presented.

zyxwv zyx

INTRODUCTION

From the past studies conducted by Chaallal et al.’, Chajes et a1.*, Sat0 et aL3, and Uji 4, it has been shown that externally bonded FRP can be used to enhance the shear capacity of RC beams. Bond of FRP sheets to concrete is of critical importance for the effectiveness of externally bonded FRP sheets. If this interfacial bond is compromised before rupture of the FRP sheets, sheet-debonding failure occurs. This study presents the shear behavior of RC beams strengthened with FRP sheets. Special focus is given for the prevention of sheet debonding to get effective utilization of FRP’s mechanical properties. Anchorage of FRP sheets at the top surface of the beam was provided in order to delay or prevent sheet debonding. Three models available in the literature by the JSCE’, Khalifa et aL6, and Triantafillou and Antonopoulos7 for computing the contribution of FRP sheet to the shear capacity of strengthened beams (5) are presented and compared with the experimental results. Two separate equations to calculate V,are presented in this paper; when failure is likely to occur due to sheet debonding and when bonded anchorage of FRP sheet is provided to the beams.

zyxw zyxw zyxw

458 FRPRCS-6: Externally Bonded Reinforcement for Shear

EXPERIMENTAL PROGRAM

a. Cross section

zyxw

b. Longitudinal section (Dimensions are in mm)

Figure 1 Details of test beams

A total of nine beams were tested. Figure 1 shows the typical dimensions and reinforcement layout for RC beams. No stirrups were provided in the potential shear failure zone. Longitudinal bars had an average yield strength of 395 MPa and elastic modulus of 196 GPa. The cross sections of beams were chamfered at 30-mm for AFRP strengthened beams, and the chamfered edges were further smoothened in round shape at 100-mm diameter for CFRP strengthened beams. Beam B-1 was kept as a control beam. Eight beams were categorized into two series as CFRP series and AFRP series. The beams were strengthened with epoxy bonded unidirectional FRP sheets applied only to the shear spans, where principal fibers were kept perpendicular to the longitudinal axis of the beams. Mechanical properties of FRP sheets are shown in Table 1. Test variables were FRP type, wrapping layout and anchorage length. Anchorage was provided by bonding a length of sheet at top of the beam. Figure 2 shows the different wrapping schemes used.

zyxw zy

Table 1. Mechanical properties of FRP sheets Sheet

CFRP AFRP

Thickness (mm) 0.167 0.286

Tensile strength WPa) 3400 2000

Elastic modulus (GPa) 230 120

Ultimate elongation (%) 1.5 1.8

z

Shear Strengthening with Bonded Anchorage 459 zyxwvut

AFRP-SERIES

zyxwvutsrqponmlkj

1zyxwv A - I (AFRP1

(U.W r a p )

€7 A 2

(80

AFRP

&ixdgs)

rl 0 A-3 (AFRP)

zyxwvutsrqponmlkj

( I 1 0 mm anchorage) 100

H

zyxwvutsrqponmlkjih

% % i% ,!F(

Figure 2 Experimental test scheme (Dimensions in mm)

TEST RESULTS AND DISCUSSION

Figure 3 Sheet debonding in beam C-1

Figure 4 Concrete splitting in beam C-2

Control beam B-1 failed in shear at a load of 224 kN. All other beams, except C-4 and A-4 failed also in shear at considerably higher loads than that of control beam B-1. The ultimate failure loads, contribution of FRP

zyxw zyxwvu z

460 FRPRCS-6: Externally Bonded Reinforcementfor Shear

sheet to the shear capacity of RC beam (Q),and increase in shear capacity are shown in Table 2. It is confirmed that the FRP sheets applied on shear spans not only increase the shear capacity, but also change the deflection characteristics of beams. It is also found that the bonded anchorage of FRP sheet not only increases the shear capacity, but also changes the ultimate failure mode from sheet debonding to concrete splitting or flexure. Beams C-1 and A-1 , which were strengthened using U-wrap of CFRP and AFRP sheets, respectively failed in shear followed by debonding of sheet (Figure 3). Beams C-2 and A-2 failed by concrete splitting (Figure 4). Beams C-4 and A-4 strengthened by full wrapping of CFRP and AFRP sheets failed in flexural mode.

zyxwvut zy Table 2. Experimental results

Beam

fc 'Shear b ~ f Increase Failure mode (MPa) strength (kN) (kN) (%) B-1 38.0 224 diagonal shear C-1 37.2 330 53 47.3 shear + debonding C-2 41.0 457 116.5 104.0 shear + splitting C-3 41.1 475 125.5 112.0 shear f splitting C-4 42.4 500 138 123.2' flexure A-I 39.6 3 10 43 38.4 shear + debonding A-2 41.8 400 88 78.6 shear + splitting A-3 43.9 490 133 118.8 shear + splitting A-4 43.5 488 132 117.9' flexure ' Shear strength of beam is equal to half the failure load; bVfisFRP contribution to the shear capacity of RC beam; is the percentage increase in failure load.

f Figure 5 Load displacement (CFRP)

Figure 6 Load displacement (AFRP)

Figure 5 and Figure 6 show the load displacement relationships for CFRP and AFRP strengthened beams, respectively. Each superior strengthening scheme showed better load-displacement characteristics than the previous

Shear Strengthening with Bonded Anchorage 461

one. Figure 7 and Figure 8 show the relationship of load and vertical strain in FRP sheet for CFRP and AFRP strengthened beams. Maximum FRP strain parallel to the fibers was measured as 6825 and 8225 microstrain for beams C-4 and A-4, which is about 45% of the ultimate strain of the sheets. Moreover, maximum FRP strain in beam C-1 is 3550 microstrain, which is only 23.7% of the ultimate FRP strain, while for beam C-2, a strain of 6045 microstrain was attained, which is 70.28% increase as compared to the control beam C-1 . This substantial increase in FRP strain in beam C-2 is due to the provision of bonded anchorage. Table 3 shows FRP strain and percentage increase in FRP strain for beams with provision of bonded anchorage as compared to U-wrapped beams.

zyxwvu

.>ooo

o

zyxwvutsrqponmlkjihg zyxwvutsrqponmlkji zyxwv zyxw zyxwv >om

>om

1000

aom

I ~ O

.?om

,o(

o

iooo

2mo

iom

moo

om

$om moo

moo

9000

FRP sheet Stran (micron)

FRPshed strain (mcron)

Figure 7 Load-FRP strain (CFRP series) Figure 8 Load-FRP strain (AFRP series) I

zyxwvuts 600

-Trend

AFRP-Series line (CFRP-series)

- - - -Trend

100

110

200

B o nd eA anchorage iengI h (m)

Figure 9 Anchorage length vs. bond stress

210

,000

4000

I000

6000

line(AFRP-series) 7000

moo

'1000

MaximmFRPsIrain (micron)

Figure 10 Failure load versus maximum FRP strain

For beams C-1 and A-1 failure was governed by to sheet debonding, therefore measured bond stress for beams C-1 and A-1 may be taken as bond strength. This was found to be 4.05 MPa. Beams with bonded anchorage did not show sheet debonding at ultimate failure due to reduction in bond stress at the interface. Influence of bonded anchorage on interface bond stresses is shown in Figure 9.

zy zyxwv zyxw zyxw zyx zy

462 FRPRCS-6: Externally Bonded Reinforcementfor Shear

Table 3. Maximum FRP strain

Beam

Anchorage Maximum FRP ahcrease in FRP Max. FRP strain / (mm) strain (p) strain (%) Ult. FRP strain c-1 0 3550 0.277 c-2 80 6045 70.3 0.403 c-3 110 6368 79.4 0.425 c-4 200 6825 92.3 0.455 A-1 0 3420 0.19 A-2 80 7063 106.5 0.392 A-3 110 7825 129.0 0.435 A-4 200 8225 140.5 0.457 a Increase compared to C-1 and A-1 for CFRP and AFRP series, respectively.

zyx

FRF' SHEET CONTRIBUTION TO SHEAR CAPACITY

The shear strength of RC beams strengthened using externally bonded FRP sheets is computed by Eq. (1). The FRP reinforcement is treated in analogy to the internal steel if it is assumed that FRP develops an effective strain (qe) that is less than the tensile failure strain (qu). Therefore, Vf for FRP sheets can be calculated by Eq. (2).

V, = pf Ef qedfbw(sin p +COS

p)

(2)

Khalifa et a1.6 proposed to use a reduction factor, R to ultimate strain to calculate the effective strain in the sheet. They suggested two equations for R to represent two possible failure modes of FRP bonded beams, namely FRP rupture and FRP debonding. The effective strain is computed using the lowest value of R. Triantafillou and Antonopoulos7 proposed three different expressions to calculate the effective FRP strain. After computing the effective strain, qe,the contribution of FRP sheet to the shear capacity can be calculated from Eq. (2) with a multiplying factor of 0.9. Japan Society of Civil Engineers (JSCE)5 proposes to use a coefficient called shear reinforcing efficiency of the FRP sheet to evaluate the ultimate mean stress of sheet and to determine the shear contribution of the sheet. Table 4 shows the comparison of predicted and experimental values of V,. It is seen that none of the models is able to predict V,correctly.

zyxwvu zyx zyx Shear Strengthening with Bonded Anchorage 463

Table 4.Comparison between the calculated and experimental V, (kN) Beam no. C-I A-I c-4 A-4

Khalifa et al. 60.4 (1.14) 67.1 (1.56)

JSCE

Triantafllou and Antonopoulos 72.3 (1.36)

Experiment

53.0 43 .O 137.6 (1.00) 154.8 (1.12) 138.0 85.3 (0.65) 156.0 (1.18) 132.0 Number in parenthesis is ratio of values from formula to that of experiment.

PROPOSED DESIGN EQUATION

Experiment showed that the U-wrapped beams failed due to sheet debonding. Bonded anchorage resulted in more than 100% increase in FRP effective strain and consequently in higher shear capacity. Shear capacity of RC beams was found to be linearly proportional to the FRP strain measured at failure (Figure lo). Past studies showed that effective strain in FRP sheet depends on the product of elastic modulus and thickness of the FRP sheet. Besides, bond strength of FRP to concrete also depends on the tensile strength of concrete, consequently to its compressive strength. 1.5

1.2

zy zyxwvutsrqp

0.6

0.3

0.1

0.2

zyxwv 0.3

0.4

0.5

0.6

P/Ed (0.2f:”)

Figure 1 1 Model calibration for effective strain

In this study, past experiment^^‘^ on shear strengthening using externally bonded FRP sheets and strips are used to calibrate an equation for estimation of effective strain in FRP sheet at debonding failure. Only those experimental results are used in which RC beams failed in diagonal shear followed by FRP debonding. The value of pf Ef/(0.2.,2/3)is plotted versus t+‘zfu as shown in Figure 11. This relation can be obtained from a best-fit power-type curve to the experimental data, given by Eq. (3).

464 FRPRCS-6: Externally Bonded Reinforcement for Shear

zyxwvu zyxwvu zyx zy zy

Eq. (3) is derived using most of the data for CFRP sheetslstrips, which is a function of axial rigidity of FRP and tensile strength of concrete. Therefore, the same equation is used for AFRP sheetdstrips using a factor of 0.9. 1

&/el

- 0.034f

EfU

-

(for AFRP Sheet)

JZ

(4)

It is seen that the additional bonded anchorage of sheet resulted in substantial increase in FRP strain. The increase in effective strain due to is related to a non-dimensional parameter ldb,, bonded anchorage (~3,~) which is given by fitting a best curve to the test results. Effective strain in FRP is equal to the sum of effective strain in FRP in debonding mode ( % I ) and increase in effective strain in FRP ( q k 2 ) due to bonded anchorage. Since four beams tested with bonded anchorage failed due to concrete splitting, the compressive strength of concrete (f',) is also considered while estimating q e 2 .

For CFRP sheet: I

10,000 Shear F,, = failure/max load; 6,, = failure/max deflection; w, = max. crack width; ~rn, = max. measured strain in flexural reinforcement; E~~~ = max. measured strain in shear reinforcement; S = steel; G = glass FRP; C = carbon FRF' values shown correspond to failure obtained with the large spacing of shear links as specified in Table 1 smcm

z zyxwv zyxwvu

520 FRPRCS-6: Externally Bonded Reinforcement f o r Shear

Strain values recorded both in the GFRP flexural reinforcement and in the externally applied shear reinforcement always exceeded the limit of 2,000/2,500 p~ assumed by most current recommendations for shear design with FRP reinforcement, thereby confirming the conservative nature of Maximum strain values, ranging from these existing 10,000 ps to around 20,000 ps for GFRP and from 9,000 ps to 10,000 p~ for CFRP, were recorded in the shear reinforcement. Decomposition of the shear resisting components was also performed on the test beams (Figures 3 and 4) in an attempt to identify the contributions of the basic shear carrying mechanisms. The component of shear resisted by the shear links was determined by considering the number of effective shear links crossing the crack that induced the ultimate failure and assuming a uniform distribution of strain within each link, equal to the maximum strain recorded in that link at each stage of loading. The concrete contribution was then determined by subtracting the contribution of the shear links from the total shear capacity.

zyxwv zyxwvut zyxwvuts

3 50

c v)

40

00 pstrain (shear rlment)

30

00 pstrain (shear rlrnent)

20

10

0

0

2

4

6

8 10 Displacement (mm)

Figure 3 . Estimate of the shear resisting components for SB40R

After initial loading of beam GB43R (see Figure 4) it became clear from strain readings that the desired shear failure might not develop. The beam was therefore unloaded and alternate shear links were cut to halve the amount of shear reinforcement, as reported in Table 1. The c.ontribution of concrete and shear reinforcement to the total shear resistance of GB43R in this latter stage of loading is represented by the shaded (dark) and un-shaded area below the dashed curves, respectively.

Shear Design Equations f o r FRP RC Beams 521 zyxwvuts

The shear load-deflection curves for the beams un-reinforced in shear (SB40 and GB43) are shown alongside those for the reinforced beams to facilitate comparison in terms of concrete shear resistance. In Figures 3 and 4, the critical values of strain recorded in both the flexural reinforcement and shear reinforcement are represented with vertical and horizontal dotted lines respectively.

zyxwvutsr zyxwv zyxwvut

e

60

9 50

.c

c/)

40

30

(shear rlrnent - 1"+2"dcycles)

(shear rlment

- 1"'+2"* cycle

20 10

0

Figure 4.

I

Estimate of the shear resisting components for GB43R during the lstand 2"d cycles (solid lines) and the 3rd cycle (dashed lines)

From the analysis of the results, it appears that the shear carrying mechanisms are mobilised in a comparable manner in GFRP and steel RC beams, and that the failure modes develop in a similar way (Figures 5 and 6). Therefore it can be concluded that the additive nature of shear resisting mechanisms can be assumed to be valid.

zyxwvut

I-igure 5 . Shear tailure in SR40 (leti) and GR43 (right)

zyxwvuts

522 FRPRCS-6: Externally Bonded Reinforcement f o r Shear

zyx

Based on the results of this research project and on previous work by various other the authors have proposed a modified approach for the design of FRP RC beams in which the strain limits are increased to the higher value of 4,500 p~ for both the shear and flexural reinforcement. This new approach, referred to as the “Sheffield approach”, has been successfully applied to various code equation^^^^^^^^^'^.

zyxwvu Figure 6. Shear failure in SB40R (left) and GB43R (right)

Table 3 reports the predicted shear capacity of the tested beams according to design code equations modified by using both the current recommendations and the Sheffield approach. It is evident that by using the Sheffield approach the total shear capacity can be predicted with much less scatter and a considerably improved level of accuracy.

zyxwv zyxw zy zyxwvu

Table 3: Comparison of predicted shear capacities for FRP RC beams implementing the Strain approach and Sheffield approach

Sheffield approach (allowable strain of 4,500 ,ud A Exp Exu Exp Exu Exu BS8110 EC2-2001 ACI-440 BS8110 EC2-2001 ACI318 1.40 1.78 1.31 1.42 1.11 1.47 SB4OR 1.98 2.49 4.64 1.50 GB43RI 1.85 1.62 1.69 2.14 3.06 1.41 1.64 1.37 Mean. 0.50 2.24 0.13 StdDev. 0.41 0.30 0.36 values shown correspond to predictions obtained with the large spacing of shear links as specified in Table 1 Beam

Current recommendations (allowable strain of 2,000-2,500 p)

Although the presence of cracks in FRP RC members does not represent a cause for concern from the point of view of durability, unlike for steel reinforced structures, a crack width limit wsL of 0.5 mm has been proposed in various design recommendations dealing with FRP RC structures for aesthetic reasons2.

zyx

Shear Design Equationsfor FRP RC Beams 523

zyxwv zyx zyxwvuts

With these considerations in mind, the predicted design loads obtained by modifying the BS 811013 shear design equations according to the Sheffield approach were compared to the results of the present study and, subsequently, the corresponding service loads were checked against the maximum shear crack widths that were observed at those load levels. The service load, SLYwas computed by dividing the previously derived predicted ultimate load, SA, by a load factor of 1.5. wmrepresents the crack width measured at a load level equivalent to the service load.

zyxwvuts zyxwvutsrq "1 140

140

4

120 100

80

SA - Sdcycle

zyxwvuts 1 Y

60

40

20

6C

-0-

w.

07 0.0

Shear crack - lstphase Shear crack - Zd phase

SL - 3" cycle

40

Zd phase - l"+Zd cycle

WSL

0.5

1.0

1.5

2.0 2.5 3.0 Crack width (mm)

0.0

0.5

1.0 1.5 2.0 Crack width (mm)

Figure 7. Shear crack width growth for beams SB40 and SB40R (left) andbeams GB43 and GB43R (right)

Figure 7 illustrates the shear crack width growth for beams SB40(R) and GB43(R). Based on the results presented in this figure and Figures 3 and 4, it can be observed that for levels of strain up to 4,500 p~ (developed both in the shear and flexural FRP reinforcement) as proposed in the Sheffield approach, shear cracks were effectively controlled and the individual shear resistance of concrete and shear reinforcementwere effectivelymobilised. DESIGN RECOMMENDATIONS

Based on the above results, the following recommendations are made for the shear design of FRP RC beams. Concrete shear resistance

Based on the Sheffield approach, modifications to code equations BS 8 11013, ACI 3 18-9914,EC-2 (EN 1992-1:2001)'' are given e l ~ e w h e r e ~ ~ ~ ~ ~

524 FRPRCS-6: Externally Bonded Reinforcement f o r Shear

zyxw

Contribution of shear reinforcement It was seen from the experimental results (Table 2), that the strain developed in the shear reinforcement was much larger than the yield strain of steel. The design approach for steel RC not only excludes the possibility that higher strains can be developed, but relies on it, since by adding the contributions from steel and concrete, plasticity is implicitly expected from both mechanisms. The adoption of the same strain limit for FRP reinforcement as for steel by the current design recommendations is therefore primarily there to control the development of crack widths. The results of this study, however, confirm that a relaxation of the strain limit for shear reinforcement to 4,500 pc, as proposed by the Sheffield approach, will not lead to undesirably large crack widths.

z

ECONOMY ACHIEVED BY THE PROPOSALS Figure 8 compares the ratio of shear reinforcement required for a given FRP RC beam according to the BS 81 10 equation modified by using both the strain approach and the Sheffield approach. Geometrical characteristics and concrete strength are kept constant, whilst the relative stiffness of the flexural reinforcement and the design applied shear stress, vd, vary. It can be observed that the reinforcement ratio required by the strain approach is always higher than that required by the Sheffield approach, and increases overall with increasing applied shear stress. A reduction in the required shear reinforcement of up to 3 times is possible when the Sheffield approach is adopted, thereby reducing reinforcement costs by up to 30%.

z

c 0

zyxwv zyxw

zyxwvutsr

2 1

3

0.0

0.2

0.4

0.6

0.8

1.o

.2

Normalized flexural stiffness ( P E/SGPa)

Figure 8. Ratio of shear reinforcement calculated according to the strain approach and the Sheffield approach plotted against normalised stifmess of the flexural reinforcement

Shear Design Equationsfor FRP RC Beams 525

z

CONCLUSIONS

The following conclusions can be drawn from the reported study: (a) The strain in both the flexural and shear FRP reinforcement can reach values that are much higher than those assumed by the current recommendations for the design of FRP RC. (b) Shear resisting mechanisms are mobilised in a similar way in both GFRP and steel RC beams and failure modes are characterised by similar behaviour. Hence, summing the contributions of the concrete and reinforcement shear resistance mechanisms remains valid. (c) For concrete shear resistance, the principle of strain control is accepted, but a new limit of 4,500 p is proposed for determining the amount of flexural reinforcement to be used in concrete shear design. For the design of shear links, the new proposed limit of 4,500 p~ also seems to lead to more appropriate and cost effective solutions.

zyxwv zy zyxw zyxwv zyxwvu z

ACKNOWLEDGMENTS

The authors wish to acknowledge the European Commission for funding the TMR Network "ConFibreCrete". REFERENCES

1. Japan Society of Civil Engineers, Recommendation for Design and Construction of Concrete Structures using Continuous Fiber Reinforcing Materials, JSCE, Tokyo, Japan, 1996. 2. Canadian Standard Association, Canadian Highway Bridge Design Code Section 16: Fibre Reinforced Structures, Final Draft, CHBDC, 1996. 3. Institution of Structural Engineers, Interim guidance on the design of reinforced concrete structures using j b r e composite reinforcement, IStructE, SET0 Ltd, London, 1999. 4. American Concrete Institute, Guide for the Design and Construction of Concrete Reinforced with FRP Bars, ACI 440.1R-0 1, ACI Committee 440, Farmington Hills, MI, USA, 2001. 5 . ISIS Canada - Intelligent Sensing for Innovative Structures, Reinforced Concrete Structures with Fibre Reinforced Polymers, Design manual No. 3, ISIS Canada Corporation, Manitoba, Canada, 2001. 6. Pilakoutas, K. and Guadagnini M., "Shear of FRP RC: a review of the State-of-the-Art", Proc. of the International Workshop Composites in Construction: a Reality, Capri, Italy, ASCE, 200 1, pp. 173-182.

526 FRPRCS-6: Externally Bonded Reinforcement for Shear

7. Guadagnini, M., Shear Behaviour and Design of FRP RC Beams, PhD Thesis, Department of Civil and Structural Engineering, The University of Sheffield, Sheffield, UK, 2002. 8. Guadagnini, M., Pilakoutas, K. and Waldron, P., “Investigation on Shear Carrying Mechanisms in FRP RC Beams”, Fifth International Symposium on Fiber Reinforced Polymer for Reinforced Concrete Structures (FRPRCS-5), Cambridge, UK, July 16-18,200 1, V01.2, pp. 949-958. 9. Guadagnini, M., Pilakoutas, K. and Waldron, P., “Shear performance of GFRP RC beams.” International Conference on FRP Composites in Civil Engineering, Hong Kong, 200 1, pp. 1 169-1 176. 10. Guadagnini, M., Pilakoutas, K. and Waldron, P., “Shear Performance of FRP Reinforced Concrete Beams”, Accepted for publication to the Journal of Reinforced Plastics and Composites, 2002. 11. Tottori, S., and Wakui, H. (1993), Shear Capacity of RC and PC Beams Using FRP Reinforcement. International Symposium on Fiber Reinforced Plastic Reinforcement for Concrete Structures, Nanni and Dolan ed., ACI, pp. 6 15-632. 12. Duranovic, N., Pilakoutas, K., and Waldron, P. (1997) Tests on Concrete Beams Reinforced with Glass Fibre Reinforced Plastic Bars, Third International Symposium on Non-Metallic (FRP) Reinforcement for Concrete Structures, Sapporo, Japan, pp. 479-486. 13. British Standard Institution, BS 8110 - Code of Practice for Design and Construction, Part 1, BSI, London, 1999. 14. American Concrete Institute, Building Code Requirements for Reinforced Concrete and Commentary ACI 318-99/R-99, ACI Committee 3 18, Farmington Hills, MI, USA, 1999. 15. European Committee for Standardization, Eurocode 2: Design of Concrete Structures - Part 1: General Rules and Rules for Buildings, prEN 1992-1 (1st draft), CEN, 1999.

zyxwvu zyxw

zyxwvutsr

FRPRCS-6, Singapore, 8-10 July 2003 Edited by Gang Hwee Tan @World Scientific Publishing Company

STRENGTHENING OF CORROSION-DAMAGED RC COLUMNS WITH FRP

zyxwv

S. N. BOUSIAS, T. C. TRIANTAFILLOU, M. N. FARDIS, L. A. SPATHIS AND B. O’REGAN Department of Civil Engineering, University of Patras GR-26500, Greece

Premature deterioration of RC structures due to corrosion of the reinforcement represents a significant problem, as it contributes considerably to the reduction of member strength and deformation capacity. Moreover, structures old enough to develop significant corrosion of the reinforcement normally belong also to the class of structures that have not been designed for earthquake resistance. In that respect, retrofitting against corrosion triggers seismic retrofitting of the structure as well. In this paper the use of fibrereinforced polymer (FFW) wraps in retrofitting RC columns with corroded reinforcement was experimentally investigated. Test results show that deformation- and, to a lesser extent, force-capacity of members with severe reinforcement corrosion can be considerably enhanced, through appropriate use of FRP.

INTRODUCTION Structures in seismic regions often suffer both from deficiencies in member strength and deformation capacity and from the effects of reinforcement corrosion due to aggressive environmental conditions. Past experience has shown that reinforcement corrosion not only reduces member strength due to steel area loss, but it also affects adversely bond and anchorage, and makes bars more susceptible to buckling and reduces steel ductility. Moreover, transverse reinforcement (for shear and confinement), being of smaller diameter and closer to the concrete surface, is more vulnerable to corrosion. Thus its contribution to the effective confinement and the resulting deformation capacity of the member decreases. For these reasons the seismic behaviour of RC members, especially of columns, is affected by steel corrosion, the problem being aggravated by the use of the more corrosion-prone tempcore S500 steel.

zyxwv

528 FRPRCS-6: Externally Bonded Reinforcement for Shear

The efficiency of electrochemical remedy measures is not commensurate to their cost. As structures old enough to develop significant reinforcement corrosion normally lack sufficient earthquake resistance, the need for measures against the on-going corrosion, often paves the way for seismic retrofitting as well. When retrofitting is realized through external (passive) confinement, fibre-reinforced polymer (FRP) wraps offer a particularly attractive solution. Corrosion is an expansive process and thus FRP jackets can act as a (passive) confining mechanism, activated by lateral expansion of member cross-section. In the past, wrapping of members with FRP jackets has been applied successfully to members without corrosion damage'.2.3z4.The performance of this scheme in retrofitting RC columns with corrosion-provoked damage is experimentally investigated in this paper. EXPERIMENTAL PROGRAMME

The experimental programme focuses on the study of the contribution of FRP wraps in the hoop direction as a means of enhancing the deformation capacity of RC columns with corroded reinforcement, by upgrading member effective confinement. Twelve cantilever-type specimens were constructed representing full scale RC columns of a length approximately equal to half a storey (1.6 m) net height and with cross-sectional dimensions 250x500 mm (Figure 1). To represent non-seismically designed and detailed members, specimens emulated old construction, as far as materials used and lack of earthquake resistant detailing. The longitudinal reinforcement comprised four 18-mm bars; transverse reinforcement was provided by 8-mm diameter smooth bars at 200-mm centres with 135O-hook at one end and a 90" hook at the other. Ready-mix C12/15 concrete was used, in which salt of 3% per weight of water was added at mixing. Table 1 summarises the characteristics of the materials used for the specimens. In all specimens longitudinal bars had a yield stress of 559.5 MPa, a tensile strength of 682 MPa and uniform elongation at failure of 13% (average of 3 coupons). The corresponding values for the transverse bars are 286 MPa and 350 MPa and 13%. The lower l-m of all specimens, except two (used as reference for comparison), was subjected to accelerated corrosion5, employing an electrochemical circuit in which each longitudinal reinforcement bar was the anode and an external galvanized steel mesh was the cathode. Stirrups were supplied by current through their contact to the main reinforcement. A 6V fixed potential was applied between anode and cathode and the

zyxw

zy

Strengthening of Corrosion-Damaged RC Columns 529

evolution of corrosion was monitored by recording the current passing and applying Faraday’s law to the integrated current. Alternating wet-dry cycles of 60 and 12 hours, respectively, were applied to the specimens using a 3% sodium chloride solution, as this has been shown in the past6 to produce corrosion products with high volumetric expansion. The accelerated corrosion conditions were maintained for about 3.5 months, at the end of which approximately 1 kg of steel mass in each specimen had been converted to oxides. The basic parameters of the retrofitting scheme studied in this research were: (a) the number of layers of the wrap material and the fibre material (carbon vs. glass), as a measure of different stiffness and strain capacity of the jacket material, (b) the effect of previous, unrepaired seismic damage, (c) the level of FRP-induced confinement in columns with cross-sectional aspect ratio other than 1.0, and (d) the effect of FRP wrapping in columns dominated by flexure or shear. As shown in Table 1, specimens were denoted after the following rules: the first letter denotes whether reinforcement was corroded (C) or not (U), the second and third signify the fibre material (C for carbon, G for glass) and the number of layers employed, while the last two define the axis of testing (W for weak and S for strong) and whether the specimen was initially damaged before retrofitting (denoted by in). Of all twelve specimens presented here, six were retrofitted with either 2 or 5 layers of FRP wraps, without any previous damage from cyclic loading. The number of FRP layers was determined as follows: the effectiveness of the FRP jacket with respect to the confinement achieved is conditioned by the deformability of the fibres and the extensional stiffness of the jacket, which is proportional to nxtfibxEf(n is the number of FRP layers, tfib is the thickness of a single layer (Le. the thickness of the fibre sheet), and Ef is the modulus of elasticity). In the tests performed the same axial stiffness was achieved with CFRP ( E ~ 2 3 0GPa, tfib=0.13 mm) and GFRP (Ef70 GPa, tfib=0.17 mm) using 2 layers for the former and 5 for the latter (axial stiffness EA 0 60 kN in both cases). Half of these first six specimens were tested along the weak and the other half along the strong column axis. Two of the other six specimens (C-C2Sin and C-C2Win) were first subjected to a number of displacement cycles beyond member yielding; then they were retrofitted with FRP wraps without restoring previous damage (e.g. cracking), except for any damage of the surface due to concrete cover spalling, which was repaired with non-shrinking mortar. Finally, the remaining four specimens (two with corroded reinforcement and

z zyxwv

zyxwv zyxwvu zy zyxwvut zy

530 FRPRCS-6: Externally Bonded Reinforcement for Shear

two without corrosion) were tested without any retrofit measures as control specimens. Table 1 . Specimen geometry and material properties Concrete FRP for retrofitting Normalised Peak Drift at axial load force failure (%) strength, f, Material Layers v=N/A,f, (m)

Specimen

u-0s

18.3 18.3 18.1 18.1 20.4 18.7 17.9 18.6 18.7 18.3 18.6 20.4

c-0s C-C2 Sin c-c2s c-c5s C G5S

u-ow c-ow C-C2Win c-c2w c-c5w C-G5 W

___

_-_

-__

___

Carbon Carbon Carbon Glass

2 2 5 5

-__

___

Carbon Carbon Carbon Glass

____2 2 5 5

0.38 0.38 0.37 0.38 0.34 0.37 0.38 0.35 0.35 0.37 0.37 0.34

190

182 167 190 182 182 72 65 67 70 67 69

2.5 2.8 4.1 5.1 3.7 4.1 4.1 4.4 7.2 7.2 7.5 7.5

A

0 10

Figure 1. Specimen cross-section and test set-up

Horizontal loading was applied at a distance of 1.6 m from the base by a servo-hydraulic actuator attached to the column head. Testing was performed by cycling horizontal displacements at increasing amplitudes along the weak or strong section axis. The tests were carried to column failure, determined either by fracture of reinforcing bars or FRP wraps, or when resistance dropped by at least 20% of its maximum previous value. An axial load of approximately 850 kN (see Table 1 for normalized axial load values) was applied through a jack placed at the top of the column. A special setup was developed to ensure that the axial load is always applied along the member longitudinal axis. The rotation and axial displacement of

zyxw

Strengthening of Corrosion-Damaged RC Columns 531

two sections 250-mm and 500-mm above the base, was also measured through displacement transducers.

Tests along Strong Axis of Column The shear span ratio of the column, when subjected to uniaxial flexure along its strong axis, was 1.60/0.50=3.2. Although (in monotonic loading) the column shear strength exceeds the ratio of the flexural capacity to shear span, for such a shear span ratio and the present low transverse reinforcement ratio column cyclic behaviour may be controlled by shear. The uncorroded control specimen, U-OS, yielded in flexure but then exhibited a mixed flexure-shear failure mode, with bar buckling, some inclined cracking and ultimate disintegration of the concrete core above the base. The corroded control specimen (C-0s) exhibits slightly lower flexural capacity than the uncorroded one - possibly due to loss of steel area due to corrosion - but higher ultimate deflection: 45 mm vs. 40 mm (drift ratio 2.8% vs. 2.5%) of the uncorroded specimen, both determined through the conventional rule of 20%-drop in resistance (Figure 2). Although the difference in the so-defined deformation capacity is small, the post-ultimatedeformation behaviour of the two specimens is very different: the uncorroded specimen suffered a sudden drop in resistance at a peak deflection of 45 mm, whereas the corroded one exhibits gradual strength degradation with increasing deflection amplitudes (up to 55 mm) and a more flexural failure mode (Figure 2b). A possible explanation is that the shear resistance - determined mainly by the contribution of concrete and of the axial load and less by the transverse reinforcement - decreases less due to corrosion than the flexural capacity, and hence the ultimate failure mode and deformation capacity of the corroded specimen is controlled less by shear than in the uncorroded control column. After retrofitting the column with two layers of CFRP (C-C2S, Figure 3b), the response changed radically: after yielding at a deflection of about 15 mm, peak resistance was maintained constant with increasing displacement amplitude (while it had dropped after 25 mm in C-0s). Peak resistance was still controlled by flexure at the base section and was about the same as in the unretrofitted specimen C-0s. After a displacement of 75 mm the cumulative lateral expansion of the compressed concrete inside the CFRP jacket caused jacket rupture when reaching the displacement of 80 mm (drift ratio 5%, see sudden drop in member resistance at 80 mm). Then rupture of a longitudinal bar took place before reaching a deflection of -80 mm (Figure 4b). Thus, the gain in deformation capacity amounts to about

zyxw zyx

zy

zyxwvu

532 FRPRCS-6: Externally Bonded Reinforcement for Shear

80%. A similar force-deformation response was exhibited by a companion specimen with uncorroded reinforcement (not presented here) subjected to the same loading history, except that the uncorroded specimen maintained its lateral load carrying capacity to a higher displacement of 90 mm. No gradually softening branch was noted in either specimen. This modification of response over that of the unretrofitted specimens (corroded or not) is attributed partly to the significant increase of flexural deformation capacity due to the confinement of concrete and also to the increase in shear strength due to the contribution of the FRP jacket (approximately proportional to the total jacket thickness and the FRP modulus of elasticity).

zyxwvutsr

Figure 2. Force-deflection loops for control specimens: (a) U-0s (uncorroded), (b) C-0s (corroded) 250 200 150

1w

zyxwvutsrqponmlk zyxwvutsr 250

200

151)

-

2 e o 9 -so

100

8 ,

g

50

-100

-100

-150

-150

-200

-260 -100

zyxwvutsrqponmlkjihgfedcbaZYXWVUTS

50

-200 -50

0

Displacement (mml

50

Iw

-250-lw

50

0

Dlrplacernsnl (mm) 5o

Force-deflection loops for specimens retrofitted with 2 CFRP layers: Figure 3. zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA C-C2Sin (with initial damage), (b) C-C2S (without initial damage)

Figure 4. Failure of (a) unretrofitted specimen C-OS, and (b) retrofitted specimen

c-c2s

zyxw zy

Strengthening of Corrosion-Damaged RC Columns 533

zyx zyx

In specimen C-C2Sin the two layers of FRP were applied to the column after it had gone through cycles of increasing amplitude of up to 25 mm. The initially damaged specimen exhibits faster strength degradation with cycling (Figure 3a) and lower ultimate deformation capacity than the companion initially damaged column (65 mm, i.e. 4.1%, vs. 80 mm or 5%). Failure was again by CFRP rupture followed by buckling and failure of a longitudinal bar. The difference in ultimate deformation, which has been consistently found in four pairs of undamaged or initially damaged specimens, may be explained by the fact that in the initially damaged specimen activation of the CFRP starts after the concrete has undergone some damage and lateral expansion; so it reaches earlier its (confined) crushing strain, triggering uncontrolled expansion and CFRP fracture. Increasing the number of layers of CFRP to 5 (specimen C-C5S) contributes marginally to member strength (Figure 5a), although it enhanced concrete confinement. Despite the increased jacket stiffness over that of specimen C-C2S, member deformability did not improve: the specimen sustained cyclic displacements of 55 mm, but at a displacement of 60 mm (drift ratio 3.75%) suffered fracture of one bar and a drop in resistance of 25% of the previous maximum value. The test was continued for two more cycles, in which neither the (reduced) lateral load capacity in the positive direction changed, nor the (unaltered after-peak) corresponding capacity in the negative direction, until another bar fractured on the positive. Fracture of the corroded steel bar preceded fracture of the CFRP and became the limiting factor. 2%,

zyxwv

.

,

zyxwvu -2%

-100

zyxwvutsrqponm .

!

-50

.

!

0

.

Displasomant(mml

!

50

a

100

Figure 5. Force- deflection loops of: (a) C-CSS (5 CFRP layers), (b) C-GSS (5 GFRP layers)

With glass FRP for the retrofitting (specimen C-GSS), the 5-layer jacket applied has a stiffness equal to that of the 2-layer carbon FRP jacket. The increased energy dissipation over the unretrofitted specimen demonstrated when retrofitting with CFRP, was also exhibited by this

zyxw

534 FRPRCS-6: Externally Bonded Reinforcement for Shear

specimen (Figure 5b). Again, the behaviour was purely flexural due to the contribution of the GFRP to the shear capacity of the member. Failure occurred at a displacement of 65 mm (drift ratio 4.1%) by fracture of one longitudinal bar, as compared to 100 mm (drift ratio 6.25%) in the companion uncorroded specimen (not included here). Compared to specimen C-C2S, which had the same jacket stiffness, this specimen had a slightly inferior performance. What is most interesting is that specimen C-CSS, with 5 layers of CFRP, did not perform better than either of these two specimens. It seems that there is no benefit in increasing the number of CFRP layers beyond a certain limit, which, for this particular specimen, is quite low.

Tests along Weak Axis of Column The specimens subjected to uniaxial flexure along their weak cross-section axis (shear span ratio 1.60/0.25=6.4) are expected to have a clearly flexural behaviour. Nonetheless, again the uncorroded control specimen, U-OW, exhibits lower deformation capacity (65 mm or drift ratio of 4.1%, vs. 70 mm or 4.4%) and a more rapid post-ultimate strength degradation (Figure 6a) than its corroded counterpart, C O W (Figure 6b). This feature of U-OW and its failure mode, which involved extensive inclined cracking, are reminiscent of shear-controlled behaviour, despite the high shear span ratio (Figure 6b). The application of two layers of carbon FRP increased member deformation capacity drastically (Figure 7b, compared to Figure 6b), while strength was not much affected (as expected, due to the marginal contribution of increased concrete confinement to member strength). Strength degradation of the retrofitted column evolved much slower than in the unretrofitted specimen (Figure 7b), permitting the column to retain large proportion of its resistance for many cycles after the peak. The conventionally defined (through the 20%-drop in resistance rule) deformation capacity increases from about 70 mm (drift ratio 4.4%) to about 115 mm (7.2%). Failure was by CFRP fracture (Figure 8b). The initially damaged specimen (Figure 7a) has in this case similar behaviour and about the same deformation capacity as the undamaged one (Figure 7b). It is noted here that for this small number of CFRP layers (i.e. for small volumetric ratio of FRP material), jacket failure by tensile fracture of fibres was observed in both specimens C C 2 W and CC2Win, as expected.

zyx

zyxw

Strengthening of Corrosion-Damaged RC Columns 535

Figure 6. Force- deflection loops for specimens (a) U-OW (uncorroded), (b) C O W (corroded) zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFE

zyxwvut , ~ "

zyxwvutsrq

Deflection (mm)

Figure 7 . Force- deflection loops for specimens retrofitted with 2 CFRP layers: (a) C-C2Win, (b) C-C2W

Figure 8. Failure of (a) unretrofitted specimen C-OW, and (b) retrofitted specimen c-c2w

For the present specimen the application of 5 layers of carbon FRP, as compared to 2 layers of the same material, does not considerably affect strength (as expected) or the conventionally defined member deformation capacity. Referring to Figures 7b, 9a and 9b, the 2-layer CFRP jacketed specimen has a conventionally defined (at 20% drop in resistance) ultimate drift of 7.2% and as compared to drift of 7.5% in the specimens with 5

zyx zyxwv

536 FRPRCS-6: Externally Bonded Reinforcement f o r Shear

CFRP or GFRP layers. In that respect, the addition of 3 more FRP layers did not prove to be very beneficial. Nonetheless, at a drift ratio of 8.45% specimen C-C2W suffered CFRP fracture, whereas the test of specimens C-CSW and C-GSW stopped at post-conventional ultimate drift ratios of 8.75% and 8.1%, respectively, without fracture of the FRP or the reinforcement. XI

Z

zyxwvutsrqp

zyxwvutsrqponmlkjihgf -

D

5 .

XI

XI 150

Deflort,on imm,

100

XI

zyxwvuts 0

I0

1w

150

o.nestlOn imm)

Figure 9. Force- deflection loops for specimens: (a) C-C5W (5 CFRP layers), (b) C-G5W (5 GFRP layers)

CONCLUSIONS

The present test results have shown that: (a) FRP wrapping of columns without earthquake resistant detailing and with corroded reinforcement does not improve strength - which is controlled by the flexural capacity at the base and affected by the loss in steel area - but increases dramatically deformation (drift) capacity to levels not easily achievable through confinement by conventional jacketing. This improvement is due to the increase in strain capacity of the compressed concrete and the restraint of bar buckling by the F W jacket, as well as to the suppression of the effects of shear on deformation capacity. (b) In the strong direction of the column, where the lower drift capacity is due to larger depth and lower shear span ratio, failure of the retrofitted element was associated with fracture of longitudinal reinforcement. This had not been observed on companion columns with non-corroded reinforcement, which exhibited larger deformation capacity. It seems that corrosion reduces the ductility of the rebars, setting therefore a limit to the improvement in deformation capacity that can be effected through FRP wraps. (c) Application of the FRP jacket to a column which had been carried to yielding of the reinforcement and to moderate damage by previous cycling, gives lower deformation capacity in comparison to an initially

zyxw

Strengthening of Corrosion-Damaged RC Columns 537

undamaged column. The difference may be due to the fact that concrete has undergone some lateral expansion in the absence of the FRP jacket and its activation. (d) In the strong direction of the specimen, in which bar fracture is the limiting factor for deformation capacity, increasing the number of the FRP layers from a low value of two to five does not offer any advantage. Some advantage is offered by such an increase in the weak direction, in which concrete confinement and FRP action control deformation capacity. Nonetheless, this advantage is disproportionately small in comparison to the additional material cost. (e) At first sight FRP wraps are expected to be more effective in the strong direction of the column, as the more narrow width of the compressed zone lends itself better to confinement by the FRP jacket and improvement in the strain capacity of the compression zone, while the confinement effect of the FRP over the wide compression zone of the weak direction is certainly smaller. Nonetheless, in the present case premature fracture of the corroded reinforcement has prevented full utilization of the larger confining effect of the FRP in the strong direction of the column. (f) Contrary to the unretrofitted specimens, which retain their axial load capacity after attaining ultimate deformation and losing their lateral load capacity, columns retrofitted with FRP wraps lose practically all their axial load capacity when they fail explosively by fracture of the FRP wrap. Failure by bar rupture does not have such severe consequences on axial load capacity.

zyxw zyxw

ACKNOWLEDGMENTS

The General Secretariat for Research and Technology (GSRT) of the Greek Ministry of Development provided partial financial support to this research. SIKA provided the FRP materials. REFERENCES 1. Federation International du Beton, Externally bonded FRP reinforcement for RC structures,fib Bulletin 14, Lausanne, 2001. 2. Matthys, S., Taerwe, L. and Audenaert, K., “Tests on axially loaded concrete columns confined by FRP sheet wrapping”, 4th International

zyxw zyxwv

zyxwvu

538 FRPRCS-6: Externally Bonded Reinforcementfor Shear

3.

4.

5.

6.

Symposium on FRP for Reinforced Concrete Structures, Baltimore, USA, 1999, pp. 2 17-228. Nanni, A. and Bradford, N. M., “FRP jacketed concrete under uniaxial compression”, Construction and Building Materials, 9(2), 1995, pp. 115-124. Seible, F., Priestley, M. J. N. and Innamorato, D., “Earthquake retrofit of bridge columns with continuous fiber jackets”, In Design guidelines, Advanced composite technology transfer consortium, 2, Report No. ACTT-95/08, Univ. of California, San Diego, 1995. Bousias, S., Triantafillou, T., Fardis, M., Spathis, L., and O’Regan, B., Use of fibre reinforced polymers in repairhetrofit of reinforced concrete elements, Research Report to the General Secretariat for Research and Technology, Greece, 200 1. Sheikh, S., Pantazopoulou,S., Bonacci, J., Thomas, M. and Hearn, N., Repair of delaminated circular pier columns by ACM, Ontario Joint Transportation Research Report, MTO Reference No. 3 1902, 1997.

zyxwvutsrq zyxwvut

FRPRCS-6, Singapore, 8-10 July 2003 Edited by Kiang Hwee Tan QWorld Scientific Publishing Company

zyxwv zyxwvu zyxw

SHEAR STRENGTHENING OF CONCRETE BRIDGE DECKS USING FRF' BAR P. VALERIO Te.i. co, via Giangiacomo Porro 18, 00197, Rome, Italy

T. J. IBELL Department of Architecture and Civil Engineering, University of Bath. Bath, BA2 7AY, United Kingdom

The shear capacity of existing concrete bridge beams is often inadequate and unable to meet current code requirements. This paper deals with a new type of shear strengthening for existing concrete bridges. It is proposed that vertical FRP bars be inserted into pre-drilled holes and fastened in place using epoxy resin. This method has the advantage that only the soffit of the concrete bridge beam (or slab) is required for access, allowing the top surface to remain undamaged during strengthening. This could allow the bridge to be used during strengthening works, with traffic relatively unhindered by work being carried out below. Ten laboratory tests are presented here to demonstrate the system, and comparisons are made against current code predictions for the strength of such concrete beams with and without transverse reinforcement. The results of this work show that the proposed strengthening scheme is effective and provides significant improvement in the shear-carrying load capacity.

INTRODUCTION Many concrete bridge elements are deteriorating, leading to a reduction in their flexural and shear strength. This deterioration may be due to poor initial design or construction (including poor material selection or poor workmanship), increased traffic loads and aggressive environments. If a concrete bridge is found to have inadequate shear strength and individual webs are inaccessible (for example in the case of many parallel closely-laid beams), one option to strengthen the bridge in shear is to insert threaded vertical steel bars through the deck and bolt on end-plates. However, this method requires access to both the soffit and top surface of the bridge. This is problematic in terms of disruption in the use of the bridge and maintenance might also be a problem, so stainless steel is often used. This adds expense and means that the stainless steel bars must be isolated

zyxw

540 FRPRCS-6: Externally Bonded Reinforcement for Shear

from the reinforcing bars in some way, in order to prevent accelerated corrosion of these reinforcing bars. The work described in this paper attempts to circumvent these problems. It is proposed that vertical holes are drilled into the bridge deck from the soffit level. FRP reinforcing bars are then inserted and embedded in place using resin. In this way, shear strength enhancement is possible. Due to its ongoing popularity, the equivalent steel-bar solution is also considered. This shear-strengthening technique implies that some pressure would need to be applied to inject high-viscosity adhesive into the drilled holes. This research project was concerned with the feasibility of the structural strengthening capabilities, rather than with the on-site practicalities.

zyxwvu

TEST PROGRAMME

In order to verify the practicality and feasibility of the proposed verticallyembedded-bar shear strengthening scheme, the following test programme was conducted. The FRF' reinforcement used was Arapree' bar whose main properties, in accordance with manufacturer's data, are a tensile strength of 1.5 GPa, Young's Modulus of 60GPa, ultimate strain of 2.4% and density of 12.5 kN/m3. Ten beams were tested under four-point loading to provide constant shear within the shear spans. Each beam had a similar cross-section and contained the same quantity of bottom steel reinforcement (2 T12 high yield bars). Figure 1 shows the typical dimensions and longitudinal reinforcement in the specimens. The first specimen contained no transverse reinforcement. The second specimen contained five manually-drilled, unfilled 12mmdiameter vertical holes in each shear span (see Figure 2). The third and fourth specimens contained respectively five similar manually-drilled vertical holes in each shear span, but this time each filled with epoxyresined Arapree (10mm diameter) or steel bars (T10 deformed bar). The fifth specimen contained three Arapree bars angled at 60" to the horizontal. The sixth and seventh specimens contained respectively three vertical Arapree or steel bars spread out over the shear span. The eighth specimen contained two vertical Arapree bars in each shear span, each of diameter lOmm as before. The ninth specimen was similar to the eighth, but contained 7.5mm diameter Arapree bars this time. The tenth specimen contained just a single l0mm-diameter Arapree bar inserted in the centre of each shear span. Table 1 summarises the reinforcement in each specimen.

zyx

zyxz zyxwv zyxw

zyxwvuts Shear Strengthening of Concrete Bridge Decks 541

Il-

& PJ2

4

PJ2

1 1

t

t

(a) Elevation of beams

(b) Typical cross section

zyxwv

Figure 1. Details of the test specimens

Table 1 . Shear strengthening details for each specimen Spec. No. Long. reinforcement

Trans. reinforcement

1

2 T12 steel

none

2

2 T12 steel

5+5 holes only

3

2 T12 steel

5+5 lOmm Arapree bars vertical

4

2 T12 steel

5+5 T 10 steel bars vertical

5

2 TI2 steel

3+3 lOmm Arapree bars angled 60"

6

2 T12 steel

3+3 lOmm Arapree bars vertical

7

2 T12 steel

3+3 T10 steel bars vertical

8

2 TI2 steel

2+2 lOmm Arapree bars vertical

9

2 T12 steel

2+2 7.5mm Arapree bars vertical

10

2 T12 steel

1+1 lOmm Arapree bar vertical

542 FRPRCS-6: Externally Bonded Reinforcement f o r Shear

zyxw I

zy

zyxw zyxwvu I

I

500

Figure 2. Positioning of vertical hole

The required concrete cube compressive strength was about 50 MPa and Figure 3 shows the typical test set-up for each of the specimens.

zyxwv Figure 3. Overall test set-up

TEST RESULTS AND DISCUSSION

Table 2 shows a summary of the results of all tests. All beams cracked in flexure around 15kN.Those beams strengthened in shear with three or five FRP bars in each shear span (Beams 3 to 7) attained full ductile flexural response. The beams which were either not strengthened or strengthened with only one or two bars in each shear span (beams 1, 2, 8, 9 and 10) all

zyx zyxw zyxw zy zyxw zy zyxwvuts Shear Strengthening of Concrete Bridge Decks 543

failed in brittle shear. Figure 4 shows the shear failure for specimen 10, containing just one bar in each shear span. Note how the shear discontinuity was constrained to occur between the single bar and the load point. This increased the shear capacity considerably compared with Beam 1, as seen in Table 2. Table 2. Test results

Beam No.

Average f,, W a )

Failure mode

Peak failure

Maximuim midspan defieccic ' "m

load (kN)

(mm\

1

51

Shear

45

12

53

Shear

42

9

60

Flexural

83

> 40

51

Flexural

80

> 40

55

Flexural

83

> 40

50

Flexural

83

> 40

60

Flexural

76

> 40

8

59

Shear

64

13

9

59

Shear

64

13

10

59

Shear

60

12

2

3 4 5

6 7

Figure 4. Shear failure of Beam 10

zyxwvuts zyxw zyx

544 FRPRCS-6: Externally Bonded Reinforcement f o r Shear

Figure 5 shows the midspan load-deflection plots for all specimens, clearly demonstrating the flexural ductility that was exhibited by those specimens containing sufficient shear strengthening.

Applied load against midspan displacement

zyx zyxwvu -Beam1

---

Beam 2

Beam 3 Beam 4

- _ _ _ _Beam 5

Beam 6

- - .-Beam

-

7

Beam 8

Beam 9

Beam 10

0

10

20

30

40

50

Displacement (mm)

Figure 5. Load-deflection plots for all specimens

zyxwvu zy

FLEXURAL ANALYSIS

In the uncracked phase, it may easily be shown that the depth to the neutral axis is x = 120 mm and the effective second moment of area I = 116x lo6 mm4. The theoretical first cracking moment M,, is given by M, = f c , (I/h-x), wheref,, = 3.9 MPa is the average measured concrete tensile strength and h is the overall depth of the beam (220 mm), so that M,, = 4.5 kNm. Therefore, it may easily be shown that the theoretical total applied load at first flexural cracking is P,, = 15.1 kN. This value matches well with that observed in all tests. At the ultimate limit state, assuming the concrete compressive cube strengthsf,, for each test and the steel yield strength&, = 635 MPa, we find that the ultimate moment of resistance of each beam turns out to be in the range of Mu, = 24.1 to 24.6 kNm (forfc, = 5OMPa to 60MPa). Therefore, the theoretical total applied ultimate load ranges from P, = 80 to 82 kN. This range too is very close to that observed in the specimens which failed in flexure (tests 3 to 7).

zyxz

zy zy

Shear Strengthening of Concrete Bridge Decks 545

SHEAR PREDICTIONS FROM CODES-OF-PRACTICE

Comparisons between the ultimate load observed and code predictions are now made for all beams. The predicted ultimate load capacity for beams 1 and 2, each of which contains no transverse reinforcement, is P,,= 2V,, where V, is the concrete contribution term with all the safety factors put equal to unity. For beams 3 to 10, with transverse reinforcement, P,, = 2 V,,, where V,, = V, + 5.The assumptions made in determining p a r e explained below. Codes-of-practice BS8 1 102, Bridge Assessment Guide BD44/953, ACI-3 184 and Eurocode EC25 are used here for comparison purposes. For the V, term from ACI-3 18, it is assumed that the equivalent cylinder compressive strengthf', = 0.80f,,. For the V , term, it is assumed that the vertically-embedded FRP bars will strain to 0.004 at the ultimate shear capacity of the beams617.As the bars contain no hooked corners, no further strain checks are made which would relate to bent portions of FRP stirrups. With a Young's Modulus of the F W bars of 60GPa, the stress in the FRP bars at shear collapse is 240MPa. Thus, the term for V, based on a 45" truss analogy, becomes, for steel and FRP transversally reinforced beams respectively:

zyxwvu V f=

635 .Ar.z sv

v/= 240.Af.z sv

zyx zyx

where Af is the cross-sectional area of each bar, z is the effective lever arm of the truss and sv is the spacing between vertical bars. Note that although the codes-of-practice adopted here limit the spacing between stirrups to various fractions of the effective depth, d, the calculations conducted here ignore this limitation. Naturally, the design of an adequate shearstrengthening scheme would require closely-spaced vertical bars in reality. The value of z is taken to be the fully-anchored length of each embedded bar, which is the overall length of each bar minus the anchorage length at each end. For purposes of analysis here, it is assumed that the average bond strength between epoxy-resined bar and concrete is of the order of 12MPa'. This translates to an anchorage length, Zb, of 50mm for the lOmm diameter bars, so that in all cases, it is assumed that

zyxw zyxwvu zyxwv zyx

546 FRPRCS-6: Externally Bonded Reinforcement for Shear

z = h-2.1, = 220-2x

50 = 120mm

(3 1

Table 4 shows details of comparisons between the various code predictions and the actual results. Where the shear capacity prediction is higher than the relevant flexural capacity prediction, the flexural capacity is used for comparison purposes. It is clear that all codes-of-practice predict the shear and flexural strengths (as relevant) reasonably accurately. This is important, as it implies that this Strengthening scheme could be used with confidence by practising engineers. However, for specimens 8 and 10, in particular, the codes-of-practice over-estimate the effectiveness of the vertical bars. This is almost certainly due to the wide spacing of the bars, which is close to one effective depth for specimen 8 and substantially more than one effective depth for specimen 10. Therefore, clearly it is essential that in order for this strengthening scheme to be used in reality, the vertically-embedded bars should be spaced sufficiently closely in order for shear predictions to be valid. It seems sensible that this minimum spacing should be in the region of 0.5 to 0.75 times the effective depth, just as recommended by present codes-of-practice for shear design. If the presence of the vertical bars is ignored entirely in specimens 8, 9 and 10 (due to the wide spacing), all codes-of-practice substantially underestimate the shear capacity, which is then based solely on the value of V,. So, in these three specimens, it seems that although the bars are spaced too widely to be fully effective, they do indeed enhance shear capacity by altering the shear discontinuity geometry. CONCLUSIONS From the test results and comparisons with code and plasticity-based predictions, the following remarks may be made. The proposed shear-strengthening approach has been shown to be feasible and successful. Such a shear-strengthening technique for concrete bridges offers many advantages over the traditional threaded-bar-and-plate approach, such as only access to the soffit being required, easier and quicker installation, and lower maintenance.

zyxw zy

Shear Strengthening of Concrete Bridge Decks 547

zyxwvu

Table 4. Correlation between code-of-practice predictions and test results Beam No.

Actual Capacity

1

(kN) 45

2

3 4 5 6

7 8 9

10

(Shear) 42 (Shear) 83 (Flexure) 80 (FI exure) 83

BS8110 Pred. (kN) 48 (Shear) 49

BD44/95 Pred. (kN)

47 (Shear) 47

ACI-318 Pred. (kN) 41 (Shear) 42

EC2 Pred. (kN) 41 (Shear) 42

(shear) (Shear) (Shear) (Shear) (Shear) 82 82 82 82 (Flexure) (Flexure) (Flexure) (Flexure) 80 80 80 80 (shear) (Shear) (Shear) (Shear) (Flexure)(Shear) (Flexure (Flexure) (FI exure) 81 81 81 81

(shear) (shear) (Shear) (Shear) (Shear) (Shear) (Shear) (Shear) (Shear) (Shear) 83

80

80

77

77

76 (F1exure) 64 (Shear) 64 (Shear) 60 (Shear)

82 (FI exure) 78 (Shear) 66 (Shear) 69 (Shear)

82 (F1exure) 76 (Shear) 64 (Shear) 67 (Shear)

82 (Flexure) 72 (Shear) 60 (Shear) 63 (Shear)

82 (F1exure) 73 (Shear) 61 (Shear) 64 (Shear)

(shear)(Shear) (Shear)(Shear) (Shear)(Shear) (Shear) (Shear) (shear) (Shear)

No particular differences were noticed between the beams reinforced with steel bars and the ones reinforced with FRP, so that the use of F W is suggested for such strengthening due to its lightness and corrosion resistance. The spacing between embedded bars should be close enough so that the shear discontinuity cannot form between bars. It is suggested that existing requirements for maximum spacing of vertical reinforcement, which vary between 0.5 and 0.75 times the effective depth, should be adequate for such strengthening. Existing codes-of-practice adequately predict behaviour of this strengthening scheme when closely-spaced vertical bars are used. It is therefore concluded that it is possible to design a shearstrengthening scheme using embedded FRP bars by assuming that the provided a value of 0.4% for the reinforcement contribution is Afifi&'~/ssy, ultimate design strain &fwd is chosen.

548 FRPRCS-6: Externally Bonded Reinforcementfor Shear

zyx

ACKNOWLEDGEMENTS

The authors gratefully acknowledge the help of the laboratory staff and the financial support from the Department of Architecture and Civil engineering at the University of Bath, and Sireg, who supplied generous discounts on the Arapree materials. REFERENCES

1. Sireg S.p.A., “Arapree-Carbopree bars”, Sireg Geotechnical Division Catalogue, Arcore, Italy, 200 1. 2. BS 8 1 10, “Structural use of concrete. Part 1: Code of Practice for design and construction”, British Standards Institution, London, 1985. 3. BD 44/95, “The assessment of concrete highway bridges and structures”, Department of transport, London, 1995. 4. ACI Committee 3 18, “Building Code Requirements for Reinforced Concrete (ACI 3 18-02) and Commentary (ACI 3 18R-02)”, American Concrete Institute, Detroit, 2002. 5. EC2, “Design of concrete structures, Part 1: General rules and rules for buildings”, 1992. 6. Arduini, M., Nanni, A., Di Tommaso, A. and Focacci, F., “Shear response of continuous RC beams strengthened with carbon FRP sheets”, Proceedings of the 3rd International Symposium on NonMetallic (FRP) Reinforcement for Concrete Structures (FRPRCS-3), Sapporo, Japan, October 1997, pp. 459-466. 7. Umezu, K., Fujita, M., Nakai, H. and Tamaki, K., “Shear behaviour of RC beams with aramid fiber sheet”, Proceedings of the 3rdInternational Symposium on Non-Metallic (FRP) Reinforcement for Concrete Structures (FRPRCS-3), Sapporo, Japan, October 1997, pp. 491-498. 8. Ibell, T.J. and Burgoyne, C.J., “The use of FRPs compared with steel for shear reinforcement of concrete”, ACI Structural Journal, 96(6), 1999, pp. 997-1003.

zyx zyxw

(shear) (Shear) (Shear) (Shear) (Shear) (shear) (Shear) (Shear) (Shear) (Shear)

This page intentionally left blank

zyxwvu

FRPRCS-6, Singapore, 8-10 July 2003 Edited by S a n g Hwee Tan BWorld Scientific Publishing Company

STRESS-STRAIN RELATIONSHIP FOR FRP-CONFINED CONCRETE CYLINDERS

zyxwv G. WU',2, Z. LU' AND Z. WU2

I

College of Civil Engineering, Southeast lJniversi& China

2

Department of Urban & Civil Engineering, Ibaraki Universi& Japan

Based on the analysis of more than two hundred specimens of concrete cylinders confined with FRP, a method for predicting the ultimate strength of FRP-confined concrete is presented. First, the calculation of the Poisson's ratio of concrete confined with a sufficient amount of FRP is suggested. Then, according to strain compatibility, the ultimate strain of FRP-confined concrete is predicted. Finally, a tri-linear model is suggested to predict the stress-strain response of FRP-confined concrete cylinders. Through comparison with existing experimental data of other researchers, the effectiveness of the proposed model is verified.

INTRODUCTION In recent years, external confinement of concrete with fiber reinforced polymer (FRP) has emerged as a popular method for the retrofitting of existing concrete columns for enhanced strength and ductility. Numerous tests on FRP-confined concrete cylinders have thus been conducted'-''. '*, ", and a number of theoretical stress-strain models have been proposed for FRP-confined concrete'"' '-16 . It is hard to accurately predict the stress-strain response of FRP-confined concrete, because parameters affecting the performance of FRP-confined are numerous, in particular, the types of FRP are diversified and the property of FRP is scattered. Based on the analysis of a large number of test databases, this paper puts forward a new method to predict the ultimate strength and strain of FRP-confined concrete cylinders and suggests a stress-strain model.

AVAILABLE EXPERIMENTAL RESULTS For the assessment of previous models and for the development and calibration of a new model, existing experimental results of more than two hundred specimens given in the references are used in this paper. The available experimental results cover a wide range of values for several parameters that affect the mechanism of confinement:

552 FRPRCS-6: Externally Bonded Reinforcement for Confinement

(1) Concrete strength between 23.0 N/mm2 and 75.4 N/mm2 were tested. (2) The majority of results were obtained from tests on the cylinders with a dimension of 0100 mmx200mm or 0 150 mmx305mm. (3) CFRP, GFRP, AFRP were used to confine the concrete, the CFRP includes common CFRP (with a modulus less than 250 GPa) and high modulus CFRP (with a modulus greater than 250 GPa). The FRP was applied in the form of a tube or sheet. (4) The tensile strength of FRP varied between 330 N/mm2 and 4433 N/mm2 , the modulus of FRP between 19100 N/mm2 and 640000 N/mm2 , and the thickness of FRP varied between 0.1 lmm and 3.Omm. ( 5 ) In order to eliminate the influence of steel, all the specimens were without internal longitudinal or transverse reinforcement. (6) This paper focuses on the test specimens without a descending stress-strain response, that is, where a sufficient confinement from FRP has been provided.

zy zyxwvu zy

ULTIMATE STRENGTH OF CONFINED CONCRETE

Many researchers 1, 6, 8-10, 15-16 have suggested some methods to predict the strength of FRP-confined concrete cylinders with reference to the well-known equation proposed by Richart et al. Other researchers", 13* l4 also presented some equations to predict the strength of FRP-confined concrete. There are two significant parameters, confinement modulus (El) and confinement strength (f;>, that will significantly influence the performance of According to previous studies, the effectiveness coefficient is related not only to the ratio of confinement strength to the strength of unconfined concrete, namely fdco, but also to the type of FRP, and it depends on the method used to determine the strength of FRP. The ultimate strength of FRP-confined cylinders can be predicted by following equations.

zyxwvu

(1) For FRP sheet-confined concrete cylinders, where the strength of FRP is determined by tensile coupon tests (Fig. 1a), the expression is:

Lc

-- 1+2.0- .A

Lo

L o

(2) For FRP sheet-confined concrete cylinders, where the strength of FRP is obtained by the value provided by manufacturers (Fig.lb), the expression is:

z zyx

Stress-Strain Relationsfor FRP-Confined Cylinders 553

--- 1+ 3.0- J; A

C

L

O

L

O

(3) For FRP tube-confined concrete cylinders, where the strength of FRP is determined by tensile coupon tests (Fig.lc), the expression is:

LC-- 1+ 2.5- J; L

(3 1

L O

O

zyxwvuts zyxwvutsrqpo zyxwvuts where f c o , and f c c are the strength of unconfined concrete cylinders and FRP-confined concrete cylinders respectively.

4

r ................................................................................

4 r

e 3 2 -

1 -

0 ' 0

0 ' 0

I

0. 5

1. 5

1

0. 5

fdco

fdco

(a) FRP sheet, strength determined by tests

(b) FRP sheet, strength obtained from manufacturers

0 ' 0

0. 5

1

fdco

(c) FRP tube, strength determined by tests Figure 1. Regression equations for the ultimate strength of confined concrete

1

554 FRPRCS-6: Externally Bonded Reinforcement f o r Confinement

zyx

ULTIMATE STRAIN OF CONFINED CONCRETE The ultimate strain of FRP-confined concrete is related not only to the value E d c , orfJfco, but also to the ultimate strain of FRP, the form of FRP, and the property of FRP (normal modulus FRP or high modulus FRP); these parameters were not all considered in existing models6, *-", 13-" . This resulted in a poor comparison between the experimental and the predicted results. Based on the analysis of the existing experimental results, some conclusions can be drawn: (1) FRP is a linearly elastic material, so the Possion's ratio of FRP-confined concrete at ultimate will tend to an asymptotic value if the confinement of FRP is significant, and the value is mainly related to the form and property of FRP, and the value offJfco. (2) The ultimate Possion's ratio of FRP-confined concrete is related to the form of FRP. For FRP tube-confined concrete columns, the FRP tube is often used as the formwork, with the concrete cast in the tube; hence the tube and concrete are in close contact. For FRP sheet confined concrete, FRP is wrapped around the cylinders, and there are voids between the FRP sheet and the concrete surface ineluctably; so the ultimate Possion's ratio of FRP sheet-confined concrete is larger than that of FRP tube-confined concrete. The ultimate Possion's ratio of concrete confined with normal modulus CFRP sheet, GFRP sheet and AFRP sheet can be approximated by [Fig 2(a)l:

z

For concrete confined with GFRP or CFRP tubes, the ultimate Possion's ratio can be approximated by [Fig 2(b)]:

zyxwv

The confinement is more effective for high modulus FRP, but as the number of test databases of cylinders confined with high modulus FRP is small, the ultimate Possion's ratio of high modulus FRP-confined concrete is approximately suggested as:

Stress-Struin Relationsfor FRP-Confined Cylinders 555

r::)"""

zyxwvut zyxwv zyxwvutsr zyxwvu v,, =0.56kf -

where kf is the influence coefficient of high modulus FRP, which is taken by 1.0 when Ef is less than or equal to 250 GPu, and d w ( u n i t GPu) when Ef is greater than 250 GPa. After obtaining the ultimate Possion's ratio of FRP-confined concrete, according to strain compatibility, the ultimate strain of FRP-confined concrete can be easily calculated by the following equation: &

=-E f i

(7)

CC

vu

where E,, is the ultimate axial strain of FRP-confined concrete, and cfiis the ultimate strain of FRP.

3 r

0 ' 0

1 1

I

0. 5

-0

zyxwvu 1

1. 5

0

fdco

(a) Common FRP sheet confined concrete

0. 5

1

fdco

(b) FRP tube confined concrete

Figure 2. Regression equations for ultimate Possion's ratio

COMPARISON OF AVAILABLE MODELS Comparison of existing models and the model proposed in this paper is showed in Table 1, some conclusions can be drawn: (1) Many existing models seem to accurately predict the ultimate strength of FRP-confined concrete; on the contrary, there is a large scatter associated with strain predictions.

zyxwvu zyxw zy

556 FRPRCS-6: Externally Bonded Reinforcement f o r Confinement

(2) The models proposed by Lam and Teng (2001), and Xiao and Wu (2001) can favorably estimate the ultimate stress and strain of FRP sheet-confined concrete, but it is not favorable for FRP tube-confined concrete. (3) The models proposed by Samaan and Mirmiran (1997), and Spoelstra and Monti (1999) can favorably estimate the ultimate stress and strain of FRP tube-confined concrete, but it is not favorable for FRP sheet-confined concrete. (4) It seems that all the existing models cannot favorably predict the ultimate stress and strain of concrete confined by high modulus FRP. ( 5 ) The model proposed by this paper can favorably predict the ultimate stress and strain of concrete confined by any type of FRP.

Researchers

zyxw zyx zyxw

Table 1. Comparison of available models Confined with high Confined with common FRP sheet Stress

modulus FRP sheet

Strain

Stress

~~~

Strain

Confined with FRP tube Stress

Strain

Mean St.dev. Mean St.dev. Mean St.dev. Mean St.dev. Mean St.dev. Mean St.dev.

[I1

1.35 0.41

0.88

0.38

1.20

0.27

3.83 3.78 1.31 0.36 0.61 0.51

[21

1.32 0.38

1.11

0.37

1.23

0.28

1.88 1.09 1.17 0.20 0.57 0.46

[61

1.11 0.20

2.03

1.30

1.09

0.15 2.08 1.57 0.99 0.08 0.93 0.27

181

1.15 0.22

1.17

0.40

1.06

0.13 2.21 1.51 1.07 0.14 0.63 0.42

191

1.35 0.39

1.53

0.71

1.22

0.27

1.56 0.76 1.26 0.29 0.16 0.86

~

~

~

~

[lo]

1.08 0.17

1.58

0.76

1.03

0.11

1.55 0.76 0.96 0.09 0.79 0.29

[I11

1.06 0.16

1.87

1.05

0.94

0.17

1.85 1.34 0.98 0.09 0.86 0.20

[I41

1.09 0.39

0.94

0.33

1.17

0.27

1.48 0.68 1.26 0.33 0.55 0.50

0.97 0.14

1.02

0.32

0.93

0.12

1.72 0.88 0.86 0.16 1.10 0.34

Thispaper 1.02 0.13

1.00

0.31

0.96

0.10

0.94 0.39 0.98 0.10 0.97 0.15

-

[I51

~

Best

This paper

Third

[I11

7his paper

This paper

~ 4 1

~ 3 1

This paper

[6]

2% paper

-

Thispaper

[I11

Stress-Strain Relations for FRP-Confined Cylinders 557

z

STRESS-STRAIN RELATIONSHIP

zyx zyxwv zyxwvuts

Tri-linear Stress-strain Model

This paper suggests a tri-linear model can be used to predict the stress-strain response of cylinders confined with a sufficient amount of FRP (Fig.3), each point in Fig.3 can be determined respectively as following: Point 1( E , ~ ocl , ):

zyxw

where E, is the modulus of concrete, and can be approximated by

Point 2( E , ~ oc2 , ):

oc2 =(1+0.0002E,)~,

(10)

+ 0.0004E,) E,,

zyx E , ~= (I

(1 1)

where E,, is the peak strain of unconfined concrete, and can be approximated by 0.002. point 3 fco

A,

( E,,

,

E,,

can be obtained from Eqs. (1)-(3) and Eqs. (4)-(7) respectively.

):

3

--c

EC Figure 3. Tri-linear stress-strain model

zyxwvu

558 FRPRCS-6: Externally Bonded Reinforcement for Confinement

Comparison with Experimental Results

Figure 4 shows the comparison between the experimental data and the proposed model. A satisfactory agreement is observed for the proposed model. 100 h

Q. v

80 60

v1

E40

zyxwvutsrqp zyxwvutsr

+

I

r

100

2 80

92 60

zyxwvutsrq E

,$ 40

v)

20 0

20

0

0

0.02

0.01

0.03

Strain

0.04

0.05

0.005

0

(a) Mirmiran et al? 120

120 r

100

100

h

h

v

zyx

5 40

0

f: 40

m

v)

20

20

0

0

0

0.005

0.01 0.015 Strain

0.02

0.025

0

(c) Miyauchi et a1.8

100

80 v

;60 al

2 40 v)

O.O1 Strain

0. 02

0. 03

(d) Toutanji 100

-

80

-

zyxwvutsrqp h

h

0.02

;60

60

120

0.015

80

80 v

2

O.O!

Strain

(b) Hosotani et a1.4

LC3

20 0

n "

0

0.01 Strain 0.02

(e) Saafi et a1.l'

0.03

0

0.01 Strain 0.02

(QXiao et aI.

Figure 4. Comparison between the proposed model to test results

0.03

zy z

Stress-Strain Relationsfor FRP-Confined Cylinders 559

CONCLUSIONS

(1) The effectiveness coefficient of FRP is related to the value of f#Lo, the form of FRP, and the method determining the strength of FRP. The ultimate strength of FRP-confined concrete can be approximated by Eqs.( 1) to (3). (2) Methods to predict the ultimate Possion’s ratio of FRP-confined concrete are proposed, from which the ultimate strain of FRP-confined concrete can be easily calculated. The method is simply and suitable for concrete confined by normal or high modulus CFRP sheet, GFRP sheet, AFRP sheet and FRP tube. (3) The stress-strain response of FRP confined-concrete can be predicted by a tri-linear stress-strain model. The model is simply, and compares favorably with many existing test results by other researches. ACKNOWLEDGEMENT Financial support of partial work of this paper from the National High Technology Research and Development Program of China (863 Program) under grant 200 1AA336010 is gratefully acknowledged.

zyxwvu zyx

REFERENCES

1. Michael, N.F., Khalili, H., “Concrete Encased in Fiberglass-Reinforced Plastic”, ACI Structural Journal, November-December, 1981, pp.440-446. 2. Ahmad, H., Khaloo, A.R. and Irshaid, A., “Behavior of Concrete Spirally Confined by Fiberglass Filaments”, Magazine of Concrete Research, N0.156, 1991, ~ ~ 1 4 3 - 1 4 8 . 3. Mirmiran, A. and Shahawy, M., “Behavior of Concrete Columns Confined by Fiber Composite”, Journal of Structural Engineering, ASCE, V01.123, No5, 1997, ~ ~ 5 8 3 - 5 9 0 . 4. Hosotani, M., Kawashima, K. and Hoshikuma, J. “A Stress-Strain Model for Concrete Cylinders Concrete by Carbon Fiber Sheets”, Concrete Structures and Pavements, JSCE, No.592, V-39, 1998-5, pp37-pp52. 5 . Nakatsuka, T., Komure , K. and Tagaki, K., “Stress-strain Characteristics of Confined Concrete with Carbon Fiber Sheet”, Concrete Research and Technology, Vo1.9, July 1998, pp65-78. 6. Samaan, M., and Mirmiran, A. and Shahawy, M., “Model of Concrete Confined Fiber Composite”, Journal of Structural Engineering, ASCE,

zyxwvuts zyxw zyxwvu

560 FRPRCS-6: Externally Bonded Reinforcementfor Confinement

zy

V01.124, No9, 1998, ~ ~ 1 0 2 5 - 1 0 3 1 . 7. Harries, K.A., Kestner, J. and Pessiki, S., “Axial Behavior of Reinforced Concrete Columns Retrofit with FRPC Jackets”, Proceeding of Second International Conference on Composite in Infiastructure (ICCI’98), USA, pp4 11-425. 8. Miyauchi, K., Inoue, S. and Kuroda, T., “Strengthening Effects with Carbon Fiber Sheet for Concrete Column”, Proceedings of the Japan Concrete Institute, Vol. 21, No.3, 1999, pp1453-1458. 9. Toutanji, H.A. “Stress-Strain Characteristics of Concrete Columns Externally Confined with Advanced Fiber Composite Sheets”, ACI Materials Journal, V.96, No.3, 1999, pp397-404. lO.Saafi, M., Toutanji, H.A. and Li, Z., “Behavior of Concrete Columns Confined with Fiber Reinforced Polymer Tubes”, ACI Materials Journal, V01.96, NO.4, 1999, ~ ~ 5 0 0 - 5 0 9 . 1 1. Spoelstra, M.R. and Monti, G., “FRP-Confined Concrete Model”, Journal of Compositefor Construction,August, 1999, pp. 143-150. 12.Xia0, Y. and Wu, H., “Compressive Behavior of Concrete Confined by Carbon Fiber Composite Jackets”, Journal of Material in Civil Engineering, ASCE, Vol. 12, N02,2000, pp139-146. 13.Vintzileou, E., “An Empirical Model for Predicting the Properties for Concrete Confined by Means of Composite Materails”, FRPRCS-5,200 1, Cambridge, UK, pp845-853. 14.Xia0, Y. and Wu, H., “Concrete Stub Columns Confined by Various Types of FRP Jackets”, Proceeding of the International Conference on FRP Composite in Civil Engineering(CICE’2001), 200 1, Hong Kong, China, pp293-300. 15.Lam, L. and Teng, J.G., “A New Stress-Strain Model for FRP-Confined Concrete”, CICE ’2001,Hong Kong, China, pp283-292. 16.Karabinis, A.I., and Rousakis, T.C. “A Model for the Mechanical Behavior of the FRP Confined Columns”, CICE’2001, Hong Kong, China, pp3 17-325. 17.And0, T. and Wu, Z.S., “Study on Strengthening effect of Compressive Concrete with Hybrid FRP Sheets”, Proceedings of the 55* Annual Conference of the Japan Society of Civil Engineering, September, 2000.

zyx

zyxwvu zyxwvu

FRPRCS-6, Singapore, 8-10 July 2003 Edited by Kiang Hwee Tan QWorld Scientific Publishing Company

zyxw

STRESS-STRAIN RELATIONSHIP FOR FRP-CONFINED CONCRETE PRISMS G. WU',2, Z. WU2 AND Z. LU' 'College of Civil Engineering, Southeast University, China 'Department of Urban & Civil Engineering, Ibaraki University, Japan

Based on the analysis of more than one hundred concrete prisms confined with FRP, an analytical method for predicting the initial peak stress and strain values of FRP-confined concrete prisms is presented. The ultimate stress and strain of FRP-confined concrete prisms can be calculatedthrough modifying the corresponding ultimate stress and strain of equivalent cylinder concrete confined with equivalent FRP. Two models which can be applied under different conditions are suggested to predict the stress-strain relation of FRP-confined concrete prisms. INTRODUCTION In recent years, several studies of concrete square or rectangular columns confined by FRP were reported in literature,2273 83 12, l 3 but the stress-strain models of FRP-confined concrete prisms are still not conducted well. Existing ones were mostly proposed based on the models of steel-confined concrete. It is realized that directly applying the models of steel-confined concrete to the case of FRP-confined concrete may result in overestimating the strength because the steel-confined concrete behaves very differently from FRP-confined concrete. This paper puts forward a new method to predict the initial peak stress, initial peak strain, ultimate stress and ultimate strain of FRP-confined concrete based on the analysis of results of more than one hundred specimens. In addition, two stress-strain models are also suggested to predict the stress-strain relation of FRP-confined concrete prisms. EXPERIMENTAL OBSERVATIONS More than one hundred experimental results from existing investigations are used in this paper. The available experimental data covers a wide range of values for several parameters including concrete strength and type of FRP. A typical stress-strain curve of FRP-confined concrete prisms is shown in Fig. 1. After point A, the curve of stress-strain relationship may be either descending or ascending due to the reinforcement level. In this paper, the

zyxwvu zyxwv zyx FJ

562 FRPRCS-6: Externally Bonded Reinforcement f o r Confinement

Point A is defined as an initial peak point, and the ultimate Point B or Point C corresponds to the fracture of FRP reinforcement.

zy

Ultimate point 0

-

C

& Figure 1. Typical stress-strain curve of FRP-confined concrete prism

In order to propose different stress-strain models, the calculations of the stress and strain values corresponding to the initial peak point and the ultimate point are first investigated based on regression analysis. During the regression analysis, some unreasonable data such as ones due to premature failure and some very scattered ones are removed, but all the data are considered in the comparison between the proposed models and experimental results. INITIAL PEAK STRESS AND STRAIN OF CONFINED CONCRETE Based on the analysis of existing experimental results, it can be found that the volumetric ratio of FRP to concrete, the modulus of FRP and the modulus of concrete will significantly affect the initial peak stress and strain. There is a direct proportion between the modulus of concrete and the square root of compressive strength of concrete; therefore, the ratio of elastic modulus

zy zyxwvu

between FRP reinforcement and concrete can be expressed as

/&

Ef

,

and a factor A can be introduced to predict the initial peak stress and strain of FRP-confined concrete prisms, as

where pf is the volumetric ratio of FRP to concrete, and

Ef the modulus of FRP,

Lothe uniaxial compressive strength of concrete cylinder.

zyxw zz

zyxw

Stress-Strain Relationsfor FRP-Confined Prisms 563

Initial Peak Stress

Through a regression analysis, the peak stress of confined concrete is approximated as follows (Fig.2):

2= 1+ 0.0008a1k1A L

where

(2)

O

zy

is the compressive strength of unconfined concrete prism, crcp the

initial peak stress of FRP-confined concrete prisms, and a, the influence coefficient of concrete strength which can be expressed as: L3tl

"1

=--L o

-

30 L o

Moreover, k, is the influence coefficient of high modulus FRP, which is taken the value of 1.O when Efis less than or equal to 250 GPa, or

d w (unit

GPa) when Efis larger than 250 GPa.

2r

0 ' 0

zyxwvut I

2 0 0 a , ~ 400

800

600

zyxwvutsr zyxwv

Figure 2. Regression equation or initial peak stress

Fig.3 shows the comparison between different test data from existing studies3-5, 7-10, 13, 14 . A good agreement is observed. 100 -

P

-

5 6 0

-

3

= Rochette

Mirmiran Nakade 0 HoE(ltm1(96) *Hasotm1(98) .Wakatarka

I. +

2 8 0

0

thao

x

svter

+I5%

-15% -15%

-

e40

3W 20 P

t;o

G

0

20

40

60

80

100

Predicted peak stress (MPa) Figure 3. Comparison of proposed model with different test results

564 FRPRCS-6: Externally Bonded Reinforcement f o r Confinement

z

zyxwvu zy zyx zyxwvu

Initial Peak Brain

The peak strain of FRP-confined concrete can be approximated by (Fig.4 ):

5 = 1+ 0.0034a,k2/Z

(4)

where E,, ,.cCp are the peak strain of unconfined concrete and FRP-confined

concrete respectively, and E,, can be taken as 0.002, k2 is the influence coefficient of high modulus FRP, which is taken by 1.O when Elis less than or equal to 250GPa, or taken by 250/E, (unit GPa) when Ef is greater than 25 OGPa. A good agreement is also observed from the comparison between experimental data and the proposed model (Fig.5).

4 j

0 ' 0

zyxwvu 400

200

600

a?

Figure 4. Regression equation for initial peak stress 0.012

0.008

0.004

0 0

0.004

0.008

0.012

Predicted peak strain Figure 5. Comparison of proposed model with test results

ULTIMATE STRESS AND STRAIN OF CONFINED CONCRETE

In order to consider the influence of FRP lateral confinement, an equivalent cylinder is defined to calculate the ultimate stress and strain of FRP-confined

zyx

Stress-Strain Relations for FRP-Confined Prisms 565

concrete prisms. The diameter of equivalent cylinder D is considered to be the length of longer side of the rectangular section (Fig. 6 ) . Ultimate Stress

zyxwv zyxwvu zy

The ultimate stress of FRP-confined prisms can be approximated through reducing the ultimate stress of equivalent cylinders confined with equivalent FRP by a factork, . fc,

=k3Lc

(5)

Lc

where f,, is the ultimate stress of FRP-confined concrete prisms, is the ultimate stress of equivalent cylinders confined with the same FRP reinforcement around the prisms, which can be easily predicted by the model proposed by the author^'^. Through the data analysis, the regression equation of k3 can be expressed as:

zyxwvu zyxw

where r is the ratio of the radius comer and h is the length of the longer side of a section (Fig. 6). Equi val ent cyl i nder 7

Figure 6 . Concept of equivalent cylinder Fig.7 shows the comparison between experimental data and the

zyxwvuts

zyxwvutsr

566 FRPRCS-6: Externally Bonded Reinforcement for Confinement

proposed analytical model, a favorable agreement is observed for the proposed model.

I

+15%

20

40

60

80

100

Predicted ultimate stress ( m a ) Figure 7. Comparison of proposed model with test data

zy zyxwvu zyxwvu

Ultimate Strain

The ultimate strain of FRP-confined concrete prisms can also be approximated by reducing the ultimate strain of equivalent cylinder concrete confined with equivalent FRP using a different factor k 4 ,that is, Em

(7)

= k4Ecc

where E, is the ultimate strain of FRP-confined concrete prisms, E,, the ultimate strain of equivalent cylinders confined with equivalent FRP reinforcement and can be easily predicted by the models suggested by the authors15.The regression equation of k4 can be expressed as:

Er I 2 5 0 (unit GPa) (8) 250

Er >250

STRESS-STRAIN RELATIONSHIP

Two models for predicting the stress-strain response of FRP-confined concrete prisms are suggested herein. Model I

A parabolic equation is used to define the stress-strain curve of FRP-confined

zyx

Stress-Strain Relations for FRP-Confined Prisms 567

zyxw

concrete in the first region before the initial peak point, a curvilinear equation is used to predict the stress-strain relationship in the second region, the curvilinear equation being first proposed for steel-confined concrete by other researcher^'^, but some coefficients are simplified here. The following expressions are obtained for Model I (Fig. 8).

zyxw zyx

"1

Inital peak p3nt

0'

E

zyxw CP

Figure 8. Stress-strain curve of Model I

oc= o c p ( 2 x - x 2 )

xll

A l-xA

x>l

o,=x

e

(9)

wherex=Ec/EcP, A=-5130B2+44B-0.778, B = o , / ( 1 0 6 ~ c p ) .

zyxwvut zyx

The ultimate stress can be calculated by Eq. (5). It will result in determination of the whole stress-strain curve of the confined concrete. Model I is valid only when the confinement effectiveness is relatively low and the post peak behavior shows a descending response. The applicable range of < 0.14 is suggested, in which J; is the lateral confinement

J/Ao

modulus of equivalent cylinder due to FRP, and f,b is the uniaxial compressive cylinder strength of unconfined concrete. Fig. 9 shows the comparison between the proposed model and some experimental results from existing investigations3,5, 9, 13.

568 FRPRCS-6: Externally Bonded Reinforcement for Confinement 70

r

50

0

0.002

zyx

r

zyxwvu

zyx zyxwvu

0.004

0.006

Strain

0.008

(a) Hosotani3 and Nakatsuka'

0.01

0

0.005

0.01

0.015

0.02

0.025

Strain

(b) Nakadeg and ZhaoI3

Figure 9. Comparison of Model I with experimental results

Model II

The first portion of stress-strain curve is considered to be parabolic, while the second one is idealized to be linear based on the experimental observations. The initial peak stress, initial peak strain, ultimate stress and ultimate strain can be calculated by Eqs. (2), (4),(5) and (7), respectively. Model I1 can be expressed as:

The predicted equations for the ultimate strain of equivalent cylinder suggested in different papersI5 is valid only when the concrete is confined

zyx zyx

zyxwvu zyxwvutsrqp zyxwvutsrqpo Stress-Strain Relations for FRP-Confined Prisms 569

with a sufficient amount of FRP reinforcement. Model I1 is considered to be 2 0.14 , and used to evaluate some experimental applicable for results4, ~ , I I 14 , , Good agreement could be seen in Fig. 11.

J;/xo

50 I

0

0.005

0.01

0.015

0.02

Strain (a) Rochette’ and Pavin”

0.025

0

0.002

0,004

0,006

0,008

Strain (b) Hosotani ( 199q4and LiI4

zyxwv

Figure 11. Comparison of Model I1 with experimental results

CONCLUSIONS

(1) A factor A is introduced to predict the initial peak stress and strain of FRP-confined concrete prisms. The model shows a satisfactory agreement with the experimental results. (2) The ultimate stress and strain of FRP-confined concrete prisms can be predicted by reducing the ultimate stress and strain of equivalent cylinders confined with equivalent FRP reinforcement using two different factors k3 and k4. (3) Two different models are suggested for the stress-strain response of FRP-confined prisms with both post strain hardening and strain softening behavior. Both models were found to agree with existing test results.. ACKNOWLEDGEMENT Financial support for partial work of this paper from the National High Technology Research and Development Program of China (863 Program) under grant 2001AA336010 is gratefully acknowledge.

REFERENCE 1. Xing, Q.S., Weng, Y.J. and Shen, J.M, “Experimental Study on the Complete Stress-strain Curve of Confined Concrete”, Proceeding of National

zyxwvuts zyxwvu

570 FRPRCS-6: Externally Bonded Reinforcement for Confinement

Conference on Application and Theory of Common and Con$ned Concrete, YinTai, China, October, 1987. 2. Saadatmanesh, H., Ehasni, M.R. and Li, M.W., “Strength and Ductility of Concrete Columns Externally Reinforced with Fiber Composite Straps”,ACI Structural Journal, Vo1.91, No.4, 1994, pp.434-447. 3. Hosotani, M., Kawashima, K. and Hoshikuma, J., “A Stress-Strain Relation of Confined Concrete Cylinders Concrete by Carbon Fiber Sheets”, Proceedings of the Japan Concrete Institute, Vol. 18, No.2, 1996, pp95-100. 4. Hosotani, M., Kawashima, K. and Hoshikuma, J., “A Stress-Strain Model for Concrete Cylinders Concrete by Carbon Fiber Sheets”, Concrete Structures and Pavements, JSCE, V-39, N0.592, 1998-5, pp37-pp52. 5 . Nakatsuka, T., Komure, K. and Tagaki, K., “Stress-strain Characteristics of Confined Concrete with Carbon Fiber Sheet”, Concrete Research and Technology,Vo1.9, No.2, July 1998. 6. Harries, K.A., Kestner, J. and Pessiki, S., “Axial Behavior of Reinforced Concrete Columns Retrofit with FRPC Jackets”, Proceeding of Second International Conference on Composite in Infrastructure (ICCI’98), USA. 7. Mirmrian, A., Shahawy, M. and et al., “Effect of Column Parameters on FRP-Confined Concrete”, Journal of Composites for Construction, ASCE, V01.2, No.4, November, 1998, pp175-185. 8. Rochette, P. and Labossiere, P., “Axial Testing of Rectangular Column Models Confined with Composites”, Journal of Compositesfor Construction, ASCE, v04, No.3,2000, ~ ~ 1 2 9 - 1 3 6 . 9. Nakade, A., Yoneoku, H. and Fuchikawa, M., “Experimental Study on the Confinement Effect of Concrete Columns Confined by Carbon Fiber Sheets”, Proceedings of the Japan Concrete Institute, Vol. 23,ko.l, 2001. 10. Suter, R. and Pinzelli, R., “Confinement of Concrete Columns with FRP Sheets”, FRPRCS-5,2001, Cambridge, UK, pp793-802. 11. Parvin, A. and Wang, W., “Behavior of FRP Jacketed Concrete Columns under Eccentric Loading”, Journal of Composites for Construction, ASCE, Vo1.5, No.3,2001, pp146-152. 12. Lam, L. and Teng, J.G., “Compressive Strength of FRP-Confined in Rectangular Columns”, Proceeding of the International Conference on FRP Composite in Civil Engineering (CICE ’2001), December 200 1, Hong Kong. 13. Zhao, T., Xie, J., “Experimental Study on Complete Stress-strian Relation Curve of CFRP Confined Concrete”, Building structure, July, 2000. 14. Li, J., Qian, J.R. and Jiang, J.B., “Study on Complete Stress-strian Relation Curve of CFRP Confined Concrete”, Proceeding of Second National Conference on Fiber-reinforced plastics for reinforced concrete structures,KunMing, China, 2002, pp157-162.

zyxw

zyxwvu zyxwvu

FRPRCS-6, Singapore, 8-10 July 2003 Edited by Kiang Hwee Tan QWorld Scientific Publishing Company

zyxw

CONCRETE CYLINDERS CONFINED BY CFRP SHEETS SUBJECTED TO CYCLIC AXIAL COMPRESSIVE LOAD

zyxwvutsrq

T. ROUSAKIS I , c. s. YOU 2, L. DE LORENZIS 3, v. T A M U ~ S', R. TEPFERS Dept. of Civil Eng., Democritus University of Thrace, Xanthi, Greece Dept. of Mechanical Eng., Pohang Univ. of Sci. and Tech., Pohang, South Korea 3 Dept. of Innovation Eng., University of Lecce, I-73100 Lecce, Italy 'Institute of Polymer Mechanics, Univ. of Latvia, Aizkraukles 23, Riga, Latvia 5 Dept. of Building Materials, Chalmers University of Tech., Goteborg, Sweden

'

The mechanical behavior of concrete confined by carbon fiber-reinforced polymer (CFRP) material is investigated. Two series of concrete cylinders are presented in this study, with compressive strength of 20.5 MPa and 49.2 MPa, and confined by CFRP sheet with 234 GPa elastic modulus. The carbon volumetric ratio ranged between 0.45% and 1.36%. Split-disk tests were performed to estimate hoop properties of CFRP sheet. The concrete cylinders were subjected to monotonic and cyclic axial compressive loading with teflon sheets inserted between concrete and steel bearing platens to reduce friction. The performance of CFRP-confined concrete in terms of strength, ductility and expansion is remarkable.

INTRODUCTION

Rehabilitation and strengthening of structures with inadequate bearing capacity has reached a level of extended design alternatives. Use of fiberreinforced polymer (FRP) materials is growing, as they tend to replace conventional steel as external reinforcement in upgrading of existing structures. Considering the mechanism of confinement in enhancement of strength and ductility of structural members, FRP materials turn to be more advantageous than steel. The linear elastic behavior of FRP reinforcement (tube or sheet) up to failure provides an ever-increasing pressure on the confined concrete core. Many experimental efforts have recently been concentrated in the investigation of the confining effect of FRP materials with a variety of mechanical and physical properties', *. The modeling of the behavior of FRP confined concrete is mostly based on semi-empirical equations3,4, while there are models based on constitutive relationships. The stress-strain response suggested by the various models is strongly dependent on the experimental data they are based on. The theory of plasticity can incorporate the unique dilation characteristics of FRP

zyxw

572 FRPRCS-6: Externally Bonded Reinforcement for Confinement

confined concrete. Recently, modified plasticity models have been developed for the reproduction of the stress-strain behavior of FRP sheet confined concrete’. In FRP-confinement, the failure criterion is the fracture of the confining reinforcement. Usually, in modeling, the ultimate strain of the jacket is taken equal to the ultimate tensile strain of the fibers. Experimental evidence testifies that the measured ultimate tensile strain ( E ~ , ) taken from FRP coupon tests is slightly lower than the ultimate strain of the fibers (given by the manufacturer). For FRP-confined specimens this strain is lower than both values6. In a recent study the reduced, effective ultimate lateral strain of FRP has been imported in modeling, improving the accuracy of semi-empirical models A two-phase experimental investigation has been performed to investigate the behavior of five concrete qualities confined by three carbon FRP (CFRP) materials (sheets and filaments) and subjected to axial compressive load*. In this paper only a part of this investigation is presented concerning two qualities of concrete cylinders, confined by CFRP sheet, in a range of volumetric ratios (pj) and subjected to monotonic and cyclic load. Split-disk tests have been performed to estimate hoop properties of FRP sheet. The enhanced mechanical behavior of CFRP sheet confined concrete is studied in terms of strength, ductility and expansion.

zy

’.

zyxwv

EXPERIMENTS Two concrete mixtures were used, with low and high concrete target strengths respectively for the construction of cylinders with 150 mm diameter and 300 mm height. The concrete composition, strength and modulus of elasticity of the plain cylindrical specimens are presented in Table 1.

zyxwvut zyx zyxwvuts zyxwvut Table 1. Concrete mixtures.

Average proportion ofthe concrete mixture ( /m3) (w+w,)/ Strength c silica k0 w Air d i t i v e s (c+silica) W a fid powder scmd GraVel (liters) % (liters))* (kd dO-8m d:8-12 ~m d : l l - l 6 ~ n (cylidical) 20.5 201.4 1205.3 364.4 364.4 189.5 2.4 0.94 500 16.5 934.5 382.3 382.3 178 2.4 9.45 0.35 49.2

*Superplasticizer 1,33% and retarder 0,5% of binder weight (C+silica).

Cylinders Subjected to Cyclic Compressive Load 573

The FRP wrapping material was BPE Composite 300S, made of Grafil carbon fibers and epoxy resin. Data provided by the supplier are cited in table 2. Moreover split-disk tests according to ASTM D 2290 were performed to estimate the hoop properties (effective ultimate lateral strain) of the FRP sheet. Results are presented in Figure 1.

zyxwvutsrq zyxw zy Table 2. Carbon fiber properties.

GraJil Inc. unidirectional 340-700 carbon BPE Composite 300s carbon sheet Tensile Tensile Tensile Thickness, Mdth, Weight, Strength2 Modulus, Elongation, Density, (dcm') ~ype mm (@a) mm (dm') % (Ma) 300 300 1.9 1.8 Unidkctional 0.17 4500 234

C F R P - S h e e t R i n g Test

zy zyx

--C 2 Layers (b)

-1

0000

0004

0008

Layerr (n)

0 012

Strain

Figure 1 . Specimens of CFRP-ring tensile test and tensile behavior of CFRP-rings.

From the comparison in Figure 2, it can be noted that the effective ultimate lateral strain ranges from 0.5 to 0.64 times the ultimate tensile strain depending on the number of sheet layers. Hoop strength of carbon sheet ranges from 0.41 to 0.61 respectively, while modulus exhibits practically no clear variation. Overlap length of the carbon sheet was taken as 150 mm based on the preliminary tests performed on carbon sheet with the same ultimate axial strength and higher elastic modulus A total of 18 specimens of the low and high strength concrete batches (labeled L and H correspondingly) were tested under axial compressive load. Four teflon 0.2 mm sheets were used between the concrete surfaces and the bearing platens to eliminate base friction. The six plain specimens were subjected to monotonic compressive load up to failure. The twelve wrapped specimens were tested under axial monotonic or cyclic load up to failure. Two identical cylinders were wrapped with one, two and t h e e layers of CFRP with fiber orientation perpendicular to their axis. One

'.

zyxw

574 FRPRCS-6: Externally Bonded Reinforcementfor Confinement

specimen was tested under monotonic load and the other under repeated load with the same loading-unloading rate. The load, axial displacement, axial strain and lateral strain were measured as shown in Figure 3.

1

1 layer (a)

1 layer (b)

zyxwvuts 2 layer (a)

2 layer (b)

3 layer (a)

3 layer (b)

Normalized values are divided by tensile properties of carbon fibers.

Figure 2. Comparison between tensile and hoop properties of carbon sheet.

LVDT #3

zyxwvu zyxw Figure 3. Instrumentation of specimens.

All specimens were conditioned at 20°C and 50% relative humidity and tested after twelve and a half months. The loading machine capacity was 5000 I&. The rate was 10 MPa I minute at monotonic and cyclic loading, similar to that of ASTM C 39/C 39M - 99 standards. The repeated loadunload cycles were related to the unconfined concrete strengthf, and the ultimate strength of confined concrete fco The load of each cycle was

zy zyxwv zyxw zyxw z Cylinders Subjected to Cyclic Compressive Load 57.5

successively increased up to 0.5&, 0.8&,f,, cfc+ 0.33cfcc-f,)),cfc+ 0.66(~&fc)} and finally up to failure. zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONML

EXPERIMENTAL RESULTS Plain concrete exhibited a brittle behavior with rather columnar type of concrete fracture due to elimination of base friction (use of teflon caps), Figure 4a. The curves of plain concrete stress-strain response are presented in Figures 5 and 6.

L batch

H batch

Figure 4.a Unconfined specimens with teflon caps after failure. Vertical cracks.

Figure 4b. Specimen of H batch with 1 layer of FRP sheet, after failure.

I

Figure 4c. Specimen of H batch with 2 layers of sheet under cyclic load after failure.

2 I D bl

i

zyxwvutsrqponm

-0.0015 -0.W Lateral smn I1

0 . W

0.0015

0.W25

0.0035

-

0.W45

Ilx,Lstra,nra

Figure 5. Axial stress - axial and lateral strain behavior of plain specimens from L and H concrete batches.

-0.001 -0.wO5

0

0 . W 0.W1 0.W15 0,002 0.WZ

0,003 0.w35

Figure 6. Stress - volumetric strain behavior of plain specimens from L and H concrete batches.

Confined specimens under monotonic or cyclic load displayed a sudden and explosive failure, after the rupture of the carbon sheet. Carbon sheet

zyxw

576 FRPRCS-6: Externally Bonded Reinforcement for Confinement

z

fracture occurred in some specimens along their height or it was focused in the upper, center or bottom part (Figures 4b and 4c). The mechanical behavior of CFRP-confined concrete for low and high strength concrete batches is illustrated through their stress-strain diagrams (Figures 7 to 10). All specimens show an ever-increasing response of expanding concrete up to failure for monotonic or cyclic mode of loading.

-0.01

-0.02 Lateral S t m n E l

0,01

0 &la

zyxwvutsrq zyxwvutsrqp 0.02

0.03

Stmn La

Figure 8. Stress-volumetric strain response of specimens from L concrete batch confined with 0, 1 , 2 and 3 layers of CFRP sheet.

Figure 7. Stress-strain behavior of specimens from L concrete batch confined with 0, 1 , 2 and 3 layers of CFRP sheet.

zyxwvut - -,-.-

I

a

5

B

-

.-

--........-

zyxwvutsrqpon zyxwvutsrqp -

0.0'

lateral strsn L I

Axm St,a,n ra

Figure 9. Stress-strain behavior of specimens from H concrete batch confined with 0, 1, 2 and 3 layers of CFRP sheet.

-0,015

-0.01 -0.W5 Volumetric Stla," 6"

0

Figure 10. Stress-volumetric strain response of specimens from H concrete batch confined with 0, 1, 2 and 3 layers of CFRF' sheet.

Up to about 0.9 of ultimate unconfined concrete load, the stress - strain response (axial, lateral and volumetric) of the composite system is dominated by the response of plain concrete. After that characteristic load, serious cracks are formed in the concrete core. The jacket interacts with expanding concrete to stabilize its disintegration and a linear stress-strain

Cylinders Subjected to Cyclic Compressive Load 577

relation is obtained up to jacket failure. An envelope-like behavior is obtained for cyclic loading-unloading mode, with gradual reduction of the concrete axial rigidity from cycle to cycle (Figures 7 and 9). However, the FRP jacket tends to prevent further concrete disintegration resulting in higher strength and ductility of confined concrete, regardless of the number of cycles (the tests included loading up to 6 cycles).

zyxw zyxwvu z zyxwvutsrq Table 3. Experimental results.

wdm

UltitmIe

Lnbels

stress

L-2 L-3t L W-lt LIC-H Lz-lt L2C-H

w-lf wc-2 H - It H-2 H-3t H HI-If HlC-H H2-lf HZC-2 H3-lf H3C-2t

17.1 23.0 20.5 41.3 48.7 57.2 63.8 63.1 73.9 50.0 50.5 47.1 49.2 79.0 75.2 83.9 79.2 100.6 108.6

Ultime (ntime W u s h i d s t r a n Lateralstran 4Ehtieiity

0.0014 0.0033 0.0026 0.00%

0.0170 0.0142 0.0191 0.0142 0.0210 0.0017 0.0019 0.0016 0.0017 0.0039 0.0050 0.0035 0.0033 0.0062 0.0076

-0.cco2

-0.ooo4 -0.0003 -0.0080 -0,0102

25.4 23.9 24.4

-0.0064 -0.M)90

-0.0058 -0.0082 -0.0003 -0.0007 -0.0012

-0.ooo7 -0.0044 -0.0074 -0.0026 -0,0022 -0.0048 -0.0062

N d i z t d Nonmliztd Nonmliztd Laad hid&ain Lateralstmin

2.02 2.38 2.80 3.12 3.08 3.62

3.69 6.54 5.46 1.35 5.46 8.08

2667 34.00 21.33 30.00 19.33 27.33

1.61 1.53

2.29 2.94 2.06 1.94 3.65 4.47

6.29 10.57 3.71 3.14 6.86 8.86

40.4 35.5 34.0 36.7

1.70

1.61 2.04 2.21

zyxw

The test results of confined specimens are presented in Table 3. The labels given are decoded as follows: target low (L) or high (H) concrete strength, number of sheet layers (1 or 20r 3), mode of loading (blank for monotonic or Cyclic), number of identical specimen and finally use of teflon or not (t or blank). The normalized values of ultimate stress and strain are presented (divided by the corresponding values of unconfined concrete). For 20.5 MPa concrete strength and 3 layers of jacket the normalized ultimate load at failure varied from 3.08 - 3.62 while for 49.2 MPa the corresponding values ranged from 2.04 - 2.21 times the load of plain concrete. The normalized axial strains were between 5.46 and 8.08, and between 3.65 and 4.47 for the two concrete strengths respectively. The

zyxw

578 FRPRCS-6: Externally Bonded Reinforcement for Confinemeni

normalized lateral strain had similar reducing tendency for higher strength concrete. In cyclic loading of confined specimens, the variation of strength and ductility was similar to that of the monotonic mode of load. In fact the failure values were higher. From the stress - volumetric response diagrams it could be noted that the carbon jacket tends to control concrete expansion. (Figures 8 and 10). For the same confinement volumetric ratio, the lower concrete strength results in better control of concrete expansion. The lateral strain is better restricted in lower strength concrete. For specimens subjected to cyclic loading, the same remarks are valid. Increasing the volumetric ratio of carbon jacket (more sheet layers), led to increase in the rigidity in fiber direction of the jacket. For plain concrete strength (20.5 MPa), the normalized values of ultimate load in Table 3 revealed an increase from 2.02 for one layer ofjacket, to 3.62 times the load of plain concrete for three layers. The increase in ductility varied from 3.69 to 8.08 correspondingly. From the normalized values in Table 3, it could be noted that the increase in both strength and ductility of the confined concrete (for every additional layer of jacket) was lower when the plain concrete strength was higher. However, high strength concrete specimens wrapped with only three layers of carbon sheet reached the strength of 108.6 MPa. No difference was noted for specimens under cyclic mode of loading. During loading, the jacket was under a tri-axial state of stress. The ultimate strain at failure of the carbon FRP jacket for the 20.5 MPa concrete strength ranged from 0.008 for one layer to 0.0058 for three layers. For the 49.2 concrete strength specimen with three layers of CFRP the strain at failure reached 0.0048. It is obvious that the strain at failure of the jacket was lower than the half of the tensile elongation at failure of the carbon fibers, (Table 2). Specimens subjected to cyclic load showed a relatively higher ultimate lateral strain.

zyxwvu

CONCLUSIONS

The response of the composite system, concrete-FRP jacket, depends on the material mechanical properties of the concrete and the carbon FRP sheet, the performance of the wrapping application performance, the overlap length, the volumetric confinement ratio as well as the mode of loading. Also, the friction between the steel bearing platens and the concrete affects the mode of failure of confined concrete. With the use of teflon, the mode of failure changes from shear cone to columnar, while in high strength concrete, a decrease in ultimate stress is observed.

Cylinders Subjected to Cyclic Compressive Load 579

The results obtained from cylindrical concrete specimens confined by BPE Composite 300s carbon FRP sheet, with 150 mm overlap length, use of four-layer teflon caps and tested under monotonic and cyclic axial compressive load, give a bilinear stress-strain relation and an ever increasing strength response for increasing imposed axial deformation of the concrete up to failure. Under cyclic loading the highest strength, 108.6 MPa at a strain of 0.0076 is obtained for 49.2 MPa concrete wrapped with 3 layers (structural thickness 0.35 1 mm) of CFRP. The CFRP-confined concrete fails when the carbon sheet is fractured. The lateral measured strain at failure of the confined concrete on the FRP surface is lower than half of the tensile elongation at failure of the carbon fibers. The ultimate strain of the jacket for the 20.5 MPa concrete shows a slight decrease with the number of carbon sheet layers, from 0.008 for one layer to 0.0058 for three layers. For high concrete strength, there is a similar tendency but it is not so clear. In general the strain at failure of the jacket fibers is lower than the nominal tensile elongation at failure of the carbon fibers.

zyxwvu zyxwvu zy zyxwvu zyxwv

ACKNOWLEDGEMENT

The investigation was sponsored by European Commission - TMR Network, ConFibreCrete “Research leading to the Development of Design Guidelines for the use of FRP in Concrete Structures” and Ake and Greta Lissheds Foundation, SEB Bank, SE-106 40 Stockholm. REFERENCES

1. Samaan M., Mirmiram A., Shahawy M.: Model of Concrete Confined by Fiber Composites. ASCE Journal of Structural Engineering, I? 124,

No 9, September 1998. pp. 1025-1031. 2. Matthys S., Taerwe L., Audenaert K.: Tests on Axially Loaded Concrete Columns Confined by Fiber Reinforced Polymer Sheet Wrapping. 41h International Symposium on Fiber Reinforced Polymer Reinforcement for Reinforced Concrete Structures, 1999. pp. 2 17-228. 3. Saafi M., Toutanji H.A., Li Z.: Behavior of Concrete Columns Confined with Fiber Reinforced Polymer Tubes. ACI Materials Journal, V: 96, NO. 4, July - August 1999. pp. 500-509.

zyx zyxw

580 FRPRCS-6: Externally Bonded Reinforcementfor Confinement

4. Spoelstra M. R., Monti G.: FRP-Confined Concrete Model. ASCE Journal of Composites for Construction, V. 3, No. 3, August 1999. pp. 143-150. 5. Karabinis A. I, Rousakis T.C.: Concrete Confined by FRP Material: A Plasticity Approach. Engineering Structures, Elsevier, 24 (2002). pp. 923-932. 6. Rousakis T.: Experimental investigation of concrete cylinders confined by carbon FRP sheets, under monotonic and cyclic axial compressive load, Research Report, Chalmers University of Technology, Goteborg, Sweden. p. 87. 7. De Lorenzis L.: A Comparative Study of Models on Confinement of Concrete Cylinders with FRP Composites. Chalmers University of Technology, Division of Building Technology, Work No 46. Publication 01 :4. Goteborg, 2001-06-30. p.73. 8. Tamuis V., Chi-Sang You, Tepfers R. (2001), “Experimental Investigation of CFRP-confined Concretes under Compressive Load”. Institute of Polymer Mechanics, University of Latvia, Aizkraukles 23, LV- 1006 Riga, Latvia and Division of Building Technology, Chalmers University of Technology, S-412 96 Goteborg, Sweden, December 2001. p. 68.

zyxwv

zyxwvu

FRPRCS-6, Singapore, 8-10 July 2003 Edited by Kiang Hwee Tan @WorldScientific Publishing Company

CONCRETE CYLINDERS CONFINED BY PRESTRESSED CFRP FILAMENT WINDING UNDER AXIAL LOAD

zyxwvutsrqponmlkjih

T. ROUSAKIS I ,

c. s. YOU 2 , L. DE LORENZIS 3 , v. TAM& ', R. TEPFERS

zyxw

Dept. of Civil Eng., Demokritus University of Thrace, Xanthi, Greece Dept. ofMechanica1 Eng., Pohang Univ. ofSci. and Tech., Pohang, South Korea 3 Dept. of Innovation Eng., University of Lecce, 1-73100 Lecce, Italy Institute of Polymer Mechanics, Univ. of Latvia, Aizkraukles 23, Riga, Latvia Dept. of Building Materials, Chalmers University of Tech., Goteborg, Sweden I

Concrete cylinders confined by CFRP and loaded axially in compression ,display a distinct bilinear stress-strain response with a transition zone around the ultimate strength of unconfined concrete where the decrease in modulus occurs. Prestressed confinement would be favorable in raising the load level of the transition zone and hence improved performance of strengthened columns. A consequent investigation was executed on concrete cylinders of 5 different strengths and confinements with 3 different prestress levels. Theoretical prediction and experiments showed that prestress of the confining device elevates the transition zone, which has importance for the stability of confined concrete columns. The ultimate load is not affected by the degree of prestress. The normalized increase in ultimate load by prestressed confinement decrease with higher concrete strength.

INTORODUCTION Most R&D and field application projects were concentrated on concrete column repair and strengthening using fiber composite materials. FRPmaterials can be used to upgrade civil engineering structures and this could be the most effective way of introducing fiber composites into widespread civil engineering use '. Experimental studies have shown that concrete cylinders confined by CFRP sheets, when loaded in uniaxial compression, display a distinct bilinear stress-strain response with a transition zone around the ultimate strength of unconfined concrete. The slope of the branch after the transition zone depends on the volumetric ratio and the mechanical properties of the confining device and is always lower than that of the initial branch 2, 5, . The use of wrapped concrete columns above the transition zone is questionable because of internal damage of concrete and 23

zyxwvuts 42

582 FRPRCS-6: Externally Bonded Reinforcementfo r Confinement

low tangential slope of load-displacement curve. The lower E-modulus above the transition zone may create problems for columns to take advantage of increased strength by confinement due to reduced Euler stability load '. Therefore it would be favorable to raise the load level of the transition zone to achieve better stability at least up to this load level. This function could be achieved by prestressing the confining device. The objective of this paper is to study the behavior of concrete confined by prestressed CFRP filament winding under axial compressive load. The investigation was executed with standard concrete cylinders of diameter 150mm and height 300mm with 5 different concrete strengths that cover nearly the whole range in practice. The parameters taken under consideration for the confinement were 3 levels of prestress. EXPERIMENTAL PROCEDURE

zy

A total of 30 confined standard cylinder specimens were prepared from five concrete batches with different strengths, C20; C40; C60; C80 and C100. Additionally 15 plain concrete cylinders (three specimens from each batch) were prepared and subjected to monotonic compressive load up to failure as control specimens. In each batch of confined concrete, three different specimens having different prestress levels were prepared and tested under increased repeated load-unload cycles up to failure. Characteristics of Confinement

The concrete cylinders were confined by winding around the cylinders prestressed carbon fiber filament impregnated with epoxy resin. The yarn of carbon fiber Zoltek Panex33 was used for confinement. The producer's data of fiber and resin are given in Table 1. Split-disk tensile test was carried out, following the ASTM standard D-2290, to estimate the properties of uni-directional composite confinement. The diameter of the split disk was 150 mm (the same as for concrete specimen) and the intensity of winding was 5 yams per cm (the same as for real confinement). The volume fraction of fibers was not determined and the strength and modulus of fiber was calculated neglecting the matrix. Results are presented on the third line of Table 1, and it is seen that strength and elongation of FRP jacket is only about half of the values of the filament.

zyxwv

Cylinders Confined by Prestressed Filament Winding 583

zy zyxwvut zyxwvu zyxwvuts Table 1. Mechanical properties of materials for confinement Tensile Strength (MPa)

Material

Carbon yarn (Zoltek PANEX33) (Producer k data) Epoxy resin (NMBPE 41 7) Fiber properties in composite (Tensile test on split disk)

Tensile Modulus (GPO)

Elongation

?A)

3800

228

1.6

50

2

3

1807

218

0.85

Cross-sectional Area

1.86 nunz (Fiber Diameter: 7.2 pm)

Specimen Preparation Five different concrete mixtures were designed to cover compressive strength (measured on cubes with 150-mm side length) ranging from 20 MPa to 100 MPa. For the fabrication of C80 and ClOO batches (corresponding to concrete with 80-MPa and 100-MPa compressive strength, respectively), silica fume and additives were used. The additives were Sikament-56 super-plasticizer and Sika retarder and both were added as a percent of binder (cement and silica fume). Table 2 shows the mechanical properties of concrete.

zyxw

Table 2. Mechanical properties of concrete from experiments Ultimate Stress (MPa) Batch Cylinder c20 C 40 C 60 C 80 c 100

20.5 40.0 44. 49.2 61.6

Cube (I50mm) 34.2 60.5 76.2 81.4 104.1

Elastic Modulus Eb

(GPO)

24.40 34.03 37.83 36.65 39.09

Poisson's Ratio V

0.14 0.17 0.19 0.18 0.19

After taking the concrete specimens out of their molds and curing them, a primer resin was applied on the concrete surface so that the proper underlay for high performance of confinement could be provided. Another type of epoxy resin was applied to fill some cavities and pores on the concrete surfaces. After finishing the treatment of the concrete surface, the confining process with prestressed carbon filaments was performed. The filament fed from the creel was received by the forcing device, where the

zyxw

584 FRPRCS-6: Externally Bonded Reinforcement for Confinemen1

zy

weight of designed prestress level was suspended from the yarn. The applied forces were 300, 600 and 800 N, resulting in prestress levels of 80, 160 and 210 MPa (indicated in the following as prestress levels 1, 2 and 3, respectively) and then the winding process was conducted around the rotating concrete cylinder. The main resin was applied on the concrete cylinder surface and filament before winding. Figure 1 shows the schematic diagram of the filament winding process and a concrete cylinder confined by filament winding. The carbon fiber ratio for specimen sections was 2.48%.

zyxwvu Winding

Tension Specimen

Figure 1 . Winding process and fabricated specimen (dimension: mm)

Two strain gauges were arranged perpendicularly to each other on each cylinder in order to measure the axial and lateral strain at the mid-height and other two were arranged on the opposite side. The specimens were centered on the platen of the loading machine to ensure that there was no load eccentricity and four teflon sheets (0.2-mm thick) were used to reduce the friction between concrete surfaces and loading platens. All specimens were tested in load-controlled mode at a loading rate of 10 MPdminute in axial compression until failure using a hydraulic 5000-kN column-testing machine. For the confined specimens, the maximum load levels in the cyclic loading were designed to be 0.5 o i p ,0.8 oip,oip,2 nip,3 oip and so forth up to failure, oipbeing the strength of plain concrete. Minimum load level at each cycle was 0 kN.

EXPERIMENTAL RESULTS The test results, compressive strength, ultimate axial and lateral strains are presented in Table 3 and the normalized values of ultimate strength are presented in Figure 2. The normalized effect of confinement on compressive strength decrease when concrete strength increases. The same tendency is observed for normalized ultimate axial strain but it is not so

zyxwv Cylinders Confined by Prestressed Filament Winding 585

clear. The normalized lateral strain decreases and then seems to increase with concrete strength.

zyxwvuts Batch

Figure 2. Normalized compressive strengths, ultimate compressive axial strains and ultimate lateral strains of filament wound concretes.

zyxw zyxwvutsrq Table 3. Filament wound concrete under cyclic compressive load

Specimen Label 20PIC 40PlC 60PlC 8OPIC l0OPlC 20P2C 40P2C 60P2C 80P2C lOOP2C 20P3C 40P3C 60P3C 80P3C 100P3C

Compressive Strength, (r Tc (MW *

I

zyxwv

.

(r I c

/ (r zp

Jltimate Axial Ultimate Lateral Strain, .zZcc/E,e Strain, 61 cc

&I c&l

c

&I cc

105.08

5.14

0.0293

11.28

-0.0080

26.53

147.12 163.91 187.78 160.77

3.67 3.70 3.82 2.61

0.0212 0.0184 0.0215 0.0112

12.47 10.81 12.63 6.24

-0.0085

-0.0079 -0.0 103 -0.0080

12.15 13.23 14.66 19.93

109.10

5.33

0.0272

10.44

-0.0074

24.72

162.73 159.30 168.12 171.91

4.06 3.59 3.42 2.79

0.0270 0.0140 0.0184 0.0131

15.85 8.25 10.85 7.28

-0.0106 -0.0064 -0.0076 -0.0082

15.18 10.71 10.88 20.61

100.15

4.90

0.0300

11.53

-0.0068

22.68

139.14 155.90 180.94 154.53

3.48 3.52 3.68 2.51

0.0196 0.0133 0.0192 0.0086

11.53 7.80 11.27 4.76

-0.0067 -0.0076 -0.0090 -0.0065

9.58 12.66 12.89 16.15

In Figure 3, the stress-strain curves are presented for specimens with C20 and ClOO concrete and all 3 prestress levels, under to cyclic loading. For concretes of higher strength, is C100, a longer initial range is observed than that of the plain concrete. The load-unload behavior is stiff with modulus about that of initial range. Prestressing by automated filament winding technique contributed to equal stretching of all fibers thereby making them effective. The confining action of prestressed carbon filament on concrete cylinders was much more effective than confinement without prestress4, The confining action was engaged from early loading stage by

'.

zyxwvuts

586 FRPRCS-6: Externally Bonded Reinforcement f o r Confinement

zyxwv

the prestressed confining device, which delayed the formation of internal cracks. As the load was increased above the transition zone, internal concrete damages were formed and developed so that the axial rigidity of concrete decreased. The monotonic behavior corresponded to the envelope curve of cyclic loading. A notable enhancement in strength and ductility of the concrete was observed as a result of confinement.

Strain

iZ,I

zyxwvutsrqp

SpPcimm Label : ZOPZC

Strain

Specimen Label : ZODC

Strain

Strain

Speeimrn Label : IOOPlC

Shin

SpeCimen Label : lOOP3C

Strain

Figure 3 . Compressive behaviors of filament wound concretes.

CONFINING EFFECT IN ELASTIC REGION AND TRANSITION ZONE Comparative investigations were carried out by varying the prestress level of carbon filament confinement and the concrete strength. The prestress level did not influence the critical parameters of specimen fracture. The

z zyxwv

Cylinders Confined by Prestressed Filament Winding 587

specimens from the same concrete batch with different prestress level showed similar ultimate compressive behavior. Another comparison was done according to the stress of transition zone 0; or the knee point (Figure 3). It is difficult in the transition zone to exactly determine where the knee point really is. The knee point was estimated as the point in the transition zone where an elastic performance ends. The normalized measured stresses are presented in Table 4. It can be noted that the high strength concretes of C80 and ClOO batches showed a more prolonged transition zone in a way improving the elastic behavior of concrete by the prestressed confinement. Table 4. Experimental effect of confinement to the transition zone Concrete Batch Values

c20

C40

C60

C80

CIOO

prestress level 1 (80 MPa) Measured 0: 1

1.41

1.32

1.40

1.38

1.30

0iP

zyxwvu

prestress level 2 (1 60 MPa) Measured 0; I

1.61

1.37

1.44

1.48

1.43

prestress level 3 (210 MPa) * I oZp * Measured u2

1.41

1.32

1.13

1.42

1.45

0iP

zyxw

Some elementary expressions’ to estimate the confinement effect on behavior of concrete are given below. Concrete confined by prestressed filament is expected to have a bilinear stress-strain response. This curve is characterized by initial modulus Econf,stress strain values of elastic limit (a knee point) 0: ,E: , tangent modulus of inclined part of stress-strain curve, and ultimate strain of composite confinement. Considering only elastic behavior of concrete and confinement, the elementary formulas can be derived for the first part of load-displacement curve (regarded as initial elastic region) and transition zone. From the deformation compatibility between the confinement and concrete surface, the lateral tensile fiber stress uI(a confining pressure arising against the concrete lateral expansion) increases continuously as the lateral strain E/ increases with the following relation:

zyxwvuts zyxw zyxwvu zyxw

588 FRPRCS-6: Externally Bonded Reinforcement for Confinement

where E, is the elastic modulus of the confining device and h and R are the thickness of confining device and radius of concrete cylinder, respectively. From Hooke’s Law and elementary transformations, the stress-strain relation of confined concrete specimen in axial direction in the initial elastic region can be expressed as follows:

and

E,.,,/ = Eb(1 - v + l / k ) / ( l - v - 2v2 + 1/ k )

( 31

k = (E,h)/(E,R)

(4)

where Eb is the modulus and v the Poisson’s ratio of concrete. Assuming that the knee point (elastic limit) on stress-strain curve of confined concrete is caused by micro cracking of concrete, and assuming that lateral strain of confined concrete at knee point equals the ultimate lateral strain of plain concrete, find the axial stress o:on knee point if found to be: (5)

o:/o;p= l + ( l - v ) k ,

zyxw (:

where oiPis ultimate compressive stress of plain concrete. Addition of prestress oc0of confinement results in following relation after some transformations:

02 = {1+ (1- v)k} +-T{ h

oco

DZP

R

ozp

(1 - v)k + 1} ---1

1

(6)

The first term on right hand side reflects the influence of confinement without prestress and the second term corresponds to prestress. The efficiency of confinement is characterized by parameter k, [Eq. (4)] ( k + 00 for completely stiff confinement, and k +- 0 in the absence of confinement). In Table 5, the main characteristics of confined concretes in the elastic region are calculated using the above equations. For a given value of k ,the initial modulus of confined concrete E,,,/ [Eq. ( 3 ) ] is negligibly larger than the unconfined value, and the axial stress at the elastic limit for the confined

zyxwv zyxw

Cylinders Confined by Prestressed Filament Winding 589

concrete without prestress [Eq. ( 5 ) ] exceeds the ultimate strength of plain concretes by only 5 to 10 percent. Conversely, the prestressed confinement [Eq. (6)] increases substantially the elastic limit load of confined concrete samples. The relatively higher increase of nonlinearity limit is predicted for low strength concrete. However this latest prediction is not confirmed by the measurements (Table 4). There could be two reasons. The assumption of maximum limit strain criterion being independent on hydrostatic pressure, which was used in derivation of above formulas, is evidently oversimplified and should be specified for concrete. Second, the fuzzy nature of transition zone makes it difficult to estimate the real knee point position.

zy zyxwvuts zyxwvu zyxwvut

Table 5. Estimated effect of confinement to the elastic region and knee point Concrete Batch Values

k

Em,

Eb

Calculated 0: / oip without prestress

Calculated a: / o:, prestress level 1 (80 MPa) . Calculated 0: / ozp prestress level 2 (160 MPa) Calculated 0: / o:, prestress level 3 (210 MPa)

c20

C40

C60

C80

ClOO

0.112

0.080

0.072

0.075

0.070

1.004

1.004

1.005

1.005

1.005

1.100

1.067

1.059

1.061

1.057

1.409

1.190

1.155

1.154

1.126

1.721

1.313

1.251

1.247

1.195

1.930

1.396

1.316

1.310

1.242

CONCLUSIONS In this paper, it is shown that the confining action of carbon filaments with prestress is more effective than confinement without prestress. Prestressing rises and prolongs the transition zone significantly with factor up to 1.4. This fact has importance for the stability of confined concrete columns, because for stresses above the transition zone, the modulus is considerably reduced. The confined cylinder ultimate compressive failure stress is not influenced by the prestress. For accurate theoretical prediction of

zyxwvu

zyxwvu

590 FRPRCS-6: Externally Bonded Reinforcementfor Confinement

nonlinearity or elastic limit stress of confined concrete samples, the yield or damage condition of concrete under hydrostatic pressure should be specified. ACKNOWLEDGEMENT The investigation is supported by European Commission - TMR - Network, ConFibreCrete “Research leading to the Development of Design Guidelines for the Use of FRP in Concrete Structures” and Ake and Greta Lissheds Foundation, SEB Bank, SE- 106 40 Stockholm. REFERENCES

zyxw

1. Jib, Cfe‘de‘ration internationale du bkton). (2001), ”Bulletin 14, Externally bonded FRP reinforcement for RC structures”. Technical report, Case Postale 88, CH-1015 Lausanne July, 2001. p. 130. 2. De Lorenzis L., Tepfers R. (2002a and b), “Performance assessment of FRPconfinement models, part I and 11”. ACIC 2002, Thomas Teyord, London, 2002. pp. 251-260 and pp.261-169. 3. Lim S. G., Hahn T. (1996), “Composite Materials in Repairing and Strengthening of Civil Engineering.” Korean Society of Composite Materials, Vol. 9, No. 4, Dec.1996, pp.1-12. 4. Rousakis T. and Tepfers R. (2001), “Experimental Investigation of Concrete Cylinders Confined by Carbon FRP Sheets, under Monotonic and Cyclic Axial Compressive Load.” Chalmers University of Technology, Division of Building Technology, Work No 44. Goteborg, 2001-03-08. p. 87. 5. Tamuis V., C. S. You, Tepfers R. (2001), “Experimental Investigation of CFRPconfined Concretes under Compressive Load”. Institute of Polymer Mechanics, University of Latvia, Aizkraukles 23, LV-1006 Riga, Latvia and Division of Building Technology, Chalmers University of Technology, S-412 96 Goteborg, Sweden, December 2001. p. 68. 6. Karabinis A. I, Rousakis T.C. (2002), “Concrete Confined by FRP Material: A Plasticity Approach”, Engineering Structures , Elsevier, 24 (2002). pp. 923932. 7. De Lorenzis, L., and Tepfers, R. (2002c), “Applicability of FRP confinement to strengthen concrete columns”, Proceedings of MCM2002, Riga, June 2002.

FRPRCS-6, Singapore, 8-10 July 2003 Edited by Gang Hwee Tan OWorld Scientific Publishing Company

zyxwvu

zyxw

CONCRETE CONFINED WITH FIBER REINFORCED CEMENT BASED THIN SHEET COMPOSITES

H.C. WU AND J. TENG Department of Civil and Environmental Engineering, Wayne State University 5050 Anthony Wayne Drive, Detroit, MI 48202, USA In this paper, a new type of fiber reinforced cement-based thin sheet composite will be presented. This innovative technology can be applied to retrofit deficient concrete structures with a superior performance to current Fiber Reinforced Polymer resin (FRP) thin sheets. It is well known that significant improvements in compressive, shear, and flexural behavior of concrete can be achieved with externally bonded FRP sheets. In an effective retrofit with external FRP sheets, a layer of dry fiber sheet (usually unidirectional tape) is placed on the top of a coat of polymer resin that will harden to bond the fiber sheet to the concrete structure. However, composite action suggests that the individual parts of a composite must work together as one. Stresses must be transferred from the FRP sheet to concrete substrate through the interface. Such good bonding requires extensive concrete surface preparation before installation of FRP sheet. Concrete surface preparation is expensive and sometimes prohibited due to dust or noise concerns. Instead of using polymeric resins, the authors have been developing innovative cement-based matrix materials for making thin composite sheets. Cement-based materials have many advantages in comparison to polymeric resins. For instance, much less or no surface preparation is needed for good bonding. Additional benefits include much higher fire and vandalism resistance, and user friendly to the construction industry. Preliminary test data suggest that the effect of using the newly developed cement sheets on concrete confinement is similar to that of using conventional FRP sheets.

INTRODUCTION

The U S . has an estimated $20 trillion investment in civil infrastructure systems. Because of aging, overuse, exposure, misuse, and neglect, many of these systems are deteriorating and becoming more vulnerable to catastrophic failure when earthquake or other natural hazards strike. It would be prohibitively costly and disruptive to replace these vast networks. They must instead be renewed in an intelligent manner. It is generally recognized that fiber reinforced polymer (FRP) sheets are one of the most

592 FRPRCS-6: Externally Bonded Reinforcement for Confinement

vital materials for repair, strengthening, and rehabilitation of existing structures. Applications involve such as externally bonded composite fabrics or jackets on beams, columns, and bridge decks. FRPs (or advanced fiber composites) have long been successfully used by the aerospace and defense industries. These materials are rapidly gaining momentum in civil engineering structural applications. The thrust is twofold: (1) an urgent call for new material to fix our nation's fast deteriorating facilities where the challenge is too great using conventional materials, and (2) properties (high strength-to-weight ratio and corrosion resistance) and easy construction (fast curing process and lightweight) that are superior to conventional concrete and steel. The greatest potential of FRPs in the near future will be in the areas of repair, strengthening, and rehabilitation of existing structures, such as externally bonded composite fabrics or jackets on beams, columns, and bridge decks. Significant improvements in compressive, shear, and flexural behavior of bonded concrete elements are obtained. Typically, increases in strength and failure strain of several times are obtained with external FRP reinforcement'-6.

zyxwvu

RETROFIT WITH FRP SHEETS

Several important considerations regarding reinforcing or retrofitting existing structures including buildings and bridges are (1) cost efficiency, (2) convenience to the occupants with minimum interference to their operations, and (3) environmentally sound for fabrication and disposal. FRPs are found to be a favored solution due to their superior properties, light weight, and easy handling. In contrast, the use of conventional materials typically requires complete shutdown of the structure for repair, or is difficult if not impossible for internal strengthening due to weight and dimensional limitation (e.g. steel truss). These constraints are particularly significant for building repair. Hence, construction costs using conventional materials are substantially increased, although the materials could be relatively inexpensive. It is typically estimated that material costs are less than 20% of the total cost of a repair project. Therefore, even from a cost viewpoint, FRPs are very competitive. In many of the repair projects, the total cost when using FRP were reported to be less than that of using conventional materials for the same project. In an effective retrofit with external FRP sheets, a layer of dry fiber sheet (usually unidirectional tape) is placed on the top of a coat of polymer resin that will harden to bond the fiber sheet to the concrete structure. Prior

zyxw

Concrete Confined with Thin Sheet Composites 593

to applying resin coating, the concrete surface must be thoroughly cleaned and smoothed, including grinding and patching that are labor intensive and sometimes require complete shutdown of the operation of the structure. When needed, multiple layers of fiber sheets can be sequentially added by repeating the same procedure. Functions of Fiber and Matrix in Composites

zyx

A typical fiber composite is primarily made of continuous fibers and matrix. The advantages of fiber reinforced polymer composites (FRP) as compared to more conventional materials are often related to the high ratios of stiffness and strength to weight. A typical FW is about 4 times lighter than steel with an equal strength. The strengthktiffness of FRPs is almost entirely attributed to the fibers7,*,since the polymeric matrix has negligible strengthktiffness in comparison to the fiber. The matrix serves three important functions: (1) it holds the fiber in place, (2) it transfers loads to the high-stiffness fiber, and (3) it protects the fiber. Typical density of common engineering fiber is 1.7 - 2.0 g/cm3 for carbon, 2.5 - 2.7 g/cm’ for glass, whereas on the matrices side, epoxy and polyester have a density between 1.2 and 1.4 g/cm3, giving a lightweight composite density between 1.5 and 2.2 g / ~ m ’ . ~It is clear from the above discussion that polymer matrix provides a negligible contribution to composite strengthktiffness that is needed for effective retrofit of concrete structures, yet polymers have many other problems such as lack of fire resistance and degradation under UV light. The authors proposed to replace polymer matrix by cement. Typical density of cementitious materials may range from 0.8 to 2.2 g/cm3 depending on their compositions, hence maintaining a lightweight of the cement composites. The in-situ applicability of cement matrix is only possible when we can control the rheological properties of cement materials that can range from water-like to dough-like. Fiber Reinforced Cement Composite

In this case, the same kind of fiber reinforcement (unidirectional tapes or woven) as in a regular FRP sheet, and we use cement materials to replace polymer resin is used. The preparation procedure is analogous to regular FRP. A fiber tape or fabric is impregnated with cement slurry to form a thin composite sheet. The current process involves the following steps (1) precut a fabric to designated dimensions, (2) submerge in cement slurry for

594 FRPRCS-6: Externally Bonded Reinforcement f o r Confinement

rapid and full penetration into the fabric, and (3) remove the impregnated fabric from the slurry tank and place in molds. Some preliminary work has been carried out to make thin plates. Plate as thin as 2 mm can be made with two layers of unidirectional fiber tapes. Both carbon and glass fibers have been used in the investigation although we are aware of potential durability problem of glass fibers in cement alkaline environment. Selected plates have been tested according to ASTM C78-75, Standard Test Method of Flexural Strength of Concrete (using simple beam with third-point loading). A picture of a 4-mm thick plate sample is shown in Figure 1 during a 3-point bending test. Figure 2 depicts the flexural load vs displacement curve showing very high flexural strength (1 00 MPa) and excellent ductility (also see Fig.1). Such high strength and high ductility are not the norm of cement materials, but it should not be too surprising since the high strength comes from the carbon fibers and the high ductility is attributed to multiple cracking phenomenon of brittle cement matrix'o"'. The current work has demonstrated that thin cement sheets having excellent strengtldductility can be produced within minutes. This fast process is a must for in-situ processibility on job sites and for achieving low cost.

zyxw zy

COMPRESSIVE BEHAVIOR OF CONFINED CONCRETE After successful production of thin cement composites, such thin sheets have been employed to retrofit concrete cylinder samples in a preliminary study at Wayne State University'*. The purpose of these tests is to evaluate retrofit effectiveness of the innovative thin Fiber Reinforced Cement (FRC) composite. Ideal FRC composites under development are expected to provide similar or improved retrofit efficiencies with lower cost and easy construction in comparison to Fiber Reinforced Polymer (FRP) composites. Unidirectional carbon fiber tapes were used in this study. Cement based matrix developed at Wayne State University and epoxy resin were used separately to make thin CFRC sheets and CFRP sheets. Both CFRC and CFRP composites contain two layers of unidirectional carbon fiber tapes. The average thickness of the CFRC jackets is 3.0 mm, whereas the CFRP between 2-3 mm. These thin composite sheets were then employed to wrap 4 inch by 8 inch concrete cylinders. The bond length of the CFRC samples is 3 inches and 2 inches for the CFRP. A 1.5 inches gap exists between the top of the cylinder and the top of the composite sheets at both ends (see Figure 3 and 4).

Concrete Confined with Thin Sheet Composites 595

zyx zyxwv zyxwvut zyxwvut

Figure 1: A thin cement infiltrated fiber plate (thickness loading

=4

mm) during flexural

zyxwv 120 r

-m.

a

100

E d 0) C

p!

G -

2 X

a,

i

-

80

-

zyxwvuts r 60 40

20

I

0

0

10

20

30

40

50

60

Deflection (mrn)

Figure 2: Flexural stress vs. mid-point deflection curve of a carbon fiber reinforced cement plate

These concrete cylinders, unconfined and confined with CFRC or CFRP composites are tested using a high-stiffness, high-capacity MTS testing machine following ASTM C39-96, Standard Test Method for Compressive Strength of Cylindrical Concrete Specimens. This equipment has sufficient capacity and stiffness, which is required for conducting such tests. The machine is also equipped with a sophisticated computer control and data acquisition system. The acquired data including the applied axial load and axial deformation of concrete are recorded automatically. Per ASTM Practice (2123 1-93, steel retaining rings and rubber pads were used without other capping during the tests.

596 FRPRCS-6: Externally Bonded Reinforcement for Confinement

zyxwvu

Figure 3: Concrete confined with CFRP composite jacket

Figure 4: Concrete confined with CFRC composite jacket

zyxw

zyxw zyxw zyxw

Concrete Confined with Thin Sheet Composites 597

Test Results

The axial stress versus axial strain relationships of the unconfined and confined concrete are shown in Figure 5. As shown in Figure 5, the initial portions of the stress-strain responses of the confined specimens essentially followed the curves of the unconfined concrete. The average compressive strength of the unconfined concrete is 54 MPa. The CFRC group shows a compressive strength two times higher than that of the control from 54 MPa to 100 MPa. In addition, the ductility is increased by 3 times from 2 mm to 6 mm. The CFRP sample has the highest compressive strength (105 MPa) and ductility. Nevertheless, the differences between CFRC and CFRP are insignificant. The CFRP jacketed samples showed explosive failure that was triggered by the complete rupture of the CFRP jacket. The remnants of the CFRP sample after failure are shown in Figure 6. The CFRC samples also show fiber rupture failure similar to the CFRP sample. The CFRC samples have a much less violent global failure than the CFRP (see Figure 7). Concrete inside the CFRC jacket was crushed completely (Figure 8).

120

zyxwvuts

0

0

2

4

6

0

10

Deflection (mm)

Figure 5 Compressive behavior o f unconfined and confined concrete

zyxw

598 FRPRCS-6: Externally Bonded Reinforcement for Confinement

Figure 6: Remnants of CFRP sample after violent failure

zyxwvuts

Figure 7: Global failure of the CFRC sample due to rupture of the CFRC jacket.

Concrete Confined with Thin Sheet Composites 599

zyxwvu

Figure 8: Concrete inside the CFRC jacket was crashed completely

CONCLUSIONS It is confirmed that the compressive strength of concrete can be significantly improved using external CFRP wraps. In addition, the ductility of the confined concrete is significantly increased. The final failure of the confined concrete is provoked by the onset of the composite rupture. The CFRC confined concrete show similar improvements in both ultimate compressive strength and ductility with the CFRP concrete. Because of the use of high strength concrete in this study, the final failure of the plain concrete is explosive. The final failure of the CFRP confined concrete shows even more violent. The CFRC concrete has a much less violent global failure than the CFRP. REFERENCES 1. McConnell, V.P. “Bridge Column Retrofit, Hybrid Woven Unifabric.” High Performance Composites, September/October, 1993,62-64 pp. 2. Seible, F. and Priestley, M.J.N. “Strengthening of Rectangular Bridge Columns for Increased Ductility.” Practical Solutions for Bridge Strengthening and Rehabilitation, Des Moines, Iowa, 1993. 3. Karbhari, V.M., Eckel, D.A., and Tunis, G.C. “Strengthening of Concrete Column Stubs Through Resin Infused Composite Wraps.” J of Thermoplastic Composite Materials, V.6, 1993,92-107 pp.

z

zyxwvutsr

zyx zyxw zyxwv zyxw

600 FRPRCS-6: Externally Bonded Reinforcementfor Confinement

4. Labossiere, P., Neale, K.W., Demers, M., and Picher, F. “Repair of Reinforced Concrete Columns with Advanced Composite Materials Confinement.” in Repair and Rehabilitation of the Infrastructure of the Americas, H.T. Toutanji (ed.), University of Puerto Rico, 1995, 153165 pp. 5. Nanni, A., Norris, M.S., and Bradford, N.M. “Lateral Confinement of Concrete Using FRP Reinforcement,” ACI SP 138, Fiber Reinforced Plastic Reinforcementfor Concrete Structures, 1992, 193-209 pp. 6. Saadatmanesh, H., Ehsani, M.R. and Li, M.W. “Strength and Ductility of Concrete Columns Externally Reinforced with Fiber Composite Straps,” ACI Structural Journal, 91[4], 1994, 434-447 pp. 7. Swanson, S.R., Advanced Composite Materials, Prentice Hall, New Jersey, 1997. 8. Bogner, B.R., “Isopolyester Pultrusion Resin Study.” Proc. SPI Composite Institute 45IhAnnu. Con$, New York, 1990. 9. Ashby, M.F. and Jones, D.R.H., Engineering Materials, Pergamon Press, Oxford, 1986. 10 Li, V.C., and Wu, H.C. “Conditions for Pseudo Strain-Hardening in Fiber Reinforced Brittle Matrix Composites,” Appl. Mech. Rev., Vol. 45, NO. 8, 1992, 390-398 pp. 11. Li, V.C. and Leung, C.K.Y., “Theory of Steady State and Multiple Cracking of Random Discontinuous Fiber Reinforced Brittle Matrix Composites”, ASCE J. of Engng. Mechanics, 118, 1 1, 1992, 2246-64 PP. 12. Wu, H.C. and Teng, J., “Innovative Cement Based Thin Sheet Composites for Retrofit”, in Proc. 3rd Inter. Composite Conf for Infrastructure, San Francisco, CA, 2002.

zyxwvu zyxwvu

FRPRCS-6, Singapore, 8-10 July 2003 Edited by Kiang Hwee Tan @WorldScientific Publishing Company

zyxw

HOOP RUPTURE STRAINS OF FRP JACKETS IN FRP CONFINED CONCRETE L.LAM AND J.G.TENG Department of Civil and Structural Engineering, The Hong Kong Polytechnic Universiv, China

One important application of fibre-reinforced polymer (FRP) composites is as a confining material for concrete in the retrofit of existing concrete columns by the provision of FRP jackets. Such jackets are commonly formed in a wet lay-up process, with the fibres being only or predominantly in the hoop direction. It has been well established in recent studies that the rupture strainshtrengths of FRP measured in tests on such FRP-confined concrete cylinders fall substantially below those from flat coupon tensile tests, but the causes are unclear. This paper presents the results of a study which is aimed at clarifying these causes. To this end, the paper reports and compares the ultimate tensile strains of two types of FRP (CFRP and GFRP) obtained from three types of tests: flat coupon tensile tests, ring splitting tests and FRP-confined concrete cylinder tests. Based on comparisons of these test results, it can be concluded that the FRP hoop rupture strains in FRP-confined concrete cylinders are reduced below the ultimate tensile strains from flat coupon tests by three factors: (a) the curvature of the FRP jacket; (b) the non-uniform deformation of the concrete; and (c) the effect of the overlapping zone which has an increased thickness.

INTRODUCTION One important application of fibre-reinforced polymer (FRP) composites is as a confining material for concrete in the retrofit of existing concrete columns by the provision of FRP jackets. Such jackets are commonly formed in a wet lay-up process, with the fibres being only or predominantly in the hoop direction. Concrete-filled FRP tubes are another important application, in which the FRP tubes have a substantial axial stiffness. This paper is explicitly concerned with the behaviour of unidirectional FRP jackets interacting with concrete only, although the conclusions are believed to be relevant to concrete-filled FRP tubes as well. The issue examined in this paper is the ultimate tensile straidstrength of FRP jackets in the hoop direction in FRP-confined concrete, versus that obtained from a material

602 FRPRCS-6: Externally Bonded Reinforcement f o r Confinement

zy

tensile test. This issue, though apparently simple, is fundamental to the proper interpretation of test results of FRP-confined concrete and the development of accurate and rationally based confinement models. In existing theoretical models for FRP-confined concrete as reviewed el~ewherel-~, it is commonly assumed that tensile rupture of the FRP occurs when the hoop stress in the FRP reaches its tensile strength from material tensile tests, either flat coupon tests' or ring splitting tests6. Test results were interpreted accordingly in the development of these models. The only exception is the model of Xiao and Wu' for which it was suggested that 50% of the FRP flat coupon tensile strength be taken as the hoop rupture strength, based on test observations. Experimental evidence from other also suggested that the material tensile strength of FRP may not be reached at hoop rupture of FRP jackets in FRP-confined concrete. The substantial difference between the FRP tensile strength or ultimate strain from material tests and that reached in FRP-confined concrete specimens has been discussed by a number of researchers*,*-". Lam and Teng" concluded that confinement models must be based on the actual tensile rupture strength of FRP achieved in FRP-confined concrete rather than that from tensile tests, and established a design-oriented stress-strain model on the basis of this conclusion. While the strength difference between FRP tensile specimens and FRP jackets confining concrete is well established", a number of uncertainties exist. In particular, the causes for this difference are unclear at the present, but must be clarified before complete confidence can be achieved in modelling such FRP-confined concrete. Several causes have been suggestedzx8-' ', including (a) deformation localization in cracked concrete leading to a non-uniform stress distribution in the FRP jacket, and (b) the effect of curvature of an FRP jacket on the tensile strength of FRP and (c) Local misalignment or waviness of fibres in the wet lay-up process leading to unequal stretching of the fibres. No specific study has been found attempting to confirm or reject these suggestions. This paper therefore presents the results from the first carefully planned study12 involving comparative experiments in an attempt to clarify these uncertainties. FLAT COUPON AND RING SPLITTING TESTS

zy

In the present study, two types of FRP were employed: carbon FRP (CFRP) and glass FRP (GFRP). The CFRP was formed from unidirectional carbon fibre tow sheets and epoxy resin. The carbon fibre sheets had a nominal thickness of 0.165 mm. The GFRP was formed from a woven fabric consisting of E-glass fibres in the longitudinal direction as the main fibres

zyxw

Hoop Rupture Strains in FRP-Confined Concrete 603

and sparsely distributed aramid fibres in the transverse direction. The woven fabric had a nominal thickness of 1.27 mm. The nominal thicknesses were used for the calculation of material proprieties as is commonly done for wet lay-up FRP. The two FRP systems were proprietary products supplied by separate companies. Two types of material tensile tests were conducted to determine the material properties: flat coupon tensile tests’ and ring splitting tests6. The purpose of these tests was mainly to determine the material ultimate tensile straidstrength and elastic modulus, and to examine the effect of curvature on tensile properties.

zyxwvu

Flat Coupon Tensile Tests

The flat coupon tests were conducted following the ASTM standard5. Dimensions of the test specimens are shown in Figure 1. The strains were average values from two strain gauges at mid-length on the two sides of the test coupon. The results of the tensile tests calculated using the normal thicknesses and the actual widths12 (about 25 mm) are shown in Table 1, where each result is the average of at least five specimens. Typical stressstrain curves of the FRPs from flat coupon tests are shown in Figure 2. Strain gauge

w 4

One-layer FRP

3

zyxwvu

Figure 1. Dimensions of flat coupons for tensile tests

(shear) (Shear) (Shear) (Shear) Table 1 . Results of material tensile(Shear) tests

Type of test

604 FRPRCS-6: Externally Bonded Reinforcement f o r Confinemenl .. . . . .

1.2 1 0.8 0.6

0.4

I

I

..... .............

I...........

. ....................

zyxw

zyxwvuts

0.2 0

0

0.2

0.4

0.6

0.8

1

1.2

Normalized tensile strain

Figure 2. Typical normalized stress-strain curves of CFRP and GFRP Overlapping zone

Strain gauge

zyx

Figure 3. Dimensions of ring specimens for tensile tests

Ring splitting tests

The ring splitting tests were also conducted following an ASTM standard6. Dimensions of the test specimens are shown in Figure 3. The GFRP rings had a nominal width of 23 mm to correspond to 10 yarns of the woven fabric instead of 25 mm which was used for the CFRP ring. For the CFRP rings, the nominal thickness and the actual widths were used in determining the tensile properties. For the GFRP, the nominal thickness and the nominal width were used as the actual widths were more difficult to measure due to the size of yarns and the presence of transverse fibres. The tensile test coupons were formed in a different way so there was no such problem'*.

zyx zyx

Hoop Rupture Strains in FRP-Confined Concrete 605

The tensile properties so obtained are shown in Table 1, where each result is again the average of at least five specimens. A comparison of the flat coupon test results and the ring splitting test results shows that the ring splitting tests lead to a much-reduced tensile strength. This may be interpreted to mean that the curvature of the ring specimens had a substantial detrimental effect on the tensile strength. However, it should be noted that the ring splitting test, due to its own limitations, may deliver a lower tensile strength due to other reasons. For example, the relative movement between the two half disks effected by an external force in a ring splitting test does not produce an axisymmetric radial expansion required to produce a uniform hoop tension in the FRP ring. COMPRESSION TESTS OF FRP-CONFINED CONCRETE CYLINDERS Test Specimens

A total of 27 concrete cylinders of 152 mm in diameter and 305 mm in height were prepared and tested in three series. Details of the test specimens are shown in Table 2. Each series consisted of six confined cylinders and three unconfined cylinders, all prepared from the same batch of concrete. The FRP jackets were formed in a wet lay-up process by impregnating a continuous fibre/fabric sheet with matching epoxy resin. In the case of the CFRP, a primer was applied first to the surface of concrete before the wrapping of FRP. Regardless of the number of FRP layers (each layer contains a single lap of sheet), a single continuous sheet with the main fibres oriented in the hoop direction was wrapped around the cylinder with the finishing end of the sheet overlapping the starting end by a prescribed length (the overlap length). The three series of tests cover CFRP and GFRP, different numbers of FRP layers and different overlap lengths. In particular, the effect of overlap length was studied using three confined cylinders which had overlap lengths of 100 mm, 250 mm and 400 mm respectively, while for all other specimens this overlap length was always 150 mm. The measured thicknesses were on average 1.03 mm for single-layer CFRP and 1.4 1 mm for two-layer CFRP respectively, while the measured thicknesses of GFRP were 1.33 mm and 2.32 mm for single- and two-layer jackets respectively.

606 FRPRCS-6: Externally Bonded Reinforcementfor Confinement

zyx

zyxwvu zyxw

Layout of Strain Gauges

For each control cylinder, four unidirectional strain gauges were bonded on the surface of concrete at the mid-height. Two with a gauge length of 60 mm, at 180" apart, were used to measure hoop strains. The other two had a gauge length of 120 mm and were used to measure axial strains. For each FRP-wrapped cylinder, 8 unidirectional strain gauges (SGl -SG8) with a gauge length of 20 mm were evenly distributed at mid-height to measure the hoop strains of the FRP jacket as shown in Figure 4, with SGl being located at 22.5" from the finishing end of the fibre sheet. For specimens with an overlap length of 150 mm, both SGI and SG2 were located within the overlapping zone, while SG3 covers the starting end of the fibre sheet (Figure 4). In addition, axial strains were determined using 4 evenly distributed strain gauges as well as two linear variable displacement transducers (LVDTs) at 180" apart covering the mid-height region of 120 mm. Overlamina zone

FW jacket

zyx z

Figure 4. Location of strain gauges for hoop strains in an FRP jacket

Results of Compression Tests

The results of compression tests are summarized in Table 2. In this table, the compressive strength of unconfined concrete fc: for each series of tests is averaged from the three control cylinders. The ultimate axial strains E,, are average values obtained using the two LVDTs. The hoop strains of FRP at rupture E ~ are, given ~ as~ average values, first of the eight strain gauges

zyxw zyxwv Hoop Rupture Strains in FRP-Confined Concrete 607

zyxwv zyx Type of test over the whole circumference and then of the strain gauges outside the overlapping zone. All 18 FRP-wrapped cylinders showed eventual failure by the rupture of FRP outside the overlapping zone and bilinear stress-strain behaviour ending at a point defined by the compressive strength fc:and the ultimate strain E,, . COMPARISON OF TEST RESULTS AND DISCUSSION

The average hoop rupture strains of FRP obtained in the compression tests (Table 2) are seen to be much smaller than the material ultimate tensile strains obtained from flat coupon tests as given in Table 1. The ratio of the FRP hoop rupture strain to the material ultimate tensile strain obtained from flat coupon tensile tests has been termed the efficiency

Type of test Type of test Type of test Type of test (shear) (Shear) (Shear) (Shear) (Shear)

a The overlap length was varied from 100 mm to 400 mm bNot available due to experimental errors

zyxw

608 FRPRCS-6: Externally Bonded Reinforcement for Confinement

For the FRP-confined cylinders with an overlap length of 150 mm, this efficiency factor is only 0.583 for the nine CFRP-wrapped cylinders and 0.669 for the six GFRP-wrapped cylinders if the FRP hoop rupture strain is taken as the average value from the eight strain gauges. These two values are very close to those found by Lam and Teng" from test results reported in the open literature, which are 0.586 for 52 CFRP-wrapped concrete cylinders and 0.624 for nine GFRP-wrapped specimens. Figures 5a and 5b show the distributions of the FRP hoop strains around the wrapped cylinder at rupture. These figures show clearly that the strains are non-uniform around the circumference, with substantially lower values within the overlapping zone. Indeed, the smallest hoop strain is always found within the overlapping zone and on average is only about 38% and 45% of the ultimate tensile strain from flat coupon tests for CFRP-wrapped cylinders and GFRP-wrapped cylinders, respectively. It is easy to understand that these lower FRP hoop strains arose because the FRP jacket was thicker in this zone. For the same confinement pressure, the strain in jacket is inversely proportional to the thickness of the jacket. Figure 6 further illustrates the effect of overlap length, where the strain distributions in confined cylinders with three different overlap lengths are compared. For the two cylinders with longer overlap lengths, the average hoop rupture strains are thus reduced (Table 2). In this connection, it is worth noting that a longer overlap length has the same effect as a reduced strain capacity of the FRP. Figures 5 and 6 also show that the strain distributions outside the overlapping zone are also non-uniform, although the variation is smaller. This non-uniformity can be attributed to the non-uniform deformation of cracked concrete which is an inhomogeneous material. The maximum hoop strain on the FRP jacket was generally observed outside the overlapping zone, except three cases in which this maximum strain was found at (Cl-3 and C2-1) or near (Cl-400) the starting end of the fibre sheet. For these three cases, the high strains measured may be partly attributable to jacket bending as a shell as a result of a thickness change. Indeed, the strains measured elsewhere on the jacket may also contain a significant bending component in the final stage as a result of non-uniform deformation of concrete.

zy

zyxwvuts zyxwvu zyxwv Hoop Rupture Strains in FRP-Confined Concrete 609

1.6

(a)

E2 3

1.2

2

1

Q

c

I

!I

I

ICFRP I

zyxw

zyxwvutsrq

._ 0.8 0.6 (0

1

1.4

8 0

.s

a

0.4

0.2

zyxwvut

E o0

60

120

180

240

300

360

60

120

180

240

300

360

Angle from the finishing end of wrapping (degree)

a 0

2

0.5

a

E

o 0

Angle from finishing end of wrapping (degree)

Figure 5 . Distributions of FRP hoop strain on cylinders with an overlap length of 150 mm (a) CFRP-wrapped cylinders and (b) GFRP-wrapped cylinders

c

m

.-K

F

*Q

c

0.5

0

0 L

a

E o 0

60

120

180

240

300

360

Angle from finishing end of wrapping (degree)

Figure 6. Distributions of FRP hoop strain on CFRP-wrapped cylinders with varying overlap lengths

zyxw

610 FRPRCS-6: Externally Bonded Reinforcement for Confinement

The maximum strain measured on the FRP in a confined concrete cylinder is found to be 1.06% on average for CFRP-wrapped specimens, and 1.91% on average for GFRP-wrapped specimens. Although these are much lower than the material ultimate strains from flat coupon tests, they are strikingly close to the ultimate strains obtained from ring tests which are 1.009% for CFRP and 1.987% for GFRP. This indicates that the effect of curvature of the FRP jacket in a ring splitting test and in a confined cylinder test is similar, although this is not yet completely certain due to the limitation of the ring splitting test as mentioned earlier. CONCLUSIONS

zyx

In an attempt to explain why FRP hoop rupture strains measured in FRPconfined concrete cylinder tests fall substantially below those from flat coupon tensile tests, this paper has presented and compared tensile strengths for two types of FRP (CFRP and GFRP) obtained from three types of tests: flat coupon tensile tests, ring splitting tests and FRP-confined concrete cylinder tests. Based on comparisons of these test results, it can be concluded that the hoop rupture strains measured in FRP-confined concrete cylinders are affected by three factors: (a) the curvature of the FRP jacket; (b) the non-uniform deformation of concrete leading to a non-uniform distribution of the strains in the FRP jacket; and (c) the effect of the overlapping zone in which the measured strains are much lower than strains measured elsewhere. These three factors combine to produce an average FRP hoop rupture strain which is much lower than that from flat coupon tests. The present study has therefore confirmed two of the causes suggested in the and listed in the introductory section of the paper. The third one mentioned there, namely the effect of misalignment and waviness of fibres, is not believed to be an important factor as these defects, if present, affect results from both material tensile tests and confined concrete cylinder tests. The third factor identified here means that a significant scatter in test results can arise as a result of strain gauge locations. It is important to note that all three factors are size related: the effect of curvature and overlap zone may be reduced, while the effect of non-uniform concrete deformation may become more important in large columns. Large concrete columns should be tested in the future to examine how these factors affect the hoop rupture strength of FRP jackets.

zyx z

Hoop Rupture Strains in FRP-Confined Concrete 611

ACKNOWLEDGMENTS The work presented in this paper forms part of a research project (Project No: PolyU 5059/02E) funded by the Research Grants Council of the Hong Kong SAR with additional support from The Hong Kong Polytechnic University provided through its Area of Strategic Development (ASD) Scheme. The authors are grateful to both organizations for their financial support. The authors also wish to thank Mr. Y.N. Tse and Miss P.Y. Fung for their valuable contributions to the experimental work. REFERENCES 1. De Lorenzis, L. and Tepfers, R., “Performance assessment of FRPconfinement models - Part I: Review of experiments and models”, Advanced Polymer Composites for Structural Applications in Construction, Proceedings of the First International Conference, Edited by R.A. Shenoi, S.S.J. Moy, and L.C. Hollaway, Thomas Telford, London, UK, 2002, pp. 25 1-260. 2. De Lorenzis, L. and Tepfers, R., “Performance assessment of FRPconfinement models - Part 11: Comparison of experiments and predictions”, Advanced Polymer Compositesfor Structural Applications in Construction; Proceedings of the First International Conference, Edited by R.A.‘Shenoi, S.S.J. Moy, and L.C. Hollaway, Thomas Telford, London, UK, 2002, pp. 261-269. 3. Monti, G., “Confining reinforced concrete with FRP: behavior and modeling”, Composites in Construction: A Realip, Proceedings of the International Workshop, Edited by E. Cosenza, G. Manfredi, and A. Nanni, ASCE, Virginia, U.S.A.,2002, pp. 213-222. 4. Teng, J.G., Chen, J.F., Smith, S.T. and Lam, L., FRP-Strengthened RC sfructures, John Wiley & Sons, Ltd., UK, 2 0 0 2 , 2 4 5 ~ ~ . 5. ASTM D3039/D3039M-95, “Standard test method for tensile properties of polymer matrix composite materials”, Annual Book of ASTMStandards, Vol. 14.02, 1995. 6. ASTM D 2290 - 92, “Standard test method for apparent tensile strength of ring or tubular plastics and reinforced plastics by split disk method ”, Annual Book OfASTMStandards, Vol. 15.03, 1992. 7. Xiao, Y . and Wu, H., “Compressive behavior of concrete confined by carbon fiber composite jackets”, Journal of Materials in Civil Engineering, ASCE, 12(2), 2000, pp. 139-146.

zyxwvu

zyx zyxw

612 FRPRCS-6: Externally Bonded Reinforcement f o r Confinement

8. Shahawy, M., Mirmiran, A. and Beitelman, A., “Test and modeling of carbon-wrapped concrete columns”, Composites: Part B, 3 1, 2000, pp.47 1-480. 9. Pessiki, S., Harries, K.A., Kestner, J.T., Sause, R., and Ricles, J.M., “Axial behavior of reinforced concrete columns confined with FRP jackets”, Journal of Composites for Construction, ASCE, 5(4), 2001, pp.237-245. 10. Lam, L. and Teng, J.G., “Design-oriented stress-strain model for FRPconfined concrete”, to be published. 1 1. Spoelstra, M.R. and Monti, G., “FRP-confined concrete model”, Journal of Compositesfor Construction, ASCE, 3(3), 1999, pp. 143-150. 12. Lam, L. and Teng, J.G., “Ultimate condition of FRP-confined concrete”, to be published.

zyxwvut zyxwvut

FRPRCS-6, Singapore, 8-1 0 July 2003 Edited by Gang Hwee Tan @World Scientific Publishing Company

zyxw zyx

EXTERNALLY CONFINED HIGH STRENGTH CONCRETE COLUMNS UNDER ECCENTRIC LOADING J. LI, M. MOULSDALE AND M. N. S. HAD1 Faculty of Engineering, University of Wollongong Wollongong, MSW 2522, Australia

Previous studies proved that the application of FRP can eliminate some unwanted properties of high strength concrete, such as the brittle behavior of high strength concrete. However, research studies conducted so far on external confinement of concrete columns have mainly concentrated on concentric loading. This paper investigates the performance of externally confined high strength concrete columns subjected to eccentric loading and evaluates the effectiveness of two confinement materials-Carbon fibre and E-glass. The contribution of external confinement with FRP to the increase of the strength of concrete columns depends on few factors, for example, the number of layers; the type of confining materials and the bond between the fibres and the concrete. The layout of fibres is another variable, which contributes much to the behaviour of confinement effectiveness when bending action is introduced. The enhancement of the strength of the plain column specimens under eccentric loading is not so pronounced as for the reinforced concrete specimens under concentric loading.

INTRODUCTION

With the development of technology, the use of high-strength concrete members has proved most popular in terms of economy; superior strength; stiffness and durability. With the increase of concrete strength, the ultimate strength of the columns increases, but a relatively more brittle failure occurs. The lack of ductility of high strength concrete results in sudden failure without warning, which is a serious drawback of high strength concrete. Previous studies have shown that addition of compressive reinforcement and confinement will increase the ductility as well as the strength of materials effectively. Concrete, confined by transverse ties, develops higher strength and to a lesser degree ductility. Studies conducted by some investigators on the improvement of the ductility of high strength concrete members have proven that the use of the spiral confinement is more effective and beneficial in the improvement of performance of concrete members'.

614 FRPRCS-6: Externally Bonded Reinforcement f o r Confinement

In recent years, FRP wrapping in lieu of steel jacket has become an increasingly popular method for external reinforcement in which FRP offers improved corrosion and fatigue resistance compared to the steel reinforcement. The high tensile strength and low weight make FRP ideal for use in the construction industry. Another attractive advantage of FRP over steel straps as external reinforcement is its easy handling, thus minimal time and labour are required to implement them2. However, research studies conducted so far on external confinement of concrete columns have mainly concentrated on concentric loading. In practice, axially compressed (ie., concentrically) structural concrete columns rarely occur. Even in a column nominally carrying only axial compression, bending action is almost always present due to unintentional load eccentricities and possible construction error. Also, there are many columns where an eccentric load is deliberately applied. Therefore, the studies for concrete columns under eccentric loading are essential for practical use. This study experimentally investigates the benefits of external confinement using FRP on high strength concrete columns under eccentric loading and compares the effectiveness of two types of external reinforcement.

zyxwvu

EXTERNAL CONFINEMENT WITH FRP

The application of FRP in the construction industry can eliminate some unwanted properties of high strength concrete, such as the brittle behaviour of high strength concrete, FRP is particularly useful for strengthening columns and other unusual shapes. Parameters that affect the strength and ductility of FRP confined concrete include concrete strength, type of fibres and resin and thickness of FRP (different layers). The experimental program conducted by Houssam and Balauru3 showed that the compressive strength improved by approximately 200 percent due to confinement, with Carbon fibre and by approximately 100 percent due to Glass fibre. Also, the shape of cross section and the spacing of FRP straps can directly affect the confinement effectiveness of FRP wrapping in the confinement and these were well known. The orientation of fibres is another factor contributing to the mechanical performance of a composite. Fibres oriented in one direction give very high stiffness and strength in that direction. If the fibres are oriented in more than one direction, such as in a mat, there will be high stiffness and strength in the direction of the fibre orientation4.

Externally Confined High Strength Concrete Clournns 615

z

EXPERIMENTAL PROGRAM

The objective of the experimental program in this study is to investigate the behaviour of external reinforced high strength concrete columns (no internal reinforcement) subjected to eccentric loading and to evaluate the effectiveness of external confinement with FRP composites. Proven by previous studies, the major parameters affecting the behaviour of concrete columns confined with external FRP are the type of fibers; the number of layers and the shape of cross-section. As the influence of cross-section is already well known, this study is limited to circular columns under eccentric loading. The test variables selected for this study are: (1) the type of reinforcement: internal and external, (2) the number of layers of FRP, (3) the type of wrapping materials: unidirectional Carbon and plain weave Eglass. Column’s Details

zyx zy

Seven high strength concrete columns were designed for testing. Each column was designed to have a diameter, D, of 235 mm for both the haunched ends and 150 mm in the test region, and an overall length, H, of 1400 mm. The clear distance between the ends was 620 mm. The dimensions were selected to be compatible with the capacity of the testing machine. There are two major amounts of reinforcement designed for the two internally reinforced specimens. Six RWlO bars were equally spaced around the inside circumference of 41 10 helix with a pitch of 60 mm through the entire length of specimens and three RW8 bars confined by circular ties are spaced in equal distances at both ends. The geometry and dimension and internal reinforcement details of column are shown in Figure 1. Five specimens wrapped continually with FRP had the following configurations: one-layered and three-layered Carbon fibres and onelayered, three-layered and five-layered E-glass. The other two specimens were internally reinforced. The only difference between these two columns was that one specimen was continually wrapped with three-layer E-glass fibres. The testing matrix is summarized in Table 1 .

z zy

616 FRPRCS-6: Externally Bonded Reinforcementfor Confinement

haunched end

zyxwvutsr zy EF0 R W 8 Circular Tie

A-A -

test region

A-A -

0 6-8 -

_81

zy 6RW10

8-B -

haunched end

AJ

(a) Column geometry

(b) Internal reinforcement details Figure 1. Column details

Table 1. Testing matrix on column specimens Column C1-1 C1-2 C1-3 C1-4 Cl-5 C2-6 C2-7

Diameter (mm) Ends Middle 235 150 235 150 235 150 235 150 235 150 235 150 150 235

Eccentric Loading

Length Internal (mm) Reinforcement 1400 1400 1400 1400 1400 1400 1400

Yes Yes No No No No No

Configurations 3-layered Carbon (ends only) 3-layered E-glass 3-lavered E-glass 5-lavered E-glass 3-lavered Carbon 1-layered E-glass I -layered Carbon

zy

In this study, all the seven columns were tested under eccentric loading, which was achieved by the introduction of haunched ends to each column. This can be seen clearly in Figure 1. When the concentric loading was applied to the top haunched ends of the column specimen, an eccentricity, e, of 42.5 mm, was achieved in the test region of each column. The large haunched ends were introduced in the configuration of the column

zyxw zyx

Externally Confined High Strength Concrete Cloumns 617

specimens in order to prevent premature failure and to allow for eccentric loading. A steel plate and a knife edge were used on the top surface of the column in order to provide an accurate concentric loading to the haunched end and to facilitate the adjustment on the direction of loading. Specimen Preparation A11 the seven column specimens were cast in the Engineering Laboratory of University of Wollongong. The target strength for both batches of concrete was 100MPa. 103.lMPa and 95.9MPa determined by compressive tests were achieved for the two batches of concrete respectively. After removal of the moulds, two internally reinforced columns were found with significant defects as shown in Figure 2, which were probably caused due to insufficient vibration. Then, column C1-1 was decided to be wrapped with three-layer of Carbon fibre at both haunched ends to prevent premature failure outside the test region.

zyxw

zyxw

(a) Column 1 (Cl-1)

(b) Column 2 (Cl-2)

Figure 2. Columns with significant defects

The resin was prepared by mixing with slow hardener according to 5: 1 ratio and firstly applied to the concrete surface. Then, the first layer of FRP was applied to the column with an overlap of 25 mm in each revolution. After wrapping the first layer, the second coating of epoxy was applied on the surface of the first layer to allow the second layer of FRP to be applied.

618 FRPRCS-6: Externally Bonded Reinforcementfor Conjinement

zyxwvu zyx

This process was repeated until the desired layers of FRP were wrapped. Finally, the final layer of epoxy resin was applied on the surface of the wrapped specimens. The wrapped column specimens were left at room temperature for about 2 weeks for epoxy system to harden adequately before the testing.

Test Specimens

Seven columns were tested to failure using 900 kN Strong Floor testing machine of the Engineering Laboratory at the University of Wollongong. The load eccentricity is 42.5 mm, which resulted in a large e/r (eccentricity/column radius) ratio of 0.57.

OBSERVED BEHAVIOUR AND TEST RESULTS All the columns showed similar behaviour under the eccentric loading. Although sounds of snapping of the fibres could be heard near the failure load, the failure of the column specimens in all cases was characterised by a very loud and explosive failure. The results from experiment conducted on the seven column specimens are shown in Table 2. Table 2. Column test results Column C1-1 (21-2 C1-3 C1-4 (21-5 C2-6 C2-7

Ultimate Axial Stress Axial Defection Max. Compressive Max. Tensile Load (kV (MPa) (mm) Stress (MPa) Stress (MPa) 836.4 525.5 601 736.8 791.5 669 644.6

47.33 29.74 34.01 41.69 44.79 37.86 36.48

0.753 1.983 2.4 1.45 1.405 1.771

--

-154.67 -97.14 -111.10 -136.20 -146.31 -123.67 -119.16

60.00 37.66 43.08 52.82 56.73 47.95 46.20 ~~

Figure 3 shows how the eccentric loading was achieved and produced an axial load combined with the bending moment. From the bending moment diagram shown in Figure 3, it can be seen the maximum moment occurred right at the joint between the haunched ends and the test region, which was exactly the same as occurred in the experiment as shown in Figure 41; all columns failed at the upper part of the test region except C1-2 due to the significant defects in it.

zyxw z

Externally Confined High Strength Concrete Cloumns 619

m = PP

Figure 3. Bending moment produced by the eccentric loading

Figure 4. Columns after failure

Internally Reinforced Columns The loading on the internally reinforced specimen wrapped with Carbon fibres at both ends resulted in the spalling of the concrete cover. The final sudden failure of this column was due to the yielding of steel reinforcement. Although defects existed in the haunched ends of this column, the failure of this column did occur in the test region as designed, which proved the effectiveness of wrapping using Carbon fibres at the ends. For the internally reinforced column with E-glass wrapping, ultimate failure occurred in the patched location. This confirmed that the final failure was marked by the fracture of the E-glass fibres as a result of lateral

zyxw

620 FRPRCS-6: Externally Bonded Reinforcementfor Confinement

expansion under axial eccentric loading, preceded by the crushing of concrete of the patching part. The results shown in Table 2 confirm that the influence of defects on the load carrying capacity, which is much lower than that of another internally reinforced column having similar configurations.

E-glass Wrapped Columns

z

The failure of all E-glass specimens was marked by the rupture of E-glass fibres. However the externally wrapped E-glass was ruptured in the hoop direction only for 3-layered E-glass column. While for 1-layered and 5 layered specimens, the fibres were torn in multi-direction and longitudinal direction besides the hoop direction, respectively. Approaching failure load, the appearance of white patches can be discerned, which indicated the imminent failure of E-glass and resin. The snapping sounds were heard before the ultimate failure, revealing the fracture of FRP composites and debonding between the layers of wrapping. Regarding the one-layered E-glass column, the inner side of wrapping was bonded with concrete even after failure, indicated that this column has achieved the best bond effect between the concrete and FRP. This is a possible explanation for this column having higher ultimate load carrying capacity than the single layered Carbon column.

Carbon Wrapped columns With a slight delamination of fibres between layers and accompanied by a simultaneous fracture of Carbon fibres and the concrete core, the final failure of both Carbon wrapped columns was more explosive and sudden when compared to the E-glass wrapped specimens. It is important to note that the ultimate load of column with a single layered carbon fibre is slightly lower than that of the single layered E-glass specimen. This can be attributed by the best bond effect achieved by the single layered E-glass specimen. Another reason is contributed the layout of fibres: the tensile strengthening by the unidirectional fibres is not as effective as that provided by plain weave fibres. COMPARISON AND ANALYSIS

zyxw

The experimental results from the six wrapped columns shows that the Carbon wrapped columns out-performed the E-glass wrapped columns. The three-layered Carbon specimen exhibited 7% and 23% increase over the five-layered and three-layered E-glass specimens, respectively. The three-

Externally Confined High Strength Concrete Cloumns 621

zy

layered and single layered Carbon columns exhibited 7.4% and 7.2% increase in ultimate load over the five-layered and three-layered E-glass columns, respectively. This proves that Carbon fibres are more effective than E-glass for external confinement. However, the single layered E-glass column achieved higher ultimate load than the single layered Carbon column due to the better bond effect and possibly smaller eccentricity. A comparison in terms of the maximum compressive stress and maximum tensile stress among columns C1-3, C1-4 and C2-6 show that increasing the number of layers leads to higher load carrying capacity of wrapped column generally. The five-layered E-glass column achieved 22% increase compared to the three-layered E-glass column. However, this is not the case for the single layered E-glass specimen due to the possible better bond effect and smaller eccentricity. As the steel plate and knife edge on the top surface of column could not easily be centred accurately, which could introduce smaller eccentricity, a higher ultimate load could be reached. The comparison between the two Carbon wrapped columns shows that increasing the number of layers from 1 to 3 increased the ultimate load by 23%. This again indicates that higher ultimate load could be achieved by increasing the number of layers. In order to evaluate the effectiveness of external confinement under eccentric loading as opposed to the internal reinforcement, comparison among Cl-1, C1-3, C1-4, C1-5 and C2-7 was made as well. Although C1-2 is one of the internally reinforced columns, it was not used here for comparison due to the significant defects that existed in this column. The three-layer Carbon wrapped column achieved the ultimate load of 79 1.5kN, which is just 5% lower than the high strength concrete column internally reinforced with high strength steel. This confirms that the external confinement with three-layer Carbon is as effective as the internal reinforcement with high strength steel. While for the E-glass wrapped columns under eccentric loading, the compressive stress and tensile stress of three-layered and five-layered columns were decreased by approximately 28% and 12% respectively. The ultimate load achieved by the column with a single layered Carbon fibre was decreased by 23% compared to the internally reinforced column.

zyxwvut

CONCLUSION

The experimental work involved in this study is mainly to evaluate the effectiveness of external and internal reinforcement as well as the contribution of two types of external reinforcement -- Carbon and E-glass to

622 FRPRCS-6: Externally Bonded Reinforcement f o r Confinement

high strength concrete columns under eccentric loading. Based on the test results of seven column specimens, it can be concluded that: (a) The experimental results clearly demonstrate that composite wrapping can enhance the structural performance of concrete columns under eccentric loading to some extent. However, the enhancement is not as significant as that of columns under concentric loading as suggested by previous studies. This was attributed to the fact that an eccentric loading once engaged, induced in the columns not only axial compression, but bending action too; (b) For the circular specimens under concentric or eccentric loading, the number of layers of FRF’ materials is one of the major parameters having a significant influence on the behaviour of specimens. However, the influence of the number of layers of FRP on the specimens under eccentric loading is not so pronounced as that of the specimens under concentric loading; (c) The fibre layout is one of major factors that affect the effectiveness of confinement especially when eccentricity is introduced. Plain weave fibres are effective both for flexural and compressive reinforcements. The behaviour of structural members can be markedly improved by using unidirectional fibres applied in right direction, which means the fibres are orientated in the direction where the higher tensile strength of FRP can be utilised; (d) Taking the expensive costs involved into consideration, external confinement with Carbon fibres is not suggested for strengthening of columns under eccentric loading at a larger eccentricity ratio.

REFERENCES

zy zyxw

1. Razvi, S. R. and Saatcioglu, M., “Strength and Deformability of Confined High-Strength Concrete Columns”, ACI Structural Journal, Vol. 9 1, November-December, 1994, pp. 678-687. 2. Demer, M. and Neale, K. W., “Confinement of Reinforced Concrete Columns with Fibre-reinforced Composite Sheets - an Experimental Study”, Canadian Journal of Civil Engineering, Vol. 26, iss.2, 1999, pp. 226-24 1. 3. Houssam, A. T. and Balauru, P., “Effects of Freeze-Thaw Exposure on Performance of Concrete Columns Strengthened with Advanced Composites”,ACI Materials Journal, Vo1.96, 1999, pp. 605-6 10. 4. Autar, K. K., Mechanics of Composite Materials, CRC Press, Boca Raton, New York, 1997

zyxwvut

FRPRCS-6, Singapore, 8-10 July 2003 Edited by Gang Hwee Tan QWorld Scientific Publishing Company

zyxwv zyx zyxwvu

CREEP PERFORMANCE OF CFRP CONFINED CONCRETE CYLINDERS

M. THERIAULT, M.-A. PELLETIER, K. KHAYAT AND G. AL CHAMI Department of Civil Engineering, Universite' de Sherbrooke 2500 b o d Universitk, Sherbrooke QC J I K 2R1,Canada Multiple rehabilitation techniques were implemented in order to increase the life cycle of deficient structures. One of the most promising techniques is the addition of fiber reinforced polymer (FRP) strengthening systems, which have proven to be quite effective in the confinement of reinforced concrete columns. However, important concerns regarding the durability of FRP-rehabilitated structures remain to be investigated. One of these concerns is the long-term creep behavior of FRP strengthening systems. In this study, 15 CFRP confined and unconfined concrete cylinders were submitted to sustained loading. Three short-term creep tests and four longterm creep tests, each performed over three replicates, were carried out. The parameters of the study include the level of confinement and the level of sustained stress. According to the results, confinement can effectively increase the creep resistance of concrete. This increase depends primarily on the level of sustained loading according to the confined concrete strength. For the same percentage of loading in term of ultimate strength, specimens with the highest confinement level exhibited the greatest short term creep resistance.

INTRODUCTION Creep is defined as a strain increase in time of a material submitted to a constant stress. Creep in concrete is a time-dependent phenomenon which can be influenced by numerous factors, such as the type of cement, the waterkement ratio, the use of admixture and the changes in humidity and/or temperature. Hence, the use of FRP-wrapped concrete columns exhibits a complex creep phenomenon, since other parameters must also be considered including the type and degree of confinement and the thickness of the resin between the FRP layers. According to the ISIS Canada Design Guidelines and with the proper FRP confinement upgrading, the ultimate load that can be applied on a column may be doubled'. This increase in capacity, while theoretically acceptable, brings forth new considerations as to the mechanical durability

624 FRPRCS-6: Externally Bonded Reinforcement for Confinement

zyxwv

of the system. In fact, for creep failure protection, the dead load applied on an unconfined concrete column should not exceed 80% of the concrete compressive strength. This value, once subjected to modification factors, is finally limited to 0.35 f,’. In the CSA reinforced concrete design manual’, the actual maximum applied load on a concrete column should not exceed 35% of its ultimate strength if all the design requirements are met; the risk of a creep failure at such stress level is therefore overlooked. However, the same cannot be said for FRP-confined columns, where the total factorized load can be as high as the unconfined concrete compressive strength3. At such high stress IeveI, the load-carrying capacity of the confined concrete columns depends on the effectiveness of the FRP wrap to restrain crack propagation, and thus prevent fragile failure. Further investigation on the durability aspects of this new technique is essential to investigate the long and safe use of this strengthening scheme. This study aims at the identification of the maximum dead load that can be sustained by an FKP-confined concrete column as a function of the applied confinement pressure.

zyxwv zyxwv

EXPERIMENTAL PROGRAM

Specimen Preparation

The concrete was prepared with coarse aggregates with 14 mm nominal size. Ready-mix concrete with a 65 mm slump and a 3.8% air content was used. Test cylinders were removed from mold one day after casting, cured in water for 7 days and then stored at room temperature until the time of testing. The mean compressive strength values at 28 days and 7 months were 24.5 and 29.1 MPa, respectively. The sample ends were ground with a diamond blade to perfectly smooth surfaces perpendicular to the cylinder axis for the secure creep testing. Unidirectional carbon fibers sheets were used for the confinement. The fiber properties are presented in Table 1. The design thickness provided by the manufacturer for a single ply is 0.165 mm. After 28 days of curing, the composite was wrapped around the cylinders in one, two or three layers corresponding to a confinement pressure of 7.4 to 22.2 MPa. A problem of insufficient anchorage length was observed during the preliminary static testing. This problem was solved by applying at the lapped joint an additional fiber sheet that had a length that did not exceed 65% of the

zyxw zy

Creep Performance of CFRP Confined Cylinders 625

cylinder circumference. Unconfined concrete cylinders were also tested and used as reference samples. Table 1. Properties of carbon fiber polymers Source

Tensile strength (N/m&plY)

Manufacturer Lab oratory

575 533

Tensile modulus of elasticity oV/mdplv) 37.5 42.5

Poisson’s Coefficient

0.22

Test Setup and Procedure Static tests

Before any creep test took place, a number of confined and unconfined concrete cylinders of 150 mm in diameter and 300 mm in height were subjected to axial compressive tests in order to obtain their ultimate strength and their stress-strain behavior. Testing procedures followed Standard CSA A23.2-9C. Short-term creep tests

zyxw

Short-term creep tests were performed on 150 x 300 mm cylinders. Standard CSA A23.2-9C loading rate was applied until the desired load was obtained. The load was then maintained until failure of the specimen. Cylinders were instrumented with an LVDT (Linear Variable Differential Transducer) and/or extensometer to monitor axial and radial strains. The short-term creep testing program started at high sustained load which was then progressively reduced for each new specimen until no failure was obtained after three days of sustained loading. This latest dead load level was then used as an upper loading level for the long-term creep testing. Long-term creep tests

The long-term creep testing was carried out according to ASTM Standard C512 using 80-ton frames (see Figure 1). Lab-built frames were mounted using upper and lower triangular trays, three huge steel bars threaded at both ends, ball-and-socket joint fixed under the upper tray and a flat jack connected to a pressurized air tank accumulator, and a manometer. Because of the ongoing danger of a fragile failure, the frames were built with two security units: Plexiglas windows installed in front of each of the triangular faces and, three large bolts and sockets systems attached at the bottom to

626 FRPRCS-6: Extenally Bonded Reinforcement for Confinemeni

the lower tray and at the top to a plate above the jack. This last security device can block the jack expansion in the case of unexpected large deformations or of a cylinder failure.

.zyxwvuts ; cr-

BALL AND SOCKET

ALUMINIUM CYLINDER 0150nn x 89nn

HUGE STEEL BARS

068nn

:i

3 CONCRETE CYLINDERS OF

0150nn x 300nn

OAD GENERATOR

zyxwvutsrq

ALUMlNIUM CYLINDER 0150"" x lOOnn FLAT JACK

LOWER TRAY

Figure 1. Frame used for the long-term creep testing

In each testing frame, three superimposed concrete cylinders measuring 150 mm in diameter and 300 mm in height were tested. Special attention was given to the superposition of the cylinders so as to secure an alternated position of the overlapping of the fiber sheets. Prior to loading, the upper and lower trays were checked for parallelism. The load was applied slowly by steps of 700 kPa (500 psi) to avoid eccentric loading. Once the desired pressure attained, it was maintained constant with a +70 @a (50 psi) margin of error. Cylinders were instrumented with two deformation acquisition systems. The first system is equipped with three displacement dials attached to the aluminum cylinders and spaced at 120 degrees around the frames. The dials (with -+0.01mm or 0.00001 m d m m precision) measure the displacement of the three cylinders all at once. The elastic deformation of the aluminum cylinders was carefully subtracted from the readings of the dials. The second measuring system consisted of four strain gauges all set on the middle cylinders of each frame. Three gauges were spaced at 120 degrees to measure axial strain; only one strain gauge was installed to measure radial strain. Temperature and relative humidity were also monitored. The axial

Creep Pe$omnce of CFRP Confined Cylinders 627

and radial strains of confined and unconfined cylinders were recorded at each loading step and at a rate corresponding to the strain increase once the target pressure level attained. Testing Parameters

zyxw zyx zy zy

Three short-term, and four long-term creep tests were carried out. The parameters of this study, which include the degree of confinement and loading level, are presented in Table 2. Table 2. Load level in function of

fi & fiC

Sustained load level Confinement scheme

Target Short-term

Long-term

0.8 $

U CI

Measured

1.2fi;0.9fic

Long-term

0.65

fi

1.2fi; 0.9fdc

0.8 fi ;0.4 fiC

fi ;0.4 f i C

c2

0.8

c2

1.2fi;0.6fic

c3

Short-term

1.7 fi ;OX5 fiC

fi ;0.5 fiC 1.7 f i ; 0.85 fiC

zyx zyxwvu

The specimen and sustained load level designations are as follows: U, unconfined concrete cylinder; C, confined concrete cylinder; 1, 2 and 3, number of CFRP layers; f i , compressive strength of unconfined concrete and; fiC, compressive strength of confined concrete.

TEST RESULTS AND DISCUSSION Compression Tests

Compression test results are presented in Table 3. Ultimate axial strains varied significantly from one specimen to another. Theoretical values were calculated according to the ISIS Canada Design Manual’. Significant differences between the experimental and expected theoretical values were found for the 3-ply specimens. The possibility of a maximum limit to

zyxw

6 2 8 FRPRCS-6: Externally Bonded Reinforcement for Confinement

confinement is not rejected. Further microscopic analysis also revealed the possibility of an insufficient resin thickness which might have led to the abrasion of the carbon fibers for the 1 and 3-ply specimens. Since these experimental results were obtained over a wide range of specimens with a standard deviation of less then 1 MPa, they were kept as reference values.

zyxwv

Table 3. Compression test results Confinement

Ultimate Axial Strains

scheme

(mm/mm) Minimum Maximum

Experimental

Theoretical

(MPa)

(MPa)

U

0.00225

0.00420

29

CI

0.01438

0.01605

38

43

c2

0.02076

0.02598

57

59

c3

0.02396

0.03048

58

73

Creep Tests When studying creep, several strain readings must be accounted for and subtracted from the recorded values. These readings include the elastic strain coming from the loading of the specimens, the dilatation or contraction of concrete due to changes in temperature or humidity and the volume variation due to the shrinkage of the concrete. In the present study, the temperature was kept in a relatively constant laboratory environment (1 S-23°C). As for humidity changes within the specimen, they were also of no concern since all specimens, including the unconfined cylinders, were covered with waterproof polymers. Shrinkage was not a relevant issue, since the testing did not start until at least one year of air drying. Therefore, only the elastic strain was subtracted from the total deformation readings. According to ASTM C512, elastic strains are obtained right after the loading for short-term creep testing and between 2 to 3 hours after loading for the long-term creep testing.

zyx

Short-term creep curves

Short-term creep test results are presented in Figure 2. According to this figure, the strain rate of concrete seems relatively constant on a log scale after about 2 minutes. Unfortunately for the C3 specimen tested at a sustained load of 170%&;85%&c, a power cut occurred after

zyxwvu zyxw zyx zyxwv zyxwv Creep Pegormance of CFRP Confined Cylinders 629

approximately 8 hours and 14 minutes of testing. The specimen was reloaded 10 hours later, and the strain readings were adjusted consequently. After 70 hours of sustained reloading, the specimen did not failed, and the loading was stopped. To give a better idea of the proper strain readings and time adjustment, the additional time and axial strain registered from the second loading was not added to the first loading until the strain rate registered before the power cut was approximately obtained. This explains the kink point in the C3 170% fi ;85% ficcurve.

-$

n

0,0001 8.64 sec 0 -200

-400 S -600 0 * -800 -1000 % -1200 .! -1400 l d 5; -1600 -1800

g s

0,001

Testing duration (days) 0,Ol 071

1.44 min

14.4 min

2.4 h

1

10

1 day

120% f ,-90% f cc

170% f,-85% f,,

1

C3 (failure at 77h)

Figure 2 . Deformations from the short-term creep tests

Given the above results, the major factor effecting creep seems to correspond to the level of sustained loading, expressed as % of fit. The lower is this level, the higher is the creep resistance (85% fit, as compared to 90%/Oc). The loading level characterized by fi does not yield any apparent relation. However, for the same loading level in terms of fiC, the specimen with the greatest load level in terms of fi ,which corresponds to a higher confinement, showed the highest creep resistance (1 80% fi ; 90%, fiC, as compared to 170% f i ; 90% fiC).

zyxwvu zy zyx zyxwvu

630 FRPRCS-6: Externally Bonded Reinforcementfor Confinement

Long-term creep curves

Figure 3 shows the on-going results of the evolution of the axial strain with time obtained for the four experiments. The axial strain readings are deducted from the mean of three dial gauges positioned around the loading frame, which is also the average reading of three specimens (three specimens/frame). The strain gauge readings on concrete and FRP became inaccurate after a few days for all the loading levels above 80% 5 . This could be due to creep and/or relaxation of gauge and of the gauge bonding polymer under high strain. Creep deformation obtained from the strain gauges also had higher standard deviations than the readings obtained from the dials. A possible explanation would be the highly localized measure obtained from the strain gauges. Locally, a concrete may present defects and different creep behavior depending on the material immediately beneath the gauge. For comparison purposes, a creep curve obtained by Neville4 for a normal concrete at a load level equal to 70% fc' is presented in Figure 3.

Time (days)

0

50

100

150

200

0

n

E

33. W

0 c,

8

gu

ru 0

.-5

m

-500

zyxwv -1000

Neville 70% f,

zyxwvu 1 -1500 -2000 -2500 -3000

-3500

C3 100% f,-50%fcc

J

C2 100% fC-50%f,

Figure 3. Long-term creep test curves

When comparing the different curves, it can be seen that the 65%& loading curve of the unconfined concrete was quite similar to the 7 0 % 5 loading curve given by Neville, with corresponding lower strains given by a lower relative loading. As expected, higher loading resulted in higher creep strain, while similar loading led to similar creep behavior.

zyxw zyxwvuts Creep Pegormance of CFRP Conjined Cylinders 631

zyxwv

Creep strains

According to Neville4, the creep failure of concrete occurs at a total strain corresponding to the ultimate concrete strain obtained from a static compression test. This statement was not confirmed for the short-term creep tests, as shown in Table 4.Short-term creep tests showed higher strains at failure than the static compression tests. Therefore, it is expected that the long-term creep tests will also show higher strains. This phenomenon can be compared to the loading rate: the slower the loading rate is, the greater the strain at failure can be. Table 4. Projected creep resistance

Expected Failure Failure (days) (Range in days)

Actual Strain @ age W / m @ days)

Strain Rate (pm/m/day)

1330@100

2.049

449-1400

CI-1.2J;?;0.95c

[email protected]

29260

0.0859-0.1280

C2- .8fd;0.4fic

25980100

5.530

3 160-4228

6 3 0 3 0 149

5.883

2457-3345

C3- J? ;0.5 fir

4331050

14.55

1349-1797

C3-1.7fC;0.85 fir

36693”@2.95

0.4240

0.0023-0.1 15

3479003.22

393.5

0.0027-0.34

Test Designation U-0.65

C2-

C3-I 8

5

;0.5 J?c

5 ;0.9 fir

z

0.1208

3.22

‘Readings interrupted by a power cut

zy

Using the ultimate static strain as a failure criterion, combined with the hypothesis of a regular strain rate, the time of failure of the long-term creep tests were roughly estimated in Table 4. Considering that higher strains were obtained for the short-term creep tests, and considering that strain rate decreases with time and that the U- 0 . 6 5 5 specimens should not fail, a longer lifetime can be expected for all the long-term creep tests. CONCLUSIONS AND RECOMENDATIONS

At a sustained load equal to the ultimate strength of concrete, the FRP confinement was found effective to limit crack propagation for several months without showing any signs of upcoming failure. With regards to the

zyx

zyxw zy z

632 FRPRCS-6: Externally Bonded Reinforcementfor Conjinemeni

limited results of this study, two hypotheses can be drawn: (1) In FRP confined concrete cylinders, creep resistance is inversely proportional to the loading level as a fraction of fd.; and (2) for the same type of concrete and identical loading level as a function of fit, the system with the highest confinement might show the greatest creep resistance. Further investigations are however needed to confirm the results of this study and to clearly identify the creep limit of confined reinforced concrete columns. ACKNOWLEDGMENTS

The authors would like to acknowledge the financial support of the Natural Sciences and Engineering Research Council of Canada (NSERC), the Network of Centres of Excellence on Intelligent Sensing for Innovative Structures (ISIS Canada), le Centre de Recherche Interuniversitaire sur le Beton (CRIB), and le Fonds pour la Formation de Chercheurs et 1’Aide Ci la Recherche (FCAR). We also gratefully acknowledged the support of Master Builders Inc. for the donation of the FRP materials. REFERENCES

1. Neale, K., Strengthening Reinforced Concrete Structures with Externally-Bonded Fibre Reinforced Polymers - Design Manual No. 4, ISIS Canada Corporation, Winnipeg, Canada, 2001,210 p. 2. CSA, Standard 23.3-94, Design of Concrete Structures, Canadian Standards Association, Rexdale, Canada. 3. ThCriault, M. and Neale, K.W., “Design Equations for Axially-Loaded Reinforced Concrete Columns Strengthened with Fibre Reinforced Polymer Wraps”, Canadian Journal of Civil Engineering, 27(5), 2000, pp. 1011-1020. 4. Neville, A.M., Propriitis des bktons, Eyrolles, Paris, France, 2000, 806 p.

zyxwvutsr

FRPRCS-6, Singapore, 8-10 July 2003 Edited by Gang Hwee Tan QWorld Scientific Publishing Company

DEVELOPMENT/SPLICE STRENGTH OF STEEL BARS IN CONCRETE CONFINED WITH CFRP SHEETS

zyxwv

M. H. HARAJLI AND B. S. HAMAD

Department of Civil and Environmental Engineering, American Universiv of Beirut, Beirut, Lebanon

The aspect of bond between reinforcing steel bars in tension and concrete confined with flexible carbon fiber-reinforced polymer (CFRP) sheets is analytically investigated. The bond analysis incorporates an experimentally derived local bond stress-slip model, applicable for both plain unconfined concrete and concrete confined with CFRP flexible sheets. It is found that confining the concrete with CFRP reinforcement increases considerably the bond strength and leads to significant improvement in the ductility of bond failure. Based on the analysis results, supported with experimental data, a design expression is proposed to evaluate the development length of reinforcing bars embedded in concrete confined with CFRP flexible sheets.

INTRODUCTION Bond strength between reinforcing bars and concrete is one of the most important aspects' that influence the structural performance and serviceability of reinforced concrete members under both static and dynamic earthquake loading. Bond strength can be improved by providing adequate bar developmentlsplice length, increasing the concrete covers, and confining the concrete at the critical locations where plastic hinges are most likely to develop. Fiber-reinforced polymer (FRP) composites are used to repair and strengthen reinforced concrete structural members, especially beams and columns. Many reinforced concrete beams have been tested, demonstrating the feasibility and efficiency of this technology to improve flexural stiffness and strength as well as seismic response.'. On the other hand, studies of the effectiveness of FRP in increasing the bond strength capacity of reinforcing bars in tension are very limited.

zyxwv

OBJECTIVES The main objectives of this investigation are to evaluate the bond performance of reinforcing bars embedded in concrete confined with carbon

634 FRPRCS-6: Externally Bonded Reinforcement for Confinement

zyxwvu

fiber-reinforced polymer (CFRP) flexible sheets, and to propose based on the analytical results and available experimental data, design expressions for evaluating the bond resistance and development/splice length of reinforcing bars, taking into account the effect of confinement provided by CFRP sheets. ANALYTICAL BOND MODEL

The analytical approach adopted in this study is based on a numerical solution scheme of the bond problem. In the analysis, the developed/spliced bar is subdivided into small elements and the stresslstrain at the loaded end is increased in increments. At any straidstress level during the response, the bar stress, bar slip and bond stress distribution along the development/splice length are determined using constitutive local bond stress-slip response and stress-strain model of the constituent materials to satisfy known or assumed stredstrain boundary conditions. More details of the analytical approach are described elsewhere3. The local bond law is shown schematically in Fig. 1. It is composed of an envelope curve applicable for pull-out bond failure in well-confined concrete and reduced bond stress-slip response corresponding to splitting bond failure for plain unconfined concrete or for concrete confined with CFRP flexible sheets. The splitting bond curve for CFRP confined concrete was developed recently by Harajli and Hamad4 based on experimental testing of beam specimens with short spliced bars at midspan having bar sizes db between 16 mm and 32 mm, and ratios of minimum concrete cover to bar diameter c / db between 0.56 and 2. I . The splice zone was confined by wrapping the beam with one or two layers of CFRP flexible sheets applied along the full splice length. The manufacturer's data of the sheets are as follows: design thickness of the fabric is equal to 0.13 mm, modulus of elasticity is equal to 230,000 MPa, tensile strength equals 3500 MPa and strain at break of the fibers is 1.5%. Referring to Fig. I, the envelope curve is expressed as follows5:

zyxwvu zyxw zyxwv z z zyxwvu

where u is the bond stress and s is the slip;

u1 (MPa) = 2 . 5 7 f i

; and sI =

0.15c0, s2 = 0.35 co and s3 = co, where co is the clear distance between the

zy zyxw zy

Development/Splice Strength of Steel Bars 635

ribs of the reinforcing bar; or equal to 1.5, 3.5, and 10 mm, respectively, if no information is available about the bar rib geometry; u f = 0 . 3 5 ~ ~ .

(c/dh)2/3I ul

urnox =

zy zyxwv zyxw

in which c is the minimum concrete cover and dh is the diameter of the steel bar, K = 0.78 for plain concrete and 0.92 for concrete confined with either one or two layers of CFRP flexible sheets. The terms a = 0.7; p = 0.65; and ufi = 0 . 3 0 ~ The ~ ~ .slip s,,, at which urn,, is mobilized, is calculated as follows: (1/ Q.3)Ln(+)

Smax

= '1'

+ s,Ln(-) UI

Umax

zyx (3)

where so = 0.15 mm for plain concrete and 0.20 mrn for concrete confined with CFRP sheets. For plain concrete, the descent in the bond stress u with increase in slip s follows the following expression: u=@ ,,

(s / s,,)-0.5

(4)

For concrete confined with CFRP sheets along the full splice length: r

1

In which k/-and k2 are equal to 0.8 and 0.13 for the beams confined with one CFRP layer, and 0.9 and 0.13, for beams confined with two layers, respectively. DISCUSSION OF RESULTS Analytical evaluation of the bond strength of plain unconfined concrete using the bond law for plain concrete in Fig. 1 was already undertaken in Ref. 3, where the results showed excellent agreement with experimental data.

zyx z zyxwv zyx zyxwvu zyxwvu zyxwv

636 FRPRCS-6: Externally Bonded Reinforcement f o r Confinement

Variation of predicted average bond strength at failure U, normalized to f ’ c 1 ’ 4 , versus ratio of developmenthplice length to bar diameter, L, / d, , corresponding to two different ratios of concrete cover to bar diameter, c / d, , and for different level of CFRP confinement are shown in Fig. 2. Typical results showing normalized steel stress versus slip response for CFRP confined concrete in comparison with plain unconfined concrete are shown in Fig. 3. The results were obtained assuming the reinforcing bar does not yield (i.e.,f, = infinite), db = 25 mm, and normal-strength concrete ( f ’ , 5 SOMPa).As shown, provided the concrete strength is within the range of NSC, the results are not sensitive to bar diameter or concrete compressive strength. Also, it should be pointed out that the choice of ‘A power instead of the conventional 54 power of f’,for normalization of the bond results is based on a recent study by Zuo and Darwin6 in which it was found that the use of ‘/4 power of concrete compressive strength leads to better representation of the effect of f‘,on bond strength. Figures 2 and 3 clearly show that confining the concrete with CFRP sheets leads to significant improvement in the bond strength at failure at all levels of L,/ d, investigated. The increase in bond strength for CFRP confined concrete increases as L,/d,or as c / d b increases (Fig. 2). Comparing the mode of failures in Fig. 3, it is clear that confining the concrete with CFRP sheets leads to a much gradual degradation in strength with increase in bar slip, and consequently much more ductile behavior in comparison with plain unconfined concrete. While using two layers, or doubling the area, of CFRP sheets may increase only slightly the ductility of bond failure in the post splitting range (Fig. 3), it does not lead to significant increase in bond strength as compared to concrete confined with one layer.

PROPOSED EQUATION FOR DEVELOPMENT LENGTH The predicted bond strength at failure of concrete confined with CFRP flexible sheets applied along the full developmenthplice length, corresponding to a wide range of c / d , and L,/d, parameters, were sorted in many different ways. The way that produces the least scatter is to plot the results as ratio of bond strength with CFRP to that of plain unconfined concrete, in function of A , = w f t , / d , ,where Af is the area, wfis the width ( w f / L, I 1.0 ), and t/- is the design thickness of one layer of the

zy

z zyxwv Development/Splice Strength of Steel Bars 637

i NIax

Bond Stress (u)

r

S3

Slip (s)

z

Figure 1 . Local bond stress-slip model used in the analysis

4.0 3.5

zyxwv 3.0

z 4

h

h

bU

W

5

2.5

2.0

1.5 1.o

0.5

0.0

0

10

20

30

40

50

60

Ld/db

Figure 2. Predicted variation of normalized bond strength U at failure with LJdh and C/dh for different levels of CFRP confinement

zyxw zy

638 FRPRCS-6: Externally Bonded Reinforcement f o r Confinement

CFRP fabric. Since for the thickness of the CFRP fabric used in deriving the experimentally based local bond law (tJ = 0.13 mm), the number of CFRP layers, or total thickness, did not influence much the results, the parameter w ft f /d, can be reduced to wf /d,. Typical variation of analytical predictions with w f / d bis shown in Fig. 4. Shown also in Fig. 4 is the

trend of experimental data obtained in Ref. 4 for L, / db = 5.0,upon which Eq. (2) is based, and the experimental results reported recently728for NSC and HSC corresponding to L, /d, = 15.0, in which CFRP sheets, similar to the one used in this analytical study, were applied in one or two layers with ratios w f / L, of 1/3,2/3 and 1.O, respectively. It is clear from the results depicted in Fig. 4 that expressing the bond results as proposed (bond ratio) leads a consistent trend in both the experimental and analytical results. Note that for concrete confined with ordinary transverse steel, the effect of the transverse reinforcement on bond strength is expressed as bond increment (above that for plain unconfined concrete) that increases linearly with parameter A,r /nsd, (Ref. 6), where A,r is the area per one stirrup that crosses the potential plain of splitting, s is the spacing of stirrups, and n is the number of splices or bars being developed. If an analogy were to be used with ordinary transverse reinforcement, a more appropriate parameter to reflect the influence of CFRP sheets on bond strength (assumed to be applied along the full splice/development length) would be to use 2NAf / w,nd, = 2Nt, / nd, , where N is the number of CFRP applications (or layers). However, because the presence of CFRP sheets altered the mode of splitting failure from sidesplitting to predominantly bottom-splitting4.7, s and also since doubling the number of layers did not lead to a noticeable increase in bond strength (see Figs. 2-4), it is believed that the mechanism by which CFRP sheets influence the bond strength is different from that when ordinary transverse reinforcement is used. This difference in the mechanism of bond resistance may justify the use of “bond ratio” instead of “bond increment” and, accordingly, a CFRP parameter w f /d, instead of 2Ntf /nd,that would have been used if analogy with ordinary transverse reinforcement were made. Based on the results of this study, the following equation is proposed to calculate the bond strength of concrete confined with CFRP flexible sheets, regardless of the number of layers used (see Fig. 4), provided the thickness of the layers is not less than 0.13 mm:

zy

zyxwvu zyxwvu zyxwvutsrqp zyxwvuts zyxwvutsrqpon zyxwvutsrqpo

zyxwvutsrqpzy *

.d

120000

“E E-

zn -3

5 n

$ 5

100000 80000

60000

4

1.9 1.8 1.7 1.6 1.5 1.4

-a- Confined conc (1 CFRP layer)

1.3 1.2

XExperiment (NSC)

1.1 1

0

10

20

30

40

50

60

$ t

w/db

40000

20000

2

1.9 1.8

4

$P

1.7

;;i 1.6 & 1.5

s$’

1.4

2

.CI

0

0

2

4

6

8

1 0 1 2 1 4

Slip at Loaded End, mm

5 %

1.3

1.2 1.1

Figure 3. Typical predicted variation of normalized bond force A& versus slip response

B

1

0

10

20

30

40

50

60

wddb Figure 4. Comparison of predicted and experimental results

s2 ta

Q

a

2 \o

640 FRPRCS-6: Externally Bonded Reinforcement f o r Confinement

where, in the absence of experimental data to justify larger increases, the term (1 + 0 . 0 1 7 /~d~, ) shall not be taken more than 1.5; p is a concrete

zyxwvu zyxwv zyxwv zyx zy zyxwv

strength power (p = !4 or %). For plain unconfined concrete, the bond strength can be calculated using the design equation proposed by Orangun, Jirsa and Breen', upon which the ACI Building code" philosophy for bond design is based, or the more recent expression proposed by Zuo and Darwin6. For instance, using the expression by Zuo and Darwin, the bond force A, f , ( A, f,= Uzd,L, ) at bond failure for plain unconfined concrete is calculated as follows: -= Abfs

f y4

[59.8L,(cm+0.5db)+2350A,]

zyxwvu

in whichh is the steel stress and f Ic is the concrete strength in psi, 4, is the bar area (in2), c, is the minimum and cMis the maximum value (cdc, < 3.5) of c, or cb (in inches) where cs is the smaller of % clear distance between bars + 0.25 in. or side cover, and cb is the bottom cover of the reinforcing bar. Combining Eqs. (6) and (7) leads to: )cFRp = (1

fY4

+ 0.017 -)[59.8L, w/

(cm+ 0 . 5 4 ) + 2 3 5 0 A , ]

db /

, (8)

1

0 . 1 2+ 0.9 [ e m

When Eq. ( 8 ) is solved for the development length L, and by considering the conservative simplifications made by Zuo and Darwin (2000), the following expression is obtained:

zyx

zyx

zyxw zyxw DevelopmentLYplice Strength of Steel Bars 641

_fy_ -

Ld = db

fY4

(

68 1+0.017--

2100

(9)

-

:)(’b)

wheref, is the yield stress of the steel bar in psi. In analogy with the effect of confinement using ordinary transverse reinforcement, and in order to safe guard against pull-out bond failure as currently the philosophy of the ACI Building code”, it is recommended to limit the maximum value of (1+ 0 . 0 1 7 /~d~b ) ( c / d b )to 4.0 as suggested by Zuo and Darwin6

CONCLUSIONS Based on the results of this study, the following conclusions are drawn: (a) Confining the concrete with CFRP flexible sheets leads to significant improvement in bond strength and ductility of bond failure. (b) While using two CFRP layers may increase the ductility of bond failure, it does not lead to significant increases in bond strength. (c) Based on the analytical results, supported with experimental data, general design expressions were proposed to evaluate the development/splice strength and developmentlsplice length of reinforcing bars embedded in concrete confined with CFRP flexible sheets. ACKNOWLEDGEMENTS This work is supported by the Lebanese National Council for Scientific Research (NCSR). The authors are grateful for that support, and to the Faculty of Engineering and Architecture at the American University of Beirut (AUB) for providing the computer facilities. REFERENCES 1. ACI Committee 440, “State-of-the-Art Report on Fiber Reinforced Plastic (FRP) Reinforcement for Structural Applications,” ACI 440R-96, 1996, American Concrete Institute, Detroit, MI.

642 FRPRCS-6: Externally Bonded Reinforcementfor Confinemen1

2. ACI Committee 440, “Guide for the Design and Construction of Externally Bonded FRP Systems for Strengthening Concrete Structures,” ACI 440.1 R-0 1, 200 1, American Concrete Institute, Detroit, MI. 3. Harajli, M. H., and Mabsout, M., E., “Evaluation of Bond Strength of Reinforcing Bars in Plain and Fiber Reinforced Concrete,” ACI Structural Journal, Vol. 99, No. 4, July-Aug., 2002, pp. 509-5 17. 4. Harajli, M., and Hamad, B., “Bond-Slip Behavior of Reinforcing Bars in Concrete Confined with CFRP sheets,” Submitted for possible Publication in the ACI Structural Journal, 2002. 5 . Harajli, M., Hout, M, and Jalkh, W., ”Local Bond Stress-Slip Relationship of Reinforcing Bars Embedded in FRC”, ACI Materials Journal, 92(4), July-Aug., 1995, pp. 343-354. 6. Zuo, J. and Darwin, D., “Splice Strength of Conventional and High Relative Rib Area Bars and High-Strength Concrete”, ACI Structural Journal, V. 97, NO. 4, July-August, 2000, pp. 630-641. 7. Hamad, B., Soudki, K, Harajli, M., and Rteil, A., “Experimental and Analytical Evaluation of the Bond Strength of Reinforcement in FFW Wrapped HSC Beams,” Submitted €or review and possible publication in the ACI Structural Journal, 2002. 8. Hamad, B., Rteil, A., Selwan, B., and Soudki, K., “Behavior of Bond Critical Regions Wrapped with FRP Sheets in Normal and High Strength Concrete,” Submitted to ASCE Journal of Composites for Construction, 2002. 9. Orangun, C. O., Jirsa, J. O., and Breen, J. E., “Strength of Anchored Bars: A Reevaluation of Test Data on Development on Development Length and Splices,” Research Report No. 154-3F, Center for Highway Research, University of Texas at Austin, 1975,78 pp. 10.ACI Committee 3 18, “Building Code Requirements for Structural Concrete”, ACI 3 18-99, American Concrete Institute, Farmington Hills, Mich, 391 pp.

zyxw zyx

zyxwvutsr

FRPRCS-6, Singapore, 8-10 July 2003 Edited by Gang Hwee Tan @WorldScientific Publishing Company

zyxw zyx

LATERAL PRESTRESSIP G OF RC CO FRP JACKETS

JMNS WITH

A. A. MORTAZAVI, K. PILAKOUTAS AND M. A. CIUPALA Department of Civil and Structural Engineering, The University of ShefJield Mappin Street, Shefield SI 350, UK

This paper presents the results e o m experiments on a new strengthening technique for concrete columns that uses expansive materials to apply lateral pretensioning. The level of pre-tensioning is controlled by using different amounts of expansive agent. The technique aims to enhance the capacity and ductility of columns as well as achieve better utilisation of the confining FRP material. It is shown that jacketing columns by pre-tensioned FRP materials can increase the load bearing capacity up to 30% compared with conventional wrapping and up to more than 2 times compared with unconfined concrete. However, the most important effect of pre-tensioning is a delay in the initiation of the hcturing process of the concrete through cracking. The paper presents details of experimental work undertaken with different confining materials (CFRP and GFRP) but having the same ultimate jacket strength.

INTRODUCTION RC columns can be very vulnerable to seismic actions, especially when they are deficient in lateral reinforcement. This can lead to shear failure, lap splice failure, buckling of the longitudinal reinforcement and premature concrete compressive failure at low displacements. All these failure modes reduce the ductility and energy dissipation potential of columns. Since the 1995 Kobe earthquake, composites have been used for the repair and strengthening of columns against seismic actions and seismic codes need to be updated to account for these new materials and techniques. One of the problems with FRP confinement of concrete is that the strength of the FRP jacket is not mobilised until the lateral strain in the confined concrete is very high. In some cases, the concrete will crush before the FRP jacket is fully ~ t i l i s e d ' ~The * ~ ~existing . designs equations for steel confined concrete, such as for Eurocode S4, assume that the confining steel is fully utilised. These equations should not be used for FRP confined concrete, since the predicted properties of the confined concrete are not necessarily achieved. Hence, several researchers have adopted concrete constitutive models developed for steel reinforced concrete for use with FRP jackets. It is possible to overcome this problem of strength utilisation by

zyxw

644 FRPRCS-6: Externally Bonded Reinforcementfor Confinement

reducing the lateral strain at failure of the concrete through pre-tensioning. This approach is particularly useful for low modulus materials (such as glass) or when relatively low amounts of confinement are used’,223. However, it is not easy to apply large pre-tensioning stresses in composites, especially when using the popular method of wrapping of sheet material. Some researcher~~’~” tried to develop lateral pre-tensioning through grout or resin injections, but had limited success in changing the concrete behaviour. The pre-stressing of the composites in this work is achieved by using an expansive agent’ (EA), normally used for concrete demolition, mixed with a cement grout in different proportions. This can ensure that the concrete is actively confined by the composite even at service loads. The advantage of this method is that, by applying additional lateral pressure in the early stages of loading, concrete dilation is delayed (as will be seen from volumetric expansion graphs). This results not only in higher confined concrete strengths, but also in higher energy dissipation. This paper will present details of a part of this experimental work with two different types of confining material (glass and carbon) and two levels of initial pre-stressing. This research is conducted at the Centre for Cement and Concrete of the University of Sheffield, UK. This research forms part of work undertaken under the EU TMR Network ConFibreCrete.

zy

PRETENSIONING METHOD In practice, the expansion of the jacket can take place through the injection of the expansion grout (EG) in preformed cavities at specific locations, such as the corners of rectangular columns. The method as applied in the laboratory differs and depends on gap size or the other parameters. In this method, a pre-formed confining tube (jacket) is placed around an existing czncrete cylinder and then the EG, comprising cement, sand and EA, is inserted between the concrete cylinder and the jacket. The jacket confines the expansion of the grout during the hardening period (3-4 days) and pressure builds-up due to the reaction of the EG against the concrete core. Once the EG sets, the jacket and grout become an integral part of the column. The expansive pressure of the grout has been shown’ to be a function of the allowable lateral displacement. If a large displacement can take place, such as when confining with very low stiffness materials, then the pressure can reduce to zero. Hence, special experiments were undertaken’,’’ to develop the understanding of the mechanical properties of the EG so as to enable accurate prediction of the lateral pre-tensioning pressure. Figure 1 shows the maximum expansive pressure (MEP)

zyxwvu zy zyxwvu zyxw Lateral Prestressing of RC Columns with FRP Jackets 645

calculated by E ,

.tcprel r where EFw is the Young’s modulus of elasticity of the jacket, t is the thickness of the FRP jacket, E~~~ is the maximum lateral expansion measured by the lateral strain gauges, and r is the radius of the jacket. MEP expected to be developed in EG having two different ratios of EA. The horizontal axis shows the confinement stiffness (CS) calculated by EFRp. t / r . It can be seen that the relation between MEP and CS is more or less linear, as shown by the trend lines. These relations are currently being developed by the authors into design equations for use with this method of pre-tensioning. 35 1

-

30 25

g

20

=

10

15

5

0 0

10000

20000

30000

40000

50000

CS (MPa)

Figure 1. MEP versus CS for two values of EA

EXPENMENTAL PROGRAMME The experimental programme comprised of testing concrete cylinders with different confinement configurations. This section gives details of material properties, specimen preparation, instrumentation and test procedure. The properties of the materials used in this study are shown in Table 1.

zyxwv zyxw zyx Table 1. Fibre properties

Fibre type

CFRP GFRP

t (mm)

0.1 17 0.135

(MPa) 240000 65000

EFRp

f,,,

(MPa)

3900 1700

EFM,

(%I

1.55

2.80

where f ,,, and cFRPu are the ultimate tensile strength and ultimate elongation in the FRP jacket. The concrete used was made with Ordinary Portland Cement, maximum aggregate size 10 mm and cylinder strength &) of 32 MPa.

646 FRPRCS-6: Externally Bonded Reinforcement for Confinement

Specimen Details

zyxw

Since the chemical pre-tensioning pressure (expansive pressure) is caused by the EG reacting against the confining jacket, this means that the magnitude of this pre-tension depends on the degree of stiffness of the jacket and percentage of EA. These two parameters were investigated in an extensive series of experimental work. A total of twenty seven lOOmm x 200mm concrete specimens were prepared without any pre-tensioning, 54 specimens were prepared with different levels of confinement pre-stressing and 18 unconfined specimens were tested under compression to determine the plain concrete strength. Four different confining materials were used (steel, glass, aramid and carbon) with different number of layers and ratios of EA. In this paper, only some results from Carbon and Glass FRF'jackets will be presented.

Instrumentation and Testing For the measurement of lateral strains, three 15mm surface strain gauges were attached horizontally at the mid-height of each specimen, 120" apart. To measure longitudinal strain, two surface strain gauges of 15mm length were mounted vertically in the middle height of the specimen. In addition to strain gauges, two other devices (DV1 and DV2) were designed to measure lateral and longitudinal strain by using displacement transducers. These devices are shown in Figures 2 and 3. The tests were undertaken in a servo-controlled hydraulic actuator under displacement control. All samples tested by monotonically loading.

Figure 2. Plan view of DVl

Figure 3. Elevation view DV2

EXPERIMENTAL RESULTS Due to the large amount of experimental data produced by each test, only a limited amount can be presented in this paper. Two pairs of specimen

zyxw zyxw

Lateral Prestressing of RC Columns with FRP Jackets 647

confined with Glass and Carbon FRP were selected. Each pair shows the effect of direct wrapping and pre-tensioning of the jacket. The properties of the fibres used as shown in Table 1. The glass fibre sheet was bi-directional (90"), hence the effective thickness of confinement is 0.0675 mm. For the selected specimens 4 layers of glass fibre were applied with an overlap of 110 mm. In the first specimen (WG4) the glass fibre sheet was applied directly onto the appropriately prepared concrete core. The second specimen of this pair (PG4-30) had the same amount of glass fibre, but the jacket was pre-tensioned with a 6 mm thick EG having 30% EA. The pair of CFRP confined specimens had only one layer of fibres with the same overlap as for the GFRP specimens. However, the pre-tensioning grout only had 20% EA (specimen PCI-20). The total confinement strength was the same for all specimens (around 460 N/mm width), but the carbon layer was 55% stiffer in the radial direction than the 4 layers of glass fibre.

zyxwv

Failure mode Failure was always explosive due to the high strain energy stored by the FRP material and it took place around the middle of the cylinder height. Figures 4(a) and 4(b) show the failure of PG4-30 and PC1-20 respectively.

Figure 4. (a) failure of PG4-30, (b) failure of PC1-20

As shown in these figures, the mode of failure of the glass fibre confinement is completely different from that of the carbon. This is due to the bi-directional nature of the glass fibre wrapping which at failure has the effect of redistributing lateral strains over the full height of the specimen. In

648 FRPRCS-6: Externally Bonded Reinforcement for Confinement

zy zyxwvutsrqponmlkjihgf

the case of carbon, the failure of the fibres at one location lead to the rapid debonding of the filaments near that location only. Stress-strain results

Figures 5 and 6 show the stress-strain relationships for all specimens. I

-25000

16 c.4

,

I

zy zyxwvut

-20000

-15000

-10000

-5000

0

5000

10000

Microstrain

Figure 5 . Stress-strain curve for WG4 and PG4-30 specimens

-20000

-14000

-8000

-2000

4000

10000

16000

22000

Microstrain

Figure 6. Stress-strain curve for WCl and PC1-20 specimens

Positive strain indicates axial compression and negative strain indicates lateral tension, as obtained from the averages of strain gauges on the surface of the jacket. The vertical axis shows the stress normalised with respect to the unconfined concrete strengthf,,.

zyx

z zyxwv zyxwv Lateral Prestressing of RC Columns with FRP Jackets 649

GFRP confinement

The lateral pre-tensioning strain developed in PG4-30 is around 7300 p& and this also led to a modest expansion in the axial direction. This expansion was restrained by the glass fibres in the axial direction. A strength of 2.23 Loand 1.88Lowas achieved by PG4-30 and WG4, respectively. The failure of WG4 took place when the average lateral strain was around 18000 pe whilst for PG4-30 the average lateral strain at failure exceeded 25000 p, which means that the strength of the glass was fully utilised. It is worth noting that in WG4 the lateral confinement was only mobilised at around 80% of Lo,whilst for PG4-30 a different behaviour can be noticed all together.

CFRP confinement The CFRP confined specimens had a similar behaviour to the GFRP confined specimens. The lateral pre-tensioning strain developed in PC 1-20 is around 5300 pe and this again led to a modest expansion in the axial direction. A strength enhancement of 2.10L0 and 1.7Of,, was achieved by PC1-20 and WC1, respectively. The failure of both specimens took place when the average lateral strain in the carbon sheet was in excess of the 1.55% specified by the supplier. Again, there is a substantial difference in the level at which the confinement is mobilised. Hence, it is worth examining the volumetric strains of these specimens. Volumetric strain Figures 7 and 8 show the normalised axial stress against the volumetric strain for all specimens. In both pairs, the volume decreases at the initial stages of loading until a critical level is reached just below Lo. At this stage, volumetric dilation begins, which means that concrete cracking is developing rapidly. In the pre-tensioned specimens, PG4-30 and PC 1-20, the volumetric dilation is delayed by almost 50% offco.This has advantages in seismic loading, since the damage in the concrete will be delayed and the reinforced concrete element will have a chance to dissipate more energy. In addition, concrete is actively confined even at service loads.

z z y x w v u t s zyxwvu zy zyxwvu

650 FRPRCS-6: Externally Bonded Reinforcement f o r Confinement 2

1- 1 PG4-30

z

-WG4

le -0.0015

0.0035

0.0085

0.0135

0.0185

0.0235

0.0285

Volumetric strain (V-V0)NO

Figure 7. Normalised axial stress versus volumetric strain for WG4 and PG4-30 specimens

r--

-0.0025

2.5

I

7

0.0005

0.0035

0.0065

0.0095

Volumetric strain (V-VO)NO

Figure 8. Normalised axial stress versus volumetric strain for WCl and PC1-20 specimens

Discussion

Even though all specimens had the same strength in the confinement jacket, they behaved differently and achieved different strengths at different lateral and axial strains. The pre-tensioning has led in both cases to higher strengths, and full utilisation of the confinement material. Though it appears that the GFRP confined specimens developed higher stresses, this may be partly due to the fact that the GFRP jacket is carrying some of the axial load. What is of interest in these two pairs is that the ultimate axial strain, E,, appears to be significantly higher in the CFRP confined specimens. The authors attribute this partly to the fact that the GFRP jackets slipped during

zyxwvu Lateral Prestressing of RC Columns with FRP Jackets 651

testing. As a result the true strain on the concrete core is higher than recorded on the jacket. In this particular case, for PG4-30 at failure, the axial concrete core strains recorded were 15000 ye. CONCLUSIONS

The results from the four specimens confined with Glass and Carbon of equal strength show different behaviour and strength enhancement. Pretensioning of the jacket led to higher strengths and a significant delay in the initiation of the fracturing process of the concrete through cracking. This is expected to lead to a better behaviour both at service loads and under cyclic loading, such as experienced during earthquakes. ACKNOWLEDGEMENTS

The authors acknowledge the financial support of the Ministry of Energy of the Islamic Republic of Iran, the financial support of the EU for TMR Network ConFibreCrete and the Marie Curie Fellowship Grant HPMF-CT-2001-01279. REFERENCES

1. Mortazavi, A.A. Pilakoutas, K. and Son, K.S. “RC column strengthening by lateral pre-tensioning of FRP”, Journal of Construction and Building Materials (approved for publication 2003). 2. Pilakoutas, K. and Mortazavi, A.A. “Laterally Pre-stressed Concrete with Composites”, Proceedings of the Fifth International Conference on Fibre-reinforced Plastics for Reinforced Concrete Structures, Cambridge, UK, 200 1, Vol. 2, pp. 855-864. 3. Mortazavi, A.A. and Pilakoutas, K. “Pre-tensioning of Composites by Lateral Pressure”, Proceedings of the International Conference on FRP Composites in Civil Engineering, Hong Kong, December 12-15, 200 1, Vol. 1, pp. 345-354. 4. EC8, Eurocode 8 - Design provisions for earthquake resistance of structures (Drafl), ENV 1998-1- 1, 1996. 5. Saadatmanesh, H. Ehsani, M.R. and Jin, L. “Seismic strengthening of circular bridge pier models with fibre composite”, ACI Structural Journal, 93(6), 1996, pp. 639-647.

z zyxwvuts

zyxwvuts

652 FRPRCS-6: Externally Bonded Reinforcement for Confinement

z zyx

6. Priestley, M. J. N. and Seible, F. “Design and seismic retrofit measures for concrete and masonry structures”, Construction and Building Materials, 9(6), 1995, pp. 365-377. 7. Harries, K.A. and Kharel, G. “Behaviour of modelling of concrete subject to variable confining pressure”, ACI Material Journal, 99(2), 2002, pp. 180-189. 8. Betonamit, The non-explosive cracking agent for universal application, Kriscut Plant Hire & Sales co. Ltd, UK, 1998. 9. Pilakoutas, K and Mortazavi, A.A. “Ductility through external lateral confinement of RC members with FRP”, Non metallic (FRP) reinforcement for Concrete Structures, Proceeding of the Third International Symposium, Sapporo, Japan, October 14-16, 1997, Vol. 1, pp. 225-232. 10. Mortazavi, A.A. “Behaviour of concrete confined with lateral pretensioned FRP PhD Thesis (expected Dec. 2002), The University of Sheffield, UK. ’I,

zyxwvu

FRPRCS-6, Singapore, 8-10 July 2003 Edited by Gang Hwee Tan @WorldScientific Publishing Company

zyx

CONFINEMENT OF RC RECTANGULAR COLUMNS USING GFRP

A. PROTA, G. MANFRED1 AND E. COSENZA Department of Structural Analysis and Design, University of Naples “Federico I/” via Claudio 21, 80125 Naples, Itah The confinement of Reinforced Concrete (RC) columns represents one of the most promising applications of Fiber Reinforced Polymer (FRP) to civil structures. The reasons for such upgrade could be generally related to durability issues, design or construction mistakes, increase of live loads or need for seismic retrofit. In all these cases, FRP laminates provide an effective and competitive strengthening technique and offers advantages such as easy and fast application, high durability, low impact on the use of the structure, negligible increase of structural mass as well as of member dimensions. The majority of studies have been performed on the confinement with FRP of circular columns, where the contribution of composites is fully exploited. A loss of effectiveness occurs in the case of square cross-sections where the presence of the corners reduces the confining action of the FRP jacket. Such problem becomes particularly critical for rectangular columns; despite that, very few studies have been conducted on them. The present paper deals with rectangular columns with high ratio between the sides of the cross-section. The experimental program concerning members subjected to axial load is herein presented and the upgrade technique using Glass FRP (GFRP) laminates is described. The effectiveness of such confining system is investigated also with respect to different fiber orientations (unidirectional, bidirectional and quadriaxial). Some preliminary experimental results are discussed in terms of column strength, failure mode and strains of the FRP jacket.

INTRODUCTION Within the applications of composites in construction, the confinement of RC columns is one of the most common. For both building columns and bridge piers, strengthening using FRP ensures an easy and fast installation, strength and/or ductility increase, high durability, low impact on the use of the structure, and almost no increase of mass and geometrical dimensions of the cross-sections. The confining action of FRP jackets gives the best performance on circular columns, whose geometrical configuration allows the fibers to be

zyxw

654 FRPRCS-6: Externally Bonded Reinforcementfor Confinement

effective on the entire cross-section. A different behavior characterizes square and rectangular columns; in these cases, due to the presence of the corners a part of the cross-section remains unconfined. Similar to the confinement with steel hoops, that loss of effectiveness is modeled with parabolic areas defined by the corners and eventually by longitudinal steel rebard. This still represents an unresolved issue even in terms of code provisions',2. Experimental studies have been carried out on the confinement of square or rectangular columns using FRP. In order to assess the effectiveness of the system, columns subjected to axial load have been examined. Square columns strengthened with Carbon FRP (CFRP) have been tested and the effect of f45-degree laminates has been investigated3. GFRP laminates have been adopted within other experiments and the square members strengthened using both unidirectional and 0/45-degree laminates4. An experimental program has been lately developed with respect to rectangular columns strengthened by unidirectional GFRP composites6. RESEARCH OBJECTIVES The goal of this research is to investigate the effectiveness of FRP confinement for rectangular cross-sections and assess the influence of different fiber textures on both global behavior and failure mode of such columns. GFRP laminates are used to strengthen the elements and the influence of diffetent fiber textures is evaluated. This represents the innovative aspect of the presented research program, since tests performed using GFRP jackets to confine rectangular columns6 were based only on unidirectional sheets. In addition to them, this research aims at evaluating also the effectiveness of bidirectional and quadriaxial textures. The experimental results are expected to provide important insights about strength, deformability and failure mode of rectangular columns confined using GFRP. Moreover, the analysis of monitored laminate strains will allow assessing the effectiveness of its confining action on crosssections with high long/short side ratio. These outcomes will be used to check the assumptions of the numerical model proposed for strength prediction of FRP-confined cross-sections5 and to further improve that with respect to rectangular columns.

zyxwz

Confinement of RC Rectangular Columns using GFRP 6.55

EXPERIMENTAL PROGRAM The specimens had rectangular cross-section with sides of 420 mm and 115 mm, respectively, and were 1.5 m high as shown in Figure 1. The forms were shaped so that the rectangular cross-section had a rounded corner with radius equal to 20 mm. This allowed reduction in time needed to prepare each specimen prior to FRP installation. In order to ensure a proper application of the axial load to such 1/3 scale columns, two bulbs with dimensions 700 mm x 350 mm x 250 mm were placed at top and bottom of each specimen. The longitudinal steel reinforcement was realized by 8 bars, half having diameter equal to 14 mm and the remaining with diameter of 12 mm (Figure l), yielding a percentage equal to 2.2% of the gross cross-section. Two stirrups having diameter of 6 mm were used and spaced at 100 mm on center along the height. A specific reinforcement layout was selected for the enlargements. The concrete cover was equal to 25 mm. Concerning material properties, for all columns concrete had a compressive strength, fc, equal to 12 MPa and steel was characterized by a yield strength, fy,equal to 420 MPa.

zy

Front view

Side view

f

I

I1

700

Section A-A

I

zy

441175 I I J 1175

I

700

I

*

Figure 1 . Specimen geometry and steel reinforcement layout (dimensions in mm)

The first part of the experimental program (herein presented) concerned the strengthening of columns using GFRP unidirectional laminates, which was accomplished by following the steps indicated in Figure 2. First, two plies with fibers parallel to column axis were installed on each short side

zyxwv zyxwv

656 FRPRCS-6: Externally Bonded Reinforcement f o r Confinement

and extended for 75 mm on each long side (Step 1 in Figure 2). Then, the other three steps followed: they aimed at obtaining a 3-ply jacket with fibers perpendicular to the member axis and avoid overlapping of laminate at the same column height. Each ply had a fiber density equal to 900 g/m’ and was characterized by modulus of elasticity, Et equal to 73,000 MPa and ultimate tensile strength, P,,, equal to 3,400 MPa. The second part of the experimental program, which is currently in progress, deals with strengthening the specimens by using bidirectional and quadriaxial laminates. Apart from the type of fiber, the upgrade has been performed by the same steps described above and summarized in Figure 2. In total, twelve columns will be tested. The repetition of experimental results for each type of column will be checked on three equal specimens. Four types of column will be analyzed: bare, strengthened with unidirectional, bidirectional and quadriaxial GFRP laminates, respectively. STEP 2

STEP 1

w

STEP 3

STEP 4

- /

zyxwvutsrq

2 plies

First ply

Second ply

Figure 2. Upgrade scheme

TEST SETUP AND INSTRUMENTATION

The adopted set-up can be observed in Figure 3. A universal machine with a maximum capacity of 5000 kN was used and tests were load-controlled. The axial load was recorded by a cell disposed between the plate of the machine and a very stiff plate placed on the top bulb in order to distribute the axial force equally on the entire cross-section. During tests both displacements and strains were recorded. Three linear variable displacement transducers (LVDT) were mounted on each

zyxw zyx

Confinement of RC Rectangular Columns using GFRP 657

specimen. Two of them measured the longitudinal relative displacement between bottom and top cross sections of the column on each long side (1 and 2 in Figure 3); such values allowed the computation of the average longitudinal deformation as well as checking the symmetry on the two sides. The third LVDT recorded the horizontal displacement of the mid-height cross-section in the direction parallel to its short side (Figure 3) in order to assess whether bending due to buckling occurred. Strain gages were installed on the GFRP jacket according to the layout as depicted in Figure 4. Twelve of them were placed at mid-height: two on the short sides (5 and 6 in Figure 4) and five on each long side (12 to 16 on side 1 and 7 to 11 on side 2 in Figure 4). Out of the remaining eight, four strain gages were installed at 375 mm (1 to 4 in Figure 4) and four at 1120 mm (17 to 20 in Figure 4) from the top of the column, respectively.

zyx

side 2 Figure 3 . Test setup and LVDT positions

side 1

PRELIMINARY EXPERIMENTAL RESULTS As mentioned, the experimental program is still in progress. The results herein presented concern two bare columns (i.e., B-1 and B-2) and three specimens strengthened using unidirectional GFRP laminates (i.e., S- 1, S-2

zyxwvutsr zyx

658 FRPRCS-6: Externally Bonded Reinforcement for Confinement

and S-3). They are discussed in the following sections in terms of strength, stiffness, axial deformation, failure mode, and jacket strains. Strength, Stgfness and Axial Deformation The failure of the virgin specimens occurred under a very similar ultimate axial force, as reported in Table 1. The GFRP strengthening provided a significant strength increase; a comparison with the average ultimate capacity of the bare columns (i.e., 1068 kN) underlines that such gain ranges between 26.2% (i.e., specimens S-3) and 30.9% (i.e., specimen S-1).

1

:: I

J

zyxwvutsrq

zyxwv zyxw I

L

-

-

Side? -

Side 1

Side 2

Figure 4.Arrangement of strain gages on the FRP jacket

Such behavior is also depicted in Figure 5 where the axial load is plotted versus the axial strain of the column (on the height of 1500 mm). Due to data acquisition problems, the LVDT values were not fully recorded for specimen B-1; this is the reason for not including its ultimate strain

zyxw zyxw zyxwvu Confinement of RC Rectangular Columns using GFRP 659

value in Table 1 as well as its curve in Figure 5. Along with the discussed strength increase, the curves of Figure 5 allow the stiffness of strengthened members due to the presence of the GFRP jacket to be evaluated. Based on the initial slope of the curves (up to 300 kN), the average stiffening has increased by about 55%.

Table 1 . Ultimate performance of tested columns Column B- I B-2 s-I 92 s-3

Ult. Axial Load (W) 1070.33 1065.45 1398.1 I 1388.32 1348.03

Avg. Strength Incr.

Ult. Axial Strain

(%!

(mdmm)

+ 26.2 Yo

0.00417 0.00618 0.00517 0.00522

----------------+ 30.9 Yo + 30.0 Yo

---------

Even though they are stiffer as compared to virgin columns, the strengthened members show an important improvement in terms of ultimate axial deformation. Such parameter was calculated by dividing the axial displacement corresponding to the ultimate axial load over the column height (see Table 1 and Figure 5). Recorded values show increases of the ultimate strain ranging between 24% (i.e. S-2) and 48% (i.e. S-1). These percentages are important to underline the beneficial effect of the GFRP jacket on the ultimate axial strain rather than to quantify the entity of such effect. In fact, it is important to recognize that, being the tests forcecontrolled, it is not possible to exactly identify the point corresponding to the ultimate axial strain (i.e., quasi-horizontal branch around the peak load). Failure Modes

The failure mode of both virgin specimens (i.e., B-1 and B-2) was due to concrete crushing occurring at mid-height of the columns (Figure 6-a). This is also consistent with the ultimate axial strain which is slightly higher than 0.004. The GFRP upgrade moved the failure of strengthened specimens from concrete to the composite jacket, as depicted in Figure 6-b. Since this crisis involved fiber breakage, the failure of strengthened columns was brittle.

zyxw

zyxw zyx

660 FRPRCS-6: Externally Bonded Reinforcement for Confinement IS00

,

0

0,WI

zyxwvutsrq zyxwvutsr 0.002

O

0.W3

W

0.WS

0.W

0.W7

e

Figure 5. Axial load vs. axial average strain for bare and strengthened columns

(a> (b) Figure 6. Failure mode of control (a) and strengthened (b) columns

Laminate strains Interesting information were provided by strain gage measurements. While data processing is still in progress, some preliminary results are herein presented with respect to strain gages located at mid-height of strengthened columns (Figure 4). Figure 7 shows strain profiles on the two sides of the rectangular cross-section for load levels of 300 kN and 600 kN, representing about 20% and 40% of the ultimate axial force.

zyxw zyx

Confinement of RC Rectangular Columns using GFRP 661

For the lower load level, the strain diagrams of columns S- 1, S-2 and S3 are very similar; strain values are not very different and range between 0.005% and 0.01%. For load equal to 600 kN, the confining action of the fibers becomes more significant with strain values up to 0.025%. The trends depicted in Figure 7 appear to be consistent with the formation of parabolic areas that are theoretically expected and are due to the presence of section corners and longitudinal steel rebars. This aspect will be further investigated in next steps of the research toward a comprehensive assessment of the effectiveness of FRP jackets on rectangular cross-sections.

ier

: : 1I .

s-3

1

Figure 7. Laminate strains at mid-height at 20% and 40% of ultimate axial load

CONCLUSIONS Preliminary results of an experimental program confirmed that the confinement with GFRP laminates could represent an effective technique for the strengthening of RC rectangular columns. Significant increase in both strength and ultimate axial strain was achieved by using unidirectional laminates. Tests now in progress will allow assessing the effectiveness of GFRP also with respect to bidirectional and quadriaxial fiber textures. The development of the experimental campaign will also provide important information about the effectiveness of the confinement of rectangular cross-sections with composites. A preliminary assessment of the strain distribution along the sides underlined very similar strains of the

662 FRPRCS-6: Externally Bonded Reinforcementfor Confinement

zyxw

jacket for low load levels (i.e., about 20%). As the load increases, peaks in the strain trends were observed; this could be due to the formation of parabolic areas which is theoretically expected. Further analysis will be performed on this aspect in the next steps of the research; the authors believe that the outcomes will be a useful reference for the modeling of rectangular cross-sections confined with FRP. ACKNOWLEDGMENTS

The authors wish to acknowledge MAPEI S.p.a., Milano, Italy, for supporting both construction and strengthening of the specimens. Thanks is extended to Messrs. Balsam0 and Zaffaroni for their contribution. REFERENCES 1. ACI Committee 440, Guide for the Design and Construction of Externally Bonded FRP Systems for Strengthening Concrete Structures, American Concrete Institute, Farmington Hills, MI, 2001. 2. fib Task Group 9.3, Externally Bonded FRP Reinforcement for RC Structures, International Federation for Structural Concrete, printed by Sprint-Digital-Druck, Stuttgart, 200 1. 3. Parretti, R. and Nanni, A., “Axial Testing of Concrete Columns Confined with Carbon FRP: Effect of Fiber Orientation”, Proceedings CD-ROM of the Third International Conference on Composites in Infrastructure, San Francisco, California, US, 10-12 June, 2002, Paper N. 8. 4. Pessiki, S., Harries, K.A., Kestner, J.T., Sause, R. and Rides, J.M., “Axial Behavior of Reinforced Concrete Columns Confined with FRP Jackets”, ASCE Journal of Composites for Construction, 5(4), 2001, pp. 237-245. 5 . Realfonzo, R., Prota, A., Manfredi, G. and Pecce, M., “Flexural Strength of FRP-Confined RC Columns”, Proceedings CD-ROM of the first j b Congress “Concrete Structures in the 21st century”, Osaka, Japan, 13-19 October, 2002, Disk B, pp. 41-50. 6. Tan, K.H., “Strength Enhancement of Rectangular Reinforced Concrete Columns using Fiber-Reinforced Polymer”, ASCE Journal of Compositesfor Construction, 6(3), 2002, pp. 175- 183.

zyxwv

zyxwvutsrq

FRPRCS-6, Singapore, 8-10 July 2003 Edited by Gang Hwee Tan QWorld Scientific Publishing Company

zyxw

BEHAVIOUR OF RC COLUMNS RETROFITTED BY FIBRE REINFORCED POLYMERS UNDER CYCLIC LOADS H. SHAHEEN AND T. RAKIB

Housing and Buifding Research Center, 87 El-Tahrir st. Giza, Egypt Y. HASHEM AND I. SHAABAN

Faculty of Engineering, Zagazig Univ., 108 Shobra St., Cairo, Egypt A. ABDELRAHMAN Structural Eng. Dept., Ain Shams Univ., Cairo, Egypt Retrofitting of reinforced concrete (RC) elements with FRP wraps is one of the techniques used successfully for the last few years. RC rectangular columns are used extensively in both residential and commercial buildings. Consequently, a comprehensive study of the different parameters affecting the seismic behavior of RC columns strengthened with FRP is vitally needed. The objectives of this research are to study the behavior of rectangular RC columns wrapped with FRP sheets under cyclic lateral and axial loading. The main parameters investigated in the research are: different anchorage systems, volumetric ratio of FRP and spacing between FRP layers. Five columns with dimensions 150x 450x2300 mm were tested under both cyclic lateral and axial loading. Different recommendations are provided for the use of FRP in strengthening rectangular columns.

INTRODUCTION

During the last three decades, considerable changes in seismic design codes were introduced. In Egypt, many existing RC structures do not comply with any of the recent seismic code provisions. Deficiencies, often found in typical moment-resisting frames, are inadequate shear strength, flexural strength and ductility of columns. During earthquake loading and at high level of compressive or shear stresses, sudden collapse may occur and there may not be enough warning signs. As a result, retrofitting of RC columns is needed for buildings located in seismic regions. Recently, attention has been focused on the use of FRP materials for structural rehabilitation. If correctly used, FRP can result in significant enhancement to both ductility and strength of RC members. Previous work was undertaken on strengthening circular and square columns'. Rectangular

zyxw

664 FRPRCS-6: Externally Bonded Reinforcement for Confinement

columns are commonly used for residential and commercial buildings. Strengthening of rectangular columns subjected to axial loading was also reported’. This paper investigates the performance of rectangular RC columns wrapped by FRP sheets and subjected to a combined axial compression and cyclic flexural loading. Rectangular RC columns of aspect ratios of 1 to 3 were tested under constant axial load and increasing cyclic lateral load up to failure. Carbon FRP (CFRP) wraps were used in different volumetric ratio, spacing and with or without anchors.

zyx

TEST PROGRAM The test program includes eleven specimens of which five of them have been tested. The columns have a rectangular cross section of 150x450 mm and a height of 2300 mm. The columns were tested horizontally under constant axial load combined with cyclic lateral load. It consisted of a right part of 1300 mm long and a left part 700 mm long with a beam stub in the middle as shown in Fig. 1. The right part of each specimen constituted the test portion. It represents a column extending from the beam-column connection to the point of inflection. The beam stub provided a point of application for the lateral load. The dimensions of the beam stub were chosen so that the failure occurs in the column rather than at the joint. The left portion was heavily reinforced and provided with two 6 mm thickness steel plates in order to force hinging into the right part. The longitudinal reinforcement ratio of the columns was 1%. Stirrups were 6 mm diameter bars spaced at 150 mm with a volumetric ratio of 0.3%. Extra stirrups were placed at the ends of each specimen where the axial load was applied to prevent crushing of concrete. CFRP laminates were applied to strengthen the columns C2 through C5 with different schemes. The thickness of CFRP laminates was 0.1 1 mm, while its tensile strength and modulus were 2400 MPa and 240 GPa, respectively. The characteristic compressive cube strength of the concrete was 25 MPa while the yield stress of the steel was 420 MPa for the longitudinal reinforcement and 3 10 MPa for stirrups.

Test Specimens The overall test program consists of eleven specimens. This paper represents the results of only five columns. The properties of the tested specimens are given in Fig. 1 and summarized as follows:

zy

RC Columns Retrofitted by FRP under Cyclic Loads 665

a. Specimen C1 is the reference specimen. A constant axial compression load of 0.15fc,Ac was applied on the specimen up to failure. This axial compression load was kept constant for all specimens. b. Specimen C2 was wrapped in the transverse direction with two layers of lateral CFRP laminates of 100 mm width .The clear distance between laminates was100 mm and the volumetric ratio of FRP was 0.2%. C. Specimen C3 was wrapped with one CFRP layer with a ratio of 0.1%. The width and spacing between the laminates were the same as C2. d. Specimen C4 was wrapped with one layer of CFRP with no spacing between the layers. The volumetric ratio of FRP was 0.2%. e. Specimen C5 was wrapped with one layer of CFRP as for specimen C3. Steel plates of 50mm width, 190mm length, 12mm thickness and 240 MPa yield stress, were used to anchor the lateral laminates. The plates were discontinuous, this is to investigate the steel plates contribution to concrete confinement and increasing the ductility of columns, independently from their contribution to the flexural capacity of columns. Two linear variable differential transducers (LVDTs) were mounted on the concrete surface at the critical section adjacent to the beam stub to measure the concrete strain and average section curvature in the plastic hinge region. The lateral displacement of the columns was also measured using LVDT’s. Electrical strain gages were attached on the longitudinal and transverse reinforcement of each specimen. As well as on the CFRP sheets. Test Setup and Procedure

Two independent reaction frames were used in the testing setup, as shown in Fig.2. The first frame was a 2000 kN capacity, large-scale testing double portal frame, while the second frame was a 3000 kN capacity, closed, horizontal, reaction frame. The closed horizontal frame was located under the cross girder of the double portal frame such that the centerline of the closed frame was oriented parallel to the line of support of the cross girder. The lateral reversed cyclic displacement was applied at the stub of the beam-column joint using a double acting hydraulic cylinder of 600 kN compression capacity and tension capacity of 400 kN. The cylinder was equipped with a tensionlcompression load cell of +/- 680 kN capacity to measure lateral load. The axial compression load was applied by a manual hydraulic cylinder of 900 kN capacity. The specimens were supported on two concrete blocks, spaced 2.10 m apart. Each block was equipped with a

zyxwv

666 FRPRCS-6: Externally Bonded Reinforcement for Confinement

zyxwvu

hinged support at its top. The upward reaction was transmitted to the cross girder of the frame through two 500 kN hydraulic cylinders, each cylinder was equipped with a threaded adjustment and an end ball bearing. At the beginning of each test, the required axial load was applied and kept constant throughout the test. The lateral load was applied in stroke control as shown in Fig. 3. two layers of CFRP for C2 one layer of CFRP for C3

2 steel plates

200

450

1 5 0

200 700

1300

11x100

700

Steel plates

One layer of CFRP

700

11x100

7

zy

one layer of =Rp

11x100

700

Fig.1 Details of test Speciments

Fig. 2Test Test set-up Set-up Fig.2

Fig. 3 Load Load History History Fig.3

RC Columns Retrofitted by FRP under Cyclic Loads 667

OBSERVED BEHAVIOR The lateral load-displacement hysteresis loops of the control specimen are shown in Fig.4. The ultimate lateral load was 237.1 kN. The loaddisplacement relationship was linear until the ultimate load was achieved. Progressive drop in the strength occurred at a lateral displacement of 4 mm. Failure load was reached at a lateral displacement of 6.75 mm. A major diagonal tension crack appeared at a lateral load of 114.0 kN and extended up to failure as shown in Fig. 5. The specimen failed in a brittle shear mode. The lateral load-displacement hysteresis loops of specimens strengthed by CFRP are shown in Figs. 6 to 9. For all the strengthed specimens, a major flexural crack initially appeared at the critical section adjacent to the beam stub and extended up to failure. Gradual decrease in the lateral load occurred after the ultimate load was reached. No shear cracks were observed as the CFRP wraps prevented diagonal tension cracks even at high lateral displacement. At onset of flexure failure, crushing of concrete, buckling of the longitudinal bars and rupture of CFRP sheets at the corners of the specimens were observed. The ultimate lateral load of the wrapped specimens ranged from 1.2 to 1.46 times the strength of the control column C 1. It should be noted that crushing of concrete was observed for the entire full depth of all specimens at the critical section adjacent to the beam stub, however, for specimen C5, crushing of concrete was limited to 100 mm of the top and bottom of the cross section. The concrete core of C5 did not crush due to the confinement provided by the steel plates anchorage. Also, it was noted that the concrete crushing for all wrapped specimens was concentrated in the first lOOmm adjacent to the stub except for specimen C2, the concrete crushing occurred at the second lOOmm adjacent to the stub, i.e. at the clear spacing between CFRP laminates. Figs. 10 and 11 show typical failure of specimens retrofitted with CFRP.

Fig.4 Lateralload-displacement for ci

Fig.5 Brittle shear failure of ci

z

668 FRPRCS-6: Externally Bonded Reinforcement for Confinemenl

'

zyxwv zyxw zyx . .

" I(.

,

I

I

1

.. .

I

"

"

,

I

.

"(.

,

,,_

.,

,

,

'

'

. .*.* .., ,,~

latered 63,s .mm

Fig.6 Lateral load-displacement of C2

Fig.10 Failure of specimen C2

Fig.7 Lateral load-displacement of C3

Fig. 11 Failure of specimen C4

ANALYSIS OF TEST RESULTS The strength envelope, which is the relationship between the peak load at each cycle and the corresponding displacement, is presented for the tested columns in Fig. 12. The lateral strength increases considerably for all the wrapped specimens. The highest strength was that of specimen C2 with two CFRP wraps, 46% higher than the control column. Specimen C3, with one layer of CFRP, produced the lowest increase, which was 20% higher than the control column. This increase in the flexural strength is attributed to the confinement provided by CFRP, which resulted in an increase in the concrete strength and strain. It should also be noted that the longitudinal

RC Columns Retrofitted by FRP under Cyclic Loads 669

reinforcement bars exhibited the strain hardening; consequently the stress in the steel bars exceeded the yield stress leading to an increase in the overall strength of the retrofitted columns. Specimen C2 also survived more cycles than all the other specimens, as shown in Table 1. Column C5 with anchored CFRP wraps had ultimate load close to C4 with zero spacing between wraps, (34% and 39% higher than the control specimen Cl), despite the CFRP ratio being 50% less than for the anchored column.

Displacement Ductility Analysis

zyx zyx zyx zyx

Displacement ductility factors were evaluated and used to compare the ductility of different columns. The yield displacement Ay of an equivalent elasto-plastic system with reduced cracked stiffness was calculated from the lateral load-displacement curve as the corresponding displacement at the intersection of the secant stiffness at a load level of 75% of the ultimate lateral load and the tangent at the ultimate load. The strength envelope is used to determine the yield displacement3. The failure load was taken equal to 75% of the ultimate load on the descending branch on the strength envelope and the corresponding displacement Af was computed. The displacement ductility factor is defined as the ratio between Af, and A, as given by Equation (1).

Displacement Ductilityfactor

No.

CI c2 c3 c4 c5

= Af/Ay

(1)

Table 1 . Lateral load-displacement test Results Visible Cracking Level Ultimate Load Level P,/ P, of

P,, (kiv) 114.0 275.0 200 230 215

A,, (mm) 3.5 8.5 4.8 6.4

5.5

P,, (kiv) 23 7.I 348.5 284.8 329.8 319.1

A, (mm) 4.04 27.80 16.09 12.10 11.85

61) I 1.46 1.20 1.39 1.34

Failure cycle 4 I1 8 9 9

The control column failed in a brittle shear mode at a low ductility factor of 2.3 as shown in Table 2. All the strengthed columns failed in a ductile mode with ductility factors more than 4.5. It should be noted that the satisfactory level of ductility is achieved by a minimum of ductility factor of three4. Specimen (C2) with double layer produced the highest value of ductility factor (3 times that of Cl). This result highlights the role of increasing the volumetric ratio of CFRP in enhancing both the ductility and lateral

zyxwvut

670 FRPRCS-6: Externally Bonded Reinforcementfor Confinement

strength of the columns. It was observed that the ductility of columns C4 and C5 were similar (2.72 and 2.25 times that of the control column) despite the latter having 50% less CFRP wraps. This is attributed to the contribution of the anchorage system used for Column C5. Table 2. Ductility Analysis of Test Specimens Yield

Failure Displacement

specimenDisplacement

Ductility Factor

Energy Index

A, (mm)

A, (mm)

CI

2.85

6.75

2.36

8.9

c2

4.40

31.91

7.25

204.7

c3

3.94

18.91

4.78

66.5

c4

4.10

26.42

6.44

122.9

c5

4.85

25.80

5.31

114.7

-40

+C1

-30

-20

-10 0 10 Lateral displacement (mm)

Control C5 one sheet with steel plates c2 double wrraping

+

Af’AY

-A-

20

30

40

C3 with one sheet c4 continous wraping

Fig. 12 Hysteresis Loops Envelop of Lateral Load vs. Displacement

Energy Dissipation The dissipated energy was computed for each cycle as the area enclosed by the lateral load-displacement hysteresis loop for the given cycle. The

RC Columns Retrofitted by FRP under Cyclic Loads 671

zyxwvu zyxw zyxw z

accumulated dissipated energy is plotted against the lateral displacement for the tested specimens in Fig. 13. A non-dimensional energy index based on Ehsani' was also used to evaluate the energy dissipated for the specimens. The index accounts for the cracked stiffness, yield load and displacement, as well as the dissipated energy of each cycle. This energy index is expressed as follows: I,

=

(24( W K J *(&'Ay) Z)4Py*AJ

(2)

where Ei is the dissipated energy at cycle (i); Ki, K, are the stiffness at cycle (i) and at yield, respectively; Ai is the average of maximum compression and tension displacements at cycle (i); and Ay, P, are the yield displacement and load, respectively. Specimen C2 possessed significantly larger energy dissipation than other columns as it had the largest index of 204. Column C3 had the lowest index of 66. The energy index of C4 and C5 were 123 and 1 14. Stiffness Analysis

The cracked stiffness of each tested specimen, Kj, is calculated for every loading cycle. The cracked cycle stiffness is computed as the ratio of the sum of the peak tension and compression loads to the sum of the corresponding tension and compression displacements. The cracked cycle stiffness is plotted against the lateral displacement to represent the stiffness degradation due to cyclic loading in Fig. 14.

s

0

10 20 Meral d i s p l m (m)

Fig. 13 Accumulated energy

30

0

10 20 30 Lateral d i s p l m (m)

Fig. 14 Stiffness degradation

It can be seen that the initial stiffness of all columns including the control one was approximately the same. This suggests that the strengthened columns will not attract more horizontal force due to seismic loading. At

zyxwvu

672 FRPRCS-6: Externally Bonded Reinforcement for Confinement

high load levels, the retrofitted columns showed higher stiffness than that of the control column. The rate of stiffness deterioration of the retrofitted columns under large reversed cyclic loading was less than that of the control column.

CONCLUSION CFRP wraps showed an excellent enhancement to the overall behavior of the strengthed columns. Conclusions are summarized as follows: 1. All the wrapped specimens failed in a ductile flexural mode instead of the brittle shear mode of the original column. 2. Increasing the CFRP volumetric ratio improved the overall behavior of the column. However, it is recommended to increase the number of CFRP layers instead of reducing the spacing between the layers. 3. A proper choice of the anchorage system may be more feasible than reducing the spacing between sheets. However, savings allowed by the anchorages should be assessed considering that the installation diminishes the ease and speed of application. 4. Unlike the conventional techniques for strengthening, the initial stiffness of the retrofitted columns was similar to that of the original one.

REFERENCES

zy

1. ACI 440.2R-02, Emerging Technology Series, “Guide for the Design and Construction of Externally Bonded FRP Systems for Strengthening Concrete Structures”, October 2002,45p. 2. Hosny A., Shaheen H., Abdelrahman A., and El-Afandy T. “Strengthening of Rectangular RC Columns Using CFRP”, MESC-3, Aswan, Egypt, December 2002. 3. Park R., and Paulay T. “Reinforced Concrete Structures”, John Wiely and Sons, New York, N.Y., 1975. 4. Priestly M., and Park R. “Strength and Ductility of Bridge Columns under Seismic Loading”, ACI Structural Journal, V84, No. 1, 1986. 5. Ehsani M., and Wieght J. “Confinement Steel Requirements for Connections in Frames”, ASCE Structural Journal, V116, No.3, 1990.

FRPRCS-6, Singapore, 8-10 July 2003 Edited by Kiang Hwee Tan OWorld Scientific Publishing Company

zyxwvu

zy zyx

PHOTOGRAMMETRICALLY MEASURED DEFORMATIONS OF FRP WRAPPED LOW STRENGTH CONCRETE A. ILKI, V. KOC, B. ERGUN, M.O. ALTAN AND N. KUMBASAR Department of Civil Engineering, Istanbul Technical University Maslak, 80626, Istanbul, Turkey

In this study, 8 specimens with square cross-section were tested under compression. The specimen dimensions were 250x250~500mm. The test program included unconfined specimens as well as specimens jacketed by 1, 3 and 5 plies of carbon fiber reinforced polymer (CFRP) sheets. The unconfined concrete compressive strength was less than 10 MPa for all specimens. Photogrammetric measurements were carried out as well as usual measurement techniques to determine the axial and lateral deformations of the specimens. With the help of photogrammetric techniques, more detailed information was collected about the deformation pattern that led to better understanding of the behavior. The performance of low strength concrete members jacketed by different thicknesses of CFRP sheets are discussed with a special emphasis on the distribution of deformations of the specimens. The effectiveness of the CFRP jackets increased significantly due to deformation characteristics of unconfined concrete with low compression strength. Very significant enhancement was obtained for compressive strength and deformability of CFRP jacketed specimens that resulted with tremendous increase in energy dissipation.

INTRODUCTION

FRP reinforcement has several advantages like its durability, electromagnetic neutrality, high strength, light weight and ease in application. Consequently, the use of FRP reinforcement in civil engineering structures has been increasing rapidly in recent years. According to Fukuyama and Sugano', the repair and seismic strengthening by continuous fiber sheet wrapping method was first developed in Japan, where research was first carried out in 1979. They presented an outline of the continuous fiber wrapping technique by comparing experimental data obtained for various rehabilitation techniques with a special emphasis on performance based engineering and effective rehabilitation techniques without hindrance of building operation. Guadagnini et a1.* reported an overview on the European Research on FRPs and their applications.

zyxw

674 FRPRCS-6: Externally Bonded Reinforcement for Confinement

Karbhari and Gao3, Toutanji4, Xiao and Wu5, and Ilki and Kumbasar6 developed experimental data for cylinder specimens, based on a variety of fiber types, orientations and jacket thicknesses. Rochette and Labossiere7, Wang and Restrepo8 , and Ilki and Kumbasar’’conducted axial loading tests on FRP jacketed specimens with square and rectangular cross-section. All of the experimental work was carried out on normal or high strength FRP jacketed concrete specimens. However, there are many existing structures, those were not built considering the up-to-date codes and recommendations. Consequently, these structures may experience severe damages due to insufficient ductility and low concrete compressive strength during earthquakes. Therefore, there is a need of research on the behavior of low strength concrete jacketed by FRP. In this study, 8 specimens of low strength concrete with square crosssection were tested under concentric compression. The standard concrete cylinder compressive strength was 6.2 MPa for all specimens. The specimens were either unconfined or jacketed by 1, 3 or 5 plies of CFRP in sets of two. Experimental results showed that the efficiency of the CFRP jackets on strength and ductility enhancement of low strength concrete is higher than that of normal or high strength concrete. Consequently, equivalent ductility or strength enhancement can be obtained with relatively smaller jacket thicknesses resulting with more economical solutions. The experimental results on similar specimens that have higher concrete compressive strength can be found elsewhere’. Photogrammetric deformation measurements were carried out as well as conventional deformation and displacement measurements done by using strain gages and displacement transducers. With the help of the photogrammetric measurements, deformation characteristics of the specimens could be analysed in more detail. When compared to deformation measurements with strain gages, photogrammetric deformation measurements have further advantages like; (a) availability of all surface deformations in three dimensions, (b) comparable precision, (c) lower cost, (d) convenience of test setup installation in a short time, (e) practically no deformation limit. The deformation patterns obtained by photogrammetric measurements are generally in good agreement with the deformations determined by the conventional techniques. Consequently, the photogrammetric measurement technique seems to be promising as an alternative or additional way of deformation measurements.

zyxwvu

Deformations of FRP Wrapped Low Strength Concrete 675 zyxwvut

TEST DETAILS

Specimen Characteristics,Mmaterials,Jacketing Eight specimens were tested under compression. The cross-section of all of the specimens were square (250x250 mm). The height of the specimens was 500 mm. General characteristics of the specimens are presented in Table 1. Concrete compressive strength was aimed to be less than 10 MPa to represent many existing structures with low strength concrete. Specially produced ready mixed concrete is used to obtain homogenous distribution of concrete in all of the specimens. The mix-proportion of ready mixed concrete is given in Table 2. As seen in this table, waterkement ratio of the mixture is 1.27.

zyxwvu

Table 1. Specimen characteristics Specimen LS-S-0-1 and LS-S-0-2 LS-S-1-1 and LS-S-1-2 LS-S-3-1 and LS-S-3-2 LS-S-5-1 and LS-S-5-2

f’c (MPa) 6.2 6.2 6.2 6.2

Dimensions (mm) 250x250~500 250x250~500 250 x 250 x 500 250x250~500

CFRP plies 0 1 3 5

Table 2. Mix-proportion for low strength concrete (kg/m3) Cement 150

Water 191

Sand 932

Gravel 1074

Total 2347

After waiting for about 28 days and rounding the corners of the specimens to radius of 40 mm, the specimens were jacketed by 1, 3 and 5 plies of CFRP in sets of two, except two specimens which were tested without external confinement (Figure 1).

Figure 1 . CFRP jacketing procedure

zyxwvutsr zyx

676 FRPRCS-6: Externally Bonded Reinforcementfor Confinement

The jacketing process included surface preparation, application of primer, putty, epoxy adhesive and wrapping. For jackets of 3 and 5 plies, the CFRP sheets were wrapped continuously and an overlap of 150 mm was formed at the end of the outermost ply. The characteristics of the epoxy system and CFRP sheets used for jacketing are given in Tables 3 and 4, respectively. Table 3. Characteristics ofthe epoxy system

Compressive strength (MPa) 80

Tensile strength (MPa) 50

Tensile elasticity modulus (MPa) 3000

Table 4. CFRP characteristics

Unit weight @g/m? 1820

Effective area Tensile strength Elas. modulus Max. elongation (mm2/mm) (MPa) (GPa) (mm/mm) 0.165 3430 230 0.015

Photogrammetric Measurements Digital photogrammetric systems have been used to solve various measurement problems in industrial applications for many years, since highresolution Charge Couple Devices (CCD) cameras and powerful computer technologies have been available. In this study, close range photogrammetric applications were carried out, where deformation measurement, analysis, and camera calibration are the most important steps. Twin Industrial Basler A302fs cameras with IEEE 1394 standard were calibrated with 16 mm fix focused Cosmicar Pentax lenses on the test setup. Then, the experimental data capturing system was designed in order to capture the images during tests in which about 20 stereo images were captured periodically. The first image pairs were captured before loading and the exterior orientation parameters were calculated. For deformation analysis, camera stations were fixed all through the test duration in order to use the same orientation parameters for other image pairs. The configuration of the signal points on the specimens was designed so that deformations could be determined by the coordinate differences recorded during loading. The derived exterior orientation parameters were obtained with 0.01 mm and the rotations were obtained with 0.001 radian accuracy.

zyxwvu

Loading and conventional data acquisition system A schematical representation of loading and conventional data acquisition system are shown in Figure 2. For loading, a 5000 kN capacity Amsler

zyxw

zy

Deformations of FRP Wrapped Low Strength Concrete 677

loading machine, for displacement measurements TML displacement transducers and for deformation measurements TML strain gages were used. The considered gage lengths for displacement transducers and strain gages are 500 and 60 mm, respectively. The axial stresses were determined by dividing applied axial load to the cross-sectional area of the specimen.

n

n

250

*------.,

I

Figure 2. Test setup and locations of strain gages TESTRESULTS

Photogrammetrically determined deformations The relative displacements of each point at the junctions of vertical and horizontal lines given in Figure 3 were determined by the photogrammetric method. Consequently, it was possible to calculate the deformations between all of these points.

Figure 3. Points considered during photogrammetric measurements

zyxw zy zyxwv

678 FRPRCS-6: Externally Bonded Reinforcement for Confinement

The original distances between horizontal lines A and F, B and E, and C and D were 450, 200 and 60 mm respectively. The axial deformations measured between lines A and F on vertical lines 1, 2, 3, 4 and 5 are presented in Figure 4 for the Specimen LS-S-0-2. This shows that, axial deformations were almost identical on each vertical line until peak stress, but then damage is concentrated around certain regions. The axial stressaxial strain relationships for the Specimen LS-$3-2 are given between A and F, and B and E in Figure 5. These show that for jacketed specimens, axial deformations are identical on all vertical lines, both between A and F, and B and E. 10

1

zyxwvu

0,0000 0.0005 0,0010 0.0015 0.0020 0.0025 0.0030 Deformation

Figure 4. Axial strains determined by photogrammetric measurements, between A and F, 450 mm, (LS-S-0-2)

0.00

0.04

0.08

Deformation

0.12

0.16

0.00

0.04

0.08

0.12

0.16

Deformation

Figure 5. Axial strains determined by photogrammetric measurements, between A F (450 mm) and B - E (200 mm), (LS-S-3-2)

The average axial stress-axial strain relationships for Specimen LS-S-3-2 considering the axial deformations between A-F, B-E and C-D are given in different deformation scales in Figure 6. In this figure, it can be seen that, the damage is more concentrated in mid 200 mm height. It should be noted in Figure 6 that, since the signal points were lost due to excessive damage, no photogrammetric data could be obtained behind point X1 in mid 60 mm and behind point X2 in mid 200 rnm.

Deformations of FRP Wrapped Low Strength Concrete 679

10

zyxwv zyx zy zyx 4 c +Between

A - F, 450 mm Between 6 - E, 200 mm Between C - D,60 mm

t-

0.04

0.00

0.08

0.12

0.16

0.00

Deformation

0.01 0.02 0.03 0.04 0.05 Deformation

0.06

Figure 6. Average axial strains determined by photogrammetric measurements (LS-s-3-2)

The stress-strain relationships obtained for mid 450, 200 and 60 mm for the Specimen LS-S-5-2 and the appearance of the damaged specimen are presented in Figure 7. The stress-strain curves indicate that, the photogrammetrically determined average axial deformations are highest in mid 60 mm height and lowest in mid 450 mm height, which is also in consensus with the damage pattern of the specimen. However, it should be noted that all three curves are quite close to each other until failure.

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

Deformation

Figure 7. Average axial strains determined by photogrammetric measurements (LS-s-5-2)

The stress-strain relationships obtained photogrammetrically for the Specimens LS-S-0-2, LS-S-1-2, LS-S-3-2 and LS-S-5-2 are presented in Figure 8. As seen in this figure, for CFRP jacketed low strength concrete significant strength and ductility enhancement can be obtained by increasing the jacket thickness.

680 FRPRCS-6: Externally Bonded Reinforcement for Confinement

zyx

zyxwvut 0.00

0.12

0.08

0.04

0.18

Dsfonnation

Figure 8. Stress-strain relationships determined by photogrammetric deformation measurements (LS-S-5-2, LS-S-3-2, LS-S- 1-2, LS-S-0-2)

Comparison of photogrammetric and conventional measurements To compare of photogrammetric and conventional deformation measurement techniques, the stress-strain relationships determined for Specimens LS-S-5-2 and LS-S-3-2 are presented in Figure 9.

zy zyxwvutsr I , -Conventional.

.,.

+-Photogrammetric. Eelween A - F. 450 mm

.

0.M

0.08

Deformation

0.12

0.18

I I GL-500 mm

-Photoprammetrk.

-Phopnmmalric. Between B . E.200 mm . . . : . . . : . . . : . . , 0.00

,

0.00

0.02

0.04

0.06

Betwaan E . 0.08

I

E.200 mm

0.10

0.12

0.14

Deformstion

Figure 9. Comparison of photogrammetric and conventional measurements

In these figures, conventional deformation measurements were obtained by utilizing the average measurements of four displacement transducers in the gage length of 500 mm, which is the entire height of the specimen. As expected, the highest deformations are obtained for conventional measurement of gage length of 500 mm, which include the deformations of capping mortar. However, it is clear that all three curves represent the behavior realistically. For the comparison of the strain gage measurements in the gage length of 60 mm at mid height of the specimens with the photogrammetric measurements, Figure 10 is presented. As seen, theje measurements are almost the same, with the exception that, strain gages lost their reading capabilities after the axial deformation level of 0.009-0.01 1.

zyx z

zyxw Deformations of FRP Wrapped Low Strength Concrete 681

In Figure 1 1, comparison of transverse deformations is given for specimen LS-S-5-2 between vertical lines 2 and 4 at mid-height and horizontal line A. 25

-

20

5

15

p

10

1

zyxwvutsrqp 25

I

3

5 0

0000

0005

0010

0015

0020

0025

OOM)

0010

Deformation

0020

0030 Deformation

0040

0050

Figure 10. Comparison of axial deformations (LS-S-3-2 and LS-S-5-2)

-

. . . . . . . . . . . .~ ~ .~ . ~ ~ ~ ~ ~ ~ ~

-Strain

gagcmid-height

Photogrammeldomid-height

+PhotcgrammBtriotoprection 0.000

0.005

0.010

0.015

0.020

Transverse Deformation

Figure 11. Comparison of transverse deformations (LS-S-3-2)

CONCLUSIONS

Photogrammetric methods can be used for deformation measurements. This may help better understanding of the behavior of the specimens by providing extensive data on all surface deformations. The photogrammetric measurement technique is easy to apply and comparably precise with the conventional measurement techniques. By using photogrammetric measurements, deformations, which are much higher than readability limits of strain gages can be obtained. Significant increase in compressive strength and ductility is obtained when low strength concrete specimens with square cross-section are jacketed by CFRP sheets. An increase in the jacket thickness results with an increase in strength and ductility. The enhancement in strength and ductility is more pronounced for jacketed low strength concrete, with respect to normal or high strength concrete. It should be noted that validity of the test results is limited in the ranges of variables considered in this study.

682 FRPRCS-6: Externally Bonded Reinforcementfor Confinement

zyx

ACKNOWLEDGMENTS

The financial support of Turkish Earthquake Foundation (Project : 0 I-AP1 IS), Research Fund of Istanbul Technical University (Project : 1607), and Yapkim Construction Chemicals Company, and the assistance of our students Mr. C. Demir and A. Karadeniz are acknowledged. REFERENCES 1. Fukuyama, H. and Sugano, S., “Japanese Seismic Rehabilitation of

2.

3. 4.

5.

6. 7.

8. 9.

Concrete Buildings after the Hyogoken-Nanbu Earthquake”, Cement and Concrete Composites, 22,2000, pp. 59-79. Guadagnini, M., Pilakoutas, K. and Waldron, P., “An Overview of the European research on FRPs and their applications”, Znt. Con$ on FRP Cornp. in Civ. Eng., Hong Kong, Dec. 12-15, 2001, Vol. 2, pp. 16991706. Karbhari, V.M. and Gao, Y., “Composite jacketed concrete under uniaxial compression-verification of simple design equations”, ASCE Journal of Materials in Civil Engineering, 9(4), 1997, pp. 185-193. Toutanji, H.A., “Stress-strain characteristics of concrete columns externally confined with advanced fiber composite sheets”, ACI Materials Journal, 96(3), 1999, pp. 397-404. Xiao, Y., and Wu, H., “Compressive behavior of concrete confined by carbon fiber composite jackets”, ASCE Journal of Materials in Civil Engineering, 12(2), 2000, pp. 139-146. Ilki, A. and Kumbasar, N., “Behavior of damaged and undamaged concrete strengthened by carbon fiber composite sheets ”, Structural Engineering and Mechanics, 13(I), 2002, pp. 75-90. Rochette, P. and Labossiere, P., “Axial testing of rectangular column models confined with composites”, ASCE Journal of Composites for Construction, 4(3), 2000, pp. 129-136. Wang, Y .C. and Restrepo, J.I., “Investigation of concentrically loaded reinforced concrete columns confined with glass fiber-reinforced polymer jackets”, ACZStructural Journal, 98(3), 2001, pp. 377-385. Ilki, A. and Kumbasar, N., “Compressive behavior of carbon fiber composite jacketed concrete with circular and non-circular crosssections”, accepted to be published in Journal of Earthquake Engineering.

zy

FRP Structural Shapes

This page intentionally left blank

zyxwvu

FRPRCS-6, Singapore, 8-10 July 2003 Edited by Gang Hwee Tan OWorld Scientific Publishing Company

RECTANGULAR FRP TUBES FILLED WITH CONCRETE FOR BEAM AND COLUMN APPLICATIONS

zyxwvu

A. Z. FAM Department of Civil Engineering, Queen’s Universiw, Kingston, Ontario, k7L 3N6, Canada D. A. SCHNERCH AND S. H. RIZKALLA Civil Engineering Department, North Carolina State Universily, Raleigh, NC, 27695-7533, U.S.A.

This paper introduces an innovative concept of FRPIconcrete hybrid structural member. This concept includes rectangular filament-wound glass fiber reinforced polymer (GFRP) tubes, totally or partially filled with concrete, and used as beams andlor columns. The paper presents the experimental program and results of three beams and five short columns tested with different eccentricities. Two of the beams were completely filled with concrete, while the third beam was partially filled to minimize the self-weight of the beam. This beam had a void within the cross-section, eccentric towards the tension side of the beam such that the remaining concrete was used to resist the internal compression and shear forces. Two of the columns were subject to zero eccentricity and the other three were subjected to various levels of eccentricity to study the combined effect of axial and flexural loading.

INTRODUCTION The application of fiber reinforced polymers (FRP) in new concrete structures started with replacement of steel bars with FRP bars. Although this direct replacement philosophy was suitable at early stages, it does not necessarily utilize the full potential of FRP materials. It is, therefore, believed that FRP could be combined with concrete through more efficient structural concepts. The proposed system in this paper consists of concrete-filled rectangular filament-wound glass-FRP tube with several layers including fibers oriented at f 45 and 90 degrees for shear resistance. The upper and lower flanges of the tube include additional uniaxial roving for flexural rigidity. The tube, which could be totally or partially filled with concrete, acts as lightweight permanent formwork and reinforcement, simultaneously. The concrete provides stability for

zyx

zyxwvu zyxwvu

686 FRPRCS-6: FRP Structural Shapes

the tube and compressive resistance. Research has been conducted on concrete-filled circular FRP tubes', however, no research has been reported on optimized concrete-filled rectangular filament wound tubes. Triantafillou and Meier * have studied hybrid rectangular sections with GFRP tubes supporting concrete flange above the section and a layer of carbon-FRP attached to the lower GFRP flange, however, premature failure occurred due to debonding of concrete.

EXPERIMENTAL PROGRAM The objective of the experimental program is to determine the interaction of axial load and flexure behavior for rectangular GFRP concrete-filled tubes as well as to optimize the section by providing a central hole to reduce the self-weight of the beam. In this case, the concrete was cast with a void offset towards the tension side of the shell such that the concrete is optimally used for compression, shear and stability of the webs.

Composite Tubes The GFRP composite shells used in this study were fabricated using a combined filament-winding and hand lay-up technique, where bidirectional glass fiber sheets were inserted into the top and bottom flanges, resulting in two longitudinal (zero degree) layers in both the tension and compression flanges of the rectangular shell. The remaining, non-zero degree laminate, were produced through conventional filament winding techniques. The final stacking sequences for the webs and flanges were [90, 45, -45, 90,45, -45, 901 and [90,45, -45,0, 90, 0,45, 45, 901 respectively. E-glass fibers used for the filament winding process have tensile strength between 1380 and 2070 MPa, and modulus of elasticity of 72.5 GPa. The E-glass fiber sheets used in the hand lay up process have a tensile strength of 798 N/mm in the warp direction and 183 N/mm in the weft direction.

Rectangular FRP Tubes Filled with Concrete 687

Hybrid Beam Specimens The GFRP concrete-filled tubes used for the flexural tests were 2200 mm in length. No reinforcement was provided other than the outer GFRP shell was provided. Two different cross-section sizes were used. The smaller tube, with a height of 271 mm and a larger tube with a height of 374 mm as shown in Figure l(a and b). The flange was thicker due to the addition of two layers of glass fiber sheets oriented in the longitudinal direction of the specimen. A second configuration was produced using the larger sized GFRP tube. In this case a void, offset towards the tension side of the member, was generated to minimize the self-weight of the member and to optimize the use of the concrete as shown in Figure l(c). Concrete was used to carry the compressive and shear forces of the beam, in addition to providing stability for the thin GFRP webs. The cross sectional area of the concrete for the optimized beam was 40% of the totally filled tube. Considering the weight of the tube, the optimized beam has 44% of the weight of the totally filled tube.

zyxw

Short Column Specimens The axial specimens tested were the same size as the small specimen used in the flexural tests, as shown in Figure l(a). However, the length of the axial specimens was reduced to 680 mm, such that the length of the column was 2.5 times its width. In total, five compression specimens were fabricated. Three of these specimens were in their original state when tested and two were obtained from the shear spans of the flexural test of the small GFRP beam, far from the failure region.

zyxwvu zyxwvut

p 9 rnm

,89 rnm

Figure 1 Cross-section configurations of test specimens

zyxwvutsr zyxwvut

688 FRPRCS-6: FRP Structural Shapes

zyxwvu

Fabrication of Test Specimens

In order to enhance the bond between the GFRP tubes and the concrete fill, the inner surface of the GFRP tubes was coated with a layer of epoxy. A thin layer of coarse silica sand was then applied on the tacky epoxy in order to provide a rough texture. For the partially filled tube, which has a void inside the concrete core, a Styrofoam prism of the same size as the inner void was fabricated and inserted inside the tube. Later, after hardening of concrete, the Styrofoam core was removed. In order to facilitate casting concrete, the tubes were braced in a vertical position to a structural wall as shown in Figure 2(a), and were filled with 53 MPa concrete from the top end. Vibration of concrete was applied during the gradual filling of the tubes. Once the specimens had cured, the Styrofoam core was removed, leaving the void. The cross-section of the optimized GFRP beam is shown in Figure 2(b). In order to prevent crippling of the thin web of the optimized beam above the supports, Concrete end blocks were cast, filling in the void over a length of 375 mm from each end.

(a) Setup used for casting the concrete into the tubes (b) Optimized beam Figure 2 Fabrication of test specimens

Test Setup and Instrumentation of Beam Specimens The three beam specimens were tested using four-point bending as shown in Figure 3. The span of the beams was 2100 mm and the distance between the loads was 300 mm. The beams were loaded using displacement control with a 500 kN capacity hydraulic actuator using displacement rate of 0.50 mdmin. The specimens were instrumented to record load, deflection and strain measurements. Longitudinal strains at midspan were measured using strain gauge type displacement transducers and electrical foil gauges attached directly to the GFRP

zyx

Rectangular FRP Tubes Filled with Concrete 689

surface at different levels along the depth of the beam. Transverse strains were also measured in the compression zone using foil strain gauges. Two displacement potentiometers were used to measure the midspan deflection and another was used to measure the support settlement. Digital displacement gauges and potentiometers were also used to measure the slip between the concrete core and the GFRP tube on the tension side at both ends.

Test Setup and Instrumentation of Short Column Specimens

zyxwvut

Figure 3 Test setup and instrumentation Figure 4 Test setup and instrumentation of test beams of short columns

Compression specimens were subjected to concentric and eccentric axial loads applied at various eccentricities using a 9000 kN capacity testing machine as shown in Figure 4. The eccentricities considered were 0, 25 mm, 50 mm and 64 mm using a setup with free rotation allowed at the ends. Another concentric test was conducted on a specimen between fixed platens, without allowing end rotation. All specimens tested were 680 mm in length. A thin layer of gypsum plaster was used to distribute the load from the steel plates to the specimens. Specimens were loaded under stroke control at a displacement rate of 0.165 mdmin. Load, deflection and strain measurements were taken for the compression specimens. Strain was measured in the axial direction at opposite faces using electrical foil gauges and strain gauge type displacement transducers with a gauge length of 200 mm. Foil strain gauges were also used in the transverse direction to measure the transverse strains.

zyxwvuts zyxwvutsr

690 FRPRCS-6: FRP Structural Shapes

RESULTS OF THE EXPERIMENTAL PROGRAM Beam Specimens The load-deflection behaviours of the totally-filled small and large beams as well as the partially-filled beam are shown in Figure 5. The figure shows that the totally-filled beam and the partially-filled beam had similar flexural stiffness. This behaviour indicates that the size of the concrete in the compression zone of the optimized beam was quite efficient. The depth of concrete in the compression zone was determined based on location of the neutral axis using strain compatibility analysis of a totally-filled tube and the partially-filled beam was designed to eliminate the concrete below the neutral axis. The flexural strength of the optimized beam was however lower than that of the totally-filled beam by about 22 %. This is attributed to the different failure modes as will be discussed. The load-axial strain behaviour at the extreme tension and compression sides of the three beams is given in Figure 6. The behavior of the totally-filled large beam is very similar to that of the partiallyfilled beam. The small GFRP beam had the largest value of compressive strain, -0.0088. Both the totally-filled large beam and partially-filled beam had lower maximum compressive strains of -0.0049 before strain reversals occurred due to buckling of the GFRP, resulting in local bending stresses in the compression flange. Buckling resulted in debonding from the concrete and not only reduced the effectiveness of the compression flange in carrying compression force, but also eliminated any concrete confinement effect. Maximum tensile strains of 0.0267 are similar for the small beam and for the large beam where rupture of the GFRP tension flange was the cause of failure. The partially-filled beam, which failed due to compression failure, had a maximum tensile strain of only 0.0213. Hoop strains measured at midspan in the constant moment region indicate significant hoop stresses are developed during loading. Figure 7 shows the hoop strains for the totally-filled large beam and the partiallyfilled beam. It should be noted that the hoop strains are a result of the Poisson’s ratio effect of the tube, the confinement effect of concrete and the flange buckling. If all the strain were attributed to the buckling, it would be expected that the hoop strains would be very low initially, followed by a sudden increase when buckling occurred at the late stages of loading. This is true for the partially-filled beam, where confinement

zy

Rectangular FRP Tubes Filled with Concrete 691

was expected to be insignificant due to the void. However, the behaviour and magnitude of the strains is greater for the large totally filled tube, indicating that a portion of the hoop strain may be attributed to confinement. The maximum slip measured between the concrete core and GFRP tube at the ends of the beams were 0.18,2.5 and 0.01 mm for the small beam, the totally-filled large beam and the partially-filled beam respectively. Test results indicate that the partially-filled beam could provide an optimized section considering the significant reduction of weight and cost of the materials. 700

6500

5

O

(Totally-fio

400

3

zyxwv zyxw zyxw zyxwv p

200

p 9

Small beam

a0

300

200

0

0

x)

20 30 40

50

zyxwv (Totally-filled)

400

Large beam

300

~z u :Fl

urge beam (Optimized)

mall beam (Totally-filled)

0.0

-x).O

60

a.0

20.0 30.0

Longitudinal Strain (6)

Deflection (mn)

Figure 6 Load-axial strain behavior

Figure 5 Load-deflection behavior 700 600

-

500

500

g 400

g400

V

0

4

A

g 3w

m 300 200

200

w

100

0 -3.0 0.0

3.0 6.0

9.0

P.0

Hoop Strain (6)

1

0 30

00

30

60

90

120

1

Hoop Strain (ms)

Figure 7 Load-lateral strain behavior at compression zone

Short Column Specimens For the two tests with zero eccentricity the load versus axial and hoop strains is shown in Figure 8. The hoop strains in each case were measured at midheight on one of the shorter sides of the rectangular tube. A higher ultimate load was achieved with the complete contact of

zyxwvuts zyxwvu

692 FRPRCS-6: FRP Structural Shapes

the cross-section with the loading plates, although with much less ductility than the pin ended condition. This is attributed to a localized premature failure in the pin-ended specimen. For the three compression specimens that were loaded eccentrically, the longitudinal strains at the extreme fibers were measured and shown in Figure 9. Towards the loaded side, strains were compressive. On other side of the specimen, strains were either compressive or tensile depending on the level of loading and the amount of eccentricity. Hoop strains were measured in more details for the pin-ended concentrically loaded column. One quarter of the circumference of the specimen was instrumented with strain gauges at mid-height as shown in Figure 10. The behavior can be categorized into three phases. Very little hoop strain, less than 0.0005, is recorded until about 1200 kN. From 1250 kN to 1550 kN, the highest strains are near the middle of the longer side and at the corner. Strains are highest near the middle of the longer side due to the outward bulging of the initially straight side due to expansion of the concrete core, resulting in bending in the plate. The strain is also high at the corner due to stress concentration. In the third

z zyx zyxwvut

1800

2000

1600 1600

-

1400

,-.

1200

g I000

1200

'0

g

U

9

800

-1

800

600 400

400

200

0

0 -12.0

4.0 a.o Axial and Hoop Strain (mS)

-8.0

-4.0

0.0

12.0

-15.0

-12.0 -9.0

-6.0

-3.0

Axial Strain (ins)

0.0

3.0

Figure 9 Load-strain behavior of columns Figure 8 Load-strain behavior of loaded with different eccentricities columns loaded with zero eccentricity

phase, from 1550 kN until failure the highest strains are at the center of the long and short sides due to the flexural strains induced by the internal expansion of the concrete and corresponding bulging of the sides of the rectangular tube. Based on the beam and column tests of the small concrete-filled tubes, a number of points on the axial load - bending moment interaction diagram have been established as shown in Figure 11.

Rectangular FRP Tubes Filled with Concrete 693

Failure Modes

The totally-filled beams failed by rupture of the GFRP on the tension side. Rupture immediately progressed up the web as shown in Figure 12(a). The partially-filled beam failed by inward buckling of the concrete compression flange.

z

Figure 10 Variation of hoop strains on one quarter of the concentrically loaded specimer

2000

h

zyxwv

zyxw e = 0 (pinned)

0

m

z

800

zyxwvut i *.i

400 0

Localized failure at one end

e=m

-c--

0

50

100

I50

200

250

300

Moment (kNm)

Figure 11 Axial load-bending moment interaction diagram of small specimens

The short column with complete surface contact with the loading plates and zero eccentricity as well as the one loaded with 64 mm eccentricity failed by rupture of the fibers along the corner of the tube, initiated at one end of the specimen and progressed towards the other end as shown in Figure 12(b). The pin-ended specimen with zero eccentricity and the one with 50 mm eccentricity failed by local shearing of one corner as shown in Figure 12(c). The specimen with 25 mm eccentricity failed by crushing of the tube at midheight, accompanied by fracture of the fibers in the hoop direction as shown in Figure 12(d).

694 FRPRCS-6: FRP Structural Shapes zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPON

CONCLUSIONS

(a)

Large totally-filled

zyxwvu (b) e = 0 (Fixed)

(c) e = 0 (Pinned)

(d) e = 25 mm

Figure 12 Failure modes of beam and column specimens

A FRP/concrete hybrid concept has been introduced. A rectangular GFRP thin tube can be totally-filled or partially-filled with concrete to optimize the concrete within the cross-section. Beams and short columns have been tested. The following conclusions are drawn: (a) The partially-filled beam showed similar stiffness to the totallyfilled beam but lower flexural strength due to the different failure mode. (b) Totally-filled beam failed by fracture of GFRP tube in tension. Partially-filled beam failed by inward local buckling of the concrete flange. (c) In short columns, higher hoop strains are developed in the middle of the straight sides of the rectangular tube than at the rounded corners due to local bending of the GFRP tube as a result of the bulging concrete core. zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGF

REFERENCES

z

1. Fam, A. Z. and Rizkalla, S. H., “Flexural Behavior of Concrete-Filled Fiber-Reinforced Polymer Circular Tubes”, ASCE Journal of Compositesfor Construction, Vol. 6, May 2002, pp.123-132. 2. Triantafillou, T. C. and Meier, U. “Innovative Design of FRP Combined with Concrete,” Proceeding of the Is‘ International Conference on Advanced Composite Materials for Bridges and Structures (ACMBS), Sherbrooke, Quebec, 1992. pp. 49 1-499.

zyxwvutsr

FRPRCS-6, Singapore, 8-1 0 July 2003 Edited by Kiang Hwee Tan @WorldScientific Publishing Company

zy zyxwv

FLEXURAL BEHAVIOUR OF GFRP-POLYMER CONCRETE HYBRID STRUCTURAL SYSTEMS M. C. S. RIBEIRO, A. J. M. FERREIRA AND A. T. MARQUES INEGI, Faculdade de Engenharia da Universidade do Port0 Rua do Barroco, 174-214, 4465-591, Leqa do Balio, Portugal

In this paper, the development of a research work undertaken in hybrid structural systems, where GFRP pultruded profiles are assembled, in an innovative way, with a layer of polymer concrete is reported. Several beams, with three different hybrid designs, were tested in four-point bending and the flexural behaviour of such structures was analyzed. Model designs produced a highly optimized flexural behaviour, with a pronounced synergetic effect. A new finite element model for GFRP-polymer concrete hybrid beams was also developed. Due to non-linearity of polymer concrete and the need to account for interaction aspects, a geometric and material nonlinear analysis was performed. A thick shell element was considered, incorporating a layered approach for laminated composite modeling.

INTRODUCTION

FRP-Polymer Concrete Hybrid Systems In the last years, an innovative structural concept was developed involving the combination of FRP pultruded profiles with conventional cement concrete, to produce lightweight, corrosion-free and yet inexpensive hybrid systems According to this new concept, a FRP profile beam is combined with a concrete layer, cast onto the top flange, which replaces the thick FRP compressive flange of traditional pultruded profiles. The method maximizes system performance using materials by combination, and can be thought as a better way of producing structural members based on pultrusion process. In the last decade, FRP researchers have developed a number of hybrid systems, either by simply replacing steel with FRP pultruded profiles in conventional steel-concrete composite construction, or by developing new structural systems Following these studies, this paper presents the development of a research work undertaken with hybrid structures composed by GFRP profiles and polymer concrete. Polymer concrete is a mixture of mineral aggregates and a polymer binder in which the polymeric resin replaces the

’* ’.

’.

zyxwvuts

zyxwv zy

696 FRPRCS-6: FRP Structural Shapes

Portland cement-water binder of hydraulic concrete. This type of concrete has very high compressive strength, good chemical resistance, and relatively low elasticity modulus 4, Since the ~ O ’ S , when it was initially applied to produce synthetic marble, its applications have increased tremendously, mainly in the field of precast components for building industry 62 Exploiting the better performances of resin concretes, it is believed that the potential of the hybrid systems can be improved: (a) The higher compressive and flexural strength of polymer concrete allow a cross-section reduction and, therefore, a lightweight beam; (b) The better chemical resistance enlarges the application field of these structural systems to highly corrosive environments; (c) The faster curing time, early high strength and good workability of resin concrete make it suitable, either for production of precast components, or for an easier in site application. In a previous investigation work small-scale models of four different hybrid beams were designed, manufactured and tested in four-point bending. Flexural behaviour of such composite structures, in terms of loaddeflection history, load capacity and failure mode was analyzed and compared with the flexural behaviour of its material constituents. The four hybrid beams models were designed considering a GFRP pultrusion profile that works in tension, and an epoxy polymer concrete filling, that works mainly in compression. Experimental results indicated the high potential of these hybrid systems, revealing a pronounced synergetic effect associated to the assembly of polymer concrete and GFRP profiles. However, some technical problems were detected. Interface debonding promoted premature failure, hindering the maximum of load bearing capacity to be reached. Further studies were necessary in order to improve bond strength between polymer concrete and GFRP profiles. In this paper, some new experiments done on GFRP-polymer concrete hybrid beams, with improved adherence between material components is reported. Three new series of small-scale hybrid beam specimens were manufactured and tested in bending. A special treatment was given to GFRP profile internal surfaces to promote the adhesion to epoxy polymer concrete. A significant improvement in the flexural performance of the hybrid beams was reached.

’.

’.

’,

Numerical Modelling Past studies showed that cracking, deflection and ultimate load behaviour of FRP-reinforced concrete beams could be predicted with the same degree of accuracy as the behaviour of regular steel reinforced concrete beams, and

GFRP-Polymer Concrete Hybrid Structural Systems 697

that a theoretical correlation is therefore possible9. Theriault and Benmoluane" proposed alternative formulas for the prediction of deflection. They concluded that usual formulas for such predictions were misleading. Most of these previous studies are related to concrete beams. Very few studies of FRP reinforced plates or shells have been proposed, if any. The present work reports a numerical model for the analysis of concrete arbitrary shell structures reinforced with composite materials, particularly with FRP pultruded materials. The proposed model is capable of predicting deflections and stresses in concrete and in FRP reinforcement, considering geometrical and material non-linear behaviour. It includes constitutive laws for concrete material, based on smeared crack concepts and is applicable for high-strength composite materials. The model is implemented in a degenerated shell element as proposed by Ahmad et al.'', Owen and Figueiras12, and more recently by Ferreira and B a r b ~ s a ' ~In. order to check the model capacities, it was applied to predict the flexural behaviour of GFRP-polymer concrete hybrid beams. A good agreement between experimental and numerical results was found. The tests were only performed in beams, but the model is generally applicable to plates or shells of arbitrary shape.

zyxwvu zy

EXPERIMENTAL PROGRAM

Component Materials

Low viscosity epoxy resin, and foundry sand with a very uniform and fine grain (d5,, of 342 microns) were used for the polymer concrete mixture. Resin content, without charge, was 20% in weight. This formulation corresponds to an optimized result from previous research 14, and its mechanical properties are already known. Compressive strength and compressive elasticity modulus are 82 MPa and 1 1.5 GPa, respectively; with ultimate strain defined to be equal to 0.01. Compression properties were obtained from uniaxial testing according to RILEM TC-113 standards. Three types of standard U-shaped GFRP pultruded profiles were used. The pultrusion profiles consisted of continuous strand mat and roving of glass fibers, impregnated with unsaturated polyester resin and having an external veil pulled through a die. Volume content of glass fibers was between 50% to d55%, and roving occupied the most part of it. The GFRP profiles cross-section dimensions and mechanical properties, obtained from 52

698 FRPRCS-6: FRP Structural Shapes

uniaxial tension tests according ISO-527/4 standard, are summarized in Table 1.

zyxw

Table 1. Dimensions and mechanical properties of GFRP profiles

GFRP

Tensile Strength (MP4 Profile A 395 Profile B 3 10 Profile C 350

Tensile Elasticity Modulus (GPa) 25.1 25.3 25.6

Cross Section (mm) 55 * 60 * 5 30 * 40 * 4 50 * 40 * 5

Hybrid Beams: Design and Manufacturing

Among the four models of hybrid beams initially designed, only three models, - those that showed better flexural performance in first test series-, were chosen for this experimental program. Cross-sections of the three models, hereinafter referred as HB I, HI3 I1 and HB 111, are illustrated in Figure 1, All the small-scale models were 600 mm in length.

zyxwv

Figure 1. Small scale models of GFRP-polymer concrete hybrid beams analyzed.

HB I and HB I11 beams have all the section full of concrete, with the profiles behaving, simultaneously, as reinforcement and as permanent formwork. HB I1 type beam is lighter, with only a thin layer of concrete positioned in the upper part of the profile. Cross-section of concrete layer, in this model, was designed in order to support only compressive stresses in the elastic range. Neutral axis was determined through the homogenization of the hybrid cross-section on GFRP material. In order to prevent premature failure due to interface debonding, adherence between GFRP profiles and polymer concrete was improved

z zyxw

GFRP-Polymer Concrete Hybrid Structural Systems 699

through both mechanical and chemical processes. Internal surfaces of GFRP profiles were previously sandpapered, and a layer of silane adhesion promoter was applied before concrete casting. In addition, in HB I1 specimens, profile ‘B’ was bonded to profile ‘A’ with a special epoxy adhesive. In the first specimen series, the resin of concrete that drained during casting process assured junction between these two profiles. For each hybrid model, three specimens were manufactured. All specimens, before being tested, were allowed to cure for seven days at room temperature, and post-cured at 80°C during 3 hours. Testing Procedure

Four-point bending tests on three specimens of each type of hybrid beams were performed. The specimens were loaded over a span of 510 mm and with constant moment zone of 100 mm. Load was gradually increased up to failure at the rate of I d m i n u t e (displacement control mode). Test procedures were identical to those applied to perform flexural tests of first series of GFRP-polymer concrete hybrid beams. In order to evaluate the synergetic effect of the assembly, four-point bending tests were also performed on each constituent material of each type of hybrid beam. NUMERIC MODEL: ASSUMPTIONS AND IMPLEMENTATION The model is based on a dual criterion for polymer concrete. Crushing and cracking of concrete is considered by modification of the material stiffness and strength. Also, composite materials are modeled with a layered approach. Each material layer is considered to be orthotropic. Elastoplasticity of the material can be used, if necessary. The approach used for these hybrid beams relies on a homogenization procedure of the distinct zones of their cross-sections, both for stiffness and strength. In Figure 2, these zones are identified. is

PC

1 GFRP

Zone 3

zyxw

Figure 2. Layered system used for finite element approach.

z zyxwvut

700 FRPRCS-6: FRP Structural Shapes

zyxwvutsrqp

In Zone 1, the equivalent material is considered to be a composite laminate only. Therefore, a typical FRP material formulation is considered 13, 15

In Zone 2, FRP walls are taken in account. Due to the shell formulation, an equivalent material in stiffness and strength is considered. In Zone 3, a combination of FRP and polymer concrete occurs. In this case, it is assumed that no contribution is given by FRP walls to support compressive solicitations. Therefore, compressive strength of the assembly layer is the compressive strength of the concrete, while tensile strength and elasticity modulus follows a law of mixtures:

where A'', AFP = stength or stiffness, of polymer concrete and FRP material, respectively; Aey = equivalent strength or stiffness of the layer; fc,AFRp = thickness of polymer concrete and FRP material in the layer, respectively; and /IToT = total thickness of the layer. EXPERIMENTAL VERSUS NUMERICAL RESULTS

For each type of beam, the average maximum load capacity and the related synergetic effects are summarized in Table 2. Synergetic effect was calculated by dividing the capacity load of the assembly by the sum of capacity load of each of its two elements. In order to evaluate the improvement of flexural strength reached through profiles surface treatment, experimental results obtained in first specimen series are also presented. Table 2. Summarized flexural tests results

Hybrid Max. Capacity Load (kN) Beam T Y P ~ I" Series T dSeries 38.07 46.30 HBI 43.09 HB II 26.96 32.96 HB 111 32.06

Synergetic Effect (%) I" Series 2ndSeries 233 284 203 3 19 258 266

Load deflection curves, obtained from both lst and 2"d test series of composite beams and correspondent components, are illustrated in Figure 3. Numerical results obtained by finite element modeling are also plotted.

zyx GFRP-Polymer Concrete Hybrid Structural Systems 701

60

-

E3 0

50

40

z

zyxwvuts

30

20

10 0 0

5

10

15

20

25

20

25

Deflection (mm)

5

0

10

15

Deflection (mm) 50

HB 111

zyxwvuts 0

5

10

15

20

25

Deflection (mm)

Figure 3. Numerical versus experimental load-deflections curves obtained from 1'' and 2"dtest series of hybrid beams, and correspondent components (GFRP and PC).

zyxwv

702 FRPRCS-6: FRP Structural Shapes

Except for HBIII type beams, a significant increase in load bearing capacity occurred for the 2ndseries of hybrid beams tested. Failure modes were also distinct. HBI and HBII type beams, which failed prematurely at the first series due to interface debonding, collapsed in the second series due to GFRP webs shear fracture, as shown in Figure 4. For HB I type beams, an explosive rupture of concrete occurred as a result of shear crack propagation and consequent gradual loss of cohesion between components. Collapse of HB I1 beams was less brittle due to slower crack propagation, thus giving some warning of eminent collapse.

Figure 4. Typical failure modes of GFRP-polymer concrete hybrid beams: lst and Ydtest series.

z

Type I11 beams, in the same way that occurred for the first series, failed due to tensile failure of lower surface of GFRP profile, followed by rupture of concrete. Cross-section of profile C is slightly narrower at the top, which made the slip of concrete more difficult. This fact explains why the improvement of bond, between concrete and GFRP profile on 2ndspecimen series, had no significant effect on its flexural performance.

GFRP-Polymer Concrete Hybrid Structural Systems 703

zyxwvu zyxw

Numerical results agree quite well with experimental results, according to the modeling hypothesis. At this stage, the model can already predict the load-displacement curves with reasonable accuracy. Although predicted initial stiffness and failure loads are reasonably modeled, the model has still to be improved to account for web shear failure and other failure mechanisms. CONCLUSIONS

Hybrid rectangular beams were designed, manufactured and tested. The assembly allows for a critical combination of polymer concrete and composite pultruded materials. A better flexural performance was reached with improved interface bonding between material components. This innovative design produces highly optimized behaviour with a pronounced synergetic effect. The highest strength to weight ratio and the highest synergetic effect associated to HI3 I1 beams, due to a more precise placement of material in stress zones, make this design model very promising for further investigations on large scale GFRP-polymer concrete hybrid beams. The numerical model used for flexural behaviour analysis of hybrid systems was implemented in a finite element code. The model predicts with very reasonable accuracy the load-displacement curves for all tested type beams. This model is still in progress. The results so far allow for a very interesting expectation in terms of a finite element code for the analysis of this kind of hybrid systems. ACKNOWLEDGMENTS

The support of Fundagilo para a CiCncia e Tecnologia under POCTVEME/42820/200 1, ‘Desenvolvimento de estruturas hibridas betilo polimerico/comp6sitos’ is gratefully acknowledged. REFERENCES

zyx

1. Deskovic, N.; Meier, U. and Triantafillou, T., “Innovative design of FRP combined with concrete: Short term behaviour”, Journal of Structural Engineering, July (1995), pp. 1069-1078.

704 FRPRCS-6: FRP Structural Shapes

zy zyxw zy

2. Hall, J.E. and Mottram, J.T., “Combined FRP reinforcement and permanent formwork for concrete members”, Journal of Composites for Construction, ASCE, 2(2), (1 998), pp.78-86. 3. Mirmiran, A., “Innovative combinations of FRP and traditional materials”, Int. Conference on FRP Composites in Civil Engineering (CICE-2001), Hong Kong, China, December 12-15, 2001, Vol. 11, pp. 1289-1298. 4. Chawalwala, A.J., “Material characteristics of polymer concrete”, M S . Thesis, University of Delaware Center for Composite Materials, 1999. 5. Ribeiro, M.C.S.; Tavares, C.M.L. and Ferreira, A.J.M., “Chemical resistance of epoxy and polyester polymer concrete to acids and salts”, Journal of Polymer Engineering, 22( l), 2002, pp. 27-44. 6. Fowler, D.W., “Polymer in concrete: a vision for the 2lSt century”, Cement & Concrete Composites, 2 I (1999), pp 449-452. 7. Dikeou, J., “Precast polymer concrete in the United States”, 5”’ Int. Congress of Polymers in Concrete, Brighton, England, 1986. 8. Ribeiro, M.C.S., Ferreira, A.J.M. and Marques, A.T., “Static flexural performance of gfrp-polymer concrete hybrid beams”, Int. Conference on FRP Composites in Civil Engineering (CICE-ZOOl), Hong Kong, China, December 12-15,2001, Vol.11, pp. 1355-1362. 9. Saadmatmanesh, H. and Ehsani, M.R., “Fiber composite bar for reinforced concrete construction”, J. Compos. Mat., 25 (2), 1991, pp. 188-203. 10. Theriault, M. and Benmokrane, B., “Effects of FRP reinforcement ratio and concrete strength on flexural behaviour of concrete beams”, Journal of Composites for Construction, 2 (l), 1988, pp.7-16. 11. Ahmad, S., Irons, B. and Zienkiewicz, O.C., “Analysis of thick and thin structures by curved finite elements”, Int. J. Num. Meth.Engng., 2 (1970), pp. 419-451. 12. Owen, D.R.J. and Figueiras, J.A., Ultimate load analysis of reinforced concrete plates and shells, in Finite Element Software for Plates and Shells, E. Hinton and D.R.J. Owen (Eds.), Pineridge Press, 1984. 13. Ferreira, A.J.M. and Barbosa, J.T., “Buckling behaviour of composite shells”, Composite Structures, 50 (2000), pp.93-98. 14. Tavares, C.M.L.; Ribeiro, M.C.S.; Ferreira, A.J.M. and Guedes, R.M., “Creep behaviour of frp-reinforced polymer concrete”, Composite Structures, 57 (2002), pp. 47-5 1. 15. Ferreira, A.J.M.; Camanho, P.P.; Marques, A.T.M. and Fernandes, A.A., “Modelling of concrete beams reinforced with frp re-bars”, Composite structures, 53( l), 200 1, pp. 107-1 16.

zyxwvu

zyxwvutsrq

FRPRCS-6, Singapore, 8-10 July 2003 Edited by Kiang Hwee Tan QWorld Scientific Publishing Company

zyx zy zy zyxwv

A NEW CONCEPT FOR AN FRP PANELIZED RAPID DEPLOYMENT SHELTER

N.M. BRADFORD AND R. SEN University of South Florida, Tampa, FL 33620

In the aftermath of a natural disaster, the aid worker is faced with three immediate tasks (1) provide protection from the environment, (2) provide food and resources for the facilitation of life and (3) provide health services for the treatment and prevention of illness. Each of these tasks depends on emergency shelters that can be rapidly deployed and assembled. This paper describes a new panelized concept for a FRP shelter developed by the University of South Florida as part of a research project sponsored by the Office of Naval Research. The size of the panels, their weight and ease of assembly were key constraints in the development of this concept. More importantly, the assembIed structure had to withstand Category 4 hurricane force winds. The lightweight, high strength and corrosion resistance of FRP makes them ideally suited for emergency structures that have to be rapidly deployed. A novel panelized concept was developed that integrated the connectors in the panels themselves thereby minimizing the need for additional connectors and greatly simplifying construction. This integration made it possible for the assembled structure to resist very high wind uplift forces. The individual panels are less than 0.914m (36”) wide and each section has four trapezoidal boxes that allow units to slide into each other and interlock. The concept is versatile and could be readily adapted for alternative applications such as bridge deck replacement.

INTRODUCTION

The goal of the project was to identify emergency shelters that utilize a “house in a box” concept in which all unassembled components can be packaged in a crate and conveyed to the site for erection by relief workers with minimal skills. The emergency shelter was to be designed to withstand hurricane force winds. Hurricane winds create pockets of wind pressure that can cause individual components to failure or the building to fail at the foundation connections, wall connections and roof connections. Further, the design is subject not only to extreme load conditions, but also non-structural parameters such as erection speed, construction simplicity and cost,

706 FRPRCS-6: FRP Structural Shapes

Structural Performance is not a simple question when it comes to emergency shelters. One must design the components to withstand extreme conditions without failure. Erection Simplicity is assessed by the speed of erection and the skill requirements of the workers. Further complicating the issue are the questions of connection and foundation requirements. The optimal solution involves an integration of several functions into a single component. Cost runs in an inversely proportional relationship to all other issues addressed during the design optimization process. This question can only be circumvented through innovation in either the areas of the materials or construction techniques used during erection. DurabilitylAdaptability addresses the possible long term usage of the final buildings as ‘safe houses’ where primary facilities can be maintained during future disasters. Based on the design issues listed above, it was decided that the optimum solution would address the problem as a material issue and a building component issue. In order to answer both questions, our focus was turned to the use of Fiber Reinforced Polymer materials, which offer design customization for specific applications. This paper presents a brief summary of the development work completed to date. The conceptual basis of the building shelter is discussed and a simple design procedure presented. A brief review of alternative buildings the shelter industry currently offers is also included.

zy zy

AVAILABLE SHELTER SYSTEM REVIEW

Subsequent to the detailed industry wide review conducted by the research team, several types of emergency shelters systems were found to provide viable alternatives. Specifically, as a result of our investigation3, it was concluded that the viable emergency shelters fell into three types of construction. (a) Standard Construction - New Materials: These systems emphasize the improved performance gained through the use of new materials. Such materials offer the user improved mechanical properties (on a localized basis), light weight, non-corrosive and non-metallic performance. Further, these systems attempt to use the new materials as direct substitutes for standard components in building systems. An example of this type of construction would be substituting FRP studs in a wood framed stud wall system or the use of styro foam molds in lieu of masonry blocks in a filled masonry wall system. Information was solicited from three manufacturers that fall into this category.

FRP Panelized Rapid Deployment Shelter 707

(b) New Construction - New Materials: These systems develop new construction systems in an attempt to best utilize the performance characteristics of the new materials. Typical examples of this construction consist of the development of panelized wall and roof systems which are fabricated using FRP systems. Information was solicited from six manufacturers in this category. (c) Alternate Systems: These systems within this category constitute a fully alternate system of construction, based on geometry, materials and construction. Typical examples of this construction include monolithic domes and Yurts. Information was solicited from two manufacturers that fall into this category. The existing emergency shelter industry was reviewed for viable candidates. Due to the stringent performance requirements, only eleven existing building systems met project requirements As a result of the existing emergency shelter review, it was concluded that while some systems appear to have met the base criteria of the project, all of the systems investigated exhibit similar weaknesses. Specifically, it was noted that the currently available rapid deployment shelters exhibited the following areas of weakness, (a) Inadequate supporting structural information (limited test data, limited detailed calculations). (b) Inadequate structural capacity of connections. (c) Reliance on supplemental steel systems for member stiffening. (d) Reliance on supplemental steel systems for connections. (e) No building system components currently being manufactured. Further, it was concluded that all of the systems are designed as one time usage buildings, since disassembly would constitute a significant amount of work and possible member damage. It was also noted that all of the systems emphasize localized member performance issues of bending and shear, while failing to fully develop the global issues of member connections and systemic performance under load.

zyx

SHELTER DESIGN - CONCEPTUAL OPTIMIZATION A multitude of structures have been developed and fabricated by the emergency shelter industry. Their geometry range from cubicle shaped boxes to monolithic dome type structures. In order to develop an optimal geometric shape for use in this project, special attention was paid to the needs of the end user. These were classified as:

zyxwvuts

708 FRPRCS-6: FRP Structural Shapes

(a) Structural Performance: Of primary concern is the capability of the structure to withstand hurricane force winds. (b) Anchorage Performance: Due to the wide variety of usage proposed for the emergency shelter, adequate foundation anchorage is required for a multitude of ground soil conditions. (c) Construction Simplicity: The emergency shelter should be simple to construct, both in terms of the constituent construction components and the technical skills of the laborers. (d) Geometric Adaptability: The emergency shelter must fit into a pod-like system of urban design. Specifically, the end user should be able to add or subtract shelter units to "build" configurations to suit specific usage needs. Initial Geometry Analysis An initial geometry based on previous shelters developed by the military was selected as a starting point for the emergency shelter. This geometry corresponded to a rectangular box shape, 7.32m x 10.97m x 2.44m (24'- 0" x 36'- 0" x 8'- 0") in shape, with a single roof ridge line and gable ends in the shorter 7.32m (24'- 0") dimension. A preliminary structural analysis of the building exposed to the design wind design pressures indicated that very large force concentrations occurred at the corners of the building and at the transition between the roof and wall members. Due to the magnitude of these forces, it was concluded that the building geometry needed to be revised. Development of Revised Geometry

In order to restrict the force concentrations to acceptable limits, the base geometry of the building was reduced to a 3.66m x 7.32m x 2.44m (12'- 0"x 24'- 0"x 8'- 0") box shape. A preliminary structural analysis of the reduced geometry showed that the force concentrations were reduced by a factor of three. Further, the reduced geometry provides for a usable area of 26.76m2 (288 ft'). This reduced footprint area allows for a greater variety of uses. Specifically, this area is more acceptable for usage as sleeping area, office space, storage area or medical facility.

FRP Panelized Rapid Deployment Shelter 709

Development of Roof Geometry The initial roof geometry called for a gable end roof system. This system was found to be unacceptable due to the inherent structural weakness that occurs at the gable end wall connections. This weakness results in the development of a hinge joint failure in the gable end wall during high wind events. A double hipped roof configuration was investigated as a first alternative roof system. While this roof configuration provides for adequate structural bracing at all wall roof transition points, it was concluded that the complexity of this construction prohibits its selection for use in a rapidly deployed emergency shelter where simplicity of the building construction is crucial to success. The geometry of the roof system is crucial to structural performance since it directly affects the magnitude and application of wind load forces'. Further, the roof system geometry affects erection speed and construction complexity. Roof optimization led to the selection of a low rise mono sloped roof configuration. This roof configuration has several positive aspects. Specifically, a mono sloped roof system provides structural stability at all roof wall transition points. A mono sloped roof is simple to constructed, requiring one basic structural component (plank member). Additional positive aspects of the mono sloped roof 'system include the ability to use the same members as used in the wall system, and the ability to align adjacent shelter units so as to create a variety of roof configurations (Fig. 1).

z

R m f CwnerZone

Wall Int. Zone i

Figure 1 Resultant Wind Design Pressures on Emergency Shelte

710 FRPRCS-6: FRP Structural Shapes

Door and Window Locations The initial design of the shelter utilized a variety of door and window sizes, placed in all elevations of the building. For the purpose of adaptability and simplicity, it was concluded that all door and window sizes should use the same opening size. It was concluded that openings should not be installed in either of the short dimension 3.66m (12'- 0") elevation wall, due to the need for lateral shear resisting members in these elevations. Subsequent design resulted in the use of a 0.914m (36") wide nominal opening for all windows and doors. Further, the placement of these openings in the long dimension elevation of the shelter facilitates the placement of adjacent shelters to create "rooms" which can be arranged into usable building complexes.

zyx

PANEL SYSTEM - LOCALIZED PANEL DESIGN The panel shape was developed to enhance structural performance. Specifically, the panel is comprised of a continuous truss system that helps stiffen the section and facilitates stress transfer between the upper and lower skins that also act to resist bending induced stresses in two directions. The panels were developed to be used in an opposing, interlocking fashion, as shown in Figure 2. The single panel configuration is detailed in Figure 3.

Figure 2 Interlocking FRP Panel System'

FRP Panelized Rapid Deployment Shelter 711

zy

Trapezoid shaped ribs method for connection aaacent members. A1 panels are interlocked, provide resistance agai twisting along member length.

Trapezoid shape reduces the stress concentration that would occur at this point in a triangle shaped rib

Members are slid together during construction

\ k

FRP skin provides the flexural andin-plane shear capacity Panel lip locks into adjacent member, providing resistance to moisture / air i n h s i m

h-

Figure 3 FRP Panel Member Cross Section’

Aside from the structural performance characteristics inherent in the continuous truss configuration of the panel, several positive attributes develop as a result of the geometry. Specifically, the truss shape of the panel ribs allow for their usage as interlocking connectors. When opposing panels are connected in this fashion, the overall panel structure acts to restrict moisture and air infiltration2. Another attribute of the panel member design are the lip connectors that run along the perimeter of each panel. These connectors, while not designed to transfer structural stresses between members, are adequate to seal the joint that occurs between adjacent panel members. Such a lip connection is required to restrict moisture and air penetration through the system. This

z

712 FRPRCS-6: FRP Structural Shapes

zyx

attribute allows the user to transfer forces between adjacent panels while ensuring that building envelop integrity is maintained. PANEL SYSTEM - SYSTEMIC CONCEPTUAL DESIGN

The interlocking connections of the panel system greatly simplify the erection process due to the lack of a need for additional connections and members. This feature contrasts with one of the main observed weaknesses in existing building systems. Specifically, it was noted that all of the available systems required separate connector members, both for member to member connection and member to support frame attachment3. It may be argued that for each supplemental connector / attachment, an increase occurs in both erection complexity and the time required to construct the building. The application of a fully interlocking assembly system greatly simplifies the construction while ensuring the quickest erection process possible (Figure 4).

Figure 4 F W Panelized Shelter System3

FRP Panelized Rapid Deployment Shelter 713

z

A critical problem that arises with respect to the design of an emergency shelter is the transfer of forces along the building's load pathways from the roof to the foundation system. During the industry review, this aspect was noted to be a significant weakness with respect to each building's structural performance. Through the development of the FRP interlocking panelized building system, this issue was addressed through the use of sets of interlocking anchor blocks, installed at the roof - wall transition, and shown in Figure 5. To allow this connection to be adjustable during the construction / erection process, each set of blocks is attached using adjustable cables. During peak load applications, these blocks act both as tension connectors (providing direct uplift resisting members to transfer roof load to wall components) and rotational stiffeners (restraining the rotation of the roof - wall transition as a result of laterally induced wind shear on the shelter).

zyxwvu Connector fits into the member's trapezoidal

zyxwvut

Triangle shaped connector - 6 length

Adjustable strap Allows the connector to be pre-tensioned

-

Wall panels overlap trapezoidal ribs interlock, forming a structural connection to adjacent members

Figure 5 FRP Roof Anchorage Detail3

714 FRPRCS-6: FRP Structural Shapes

zyxwv

CONCLUSIONS

The development of the panelized shelter building system started with a conceptual design to address non-structural issues such as building system simplicity and the ability to disassemble and rebuild the structure with minimal work or member damage. The new system reduced the number of member types required during construction and facilitates systemic strength through interlocking of component members. Moreover, the use of FRP materials provided the greatest amount of design flexibility. We would recommend that a full size proto-type be built and tested to further develop the useful applications for this technology. ACKNOWLEDGMENTS

This work was supported by the Center for Disaster Management and Humanitarian Assistance, University of South Florida under contract from the Office of Naval Research. Further, we wish to thank Dr. Ayman Mosallam for his expertise in the areas of FRP composite materials and structural performance. REFERENCES

zy

1. ASCE 7-98, (1998). “Minimum Design Loads for Buildings and Other Structures”, American Society of Civil Engineers, New York, NY. 2. Bradford, N., Sen, R. and Mosallam, A. (2001). “Development of New Modular Composite Panel”. Proceedings of the 46‘h International SAMPE Symposium and Exhibition-Science of Advanced Materials and Process Engineering Series, Vol. 46, Long Beach, CA, May 6-10. Society for the Advancement of Material and Process Engineering, pp 93 1-942. 3. Bradford, N., Sen, R., Cooke, S. and Crespi, R. (2000). “Rapid Deployment Emergency Shelter”, Final Report submitted to Center for Disaster Management and Humanitarian Assistance/OfJice of Naval Research, December, pp. 169.

zyxwvut

FRPRCS-6, Singapore, 8-10 July 2003 Edited by Gang Hwee Tan QWorld Scientific Publishing Company

EXPERIMENTAL INVESTIGATION OF PULTRUDED FRP SECTION COMBINED WITH CONCRETE SLAB

zyxwv zyxwvu

A. BIDDAH Department of Civil and Environmental Engineering, UAE University Al-Ain, P. O.B. 17555, United Arab Emirates

The use of Fiber Reinforced Polymer (FRP) Composites is one of the latest significant developments in the field of bridge construction. FRP composites have been the material of choice in the aerospace industry since 1960s. However, only recently glass FRP composites have been gaining both popularity and acceptance as one of the structural materials of choice because of their high strength and stiffness to weight ratio, and corrosion resistance. This paper demonstrates the possibility of combining composite materials with a low-cost construction material (i.e. concrete) resulting in a new concept of designing lightweight, corrosion immune, yet inexpensive beams having acceptable structural properties. Pultruded FRP beam section-to-concrete slab is proposed to behave under bending as a composite beam. The objectives of the proposed research are to investigate the composite behaviour of FRP members with reinforced concrete slabs and comparing the behaviour of fully encased FRP beams with that of FRP beams mechanically anchored to the concrete. An experimental program was conducted to demonstrate the behaviour of the pultruded FRP beam-toconcrete slab in composite action. Three large scale specimens of 2.25 m length were tested under four-point loading. The first specimen is a pultruded FRP beam used as a control beam. The second specimen consists of pultruded FRP beam-to-concrete slab acting as a composite beam. The third specimen is a fully concrete encased FRP beam. The test results indicated the feasibility of using hybrid FRP-concrete beam to increase the load carrying capacity in flexure as well as beam stiffness. The outcome of this research provides substantial information for both designers and researchers in the field of FRP composites.

INTRODUCTION The use of FRP composites as structural materials to replace the more traditional steel and concrete materials has gradually increased in the construction industry. Properties such as high corrosion resistance, low densities, high durability, high strength, good stiffness to weight ratios and ease of handling and installation make composites far more desirable.

zyxwvu

716 FRPRCS-6: FRP Structural Shapes

However, applications in current practice of this type of structural material are still very rare. Deformability and cost are the main obstacles. To overcome these disadvantages, many authors have proposed structural elements in which GFRP is combined with other (less expensive and more massive) structural materials. For instance, GFRP wires and rods have been proposed as reinforcing bars for concrete and GFRP box and tubular sections have been used as permanent forms (having a structural effect as confinement and reinforcement) for concrete columns. GFRP plates (bonded to an external surface of reinforced concrete beam or slab) have been used as a repair or strengthening measure. However, the associated structural deformation in these applications is comparatively high. To overcome the disadvantage of high deformability of GFRP, pultruded FRP-to-concrete slab is proposed in this research as a composite beam. The composite action between the FRP structural shapes and concrete is supposed to reduce the previously observed high deformation when FRP rebars were used in concrete. The results of the proposed research will provide an alternative, practical construction system. Shear connectors are installed between the pultruded FRP beam and the concrete slab. These connectors resist the horizontal shear at the pultruded FRP beam-to-concrete slab interface caused by composite action.

zyx

BACKGROUND Pultruded fiber reinforced polymer (FRP) shapes are either thin-walled or moderately thick-walled open or closed sections consisting of assemblies of flat panels. Due to the high strength-to-stiffness ratio and thin-walled sectional geometry of FRP shapes, problems of global instability, global and local buckling mode interaction, and excessive local deformations are common in current structural shapes. Buckling prior to attainment of the ultimate material strength as well as deflection limits generally control the design of current FRP beams. For long-span FRP beams, Euler (overall) buckling is more likely to occur than local buckling. As for short-span FRP beams, local buckling or distortional buckling (a combination of local and lateral buckling) may occur first and may finally lead to large deflections or material degradations (crippling). The web element of FRP shapes may buckle locally due to shear loading when the beams are under bending. When the web is deep and thin, local buckling of the web under shear loading may occur before the flange buckles. The first application of FRP composites in the USA for road bridges was the structural composite honeycomb sandwich panel deck in Russell,

zyxwv Pultruded FRP Section with Concrete Slab 717

Kansas in 1996. Following this, several demonstration bridges were installed throughout the USA. Currently over 40 FRP composite vehicle bridges have been installed. Some FRP composite examples include2: (a) Tom's Creek Bridge, Blacksburg, Virginia - Strongwell Product constructed in 1997, and (b) Troutville Weigh Station, Troutville, Virginia- Strongwell constructed in 1999. In most bridges, composite action with the deck was not considered. The design of GFRP beams is usually governed by the low stiffness, resulting in a need for excessive use of composite material to satisfy certain displacement requirements. In view of this, a novel and more efficient design of composite pultruded GFRP beam is necessary. Since concrete provides the highest compressive strength and stiffness to cost ratio, the GFRP flange could be substituted by a layer of concrete. Such a layer will prevent the local buckling of the compression flange of the GFRP. A good bond between the concrete and the GFRP can be achieved by either using epoxy adhesives or installing mechanical shear connectors to the top GFRP surface. EXPERIMENTAL PROGRAM

The work proposed in this research focuses on the combination of FRP members with concrete to provide a composite action. Two types of composite construction will be investigated as shown in Figure 1. The first type is a fully encased beam which rely on the natural bond between concrete and FRP without additional anchorage. The second type is a FRP beam with mechanical anchorage to concrete slabs in the form of double nut bolts as shear connectors. Composite action reduces the flexural stresses and increases the flexural stiffness. Consequently, the proposed composite action eliminates the disadvantage of high deformation obtained in the case when FRP rebars are used.

FRP beam mechanically anchored to concrete slab

Fully encased FRP beam without additional anchorage Figure 1. Types of composite construction

718 FRPRCS-6: FRP Structural Shapes

Test Specimens

Three large scale specimens were tested. The overall length of the specimens was 2.25 m. Specimen SPl, the control specimen, is a pultruded FRP wide-flange I-beam with flange dimensions of 150 x 6 mm and web dimensions of 150 x 6 mm. Specimen SP2 consists of a pultruded FRP wide-flange I-beam similar to that of SP1 but with a top concrete slab of thickness 60 mm and width of 600 mm. The slab reinforcement consists of 6 mm diameter steel bars at 200 mm in both directions. The slab was connected to the FRP section using two rows of 16 mm diameter double nut bolts at a uniform spacing of 150 mm throughout the span of the FRP beam. The height of the bolts above the FRP flange is 40 mm. A total of four FRP stiffeners (6 mm thick) were mounted on the FRP beam using polyester resin; two stiffeners at each support point. The third specimen SP3 consists of a pultruded FRP wide-flange I-beam similar to that of SPl fully encased in a concrete T-beam of beam dimensions 210 x 210 mm and top slab of dimensions 60 x 600 mm. The slab reinforcement consists of 6 mm diameter steel bars at 200 mm in both directions. A concrete slab similar to that used in specimen SP2 but without the FRP beam was tested to study the effect of the composite action on it. Figures 2 and 3 show the dimensions, loading and support system for Specimens SP2 and SP3, respectively. Figure 4 shows the test setup for specimens SPl and SP2. The specimens were tested in four-point bending over a 2.05 m simply supported span. Load control was employed and the loading was applied in increments of approximately 7% of the expected failure load. The loading was continued up to failure of the specimen. In all specimens, electrical strain gauges were mounted on the top and bottom surfaces of the FRP beam at mid-span section while in specimens SP2 and SP3, extra strain gauges were mounted on the concrete top surface. The deflection was measured by means of LVDT placed at the mid-span section.

zyx

Section x-x

100 1

683mm

zyxwv zyxw zy zy u

684

V

4

683

JQ

114

Figure 2. Specimen SP2 dimensions and loading and supporting system

zy

Pultruded FRP Section with Concrete Slab 719

PI2 zyxwvutsrqponmlkjihgfed

zyxwvut

zyxw zyxwvut J

I ,

f

195 c 210

l

4

,

1

195

4

Sectionx-x

1%'

683 mm

zyxwvu u

684

Y

4

687

JW

4 4

Figure 3. Specimen SP3 : dimensions and loading and supporting system

Figure 4. Test set-up of specimens SPl and SP2

Material Properties The reinforcing bars used as longitudinal and transverse reinforcement in the slab are of yield strength of 275 MPa. The concrete cylinder compressive strength is 28 MPa. The pultruded GFRP structural shape is a wide-flange 150 x 6 mm beam. The mechanical properties of the FRP structural shapes are given in Table I , Table 1. Mechanical and Physical properties of FRP structural shapes

Mechanical properties

Physical properties

Tensile and compressive strength (MPa) Shear strength (MPa) Flexural strength (MPa) Tensile and compressive modulus (MPa) Flexural modulus (MPa) Full section Modulus of Elasticity (MPa) Poisson's ratio (MPa) Density ( p / m m ' ) Coeflcient of thermal expansion (mm/mmPC)

207 31 207 17200 11000 17200 0.33 0.002 8x

720 FRPRCS-6: FRP Structural Shapes zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQP

zyxwv zy

TEST RESULTS AND DISCUSSIONS

The mid-span deflection, strains and crack propagation were recorded at different stages of loading till the failure of the specimens. Behaviour of Test Specimens

A summary of the experimental results is presented in Table 2. It includes the loads and deflections at ultimate (failure) stage in addition to the modes of failure and the weight of the specimens per unit length. Figure 5 shows the local buckling of the compression flange of Specimen SPl and Figure 6 shows the failure of Specimen SP2. Table 2 . Failure loads, deflections and mode of failure of test specimens

Ultimate Specimen Specimen ultimate deflection weight designation load (IN) (mm) (kN/m) SPl

42.1

36.8

0.06

SP2

98.3

33.6

0.92

SP3

111.3

63.4

1.68

Figure 5. Local bucking of the compression flange of specimen SPI

Modes of failure local bucking of the comuression flanpe web-flangejunction delamination and local web buckling due to shear concrete crushing

Figure 6. Failure of specimen SP2 showing delamination in flange and buckling of web

From Table 2, Figures 5 and 6 as well as the observed behaviour of the test specimens, the following remarks could be concluded:

Pultruded FRP Section with Concrete Slab 721

(a) In the case of Specimen SP1, the FRP beam was loaded until the

(b)

(c)

(d)

(c)

zyxwv zyxw

compression flange developed wave like deformations along the length due to local buckling. The flange deformations were large, often greater than the thickness of the flanges. The local buckling load was taken as the failure criterion. In the case of specimen SP2, cracking in the concrete slab started at a taken load of 45 kN. A slight noise was heard at a load of 58 kN. Separation between the FRP top flange and the FRP web started at a load of 70 kN. Specimen SP2 failed due to delamination at the interface between the top flange and web in the shear zone followed by buckling and compression failure of the web due to shear. Shear connectors eliminated the possibility of bond failure between concrete and FRP beam. In the case of Specimen SP3, primary flexural crack was first visible at the mid-third region at about 10 % of the ultimate load. A slight noise was heard at a load level of 52 kN. Specimen SP3 failed due to crushing of concrete near the slab bottom surface. Wide cracks appeared at a load of 70 kN in Specimen SP3. This is attributed to the sliding between the concrete slab and the FRP beam which prevented the development of monolithic flexural action in the longitudinal direction. In order to achieve monolithic flexural action between the concrete slab and the FRP beam, a mechanical bond needs to be established. This proposed bond improvement is under current investigation.

Load-Deflection Behaviour The applied load versus central-deflection of the test specimens is presented in Figure 7 from which the following remarks could be concluded: (a) The total applied load in Specimen SP1 increased linearly with deflection till a load level of 27 kN where the compression flange started to buckle. A linear behaviour is observed even after local compression flange buckling occurred. (b) The stiffness of the composite beam Specimen SP2 is higher than that of the control Specimen SP1 at all loading stages. This is attributed to the composite action between the concrete slab and the FRP beam which improved the strength and the stiffness of the beam. The addition of shear connectors does not only serve to improve the strength of the FRP beam but also leads to a stiffer specimen.

zyxwvutsr

722 FRPRCS-6: FRP Structural Shapes

(c) Although the initial stiffness of Specimen SP3 before cracking is higher than that of Specimen SP2, the latter shows higher stiffness than Specimen SP3 after cracking until failure. (d) At a load level of 53 kN, the stiffness of Specimen SP3 decreased due to the slippage between the FRP beam and the concrete. (e) At a load level of 70 kN, the stiffness of Specimen SP2 decreased due to the start of separation between the FRP top flange and the FRP web in the shear zone.

zyxwv

0

10

20

30

40

50

60

Central deflection (mm)

Figure 7. Load-central deflection relationship of the tested specimens

Concrete and FRP Strains The longitudinal strain was recorded at various locations in the specimens at different load levels. The total applied load versus strain curves, at midspan section, is displayed in Figures 8 and 9 of specimens SPl and SP2 and SP3, respectively. The longitudinal strain profile at various load levels at mid-span section is shown in Figures 10 and 11 for Specimens SP2 and SP3, respectively. From these figures the following remarks could be concluded: (a) In Specimen SP1, the strains in the bottom flange and top flange were approximately equal, indicating that the neutral axis is in the middle. At a load level of approximately 27 kN, local buckling started in the top flange causing larger strains which can accelerate the failure of the compression flange. (b) At all load levels, lower longitudinal FRP strains are noted in the composite Specimen SP2 to those in the control specimen SP1. This significant reduction in strain corresponds to a higher load capacity, which arises due to the composite action of the slab with the FRF'

Pultruded FRP Section with Concrete Slab 723

z

beam. The composite actions increased the effective depth of the specimen and shifted the neutral axis towards the concrete slab. c) Significant increase in strains in the FRP upper flange in Specimen SP3 occurred after the slippage of the FRP beam at a load level of 53 kN. (d) In Specimen SP2, the neutral axis remains roughly at the same location near the concrete slab bottom surface until failure took place. An approximate linear profile is observed at lower load levels. Non linearity in the strain profile is noticed at higher load levels, which indicates that the shear connectors could not provide full composite action between the concrete slab and the FRP beam. (e) In Specimen SP3, an approximately linear profile is observed, showing a composite action between the FRP beam and the concrete slab before slippage took place at a load level of 53 kN. After slippage, the strains at the concrete surface slightly increased, shifting the neutral axis of the specimen away from the concrete slab.

-8 v

-a

120 110 100 90 80

70

zyxwvutsrqpon zyxwvutsrqpon zyxwvutsrqpo

60 so

40 30 20 10 0 4.006

4.004

0

4.002

0.002

O.OM

0.006

4.006

Strains of specimens SP1 B SPZ(mm/mm)

4.004

4.002

0

0.002

0.004

0.006

Strains of specimen SP3 (mmlmm)

Figure 8. Load-strain relationship of the specimens SPl and SP2

Figure 9. Load-strain relationship of the specimen SP3

210

210

180

180

g I50 E ; 120

E ; 120

154

0

0

290

g

90

$

$

60

>

60

30

30 0 -4000

-2000

0

2000

4000

6000

Longitudinal strain (microstrain)

Figure 10. Longitudinal strain profile at mid-span section of specimen SP2

0 4000

-2000

0

2000

4000

6000

8000

Longitudinalstrain (microstrain)

Figure 11. Longitudinal strain profile at midspan section of specimen SP3

724 FRPRCS-6: FRP Structural Shapes

zyxwv

CONCLUSIONS In this study, FRP-to-slab hybrid system is considered as an outstanding alternative to conventional slab systems. The system exhibited many advantages such as lightweight, easy and fast assembly without heavy lifting equipment, corrosion resistance and significant over-strength in the assembled system. However, the performance characteristics of the system need further investigation and development. The following conclusions can be deduced: (a) The tests revealed noticeable difference in flexural stiffness between the FRP beam and the FRP-to-slab hybrid systems. The difference is attributed to the composite action between the FRP beam and the concrete slab. (b) The overall response of the FRP-to-slab hybrid specimen was essentially elastic all way till reaching failure. The local web buckling failure mode indicated the possibility of further capacity increases when adding more web stiffeners to the FRP beam. (c) Despite the longitudinal debonding between the FRP beam and concrete in the FRP encased beam, the system capacity increased, thus, demonstrating the effectiveness of FRP encasement in concrete. The concrete surrounding the FRP beam prevents local buckling of its compression flange and web. ACKNOWLEDGEMENTS

zy

This research was funded by the scientific research council (grant No. 7-711/02), United Arab Emirates University and was conducted in the Department of Civil and Environmental Engineering laboratories. The efficient cooperation of Eng. Tarek Shaikhoun, Laboratory Engineer, in conducting the tests, is acknowledged. REFERENCES

1. Qiao P., Davalos J.F. and Wang J., "Local buckling of composite FRP shapes by discrete plate analysis," Journal of Structural Engineering, ASCE, Vol. 127, No. 3, March 2001, pp. 245-255. 2. C. Waldron, M. Hayes, E. Restrepo, T. Cousins and J. Lesko, "Determining the design parameters for an FRP girder bridge in Virginia," Proceedings of the 6* International Conference on short and medium span bridges, Vancouver, Canada, August 2002, pp. 1361-1368.

Author Index Abdelrahman, A. 663 Adhikary, B.B. 457 Agyei, B.B. 935 fit-Mokhtar, A. 833 AlChami,G. 623 Alhozaimy, A.M. 823 Al-Mahaidi, R. 247 Almusallam, T.H. 823 Al-Saidy, A.H. 1269 Al-Salloum, Y .A. 823 Alsayed, S.H. 823 Altan,M.O. 673 Alwis, K.G.N.C. 111 An,L. 995 Anderson, A.H. 1301 Araujo, A.F. 1003 Ara~jo,A.S. 477 Arora,D. 1067 Asakura, T. 1157 Ashraf,M. 457 Bakht, B. 923,945 Balafas, I. 1391 Balaguru, P. 367 Balendra, T. 1127 Balendran, R.V. 1047 Bank, L.C. 1067,1301 Banthia,V. 945 Barbato, M. 387 Benlloch, J. 337 Benmokrane, B. 737, 1291,1311, 1341 Biddah,A. 715 Bittencourt, T.N. 173

Blaschko, M. 205 Boulay, C. 913 Bousias, S.N. 527 Bradford, N.M. 705 Brikre, F. 1341 Burgoyne, C.J. 111,1013, 1391 Camata, G. 267,307 Carolin, A. 467, 1371 Carter, J.W. 1301 Casadei, P. 1097 Choi, M.C. 955 Chu,W. 759 Ciupala, M.A. 643, 1117 Clement, J.L. 913 Codato, D. 1239 Cosenza, E. 653,1361 Crawford, J.E. 1199 Dai, J.G. 143 Davies, J.M. 217 Davies, P. 347 De Lorenzis, L. 57 1,581,795, 975, 1351,1455 Dejke,V. 833 Delmas, Y. 497 Delpak,R. 347 Deng,Y. 875 Denton, S.R. 1147 DesgagnC, G. 1311, 1341 Desiderio, P. 843 Diagana, C.497 Dieter, D.A. 1301 Dietsche, J.S. 1301 Dos Santos, A.C. 173

Volume One: 1-724; Volume Two: 725-1464

zyxwvu

Ebead, U.A. 427,437 Ehrlacher, A. 407 El Maaddawy, T. 855 El-Hacha, R. 895 Elremaily, A.F. 79 El-Salakawy, E.F. 737, 1291, 1311, 1341 Ergun, B. 673 Erki,M.A. 895 Fakhri,P. 913 Fam,A.Z. 685 Fardis, M.N. 527 Feng,P. 1401 Ferracuti, B. 163 Ferreira, A.J.M. 695 Foret, G. 407 Foster, S.J. 1177 Fujisaki, T. 1435 Fukai,S. 1435 Fukuyama, H. 133,507,1435 Furuta, T. 133,507 Galati, N. 1219 Gale, L. 955 Gallagher, B. 1301 Gedalia, B. 497 Gettu, R. 173 Gonenc, 0. 1067 Gottardo, R. 1239 Grace, N.F. 1281 Grando, S. 1229 Gremel, D. 1067 Gu, X.L. 1107, 1259 Guadagnini, M. 517 Guan,H. 1381 Guglielmo, E. 1361 GuimarZes, G.B. 1003 Hadi, M.N.S. 487,613 Hamad, B.S. 633 Hanamori, N. 885 Harada,T. 89

Harajli, M.H. 633 Hashem, Y. 663 Hassan,T. 123 Hattori, A. 815,995 Hayashi, K. 885 He, W. 1157 Heffernan, P.J. 895 Hejll, A. 1371 Higuchi, T. 885 Hill, R.A. 1301 Hong, W.H. 1401 Huang, Y.H. 1107 Ibe,H. 227 Ibell, T.J. 539,955, 1097, 1147 Ichiryu,T. 885 Iervolino, I. 1361 Ikeda,A. 885 Ilki,A. 673 Ishikawa, T. 885 Ishiyama, S. 1037 Iwashita, K. 885 Janssens, J. 297 Jia, M. 875 Joh,O. 227 Kaku,T. 1445 Kanakubo, T. 133,507,1435 Karbhari, V.M. 759, 1381 Kassem, C. 1291 Keller, T. 1331 Kesse, G. 447 Khayat, K. 623 Khin, M. 89 Khomwan, N. 1177 Kirikoshi, K. 1037 Kishi, N. 287,327 Kishimoto, M. 865 Klaiber, F.W. 1269 Kobayashi, A. 865 Kobayashi, K. 1435 Koc,V. 673

Volume One: 1-724; Volume Two: 725-1464

Kojima, Y. 1157 Kong, K.H. 1127 Kubo,Y. 815 Kumbasar, N. 673 Kurihashi, Y. 287 La Tegola, A. 749,795,975, 1351 Labossikre, P. 779 Lackey, T. 1311 Lam, L. 99,601 Laoubi, K. 737 Lee,K. 247 Lees, J.M. 447,935 Leong, K.S. 257 Leung, H.Y. 1047 Li,A. 497 Li, J. 613 Liew, Y.S. 769 Ligozio, C.A. 79 Limam,O. 407 Limkatanyu, S . 307 Lin,L. 1401 Lopez, M.M. 317 Lu,M. 193 Lu, Z. 551,561 Luciani, P. 183 Maalej, M. 257 Manfredi, G. 653, 1209 Maqsood, T. 1047 Marcari,G. 1209 Marques, A.T. 695 Marzouk, H. 427,437 Masmoudi, R. 1341 Masuo, K. 1445 Matsui, S . 865 Matsuzaki., Y. 1445 Matthys, S. 297 Mazzoti, C. 163 McMonigal, D. 1067 Mehrabi, A.B. 79 Meier, U. 153, 1321

Melo, G.S. 477 Memon, A.H. 923 Micelli, F. 749,795, 1351 Migayama, T. 995 Mihashi, H. 1037 Mikami, H. 287,327 Miyagawa, T. 815 Modena, C. 1249 Monti, G. 183,387,1057 Morais, M.M. 1013 Morrill, K.B. 1199 Mortazavi, A.A. 643 Moulsdale, M. 613 Mufti, A.A. 923,945 Murakami,S. 885 Mutsuyoshi, H. 457 Myers, J.J. 749 Naaman, A.E. 3,25,317 Nagato, Y. 477 Nakai,H. 785 Nakan0.K. 1445 Nanni, A. 417,1097, 1147, 1219, 1229, 1455 Neale, K.W. 427,437,779 Nelson, B. 1067 Nilsson, L.O. 833 Nishimura, T. 785 Nishizaki, I. 779 Niu,H. 377 Nordin, H. 1077 Nurchi, A. 297 O’Regan, B. 527 Oh, H.S. 905 Oliva, M.G. 1301 Ouyang, Y. 1107,1259 OzeLM. 1067 Parra, C.J. 337 Patoary, M.K.H. 1189 Pecce, M. 1209 Pelletier, M.-A. 623

zyxwvu

zyxwvu Volume One: 1-724; Volume Two: 725-1464

Pigeon, M. 737 Pilakoutas, K. 5 17,643, 1117 Pimanmas, A. 277 Pornpongsaroj, P. 277 Porter, A.D. 1147 Poupard, 0. 833 Prota, A. 653 Qian, Z.Z. 1107 Rakib,T. 663 Renzelli, M. 183 Ribeiro, M.C.S. 695 Rizkalla, S.H. 123,685 Rousakis, T. 571,581 Russell, J.S. 1301 Russo, S. 1239 Sakai,H. 785 Santini, S. 1057 Saouma, V. 267 Sato, Y. 237,965 Savoia, M. 163 Sawada, S. 287 Sayed, G.A. 1281 Scarpa,M. 297 Schnerch, D.A. 685 Sen,R. 705 Shaaban, I. 663 Shaheen, H. 663 Sherping, R. 1137 Sim, J. 905 Smith, S.T. 193, 1023 Soudki, K. 855,1137 Spacone, E. 267,307 Spathis, L.A. 527 Stoecklin, I. 1321 Sugiyama, M. 727 Svecova, D. 945 Taerwe, L. 297 Tailhan, J.-L. 913 Takahashi, Y. 237 Takeda, T. 885

Taljsten, B. 467, 1077, 1167, 1371, 1425 Tamuis, V. 571,581 Tan, K.H. 769,985, 1087, 1127, 1189 Tan,K.Y. 417 Tanaka,M. 89 Tann, D.B. 347,357 Taranu, N. 1117 Teng,J. 591 Teng, J.G. 99,193,601 Tepfers, R. 571, 581, 833 Terrasi, G.P. 935 ThCriault, M. 623 Thomsen, H. 307 Tinazzi, D. 1249 Tjandra, R.A. 985 Toutanji, H. 367, 875 Triantafillou, T.C. 527 Tuladhar, R. 965 Tumialan, J.G. 417, 1219, 1229 Turco, V. 1219 Ueda, T. 51, 143,965 Ulaga, T. 153, 1415 Uomoto, T. 37,727,785 Utsunomiya, Y. 965 Valcuende, M. 337 Valerio, P. 539 Valluzzi, M.R. 1229, 1249 Van Zwol, T. 1137 Vanderpool, D.R. 79 Venkataramana, K. 89 Vogel, T. 153, 1415 Waldron, P. 5 17 Wang,Z. 227 Wight, R.G. 895 Wigum, B.J. 805 Wipf, T.J. 1269 Woods, S. 1301 Wu, G. 551,561

zyxwvut

Volume One: 1-724; Volume Two: 725-1464

zyx zyxwvut

Wu, G. 551,561 Wu,H.C. 591 Wu, Z. 377,551,561,885,1157 Wu, Z.J. 217 Wu,Z.Y. 913 Xiong, Z.H. 487 Yamada, K. 1037 Yamamoto, S. 815 Yamamoto, T. 995 Yang,T. 1401 Ye, F.F. 1259 Ye, L.P. 1401 Yin, J. 1157

You, C.S. 571,581 Yue, Q.R. 1401 Zehetmaier, G. 397 Zhang, G.F. 327 Zhang,K. 1401 Zhang, N. 1401 Zhang, W.P. 1259 Zhang,Y. 367 Zhao, H.D. 1087 Zhao,L. 1381 Zilch, K. 397 Zou, P.X.W. 1023

Volume One: 1-724; Volume Two: 725-1464

Edited by  Kiang Hwee  TAN 

Reinforcement for Concrete Structures VOLUME 2

Proceedings of the Sixth International Symposium on FRP Reinforcement for Concrete Structures (FRPRCS-6) ^FRPRCS 1003 Singapore

World Scientific

Reinforcement for Concrete Structures VOLUME 2

This page is intentionally left blank

Edited by  Kiang Hwee  TAN  National University of Singapore, Singapore

Singapore

8-10 July, 2003

Reinforcement for Concrete Structures

Proceedings of the Sixth International Symposium on F1P Reinforcement for Concrete Structures

VOLUME 2

(FRPRG-6)

Ililli •2#PJ Si^^pum i

^

World Scientific New Jersey • London

• Singapore

• Hong

Kong

Published by World Scientific Publishing Co. Pte. Ltd.  Toh Tuck Link, Singapore 596224  USA office:  Suite 202, 1060 Main Street, River Edge, NJ 07661  UK office: 

Shelton Street, Covent Garden, London WC2H 9HE 

British Library Cataloguing-m-Publication Data catalogue record for this book  available from the British Library. 

FIBRE-REINFORCED POLYMER REINFORCEMENT FOR CONCRETE STRUCTURES — (In Volumes) Proceedings the Sixth International Symposium on FRP Reinforcement for Concrete Structures (FRPRCS-6) Copyright 

2003 by World Scientific  Publishing Co. Pte. Ltd. 

any means, All rights reserved. This book, parts thereof, may not reproduced any form electronic or mechanical, including photocopying, recording any information storage and retrieval invented, without written permission from the Publisher. system now known

For  photocopying  material  this  volume,  please  pay  copying  fee  through  the  Copyright  Clearance Center, Inc., 222 Rosewood  Drive, Danvers, MA 01923, USA.  this case permission  photocopy  not required from  the publisher. 

ISBN  981­238­401­4  (Set)  ISBN  981­238­973­  (Vol.  ISBN  981­238­972­  (Vol.11) 

Printed  by  Fulsland  Offset  Printing  (S)  Pte  Ltd,  Singapore 

Preface Research  on  the  application  of  fibre­reinforced  polymer  (FRP)  as  reinforcement  for  concrete  structures  appeared  m as  early  as  the  1960s.  However,  was not until the late  1980s that such research has  escalated,  leading  to field  applications.  The  interest  in non­metallic  reinforcement  was fuelled  by the corrosion problem associated with steel  reinforcement  that  surfaced  around  the  world  at  that  time,  and  the  downturn  of  the  aerospace  industry  where  fibre­reinforced  polymers  have  been  widely  used  due  to  its  high  specific  strength  and  modulus,  and  other  superior  characteristics.  I was fortunate  to spend my sabbatical  with Professor  Naaman at the  University  of  Michigan,  USA,  during  the  Fall  and  Winter  of  1991  and  with Professor  Okamura at the University of Tokyo, Japan, during Spring  and Summer of  1992.  The former  introduced to me this new material that  has  since  fascinated  many  in  the  research  community  and  construction  industry  In  okyo, in particular,  was overwhelmed by the mountains of  research  that  were  embarked  by  universities,  public  institutions  and  private  companies  on  the  development  and  application  of  FRP  rods  as  reinforcement  for  concrete  structures.  There  were  round  bars,  flat  bars,  square bars, braided bars, sanded bars, strands, grids  and links, and  even  three­dimensional  reinforcement.  Several  applications  m  footbridges,  foundation  beams,  tunnel  linings,  and  floating  structures  suddenly  mushroomed  all  over  Japan  and  the  rest  of  the  world.  That  probably  constituted  the  first  era  in  the  application  of  FRP  reinforcement  in  concrete structures.  The FRPRCS Symposia Series was initiated in  1993, and subsequently  held every two years in the continents of America, Europe and Asia, on  rotational basis.  The previous symposia were held in Vancouver,  Canada  (1993), Ghent, Belgium (1995), Sapporo, Japan  (1997), Baltimore,  USA  (1999), and Cambridge, UK (2001).  This year marks the  10th anniversary  of the FRPRCS Symposia Series, and the Department of Civil Engineering  the  National  University  of  Singapore  honored  to  host  the  6th  International  Symposium on FRP Reinforcement  for  Concrete  Structures  (FRPRCS­6) in Singapore. 

The planning  and  preparation  works  for  FRPRCS­6  in  effect  began  almost  six  years  ago  in  1997  when  was  asked  in  Sapporo,  to  be  the  second  Asian  host  for  the  FRPRCS  International  Symposium.  At  that  time,  there  was  still  little  awareness  of  the  material  known  as  FRP  reinforcement  in  Asia  outside  Japan,  and  if  any  the  interests  were  centered  mainly  on  externally  bonded  FRP  systems  rather  than  FRP  reinforcing  rods.  The  Kobe  earthquake  in  1995  has  brought  about  rapidly  increasing  interests  in  the  use  of  FRP  systems  in  structural  rehabilitation,  and  that  marked  the  beginning  of  the  second  era  in  FRP  applications in concrete structures.  T  promote  awareness  and  interests  in  the  development  and  application  of FRP reinforcement  in Singapore and the region, the Fibre­ Reinforced  Society  (Singapore)  was formed  in  September  2002  and  has  since been  co­organizer of this Symposium.  The FRPRCS­6  International  Symposium  will  signify  the  beginning  of  the  third  era,  in  which  one  could  witness  global  interests  in  FRP  reinforcement,  as  well  as  the  use  of  FRP  reinforcements  as  structural  shapes,  and  in  masonry  and  steel  structures.  This  set  of  proceedings  contains  total  of  140 papers from  26 countries,  in two volumes.  Each  technical  paper  had  been  reviewed  and  selected  for  presentation  by  at  least two members  of the International  Scientific  Committee, to whom  I  would like to express my gratitude.  Volume  of  the  proceedings  contains  four  invited  keynote  papers  and  63 technical  papers  dealing  with:  (i) FRP  Materials  and  Properties;  (ii)  Bond  Behaviour;  (iii)  Externally  Bonded  Reinforcement  (EBR)  for  Flexure,  Shear  and  Confinement;  and  (iv)  FRP  Structural  Shapes.  The  topics  covered  in  Volume  2  are:  (v)  Durability  and  Maintenance;  (vi)  Sustained  and  Fatigue  Loads;  (vii)  Prestressed  FRP  Reinforcement  and  Tendons;  (viii)  Structural  Strengthening;  (ix)  Applications  in  Masonry  and  Steel  Structures;  (x)  Field  Applications  and  Case  Studies;  and  (xi)  Codes and Standards.  Seventy­three papers are included m Volume 2.  The  FRPRCS­6  International  Symposium  also  witnessed  the  formation  of the  International  Steering  Committee,  which comprises  the  chairmen  of  the  current  and  previous  FRPRCS  Symposia.  The  mam  purpose  of  this  Committee  to  chart  the  future  directions  for  the  Symposia Series,  has appointed  three­man taskforce to determine the  Best Paper (Research), Best Paper (Application)  and Honorable  Mention  Awards, which were first  introduced  at FRPRCS­6.  The three  gentlemen  in  the  taskforce  were  Professor  C.W  Dolan  from  USA,  Professor  F.S. 

VII 

Rostasy  from  Germany,  and  Professor  H.  Okamura  from  Japan.  All  of  them  are  well  known  in  the  areas  of  FRP  reinforcement  and  structural  concrete.  The  organization  of  the  Symposium  would  not  have  been  possible  without  the  generous  contributions  from  the  sponsors,  who  are  degussa­ MBT (S) Pte Ltd, Fyfe Asia Pte Ltd, Mapei Far East Pte Ltd, Sika (S) Pte  Ltd,  Lee  Foundation  (Singapore)  and  Defence  Science  &  Technology  Agency.  I would  also like to express my sincere thanks to the  American  Concrete  Institute,  USA,  Institution  of  Engineers,  Singapore,  Japan  Concrete  Institute, Japan, and The  Concrete  Society, UK,  for  supporting  the event.  Last,  but  not  least,  I  would  like  to  acknowledge  the  help  of  my  colleagues,  in  particular,  Balendra,  Mansur  and  Maalej,  and  the  Secretariat,  comprising  Christine,  Siti  and  Sarimah,  who  have  devoted  many hours in getting the Symposium organized. 

Kiang Hwee Tan Singapore  July 2003 

FRPRCS­6  Organizing  Committees  National University of Singapore Organizing Committee Chairman  Members  Secretariat 

K.H. Tan  T. Balendra, M.A. Mansur, M. Maalej  C.S. Tan, Siti Rohani, Sarimah 

International Steering Committee S.H. Rizkalla, USA  L. Taerwe, Belgium  K.H. Tan, Singapore  T. Uomoto, Japan 

CJ.  Burgoyne, UK  C.W. Dolan, USA  A. Nanni, USA  H. Okamura, Japan 

International Scientific Committee K.H. Tan, Singapore  (Chairman)  K.W. Neale, Canada  L.C. Bank, USA  K. Pilakoutas, UK  B. Benmokrane, Canada  S.H. Rizkalla, USA  C.J. Burgoyne, UK  J. Sim, Korea  E. Cosenza, Italy  R.N. Swamy, UK  C.W. Dolan, USA  L. Taerwe, Belgium  G.B. Guimaraes, Brazil  J.G. Teng, China  M. Harajli,  Lebanon  R. Tepfers,  Sweden  P. Hamelin, France  T. Ueda, Japan  L. Hollaway, UK  T. Uomoto, Japan  G. Manfredi,  Italy  P. Waldron, UK  K. Maruyama, Japan  Z. Wu, Japan  U. Meier,  Switzerland  Q.R. Yue, China  A.E. Naaman, USA  A. Nanni, USA 

VIM

Contents VOLUME 1 KEYNOTE PAPERS  FRP Reinforcements in Structural Concrete: Assessment, Progress, and Prospects A.E. Naaman 

3 

Progress and Prospects of FRP Reinforcements: Survey of Expert Opinions A.E. Naaman 

25 

Durability Design of GFRP Rods for Concrete Reinforcement T. Uomoto 

37 

New Types of Continuous Fiber Reinforcements for Concrete Members T.Ueda 

51 

FRP  MATERIALS AND  PROPERTIES  Performance of Thermoplastic Fiber Reinforced Polymer Rebars A.B. Mehrabi, C.A. Ligozio, A.F. Elremaily and D.R.  Vanderpool 

79 

Experimental Study on Poisson's Ratio for FRP Tendons M. Tanaka, M. Khin, T. Harada and K. Venkataramana 

89 

Stress-Strain Model for FRP-Confined Concrete for Design Applications L. Lam and J.G. Teng 

99 

IX

Accelerated Techniques to Predict the Stress-Rupture Behaviour of Aramid Fibres (Best Paper - Research) K.G.N.C. Alwis and C.J. Burgoyne 

111 

BOND BEHAVIOUR  Bond Characteristics of Various FRP Strengthening Techniques S.H. Rizkalla and T. Hassan 

123 

Bond Strength between Fiber-Reinforced Polymer Laminates and Concrete T. Kanakubo, T. Furuta and H. Fukuyama 

133 

Local Bond Stress-Slip Relations for FRP Sheets-Concrete Interfaces (Best Paper - Research) J.G. Dai and T. Ueda 

143 

Bilinear Stress - Slip Bond Model: Theoretical Background and Significance T. Ulaga, T. Vogel and U. Meier 

153 

Non Linear Bond-Slip Law for FRP-Concrete Interface M. Savoia, B. Ferracuti and C. Mazzoti 

163 

Experimental Analysis of Interface between CFRP and Concrete using Cylindrical Specimens A.C. Dos Santos, T.N. Bittencourt and R. Gettu 

173 

FRP Adhesion in Uncracked and Cracked Concrete Zones G. Monti, M. Renzelli and P. Luciani 

183 

Neural Network Prediction of Plate End Debonding in FRPPlated RC Beams S.T. Smith, J.G. Teng and M. Lu 

193 

Bond Behaviour of CFRP Strips Glued into Slits M. Blaschko 

205 

XI

EXTERNALLY BONDED REINFORCEMENT FOR  FLEXURE  Load Capacity of Concrete Beams Strengthened with External FRP Sheets Z.J. Wu and J.M. Davies 

217 

Reinforcing Effects of CFRP and AFRP Sheets with Respect to Flexural Behaviour of RC Beams O. Joh, Z. Wang and H. Ibe 

227 

Flexural Behaviour of RC Beams Externally Reinforced with Carbon Fiber Sheets Y. Takahashi and Y. Sato 

237 

Strength and Failure Mechanism of RC T-Beams Strengthened with CFRP Plates K. Lee and R. Al­Mahaidi 

247 

Effect of Beam Size on Interfacial Shear Stresses and Failure Mode of FRP-Bonded Beams K.S. Leong and M. Maalej 

257 

Debonding Failure of RC Structural Members Strengthened with FRP Laminates G. Camata, E. Spacone and V. Saouma 

267 

Effect of End Wrapping on Peeling Behaviour of FRP-Strengthened Beams P. Pornpongsaroj  and A. Pimanmas 

277 

An Experimental Study on Debond-Control of AFRP for Flexurally Strengthened RC Beams S. Sawada, N. Kishi, H. Mikami and Y. Kurihashi 

287 

Tests on RC T-Beams Strengthened in Flexure with a Glued and Bolted CFRP Laminate A. Nurchi, S. Matthys, L. Taerwe, M. Scarpa and  J. Janssens 

297 

XII

Parametric Studies of RC Beams Strengthened in Flexure with Externally Bonded FRP S. Limkatanyu, H. Thomsen, E. Spacone and G. Camata 

307 

Concrete Cover Failure or Tooth Type Failure in RC Beams Strengthened with FRP Laminates M.M. Lopez and A.E. Naaman 

317 

Influence of Material Properties of FRPs on Strength of Flexural Strengthened RC Beams G.F. Zhang, N. Kishi and H. Mikami 

327 

Ductility of Reinforced Concrete Beams Strengthened with CFRP Strips and Fabric M. Valcuende, J. Benlloch and C.J. Parra 

337 

A Review of Ductility Determination of FRP Strengthened Flexural RC Elements D.B.Tann,P. DaviesandR.  Delpak 

347 

A Semi-Empirical Approach for the Prediction of Deflections of FRP Strengthened RC Slabs D.B. Tann 

357 

Crack Widths in RC Beams Externally Bonded with CFRP Sheets Y. Zhang, H. Toutanji  and P. Balaguru 

367 

Numerical Simulations for Strengthened Structures with Hybrid Fiber Sheets H. Niu and Z. Wu 

377 

Fibre-Section FE of FRP-Strengthened RC Beam in Flexure, Shear and Confinement G. Monti and M. Barbato 

387 

Interaction between Internal Bars and External FRP Reinforcement in RC Members G. Zehetmaier and K. Zilch 

397 

XIII

Strengthening of RC Two-Way Slabs with Composite Materials O. Limam, G. Foret and A. Ehrlacher 

407 

Evaluation of Externally Bonded CFRP Systems for the Strengthening of RC Slabs K.Y. Tan, J.G. Tumialan and A. Nanni 

417 

Flexural Strengthening of Two-Way Slabs Using FRPs H. Marzouk, U.A. Ebead and K.W. Neale 

427 

Tensile Properties of Concrete in FRP Strengthened Two-Way Slabs H. Marzouk, U.A. Ebead and K.W. Neale 

437 

EXTERNALLY BONDED REINFORCEMENT FOR  SHEAR  Shear Critical RC Beams Strengthened with CFRP Straps G. Kesse and J. M. Lees 

447 

Effective Shear Strengthening of Concrete Beams using FRP Sheets with Bonded Anchorage B.B. Adhikary, H. Mutsuyoshi and M. Ashraf 

457 

Behaviour of Concrete Structures Strengthened in Shear with CFRP A. Carolin and B. Taljsten 

467 

Strengthening of RC T Beams in Shear with Carbon Sheet Laminates (CFRP) G.S. Melo, A.S. Araujo  and Y. Nagato 

477 

Strength Analysis of Sheared Beams Retrofitted with Strengthening Materials Z. H. Xiong and M.N.S. Hadi 

487 

Shear Performance with Externally Bonded Carbon Fibre Fabrics A. Li, C. Diagana, Y. Delmas and B. Gedalia 

497 

XIV 

Evaluation of Shear Capacity of RC Columns Strengthened by Continuous Fiber T. Furuta, T. Kanakubo and H. Fukuyama 

507 

Shear Design Equations for FRP RC Beams M. Guadagnini, K. Pilakoutas and P. Waldron 

517 

Strengthening of Corrosion-Damaged RC Columns with FRP S.N. Bousias, T.C. Triantafillou,  M.N. Fardis,  L.A. Spathis and B. O'Regan 

527 

Shear Strengthening of Concrete Bridge Decks using FRP Bar P. Valerio and T.J. Ibell 

539 

EXTERNALLY BONDED REINFORCEMENT FOR CONFINEMENT  Stress-Strain Relationship for FRP-Confined Concrete Cylinders G. Wu, Z. Lu and Z. Wu 

551 

Stress-Strain Relationship for FRP-Confined Concrete Prisms G. Wu, Z. Wu and Z. Lu

561

Concrete Cylinders Confined by CFRP Sheets Subjected to Cyclic Axial Compressive Load T. Rousakis, C.S. You, L. De Lorenzis, V. Tamuzs and  R. Tepfers 

571 

Concrete Cylinders Confined by Prestressed CFRP Filament Winding under Axial Compressive Load T. Rousakis, C.S. You, L. De Lorenzis, V. Tamuzs and  R. Tepfers 

581 

Concrete Confined with Fiber Reinforced Cement Based Thin Sheet Composites H.C. Wu and J. Teng 

591 

XV

Hoop Rupture Strains of FRP Jackets in FRP-Confined Concrete L. Lam and J.G. Teng

601

Externally Confined High Strength Concrete Columns under Eccentric Loading J. Li, M. Moulsdale and M.S.N. Hadi 

613 

Creep Performance of CFRP Confined Concrete Cylinders M. Theriault, M.­A. Pelletier, K. Khayat and G. Al Chami 

623 

Development/Splice Strength of Steel Bars in Concrete Confined with CFRP Sheets M.H. Harajli and B.S. Hamad 

633 

Lateral Prestressing of RC Columns with FRP Jackets A.A. Mortazavi, K. Pilakoutas and M.A. Ciupala 

643 

Confinement of RC Rectangular Columns Using GFRP A. Prota, G. Manfredi  and E. Cosenza 

653 

Behaviour of RC Columns Retrofitted by Fibre Reinforced Polymers under Cyclic Loads H. Shaheen, T. Rakib, Y. Hashem, I. Shaaban and  A. Abdelrahman 

663 

Photogrammetrically Measured Deformations of FRP Wrapped Low Strength Concrete A. Ilki, V. Koc, B. Ergun, M.O. Altan and N. Kumbasar 

673 

FRP STRUCTURAL SHAPES Rectangular FRP Tubes Filled with Concrete for Beam and Column Applications A.Z. Fam, D.A. Schnerch and S.H. Rizkalla 

685 

Flexural Behaviour of GFRP-Polymer Concrete Hybrid Structural Systems M.C.S. Ribeiro, A.J.M. Ferreira and A.T. Marques 

695 

XVI 

A New Concept for an FRP Panelized Rapid Deployment Shelter N.M. Bradford  and R. Sen 

705 

Experimental Investigation of Pultruded FRP Section Combined with Concrete Slab A.Biddah 

715 

VOLUME TWO DURABILITY AND MAINTENANCE  Research on Strength and Durability of GFRP Rods for Prestressed Concrete Tendons M. Sugiyama and T. Uomoto 

727 

Durability of Concrete Beams Reinforced with GFRP Bars under Different Environmental and Loading Conditions K. Laoubi, E.F. El­Salakawy, B. Benmokrane and  M. Pigeon 

737 

Environmental Effects on RC Beams Strengthened with Near Surface Mounted FRP Rods F. Micelli, A. La Tegola and J.J. Myers 

749 

Synergistic Hydrothermal Effects on Durability of E-Glass Vinylester Composites W. Chu and V.M. Karbhari 

759 

Durability of GFRP Composites under Tropical Climate Y.S. Liew and K.H. Tan 

769 

Effects of Different Long-Term Climatic Conditions on FRP Durability P. Labossiere, K.W. Neale and I. Nishizaki 

779 

XVII

Durability of Aramid and Carbon FRP PC Beams under Natural and Accelerated Exposure H. Nakai, H. Sakai, T. Nishimura and T. Uomoto 

785 

Effects of Wet Environment on CFRP-Confined Concrete Cylinders F. Micelli, L. De Lorenzis, and A. La Tegola 

795 

Alkali Aggregate Reactive Mortar Cylinders Partly Restrained by External CFRP Fabric B.J. Wigum 

805 

ASR Expansion Reduction and Ductility Improvement by CFRP Sheet Wrapping A. Hattori, S. Yamamoto, T. Miyagawa and Y. Kubo 

815 

Durability of GFRP Rebars in Concrete Beams under Sustained Loads at Severe Environments T.H. Almusallam, Y.A. Al­Salloum, S.H. Alsayed and A.M.  Alhozaimy 

823 

Influence of Sustained Stress on the Durability of GFRP Bars Embedded in Concrete V. Dejke, O. Poupard, L.O. Nilsson, R. Tepfers  and  A. Ait­Mokhtar 

833 

A Maintenance Strategy for FRP Strengthening Systems P. Desiderio 

843 

SUSTAINED AND FATIGUE LOADS  Viability of using CFRP Laminates to Repair RC Beams Corroded under Sustained Loads T. El Maaddawy and K. Soudki 

855 

Fatigue Bond of Carbon Fiber Sheets and Concrete in RC Slabs Strengthened by CFRP A. Kobayashi, S. Matsui and M. Kishimoto 

865 

XVIII

Fatigue Performance of RC Beams Strengthened with CF Sheets Bonded by Inorganic Matrix H. Toutanji, Y. Deng and M. Jia 

875 

Fatigue Performance of RC Beams Strengthened with Externally Prestressed PBO Fiber Sheets Z. Wu, K. Iwashita, T. Ishikawa, K. Hayashi,  N. Hanamori, T. Higuchi, A. Ikeda, T. Takeda,  S. Murakami and T. Ichiryu 

885 

Prestressed CFRP Sheets for Strengthening Reinforced Concrete Structures in Fatigue R. El­Hacha, R.G. Wight, P.J. Heffernan  and M.A. Erki 

895 

Fatigue Behaviour of Bridge Deck Specimen Strengthened with Carbon Fiber Polymer Composites J. Sim and H.S. Oh 

905 

Static and Fatigue Tests on Precracked RC Beams Strengthened with CFRP Sheets Z.Y. Wu, J.L. Clement, J.­L. Tailhan, C. Boulay and  P. Fakhri 

913 

Fatigue Investigation of Concrete Bridge Deck Slab Reinforced with GFRP and Steel Strap A.H. Memon, A.A. Mufti  and B. Bakht 

923 

PRESTRESSED FRP  REINFORCEMENT AND TENDONS  Fatigue of High Strength Concrete Beams Pretensioned with CFRP Tendons B.B. Agyei, J.M. Lees and G.P. Terrasi 

935 

Transverse Confinement of Deck Slabs by Concrete Straps V. Banthia, A.A. Mufti, D. Svecova and B. Bakht 

945 

Design of Anchorage Zones for FRP-Prestressed Concrete T.J. Ibell, L. Gale and M.C. Choi 

955 

XIX

A Simple Continuous System of Shear Reinforcement with Polyacetal Fiber R. Tuladhar, Y. Utsunomiya, Y. Sato and T.Ueda 

965 

Analytical Modeling of Splitting Bond Failure for NSM FRP Reinforcement in Concrete L. De Lorenzis and A. La Tegola 

975 

Strengthening of RC Beams with External FRP Tendons: Tendon Stress at Ultimate R.A. Tjandra  and K.H. Tan 

985 

Comparative Analysis on Stress Calculation Methods for External FRP Cables L. An, T. Yamamoto, A. Hattori and T. Migayama 

995 

Moment Redistribution in Continuous Monolithic and Segmental Concrete Beams Prestressed with External A ramid Tendons A.F. Araujo  and G.B. Guimaraes 

1003 

Experimental Investigation on the Ductility of Beams Prestressed with FRP M.M. Morais and C.J. Burgoyne 

1013 

Time-Dependent Flexural Crack Width Prediction of Concrete Beams Prestressed with CFRP tendons P.X.W. Zou and S.T. Smith 

1023 

STRUCTURAL STRENGTHENING  Multiscale Reinforcement Concept for Employment of Carbon Fiber Woven Mesh K. Yamada, S. Ishiyama, H. Mihashi and K. Kirikoshi 

1037 

Woven Composite Fabric to Strengthen Structurally Deficient RC Beams H.Y. Leung, R.V. Balendran and T. Maqsood 

1047 

XX

Calibration of Partial Safety Coefficients for FRP Strengthening G. Monti and S. Santini 

1057 

Comparison between FRP Rebar, FRP Grid and Steel Rebar Reinforced Concrete Beams M. Ozel, L.C. Bank, D. Arora, O. Gonenc, D. Gremel,  B. Nelson and D. McMonigal 

1067 

Concrete Beams Strengthened with Pre-Stressed Near Surface Mounted Reinforcement H. Nordin and B. Talj sten 

1077 

Strengthening of One-Way RC Slabs with Openings using CFRP Systems H.D. Zhao and K.H. Tan 

1087 

Experimental Results of One-Way RC Slabs with Openings Strengthened with CFRP Composites P. Casadei, T.J. Ibell and A. Nanni 

1097 

Seismic Behaviour of Reinforced Concrete Beam-Column Joint Strengthened with GFRP Y. Ouyang, X.L. Gu, Y.H. Huang and Z.Z. Qian 

1107 

FRP Seismic Strengthening of Columns in Frames M.A. Ciupala, K. Pilakoutas and N. Taranu 

1117 

Retrofitting of Shear Walls Designed to BS8110 for Seismic Loads using FRP K.H. Kong, K.H. Tan and T. Balendra 

1127 

Strengthening of Interior Slab-Column Connections with CFRP Strips K. Soudki, T. Van Zwol and R. Sherping 

1137 

Effectiveness of FRP Plate Strengthening on Curved Soffits A.D. Porter, S.R. Denton, A. Nanni and T.J. Ibell 

1147 

XXI

Strengthening Performance of FRP Sheets Bonded to Concrete Tunnel Linings Z. Wu, W. He, J. Yin, Y. Kojima  and T. Asakura 

1157 

Strengthening of Concrete Structures in Torsion with FRP B.Taljsten 

1167 

FE Modelling of FRP-Repaired RC Plane Stress Elements N. Khomwan and S.J. Foster 

1177 

APPLICATIONS IN MASONRY AND STEEL STRUCTURES  Blast Resistance of Prototype In-Built Masonry Walls Strengthened with FRP Systems (Honourable Mention) M.K.H. Patoary and K.H. Tan 

1189 

Retrofit Techniques using Polymers and FRPs for Preventing Injurious Wall Debris J.E. Crawford  and K.B. Morrill 

1199 

Experimental Behaviour of Masonry Panels Strengthened with FRP Sheets G. Marcari, G. Manfredi  and M. Pecce 

1209 

Flexural Strengthening of URM Walls with FRP Systems V. Turco, N. Galati, J.G. Tumialan and A. Nanni 

1219 

Shear Strengthening of URM Clay Walls with FRP Systems S. Grando, M.R. Valluzzi, J. G. Tumialan and A. Nanni 

1229 

Effect of FRP Mesh Reinforcement on Shear Capacity and Deformability of Masonry Walls S. Russo, R. Gottardo and D. Codato 

1239 

Strengthening of Masonry Structures under Compressive Loads by using FRP Strips M.R. Valluzzi, D. Tinazzi and C. Modena 

1249 

Seismic Behaviour of Masonry Structural Walls Strengthened with CFRP Plates X.L. Gu, Y. Ouyang, W.P. Zhang and F.F. Ye 

1259 

Advanced Composite Materials for the Repair of Steel Structures A.H. Al­Saidy, T.J. Wipf and F.W. Klaiber 

1269 

FIELD APPLICATIONS AND CASE STUDIES  Construction and Evaluation of Full-Scale CFRP Prestressed Concrete DT-Girder N.F. Grace and G. A. Sayed 

1281 

Flexural Behaviour of Bridge Deck Slabs Reinforced with FRP Composite Bars E.F. El­Salakawy, C. Kassem and B. Benmokrane 

1291 

Details and Specifications for a Bridge Deck with FRP Framework, Grid and Rebar L.C Bank, M.G. Oliva, J.S. Russell, D.A. Dieter,  J.S. Dietsche, R.A. Hill, B. Gallagher, J.W. Carter,  S. Woods and A.H. Anderson 

1301 

Construction, Testing and Monitoring of FRP RC Bridges in North America B. Benmokrane, E.F. El­Salakawy, G. Desgagne and  T. Lackey 

1311 

Strengthening of Concrete Structures with Prestressed and Gradually Anchored CFRP Strips {Best Paper - Application) I. Stoecklin and U. Meier 

1321 

Strengthening of Concrete Bridges with Carbon Cables and Strips T.Keller 

1331 

New Corrosion-Free Concrete Bridge Barriers Reinforced with GFRP Composite Bars E.F. El­Salakawy, R. Masmoudi, B. Benmokrane,  F. Briere and G. Desgagne 

1341 

XXIII

Strengthening of Steel Silos with Post-Tensioned CFRP Laminates L. De Lorenzis, F. Micelli and A. La Tegola 

1351 

Seismic Performance Improvement of the Bell Tower in Serra S. Quirico by Composites E. Cosenza, I. Iervolino and E. Guglielmo 

1361 

Strengthening with CFRP under Simulated Live Loads A. Hejll, A. Carolin and B. Taljsten 

1371

Composite Structural Systems - From Characterization to Field Implementation V.M. Karbhari, H. Guan and L. Zhao 

1381 

Optimal Cost Design for Beams Prestressed with FRP Tendons I. Balafas  and C.J. Burgoyne 

1391 

FRP in Civil Engineering in China: Research and Applications L.P. Ye, P. Feng, K. Zhang, L. Lin, W.H. Hong, Q.R. Yue,  N. Zhang and T. Yang 

1401 

CODES AND STANDARDS  Design Concepts of the New Swiss Code on Externally Bonded Reinforcement T. Vogel and T. Ulaga 

1415 

Design Guideline for CFRP Strengthening of Concrete Structures B. Taljsten 

1425 

Design Practice of Framed Building Structure Based on AIJ Design Guideline 2002 K. Kobayashi, H. Fukuyama, T. Fujisaki,  S. Fukai and  T. Kanakubo 

1435 

XXIV

Evaluations of Continuous Fiber Reinforced RC Members based on AIJ Design Guildeline 2002 K. Nakano. Y. Matsuzaki., T. Kaku and K. Masuo 

1445 

Design Procedure of NSM FRP Reinforcement for Strengthening of RC Beams L. De Lorenzis and A. Nanni 

1455 

Durability  and  Maintenance 

This page is intentionally left blank

FRPRCS-6, Singapore, 8­10  July 2003  Edited by Kiang Hwee Tan  ©World Scientific Publishing Company 

RESEARCH ON STRENGTH AND DURABILITY OF GFRP  RODS FOR PRESTRESSED CONCRETE  TENDONS  M. SUGIYAMA  ARG Development Department, Nippon Electric Glass Co.,Ltd. 906 Ima, Notogawa, Kanzaki, Shiga, Japan T. UOMOTO  Institute of Industrial Science, University of Tokyo 4-6-IKomaba, Meguro, Tokyo, Japan Many  concrete structures  affected  by chloride  ions have been  deteriorated  remarkably  due  to  the  corrosion  of  steel  bars  embedded  in  concrete  structures. This problem occurs not only in Japan but also in other countries.  The  deterioration  of  concrete  structures  located  at  the  coast  and  bridges  spread  with  de­icing  salt  is  a  big  concern.  To  deal  with  the  problem  of  prestressed  concrete  structures constructed  under environment  of chloride­ induced  corrosion,  many  attempts  were  performed  to  utilize  cathodic  protection,  epoxy  coated  reinforcing  steel  etc.  Among  many  attempts,  utilization  of  fiber  reinforced  plastics  (FRP)  was  evaluated  as  a  good  method to deal with the problem drastically, because it does not corrode in  a chloride environment. In Japan, many researches have been performed  to  utilize  FRP  rods  as  concrete  reinforcement,  especially  for  prestressed  tendons  since  1980. Many  researches  have  been  reported  on  strength  and  durability of FRP rods made of glass fiber, aramid fiber, and carbon fiber.  In  this  study,  ultimate  goal  is that  glass fiber reinforced  plastics  (GFRP)  rods  are  utilized  as  prestressed  concrete  tendons.  Regarding  GFRP  rods  made  of  new  glass  fiber,  physical  properties  and  durability  were  investigated.  INTRODUCTION There  is  a  high  possibility  of  the  utilization  of  FRP  rods  in  the  field  of  concrete.  They  have  characteristics  such  as  high  strength,  high  corrosion  resistance, light weight and  non­magnetic property. However,  in the case  of  GFRP  rods,  alkali  resistance,  cyclic  fatigue  properties,  and  static  fatigue  properties  were  inferior  to  AFRP  rods,  CFRP  rods;  hence  there  were  many  problems  in the practical  application. 

728 FRPRCS-6: Durability and Maintenance In this study, GFRP rods made of  conventional glass fiber (T type) and  new glass fiber (New type) were used. The strength and durability of GFRP  rods were investigated.  In particular, cyclic fatigue  properties, static  fatigue  properties  to  be  considered  important  items  in  the  case  of  tendons  were  evaluated.  OUTLINE OF EXPERIMENTS The composition of glass fibers (T type, New type) is shown in Table  1. The  new type has Ti, Zr in its composition. Also, vinyl  ester  resin  was used as  matrix. The properties of resin are shown in Table 2. The tested GFRP rods  are round bar, one­dimensionally  reinforced  rods with fiber content of 66%  by volume and the rods are 6 mm in diameter, and 40 cm in length. The rods  were  gripped  using  two  conical  steel  wedges,  which  were  developed  by  Kobayashi1^  To  ensure  sufficient  gripping,  the  ends  of  rods  were  coated  with a mixture of  unsaturated polyester resin and iron powder.  Table 1. Composition of  glass fibers  others

Si02

Al203

MgO

CaO

BaO

Ti02

Zr02

T type

65 

24 

10 

—

—

—

— 

1 

New type

44 

5 

4 

6

16 

10 

3 

12 

Table 2.  Properties of resin 

Average

Tensile Strength (MPa) 83.2 

Elastic Modulus (MPa) 3048 

Maximum Strain (%) 5.22 

S.D.

1.13 

36.3 

0.11 

C.ofV.

0.014 

0.012 

0.021 

The  tensile  tests  were  performed  using  displacement  control  type  autograph  (98kN)  at room temperature(20±5°C).  The cross  head  speed  of  the autograph  was kept to 5 mm/min and the number  of tests  for  each case  was fixed at 100.  The  cyclic  fatigue  tests  were  performed  using  load  control  type  servopulser  (98kN)  at  room  temperature  (20 ±  5 °C).  Maximum  stress 

Strength and Durability of GFRP Rods 729 (upper  stress) in the cyclic fatigue  tests was set from  30% to 80% of static  tensile  strength.  Stress  amplitude  was varied  as 50,  100, 250, 500  MPa,  frequency  was set from  1 to 10 Hz and the number of tests for each case was  fixed at 8.  The  static  fatigue  tests  were  performed  using  load  control  type  servopulser  (98kN) at room  temperature  (20±5°C).  Sustained  tensile  load  in the static fatigue tests was set from 70% to 95% of static tensile strength.  The number of tests for each case was fixed at 8.  EXPERIMENTAL RESULTS AND DISCUSSION Tensile tests Table  3  shows  the average,  standard  deviation  (S.D.),  and coefficient  of  variation (C. of  V.) of tensile strength and elastic modulus of GFRP rods. It  is  clear  that  the tensile  strength  of GFRP  rods  made  of New  type  is low  compared with that of GFRP rods made of T type and elastic modulus of the  two types is almost the same.  Table 3.  Properties of GFRP rods  Type

T type

New type

Tensile Strength

Average

1735 

1192 

(MPa)

S.D.

119.6 

73.1 

C.ofV.

0.069 

0.061 

Elastic Modulus

Average

59035 

58839 

(MPa)

S.D.

992 

462 

C.ofV.

0.017 

0.008 

Table  4  shows  the average,  standard  deviation  (S.D.),  coefficient  of  variation  (C. of V.) of tensile  strength and elastic modulus  of glass  fibers.  The tensile tests were performed  using displacement  control type autograph  (49N)  in  accordance  with  JIS  R­7601.  The  cross  head  speed  of  the  autograph was kept to 2 mm/min and the number of tests for each case was  fixed at 50. It is clear from Table 4 that  tensile strength of New type is  low 

730 FRPRCS-6: Durability and Maintenance

compared  with that of T type. It is because the New type has low Si content  in glass composition  compared  with T type.  Table 4. Properties of glass fibers(monofilament)

Type

T type

New type

Tensile Strength

Average

2460 

2044 

(MPa)

S.D.

853 

497 

C.ofV.

0.347 

0.243 

Elastic Modulus

Average

83780 

86091 

(MPa)

S.D.

15964 

21063 

C.ofV.

0.191 

0.245 

Usually,  in the  case  of  composite  materials  such  as  FRP, the  law of  mixtures can be applied  as shown in Equation  (1), (2).  E = EjVf+Em{\ - Vf)  a=afVf+am{\-Vj) 

(1)  (2) 

in  which E =  elastic  modulus  of rods, Ef=  elastic  modulus  of  fiber, Em = elastic  modulus  of  matrix, a =  tensile  strength  of  rods, aj =  tensile  strength  of  fiber, am  =  tensile  strength  of matrix,  and Vf =  fiber  volume  fraction.  The tensile  strength  and  elastic  modulus  of  matrix  are very  low  compared  with  those  of  fiber,  hence  one  can  neglect  the  matrix  in  the  calculation  as shown in Equation (3), (4)  E = EfVf  a ±F ajVf 

(3)  (4) 

Table  5  shows  the  experimental  values  and the  calculated  values  for  GFRP rods and their  ratios.  The experimental  values are 0.88 to 1.07 that of  calculated  values;  hence, the tensile  strength  and elastic  modulus  of  GFRP  rods can be obtained  using the law of mixtures. 

Strength and Durability of GFRP Rods 731 Table 5.  Comparison of experimental values and calculated values  Type

T type

New type

Tensile Strength

Experimental  (1) 

1735 

1192 

(MPa)

Calculated  (2) 

1624 

1349 

(l)/(2) 

1.07 

0.88 

Elastic Modulus

Experimental  (1) 

59035 

58839 

(MPa)

Calculated  (2) 

55295 

56820 

( l ) / ( 2 ) 

1.07 

1.04 

Cyclic fatigue tests The  result  of  cyclic  fatigue  test  (New  type,  stress  amplitude: lOOMPa)  is  shown  in  Figure  1.  The  number  of  cycles  at  fatigue  failure  varied  with  the  same  stress  ratio,  hence,  average  number  of  cycles  of  8  specimens  was  defined  as the number of cycles at fatigue  failure.  80  70 

rf

H. LXJT"

60 

2 50

:

hti K

^V-S-lJ

:

;

1 40 H 30 on

,\iJ&iiii

20  10  L_L i i .111-i ill. i ! i;.,! i I i iiii 0  l.E+01  l.E+02  l.E+03  l.E+04  l.E+05  l.E+06  Number of cycles at fatigue  failure  Figure  1.  Result of cyclic fatigue test (New type,  stress amplitude: lOOMPa) 

732 FRPRCS-6: Durability and Maintenance

Figure  2  shows  the  relation  between  the  stress  amplitude  and  the  number  of  cycles  at  fatigue  failure  of  GFRP  rods. The  maximum  stress  was  set 30%, 50%, 70% of  static tensile  strength  and  the  stress  amplitude  varied  from  50  MPa  to  500  MPa  for  each  condition.  The  tests  were  performed  up  to 4 million  cycles.  In Figure 2, N30%, N50%, N70% represent  30%, 50%,  70%  stress  ratio  using  GFRP  rods  made  of  New  type  respectively,  T30%,  T50%, T70% represent  30%, 50%, 70% stress ratio using GFRP  rods  made  of T type  respectively.  As  the  maximum  stress  or  stress  amplitude  increases,  the  number  of  cycles  at  fatigue  failure  reduces.  In the  case  of  30% stress  ratio, the  number  of cycles  at fatigue  failure  is almost  same  in both rods. However,  in the  case  of  50%  and  70%  stress  ratio,  the  number  of  cycles  at  fatigue  failure  of  GFRP  rods  made  of  New  type  increases  compared  with  that  of  T  type.  In  particular,  in the  case  of 70% stress  ratio, there  is  a difference  of  1 order  or  more.  — •  N30%  — — • N 5 0 %  — * — N 7 0 %  ­ ­ O  ­ ­ T30%  ­ ­ a  ­ ­T50%  ­ ­ ­A ­ ­T70% 

1000 

PL,

T3

S  100  D-

B a

in in ,„,™;, 

80  C 

20  30  40  Time  (weeks) 

20  30  40  Time  (weeks) 

(a) "Wet" condition  (b) After redrying  Figure 2. Effect  of Immersion in simulated alkaline solution (CaC03 and Ca(OH)2 in  deionized water) on tensile strength 

764 FRPRCS-6: Durability and Maintenance A  comparison  of  results  due  to  immersion  in  the  simulated  alkali  solution (Figure 2) indicates that the decrease in performance  is greater than  due to immersion in deionized water at the same temperature levels, both in  the  wet  and  redried  cases.  Micrographic  examination  of  the  specimens  shows that the exposure to the simulated alkali solutions results in a greater  level of degradation at the level of the fiber and the interface.  It is of interest to compare the effects  of the three solutions ­  deionized  water,  simulated  alkali  solution, and the concrete  leachate  on properties at  23°C so as to assess effects  that could take place when a FRP composite was  embedded  in  concrete  (rebar  for  example)  or  constantly  in  contact  with  concrete  and hence subject to leachate due to moisture migration  (as in the  case  of  external  strengthening  or  structural  formwork).  A  comparison  of  effects  in terms of percentage retention of properties is given in Table 2.  It  can  be  seen  that  the  exposure  to  concrete  leachate  causes  the  maximum  level  of degradation  emphasizing that the choice of solution used to assess  actual  field  level  degradation  is  important  as  the  conventionally  used  simulated  exposure  conditions  (deionized  water  and  CaC03  and  Ca(OH)2  solution) may not duplicate the actual effects  of concrete salts.  At the end  of  the  year­long  exposure  the  percentage  recovery  between  the  wet  and  redried  specimens  was 7.46%,  3.84% and 6.12%, respectively  for the three  solutions.  It  is  noted  that  even  at  23°C  the  composites  did  not  regain  a  substantial  level of their  properties  after  redrying indicating a high  level of  irreversible damage in this class of composites.  Table 2. Comparison of aqueous immersion effects based on solution  Time Period of Exposure 5 weeks 10 weeks 15 weeks 20 weeks 30 weeks 50 weeks

Percentage Retention of Tensile Strength Deionized Simulated Concrete Water Alkali Leachate 93.58  90.29  85.15  92.35  86.33  84.17  91.33  83.20  82.45  79.72  88.38  83.18  87.51  76.73  77.73  72.68  60.69  66.35 

Percentage Retention of Tensile Modulus Deionized Simulated Concrete Water Alkali Leachate 88.65  89.23  89.77  91.68  92.36  92.77  92.52  86.80  90.61  91.14  91.73  92.74  91.69  91.05  92.63  90.89  90.89  90.29 

Since microscopic studies conducted on the tension specimens indicated  degradation  at  the  level  of  the  interface  it  is  of  significant  importance  to  assess changes in short­beam­shear properties under similar conditions since  the tension  tests only  indicate  changes  in fiber dominated  properties rather  than  the  other  properties  such  as  off­axis  strength  and  shear  properties 

Durability ofE-Glass Vinylester Composites 765

which  are  significantly  affected  by  changes  in  the  resin  state  and  the  interface. 

V\ 

 •

^ 

w>-

^ 

­ 

" •­

~23C

. 

a

so

\ \  \ 

. 

"~~­ ­ ­ ­ . 

« \ . 

M 

X

\ 

20

1

70

\ 

30

•\

' " "• " ­ .• .

I

40C

, 

-80C

40

-23 C

v

V 

-60C

—  10

40C

\  \  \  \ 

20

30

40

. 

-80C

5

Time (Weeks)

Time (Weeks)

(a) Deionized water  (b) Alkali solution  Figure 3. Effect  of Immersion on short­beam­shear strength 

Figure  3  provides  a  comparison  of  changes  in  short­beam­shear  (sbs)  strength  for  the  exposures  related  to  deionized  water  and  the  simulated  alkali  solution.  As can be seen the  immersion  in simulated  alkali  solution  results in greater levels of degradation due to deterioration at the level of the  interface  and  the  fiber,  in  addition  to  microcracking  between  fibers  coalescing to form a macrocrack.  It is of interest to note that insofar  as sbs  strength  is  concerned  the  use  of  concrete  leachate  actually  results  in  the  least amount of degradation of all three aqueous solutions as shown in Table  3.  Table 3. Comparison of aqueous immersion effects  based on solution  Time Percentage Retention of ShortPeriod of Beam Shear Exposure 15 20 30 50

weeks weeks weeks weeks

Deionized Water 95.91  92.00  90.05  87.12 

Simulated Alkali 94.02  91.78  89.22  85.56 

Concrete Leachate 96.15  95.35  90.64  88.96 

It should be remembered that  interlaminar  shear strength is affected  by  the distribution of moisture in a composite.  In a sbs test the center plane of  the  composite  specimen  develops  the  highest  level  of  shear  stress.  Thus 

766 FRPRCS-6: Durability and Maintenance

degradation  due  to  moisture  absorption  is not  fully  operative  at this  level  until the moisture has diffused  to the midplane.  Using the calculated  rates  of  diffusion  it  can  be  seen  that  moisture  does  not  penetrate  to  the  center  plane  till  after  the  15th  week  of  immersion  in  deionized  water  at  23°C.  Silvergleit  indicated  that  for  glass/epoxy  composites  immersed  in  water  shear  strength  initially  decreased  with  increasing  water  content4.  The rate  of  shear  strength  reduction  was  noted  to  change  slope  and  decrease  more  slowly  over  time  as  moisture  uptake  reached  saturation.  It  is  interesting,  however, to note that  as shown in Table 4 in the current  samples, with the  exception  of  data  corresponding  to  deionized  water  at  40°C  the  initial  decline  in  shear  strength  is  greater  than  that  for  the  interval  between  15  weeks and 50 weeks (the  15 week period is theoretically after  moisture has  reached  the midplane).  It can also be seen that the  initial  decline in shear  strength increases with increase in temperature.  Table 4. Comparison  of aqueous  immersion  effects  based  on  solution  Temp. of Slope of Maximum Shear Strength Slope of Maximum Shear Strength Exposn ire Change Between 0-15 weeks Change Between 15-50 weeks Concrete Deionized Simulated Concrete Deionized Simulated CQ Leachate Water Water Alkali Leachate Alkali ­0.112  20 ­0.1179  ­0.1726  ­0.1088  ­0.1045  ­0.0888  ­0.1144  NA  40 ­0.2879  ­0.3505  ­0.2838  NA  ­0.4760  ­0.4827  NA  ­0.2618  NA  60 ­0.2996  ­0.8412  NA  ­0.2048  NA  80 ­0.9165  ­0.1778 

DMTA CHARACTERIZATION As  noted  previously  for  long­term  response,  in ambient  temperature  cured  vinylester  composites, there are often  simultaneous effects  of post­cure and  leaching/degradation.  Whereas  the  former  causes  an  increase  in  performance  attributes the latter results in a decrease.  However, this clarity  in differentiation  is not seen in reality due to competition between the two.  Further,  it  is  known  that  glass  transition  temperature  can  decrease  due to  moisture  uptake  and  hydrolysis2,  whereas  it  can  increase  due  to  both  residual  cure  and the  leaching  of  low molecular  weight  species.  Figure 4  provides  a record  of  the  change  in  glass  transition  temperature  with  time  under  deionized  water exposure.  Through additional tests such as ICP and  FTIR  it  was  determined  that the  fluctuations  are  due to  competing  effects  especially  of  post  cure  due to  elevated  temperatures  and that  the  apparent  higher  levels  of  Tg  attained  after  immersion  in  60°C  and  80°C  deionized 

Durability ofE-Glass Vinylester Composites 767

water is actually due to leaching effects  of low molecular weight species.  It  is noted that the glass transition  temperature  (determined  from  the peak of  the tan  8 curve)  of the material  prior to immersion  was  148°C.  As can be  seen  the  effect  of  the  alkali  solution  is  substantially  less  than  that  of  deionized water. 

95 c e

c•

w\ '

V

V

/

r ''•

V . N. v•

.

>

a

§

90

a

'

s"

v

!

80 C

;

^ y ^ 

; ! 5 * ; 

60 C

\

& 3 85 u ft. 80

\

^S,40C

23 C

20

30

40

Time (Weeks)

50

60

10

20

30

40

50

60

Time (Weeks)

(a) Deionized water  (b) Alkali solution  Figure 4. Effect of Immersion on glass transition temperature  A comparison of the DMTA characteristics  for the exposure conditions  at 23°C is given in Table 5 and it can be noted that as with the change in sbs  strength the alkaline environments have a reduced  effect.  Table 5. Comparison of aqueous immersion effects  based on solution  Percentage Retention of Glass Time Percentage Retention of Storage Transition Temperature Period of Modulus at 25 °C Exposure Deionized Simulated Concrete Deionized Simulated Concrete Alkali Leachate Water Water Alkali Leachate 5 weeks 95.65  93.63  117.10  108.55  88.20  100.00  95.55  10 weeks 98.95  95.13  129.06  136.75  104.27  15 weeks 96.67  90.06  243.59  194.02  170.94  93.19  95.25  98.08  105.13  113.68  116.24  94.07  20 weeks 92.56  88.21  102.56  112.82  30 weeks 87.17  99.15  88.11  88.33  50 weeks 78.06  85.47  100.86  94.87 

This  points  to  the  fact  that  the  alkaline  salts  in  the  solution  have  a  greater  effect  on  the  fiber  and  the  local  fiber­matrix  interphasial  region  rather  than  the  matrix  itself  indicating  that  degradation  is  more  fiber  dominated  than  resin  dominated.  While this  points  out  a weakness  in the  composite in its primary load bearing constituent it also indicates a potential 

768 FRPRCS-6: Durability and Maintenance

means  of  increasing  durability  through  the  use  of  resins  that  can  serve  as  greater  diffusion  barriers  for  the  salts  causing  fiber  and  fiber­matrix  interphase level degradation.  CONCLUSIONS Mechanisms  of  degradation  are  seen  to  differ  substantially  based  on  aqueous  solution  and  although  acceleration  can  be  conducted  through  the  use of higher temperature  levels this must be conducted  with caution  since  changes  in  temperature  can  result  in  dramatic  changes  in  modes  and  mechanisms  of  failure  thereby  nullifying  the  primary  principles  of  time­ temperature  superposition  for  acceleration.  The  use  of  time­temperature  superposition on the deionized water exposure data provides a conservative  estimate  of  42%  retention  of  tensile  strength  after  10  years  and  35%  retention after  30 years.  ACKNOWLEDGMENTS The support by the California  Department of Transportation, the  California  Department  of  Water  Resources,  and  the  National  Science  Foundation  through a CAREER award to the second author is gratefully  acknowledged  REFERENCES 1.  Burrell, P.P., Herzog, D.J. and MacCabe, R.T., "A Study of Permeation  Barriers  to  Prevent  in  Marine  Composites  and  a Novel  Technique  for  Evaluating  Blister  Formation," Proceedings of the 42nd Annual SPI Conference, Session  15E, 1987, pp. 15.E: 1­13.  2.  Ghorbel,  I.  and  Valentin,  D.,  "Hydrothermal  Effects  on  the  Physico­ Chemical  Properties  of Pure and Glass Fiber Reinforced  Polyester and  Vinylester  Resins," Polymer Composites, Vol.  14[4],  1993,  pp.  324­ 334.  3.  Karbhari, V.M., Murphy,  K. and Zhang, S., "Effect  of Concrete Based  Alkali  Solutins  on  Short­Term  Durability  of  E­Glass/Vinylester  Composites," Journal of Composite Materials,  36[17], 2002, pp.  2101­ 2121.  4.  Silvergleit, M., Macander, A.B., and Cardamone, J.A., "Effect  of Long­ Term  Water  Immersion  on Properties  of Graphite/Epoxy  Composites,"  Report MAT­76­16, DTNSRDC, June 1976. 

FRPRCS-6, Singapore, 8­10 July 2003  Edited by Kiang Hwee Tan  ©World Scientific Publishing Company 

DURABILITY OF GFRP COMPOSITES UNDER  TROPICAL CLIMATE  Y. S. LIEW AND K. H. TAN  Department of Civil Engineering, National University of Singapore 1 Engineering Drive 2, Singapore 117576 In  this  study,  two  types  of  glass  fiber  reinforced  polymer  (GFRP)  composites were weathered outdoor and in accelerated weathering chamber  to  evaluate  the  durability  of  the  composites  under  tropical  climate.  Engineering  properties  of GFRP  laminates  were evaluated  after  1, 3, 6, 9  and  12 months of outdoor  exposure,  as well  as  in  accelerated  weathering  chamber  after  equivalent  exposure  times  to  correlate  the  effects  of  two  exposure  conditions  on  the  composites.  In  addition,  the  mechanical  properties of GFRP laminates obtained  from  weathering tests were used to  predict  the  ultimate  moment  capacity  and  failure  mode  of  small­scale  GFRP­strengthened  RC  beams which  were  also  exposed  under  the  same  conditions. Test results showed that the tensile and bond strength of GFRP  laminates  decreased  with exposure time.  The reduction  in tensile strength  could  be due to either  a drop in ultimate  strain  or modulus depending on  the type of resin.  Failure  modes of small­scale  GFRP­strengthened  beams  were  affected  by  exposure  to  outdoor  weather.  Such  variations  were  accurately  predicted  using  appropriate  mechanical  model  incorporating  weathering test data of the material.  INTRODUCTION Externally  bonded  fiber  reinforced  polymer  (FRP)  system,  either  by  wet  lay­up  of  fiber  sheets  or  adhesive  bonding  of  composite  strips/panels,  has  gained  its  popularity  in  structural  retrofitting  and  rehabilitation  of  deteriorated  structures. However, the susceptibility  of FRP reinforcement  to  individual,  as  well  as  synergistic  effects  of  ultra­violet  (UV) rays,  moisture  and  heat  were  reported  by  recent  studies.1  The  high  average  annual  temperature,  humidity  and  relatively  constant  UV  dosage  for  tropical  countries  in  equatorial  countries  like  Singapore2  is  believed  to  have  detrimental  effects  on the  mechanical  performance  of the externally  bonded  FRP  composites,  which  may  further  affect  the  performance  of  FRP­ strengthened  structures.  Therefore,  the  durability  of  FRP  composites  under  tropical  climate  is an important  issue to be  addressed. 

770 FRPRCS-6: Durability and Maintenance

In  light  of  this,  a  durability  study  was  conducted  to  investigate  the  effects  of  local  weathering  on  engineering  properties  of  glass  fiber  reinforced  polymer  (GFRP)  laminates  and  GFRP­strengthened  beams.  Specimens  were  subjected  to both  natural  and  accelerated  weathering  tests.  The  reproducibility  of  accelerated  weathering  tests  was  verified  and  the  performance  of  weathered  GFRP  laminates  was  evaluated.  The  weathering  test  results  of  GFRP  laminates  were  then  used  to  predict  the  changes  in  failure  mode  of  small­scale  GFRP­strengthened  beams  subjected  to  the  same exposure  condition.  FAILURE MODE OF GFRP-STRENGTHENED BEAMS With  proper  anchorage  at  the  sheet  cut­off  points  at  beam  ends  including  the  use  of  adequate  transverse  reinforcement,  failure  modes  of  beams  strengthened  with  GFRP  laminates  can  be  reduced  to  (i)  compression  crushing  of  concrete,  (ii)  rupture  of  GFRP  laminates  and  (iii)  debonding  of  GFRP  laminates  due  to  flexural  cracks.  Referring  to  Fig.  1,  the  internal  resisting moment of the beam at failure  can be calculated  as 

M  = ApspEp(dp -x)+A,e,E,{d,

-x)-A,'e,'E,'(d,'-x) 

(1) 

where Ap, As  and As'  =  area  of  GFRP,  tensile  and  compressive  steel  reinforcement,  respectively; Ep, Es  and Es'  =  modulus  of  GFRP,  tensile  steel reinforcement  and compressive  steel reinforcement,  respectively; ep, es and ss'  = strains  in GFRP, tensile steel reinforcement  and compressive  steel  reinforcement,  respectively; ds  and d/ =  distance  from  the  top  concrete  fiber  to the centroid of the tensile steel and compressive  steel  reinforcement,  respectively; x  =  neutral  axis  depth, b  =  beam  width, h = beam  depth  (=  distance  from  the  top  concrete  fiber  to  the  centroid  of  GFRP).  For  tensile  and  compressive  steel  reinforcement, esEs  and ss'Es'  are  taken  to  be 

A,' •

JpK:T-7' h ds

N.Ar •  " 

(a) Beam section 

(b) Strain distribution 

(c) Stress distribution 

Figure  1. Stress and strain distribution of a beam  section 

GFRP Composites under Tropical Climate 771

less  than fy  and fy',  that  is,  the  yield  stress  of  tensile  and  compressive  reinforcement,  respectively.  The  beam  moment  capacity  corresponding  to  concrete  crushing  and  GFRP rupture can be found by substituting values for sc or sp, that is,  EC = ecu for concrete crushing  (2)  or  ep=0.iepu  for GFRP rupture  (3)  where scu  and epu  are  the  ultimate  compressive  strain  of  concrete  and  ultimate  tensile  strain  of  GFRP  laminates  respectively,  and  solving  other  terms by iteration until the following equilibrium condition is achieved  b jfjxjdx  + A.'e.'E,1 = AsssEs  + ApEpEp 

(4) 

X

The  coefficient  of 0.8  in Eq. (3) accounts  for  the average  lower  strains of  FRP rupturing  when  bonded to beams  compared  to strains  measured  from  material tensile test.3  To predict the flexural  crack induced debonding, the  associated strain in GFRP laminates can be taken as4  P

~e-db- Ep ~ Ep i

tp

(5)

where a -  calibration factor, flp = width coefficient, f5L = length  coefficient,  fc'=  concrete  cylinder  compressive  strength, tp  =  thickness  of  GFRP  laminates.  The  failure  mode  and  ultimate  beam  capacity  are  then  determined  from  the  minimum  moment  capacity  of  all  the  three  failure  modes,  that  is,  Mu =  min(Mcc, Mfr, Mdb)  where Mcc = moment  capacity  corresponding  to  concrete  crushing, Mfr = moment  capacity  corresponding  to  GFRP  rupture  and Mas, =  moment capacity corresponding to debonding.  To predict the time­dependent beam capacity, the properties of internal  steel reinforcement  and concrete are assumed unchanged since the former is  protected by concrete while the latter gained most of its strength with good  initial curing and is not susceptible to moisture, heat and UV. Therefore, the  variation  in  the  failure  modes  and  moment  capacity  is  associated  with  changes  in  properties  of  GFRP  laminates.  With  adequate  field  or  well  represented  accelerated  weathering  test  data,  the  time­dependent  GFRP  prosperities of GFRP laminates can be expressed as  Q{t) = 4>p(f)P0  (6)  where P0  is  the  initial  property  (for  example,  the  as­received  material  engineering properties), p(t)  is the material property variation  function  and 

772 FRPRCS-6: Durability and Maintenance

Q(t)  is the predicted property at time t. The property variation  function  can be derived by regression  function  of outdoor or accelerated weathering  data, that is, 

i n\ = I^L

(7a)

fpU

Pit.) where P(t) is the regressed  material property  function,  from  either  outdoor  or  accelerated  weathering  tests,  and t0  is  the  time  where  no  weathering  effects  have taken place on the material properties. If regression is obtained  from accelerated test data with a factor of ka, then Eq. (7a) becomes  *P(/k) A(f) = LLL.  (7b)  p *P(0 where *P is the  regressed  material  property  function  based  on  accelerated  weathering  tests.  With  the  above  considerations,  the  beam  capacity  and  failure  mode after  a period of weathering can be estimated  from  Eq. (1) by  incorporating Eq. (7), with Eq. (2), (3) and (5) taking the form  concrete crushing (at age t): sc (t) = scu  (8a)  (8b)  GFRP rupture (at age t): ep(t) = (j)e __ (t) epu  GFRP debonding (at age t): ep(t)  = 0e

M

(t) spdb 

and the modulus of GFRP laminates in Eq. (1) taken as  EJt)=tE (t)Ep 

(8c)  (9) 

TEST PROGRAM Tropical Climate and Accelerated Weathering Chamber Table  1  shows summarizes the past  11 years (1987­1997) of meteorological  data of Singapore. Based on this, an accelerated test chamber was designed  by  intensify  the  UV­A  irradiance  to  six  times  that  of  outdoor,  while  maintaining the proportion of light/dark, wet/dry period, as well as relative  humidity (RH) level.  GFRP Tensile Coupon Two  E­Glass  composite  systems,  denoted  as  Gl  and  G2,  with  different  fiber  weaving  configurations  and  resin  systems,  were  studied.  Their  properties  are  shown  in  Table  2.  Tensile  coupons,  as  shown  in  Fig.  2(a),  were  fabricated  in  accordance  with  JSCE­E­541­2000.5  The 

GFRP Composites under Tropical Climate 773 Table  1. Average of  Singapore outdoor weathering factors (1987­1997)  Weathering Factor Solar Irradiance (m Wh/cm2) Precipitation (mm) Temperature

Relative Humidity (%) Sunshine Rainfall

Yearly

Monthly

Daily

Mean

—

13875.90 

462.53 

Mean

2044.80 

170.40 

—

Average Max Min Average Max Min Total hours % Total days %

­

­ — 

27.47  33.50  23.40  83.11  98.70  54.10  — 

223.54  61 

­

­ 5.6  23  — 

Table 2. Properties of Fiber sheets and Resins '  Gl E-Glass Unidirectional roving Tensile strength (MPa) 1700  Elastic modulus (Gpa) 71  Ultimate strain (%) 2.0  Two part, 100% Type Resin solid, low viscosity aminecured epoxy Tensile strength (MPa) 54  Elastic modulus (Gpa) 3.00  Ultimate strain (%) 2.5  " Based on manufacturers' product specifications Fiber

Type Tow sheet form

G2 E~Glass Bidirectional woven roving 130  11  1.25  Orthophthalic unsaturated polyester 30  0.67  4.4 

774 FRPRCS-6: Durability and Maintenance

coupons were allowed to cure in the laboratory  for  at least 2 weeks  before  being  subjected  to  both  outdoor  and  accelerated  weathering.  All  the  coupons  were tested  after  0,  1, 3, 6, 9,  12 months  for  outdoor  weathering  and after  5, 15, 30, 45, 60 days in the case of accelerated weathering.  GFRP-Concrete Plates The  bond  strength  of  GFRP  reinforcement  was  evaluated  by  using  pull­ apart  concrete  plates  bonded with GFRP laminates, as shown in Fig. 2(b).  The average compressive cube strength of the hosting plates was around 40  MPa. All the bonded plates were cast, bonded and cured at least 2 weeks in  laboratory and subjected to accelerated weathering and tested at 0,  1 week,  and 1, 3 and 6 months.  GFRPstrengthened Beams Small­scale reinforced  concrete beams bonded with Gl  and G2  (as shown  in. Fig.  2(c))  were  fabricated  and  exposed  outdoor  and  in  the  weathering  chamber, and tested after 0, 1, 3, 6, 9, and 12 months of exposure. 

Lateral  View  Section  A­A  (c) Schematic view of GFRP­strengthened RC beam  Figure 2. Types of specimen for weathering tests 

TEST RESULTS AND DISCUSSION  Preliminary  results  up to  6 months  have  been  obtained  and  are  presented  herein. 

GFRP Composites under Tropical Climate 775 Reproducibility of Outdoor Weathering Effects Fig.  3(a)  shows  the  measured  temperatures  and  relative  humidity  (RH)  for  outdoor  condition  and  in  the  accelerated  weathering  chamber  at  different  time scale, in which  1 day in chamber is being compared to 6 days  outdoors.  The  observed  outdoor  temperature  and  RH  were  well  reproduced  in  the  chamber  while  the  UV­A  intensity  was  kept  6 times  that  of  outdoors.  Fig.  3(b)  shows  the  regressed  ultimate  strain  and  elastics  modulus  for  both  outdoor­  and  chamber­exposed  Gl  tensile  coupons  on  logarithmic  time  scales, that  is  1 to  180 days  for  outdoor  and  1 to  30  days  for  chamber.  It  is  also  obvious  that  both  the  outdoor­  and  chamber­weathered  coupons  exhibited the same trend  in their property variations.  Chamber Elapsed Time 8 hrs

16 hrs

21Jul'02 12:00pm

24 hrs

23Jul'02 12:00pm

Outdoor Elapsed Time

(a) Temperature and RH  Logarithmic chamber age, ln[ t ] (day) 1.00 O

3.

2.72

7.39

20.09

20

W 15

Outdoor Chamber 1.00

2.72

strain modulus strain modulus 7.39

• © a A-

test test test test -~.o» ""I ' ' 20.09 54.60

148.41

Logarithmic outdoor age, ln[ t ] (day)

(b) Properties of Gl  Figure 3. Comparison of outdoors and accelerated effects on specimens 

776 FRPRCS-6: Durability and Maintenance

Tropical Climatic Effects on GFRP Tensile Properties For  all  the  virgin  and  weathered  specimens,  Gl  tensile  coupons  typically  ruptured  with  longitudinal  splits  between  unidirectional  fiber  roving  whereas  G2  coupons  ruptured  with  cracks  perpendicular  to  direction  of  tensile  load.  The  ultimate  strain  and  modulus  variation  of  Gl  and  G2  specimens  after  outdoor  weathering  are  depicted  in  Figs.  4(a)  and  4(b),  respectively.  Under  the  same  exposure  condition,  the  variations  were  different  for  composites  with  different  resins.  The  ultimate  strain  of  Gl  decreased  significantly  more  than  the  modulus.  However,  the  decrease  in  ultimate  strain  in  G2  specimens  was  relatively  small  compared  to  the  decrease in modulus.  The ultimate strength of both Gl  and G2 composites  decreased as a result of outdoor exposure. Reduction in ultimate strain  after  environmental  exposure  was  also  observed  in  other  durability  tests  conducted on GFRP composites.6 By plotting the product of regressed strain  and  modulus, spu(t)Ep(t),  in  Fig.  4(c),  it  is  seen  that  the  product  matches  well  with  the  regressed  ultimate  stress  for  both  composites.  This  implies  that  the  strength  reduction  in  weathered  composites  could be due to either  loss in ultimate  strain (embitterment)  or  modulus, or both, depending on the composite constituents.  Accelerated Weathering Effects on GFRP-concrete Bond The  GFRP­concrete  interfacial  bond  strength  after  6  months  in  chamber,  which is believed to be equivalent to that of 3 years outdoors, is depicted in  Fig. 4(d). The initial bond strength was governed by the type of resin matrix  used.  However,  both  types  of  composites  bond  strength  decreased  at  the  same rate throughout the accelerated weathering period.  Prediction and Test Result of' GFRP-strengthened Beams Material  property  variation  functions  (s ^ (t),   25,8  25,4 

Average 41,1  35,4  31,8  41,1  32,4  31,3  35,0  31,7  29,3 

(a), (b), (c) Same superscript indicates same set of five specimens 

The  evolution  of  shear  strength  shows  more  consistency  than  that  of  tensile  strength;  Table  2  indicates  that  it  diminished  consistently  for  all 

784 FRPRCS-6: Durability and Maintenance

series  of  specimens.  Once  again,  the  coating  does  not  appear  to  exert  a  positive effect  on the properties of the specimens.  CONCLUSIONS The  outline  of  a  long­term  project  aimed  at  measuring  the  long­term  properties  of  CFRP  laminates  was  presented.  Initial  results  show  that  climatic  conditions  actually  have  an  adverse  effect  on  the  product  properties.  ACKNOWLEDGMENTS This research program benefits  from the long­term funding  of the  following  sources  which  are  gratefully  acknowledged:  the  Natural  Sciences  and  Engineering  Research  Council  of  Canada  (NSERC)  through  the  ISIS  Canada Network of Centres of Excellence for the Canadian co­authors; the  Public  Works  Research  Institute  of  Japan  for  the  Japanese  co­author.  The  sabbatical  stay  of Professor  Labossiere  in Japan  was  financially  supported  by  the  Japan  Science  Foundation:  this  invaluable  contribution  to  the  establishment  of  a  long­term  international  collaboration  is  gratefully  acknowledged.  In addition, the authors wish to thank Mr. Marc Demers and  Mr.  Iwao  Sasaki  for  their  technical  advice  and  their  support  in  the  maintenance of the exposure sites. 

REFERENCES 1.  Mufti,  A.A., Labossiere, P., Neale, K.W.,  "Recent Bridge Applications  of FRPs in Canada," Structural Engineering Int., 12(2), 2002, 96­98.  2.  Labossiere,  P.,  Neale,  K.W.,  Rochette,  P.,  Demers,  M.,  Lamothe,  P.,  Lapierre, P., and Desgagne, G.,  "FRP Strengthening  of the Ste­Emelie­ de­1'Energie  Bridge:  Design,  Instrumentation  and  Field  Testing,"  Canadian Journal of Civil Engineering,  27(5), 2000, 916­927.  3.  Japan  Society  of  Civil  Engineers,  "Code  for  Repair  and  Strengthening  of Concrete  Structures Using Continuous Fiber Sheet, Concrete Library  no.  101, July 2000 (in Japanese).  4.  ISIS  Canada,  "Strengthening  Reinforced  Concrete  Structures  with  Externally­Bonded  Fibre Reinforced  Polymers (FRPs)," Design Manual  #4, Winnipeg, Manitoba, 2001. 

FRPRCS-6, Singapore, 8­10  July 2003  Edited by Kiang Hwee Tan  ©World Scientific  Publishing Company 

DURABILITY OF ARAMID AND CARBON FRP PC BEAMS UNDER NATURAL AND ACCELERATED EXPOSURE H. NAKAI  Manager, Sumitomo Construction Co., Ltd. 13-4 Araki-cho Shinjyuku-ku, Tokyo 162-5788, Japan H. SAKAI  Chief Engineer, P.S. Corporation Otuka-bld. 3F, 1-3-17 Kita-otuka, Tosima-ku, Tokyo 170-0004, Japan T. NISHIMURA  Technician, IIS, The University of Tokyo 4-6-1 Komaba, Meguro-ku, Tokyo 153-8505, Japan T. UOMOTO  Professor, The University of Tokyo, Japan In  an  effort  to  better  understand  the  durability  characteristics  of  fiber  reinforced  polymer (FRP) bar in severe environment, a long­term study has  been  undertaken  using  tendons  made  of  aramid  and  carbon  fibers,  in  pretensioned  concrete  (PC)  beams.  Reference  beams  have  also  been  cast  using  normal  steel  reinforcement.  The  results  obtained  after  3  years  of  exposure  show  that  there  is  no  noticeable  deterioration  in  the  FRP  reinforced  specimens,  though  some  differences  are  found  in  the  flexural  behavior of specimens that use steel strands as reinforcement.  INTRODUCTION Fiber  reinforced  polymer  (FRP)  bars  have  attracted  a  lot  of  interest  as  an  alternative  reinforcement  material  for  concrete  structures,  due  to  their  high  strength,  light  weight,  non­magnetic  nature  and  corrosion­free  properties.  Since  1988, FRP bar has been used  in over  180 structures  throughout  Japan1,  and  no  reports  of  degradation  in  any  of  these  structures  has  been  reported.  Also,  in  large­scale  exposure  tests  undertaken  by the  government,  academic  institutions  and  the  private  sectors,  no  degradation  has  been  reported  2.  However,  it  has  been  pointed  out  that  the  bond  between  FRP  and  concrete  could  be  a  potential  problem  due  to  the  difference  between  the  thermal 

786

FRPRCS-6: Durability and

Maintenance

expansion  coefficient  of  FRP  and  that  of  concrete;  and  secondly,  the  fact  that organic fibers are hygroscopic3.  The  objective  of this  study  is to  better  understand  the  degradation  over  time  of  pretensioned  concrete  beams  (reinforced  with  FRP  or  steel)  subjected  to  repeated  drying/wetting  cycles  and  variations  in  temperature.  Exposure tests were carried out in two environments ­  (a) a natural outdoor  exposure test at two sites: one in the splash zone, and the other inland, and,  (b) accelerated test, where the beams were stored in a tank with a controlled  temperature and saturation regime. This paper reports the results after  about  40 months of exposure tests, though  specimens  have  been prepared  with a  15­year test program in mind.  THE PARAMETER OF EXPOSURE The  parameters  of this  study  are the type  of tendon material, the  exposure  test  method  and  exposure  time.  The  extent  of  degradation  was  assessed  through static testing at the end of the exposure period, fatigue tests in some  cases to study any changes in the bond between the tendon and the concrete,  observation  of  the  FRP  tendons  using  an  electron  microscope,  and  observation of the onset of rusting in the case of steel reinforcement.  In  the  natural  outdoor  exposure  test  (N  series),  the  exposure  locations  were splash zone (I) and inland  (T), the tendon materials were aramid FRP  (A),  carbon  FRP (C) and  steel (S). A total  of 24 beams  were  used.  In the  accelerated exposure test (A series), the tendon materials were (A), (C) and  (S), with 5, 5, and 3 beams, respectively, making a total of  13 beams.  ' 

fable  1. Parameter  of  exposure  ;  Destru 40 2nd Fatigue Micro [ ReQuanInitial Name of specimen : months loading [loading ctiv e : loading test scope ' exposure  ; tity [exposure yes yes yes 3  N[A,C,S]Control yes 6  N[I,T][A,C,S]1 yes yes ; yes )yes  *'  yes N[I,T][A,C,S][2,3]  !  yes yes yes 12  NT[A,C,S]4 yes yes yes yes yes 3/24  i A[A,C,SJControl yes 3  ; yes yes 3  yes yes A[A,C,S]1 : yes yes  *'  ] yes yes A[A,C,SJ[2,3,4] yes yes   7* 3 /13  Initial  letter  designates  type  of  exposure:  N  =  N a t u r a l ;  A  =  Accelerated.  [I,T]  designates  exposure  site:  I  =  Splash  zone  on  t h e  Izu  P e n i n s u l a ;  T  =  I n l a n d  site  at  Chiba  Prefecture.  [A,C,S]  describes  the  tendon:  A  =  A r a m i d ;  C  =  Carbon;  S  =  Steel.  * 1 :  without  Steel  tendon  ;* 2: with  steel  tendon  specimen  j * 3 : S h a s 2 o n l y 

Durability ofAramid and Carbon FRP PC Beams

787

Materials Table 2 shows the properties of the materials  used as tendons. The  aramid  FRP  was Technora®; the carbon  FRP was CFCC®; and the steel  tendon  material was steel wire strand. Whereas the coefficient  of thermal expansion  of steel is about the same as that of concrete, that of the FRPs is about zero,  or  even  negative  in some  cases.  Comparing  the Young's  modulus  of the  tendon  materials  reveals  a  ratio  of  1:2:4  for  aramid,  carbon  and steel,  respectively.  Table 3 shows the mix proportions of the concrete used ­  the proportions  were chosen so that the concrete has a strength of 35 N/mm2 at an age of 12  hours with steam curing, and had a W/C ratio of 0.372. Table 4 shows the  compressive  strengths  of concrete at the different  ages. It is apparent  that  the  compressive  strength  of the  concrete  increases  substantially  over  time  compared with the initial values. Creep and drying shrinkage  measurements  were also carried out using the inland exposed beams in the N series. Figure  1  shows  the  shrinkage  strain  observed.  At  an  age of  44  months  after  prestressing  (when  the  tests  were  carried  out  for  the  series  of  results  reported in this paper), a shrinkage strain of about 0.08% was observed.  Table  2. Properties  of tendons  and  reinforcements  Material

Aramid

cross section (mm 2)

Nominal

capacity

area

Tensile Capacity

(kN) (kN)

Young's modulus (kN/mm Elongation

Rebar

CFCC

SWPR-7A

SWM-P

failure* strains load strains '""''W (micro)'' (kN) (micro)*

no-preslressed PFS (2 layers)

12 

87 

119 

25%-prestressed

30 

121 

141 

36 

112 

132 

PFS (2 layers) 33%-prestressed PFS (2 layers) no-prestressed PFS (2 layers)  ;  25%-prestressed PFS (2 layers) 25%-prestressed PFS (2 layers) 33%-prestressed PFS (2 layers) 33%-prestressed:  PFS (2 layers)

RCF­0­ 59.5  RCF­25­ 64.0(1)  RCF­25­ 76.8(1)  RCF­33­ 55.2  RCF­33­ 66.0  RCF­25­ 25%-prestressed 64.0(2)  PFS (2 layers) RCF­25­ 76.8(2)  25%-prestressed PFS (2 layers)

59.5  64.0  76.8  55.2  66.0  64.0  76.8 

1250  [1271]  1250  [1138]  1500  [1537]  1000  [1008]  1250  [1260]  1250 

....El?.?.!]....  1500  [2294] 

Failure mode

PFS Debonding PFS Debonding

9904  [4200]  8067  [5544] 

13.5 

140  [211] 

12 

94 

120 

8733 

16.0 

140(138] 

28 

141 

160 

7957  [4200] 

16.0 

140  [138] 

30 

152 

161 

6445  [4200] 

15.0 

140(160] 

36 

134 

143 

15.0 

140  [165] 

36 

121 

146 

16.0 

140(145] 

28 

120 

134 

9749  [4200] 

16.0 

140 [142] 

30 

136 

146 

7505  [4200] 

6640  [5544]  7910  [5544] 

PFS Rupture PFS Debonding PFS Debonding PFS : Debonding PFS Rupture PFS Rupture PFS Debonding :

PFS Debonding

*  The  values  in  [ ]  show the  measured  PFS  strains  at the  specimen's  mid­point  due to  upper or lower limit of fatigue loading.  ** Under static loading  * * * The values in [  ] show the PFS strains due to PFS prestressing  and the average PFS  strain is average value of 5  gages within the range of equal moment section. 

were  designed  with  0%,  25%  and  33%  prestress  levels  of  PFS  tensile  strength.  These  specimens  are  initially  tested  under  static  loading  in  four­point  bending at a loading rate of 2kN/min. Seven specimens reinforced  with  non­prestressed  (one  specimens)  and,  two  layers  of  25%  (four  specimens)  and  33% (two  specimens)  prestressed,  internal  PFS  are  tested  under  fatigue  loading  at  a  frequency  of  2.4Hz.  An  upper  load  is  set,  assuming PFS strains at beam mid­span  of about 0.0125 (in  all­prestressing  level), 0.015 (in 25%­prestressing) and 0.01  (in 33%­prestressing) at the first  cycle of fatigue  loading, and the lower load is set at a strain of 0.0014 in all  investigations.  The measured  PFS  strains  are  shown  in Table 2. Where the  beam specimens did not fail  under 2 million cycles of fatigue  load, they are  subsequently tested under static loading. Design valuables are displacement,  reinforcement  bar strain, PFS strain, concrete strain at beam mid­span, crack  width,  number  and  spacing  of  cracks  as  shown  in  Figure  2.  The  test  temperature is about 26­29°C. 

890

FRPRCS-6: Sustained and Fatigue Loads

Initiation  of crack A  2500 nHtfatioTToT crackl B 

EESjuuture. 

2000 1500 l< g J E J fl B7JM&>A• M

­RCS­0  RCS­2 • 5  RCS­3 • 3  60 20  40  D isplacem ent  rfjm )  Figure 3 Static Performance of  PFS­strengthened Beams with  Different  Prestressing Levels 

•*

B

5 1000

—B—No.4:RCF­0­59.5  — • —  No.5:RCF­25­64.0  •'  No.6:RCF­25­76.8  1  No,7:RCF­33­55.2  —A—NO.8RCF­33­66.0 

500

1000000

2000000

Number of cycles 

Figure  4  PFS  Strain  Versus  Loading  Cycles Relationships 

Initiation of  30  25  20 

­H—No.4:RCF­0­59.5  0  No.5:RCF­25­64.0  —No.6:RCF­25­76.8  —I  No.7:RCF­33­55.2  ­A—No.8:RCF­33­66.0 

15  10  5  0& 0 

ration  of  crack)  Initiation!  rackB  1000000  Number of Cycles 

2000000 

Figure 5 Displacement Versus  Loading Cycles Relationships 

0.35  0.3  0.25  0.2  0.15  0.1  0.05  0 

ckB 

­No.4:RCF­0­59.5  ­No.5:RCF­25­64.0  ­No.6:RCF­25­76.8  ­No.7:RCF­33­55,2  ­No.8:RCF­33­66.0 

1000000 

2000000

Number of Cycles 

Figure 6 Crack Width Versus  Loading Cycles Relationships 

EXPERIMENTAL RESULTS Typical  data  of  the  experimental  results  are  summarized  in  Table  2  and  the  load  versus  mid­span  deflection  curves  under  static  loading  is  shown  in  Figure  3.  Specimens  RCS­0  and  RCS­25  exhibited  debonding  failure  of  the  PFS, which  initiated  from  a flexural  crack within the beam  mid­span  region.  Only  specimen  RCS­33  showed  a  tensile  rupture  of  the  PFS  at  beam  mid­span.  The  ultimate  load  of  the  RCS­25  is  about  18.5%  higher  than  RCS­0,  and  cracking  load  and  yielding  load  of  reinforcing  bars  major  increases  due  to  prestressing  of  the  internal  PFS.  However,  comparing  RCS­33  with  RCS­25, ultimate  load  does  not  increase  although  prestress  of  internal  of  PFS  is  larger.  It  is  considered  to  be  due  to  a  premature  PFS  rupture  in RCS­33. Under fatigue  load, the relationships of  displacement, 

RC Beams with Externally Prestressed PBO Fiber Sheets 891 i  \ > i

i

1

/ 

\\

\ \ / ,  \ \\ (c)  No.3RCS­33 

1 

Crack A 

/ C r a c k B 

V\\  11

* 

K\ IX

!

x

n\ \\

Wtittf

\

(f)  No.6 RCF­25­76.8  Crack B 

Crack B  Crack A  Crack B  \

\

(d)  No.4 RCS­0­59.5  Crack B  Crack A  Crack B  r^cKBiracKA  crack a

f A\< Ah (e)  No.5  RCF­25­64.0  \

(

f 1 1/ {\

(b)  No.2RCS­25  Crack  B  Crack A 

(a)  No.l  RCS­0  i i 

1 \

irr A#\ \ III  tTlll 

tS&

(g)  No.7 RCF­33­55.2  (h)  No.8 RCF­33­66.0  Figure 7 Cracking Distributions before Yielding of Reinforcing Bars 

PFS rupture 

RCF­25­64.0(2)  RCS­25 

12000 

150 

2)

10000  8000 

100 

' \.J/

PFS debonding 

W"'­­ AB ­­ RR CC FF ­­ 20 5­ 5­ 694. 5 . 0 

50 

—©­RCF­33­66.0 

0  0

i 

i 

20 

40 

6000' X0%­prestressedPFS  strengthened  beam  s 

4000 m  2000  ­

0  Displacement  (mm) 

D25%­prestressedPFS  strengthened  beam  s  50 

U p p e r  L i m j t  o f  F a t i g u e  L o a d i n g 

100  ( k N ) 

Figure  8 Fatigue Performance  of  Figure  9 Relationship  of  PFS  Strain  at  PFS­strengthened  Beams  with  Failure under Following  Static Loading  Different  Prestressing Levels  Versus Maximum Fatigue Loading  PFS  strain  and  crack  width  at  beam  mid­span  versus  number  of  cycles  are  shown  in  Figures  4­6,  and  the  crack  patterns  before  steel  bar  yielding  are  shown  in  Figure  7,  where  crack  A  indicates  cracks  occurring  in  0­1  million  cycles  and  crack  B  those  in  1­2  millions  cycles.  Mid­span  displacements  appear  to  increase  temporarily  with  increasing  number  of  cycles  in  all  strengthened  beam  specimens.  However,  PFS  strain  and  crack  width  rapidly  decreased  when  new cracks occurred  under  fatigue  loading.  All  strengthened  beam  specimens  did  not  fail  under  2  million  cycles  of 

892 FRPRCS-6: Sustained and Fatigue Loads

Concrete surface debondins 5 Omni 

(a) No.2 RCS­25  (b) No.5  RCF­25­64.0  (c) No.6 RCF-25-76.8 Figure  10 Photos of Concrete  Surface  after  Debonding  Failure 

Non­fatigue  specimens  are  stiffer  than  Fatigue  specimens  180  j^­~»»—­—­­—  ­pFSndebonding i  M i n g  160  140  _ „ _ _  A.EES..debQiKiing...  140  120  120  100  80  60  RCF­25­64.0(l)  ­RCS­0  40  RCF­25­76.8 (L)  RCF­25­64.0G)  20  RCF­0­59. • 5  RCF-25-76.8 C)  0  20  40  0 80 20  40  60  Displacement  (mm)  Displacement  (mm)  Figure  11 Fatigue Performance  of  Figure  12 Fatigue Performances  of  25%­ no­prestressing  PFS  strengthened  prestressing PFS  Strengthened  Beams  with  Beams  Different  Upper Limit of Fatigue  Loading 

loading and were subsequently tested under static loading. The relationships  of  load  versus  beam  mid­span  deflection  under  static  load  are  shown  in  Figure  8. Compared  with result of static test  in Figure 3, ultimate  load and  steel yielding load of fatigue'damaged  specimens are the same or higher than  no­fatigue  damaged  specimens.  The  relationship  between  PFS  strain  at  failure under static loading and the upper limit of fatigue loading is shown in  Figure  9.  The  average  PFS  strain  at  beam  mid­span  at  failure  appears  to  decrease linearly with increasing upper limit of fatigue  loading. Appearence of the  concrete  surface  after  PFRP sheets  debonding and  cracking  patterns  are shown in Figure 10. The cover concrete debonded with PFRP sheets, and  the  depth  is  about  50mm  in  RC8­25,  about  2mm  by  40mm  width  in  RCP­25­64.0  and  about  2mm  in  RCF­25­76.8.  These  depths  decrease  linearly  with  increasing  upper  limit  of  fatigue  load.  It  is  considered  that  PFS­concrete interface is damaged 

RC Beams with Externally Prestressed PBO Fiber Sheets 893

150  130  110  90  70  50  30  10  ­10 

Hcrackhg  hsiie  of state  badhg  before  fatigue  badhg  EDcrackhg  hsiie  offatgue  badhg  RCF­25­76.8© 

PES rupture 

riifc^l

RCF­25­64.00 

[fir

RCF­33­66.0 

if

RCF­33­55.2 

­ * ­ R C S ­ 3 3  —1—  RCF­33­55.2  ­ A ­ R C F ­ 3 3 ­ 6 6 . 0 

RCF­25­76.8  RCF­25­64.0 

\

60 20  40  D ispfecem ent rfjm )  Figure  13 Fatigue Performances of  33%­Prestressing PFS Strengthened  Beams with Different  Upper Limit  of Fatigue Loading  0

ra ••

m

RCF­0­59.5  RCS­33  RCS­25  RCS­0  5  10  Numberof  crack 

15 

Figure 14 Comparison of Crack Number  among Different  Specimens before The  Yielding Steel  Reinforcement 

Table 3 Crack Spacing  No.

! 

Specimens

Crac k spacing after 2 millions of eye les of fatigue loading (mm)

1 

RCS­0 

84 

static

loading

2 

RCS­25 

84.5 

static

loading

79 

static

I  3  •'•  

;

C rack spacing at first cycle of fatigue loading (mm)

4 

5  6  7  :   8 

RCS­33  RCF­0­59.5  !  RCF­25­64.0  !  RCF­25­76.8  ;  RCF­33­55.2  RCF­33­66.0 

90.0  88.9  110.0  83.8  88.0 

loading 80.8  68.5  83.3  70.8  86.4 

according  to  the  fatigue  load.  Figures  11  to  13  show  the  load  versus  deflection  relations  at beam  mid­span  for  each  prestressing  level.  It  is  found  that the load­carrying  characteristics  both for the yielding load  of  reinforcing  bars  and  ultimate  load  increase  due  to  the  additional  cracking  occurring  under  fatigue  loading.  Moreover,  it  also  alters  some  other  characteristics  such  as  stiffness.  Comparison  of  number  of  cracks  among  different  specimens  is  shown  in  Figure  14  and  crack  spacing  is  shown  in  Table  3.  Crack  number  is  increased  and  crack  spacing  decreased  linearly  with  increasing loading cycles. 

894 FRPRCS-6: Sustained and Fatigue Loads

CONCLUSIONS Based  on  the  results  of  this  experimental  investigation,  the  following  conclusions can be  obtained:  (a)  It  is  shown  that  RC  beam  strengthened  with  prestressed  PFS  following  the P­PUT method  exhibits satisfactory  fatigue  performance.  (b)  The capacity of PFS­concrete  interface generally weaker with an increase in  the  upper  limit  of  fatigue  load;  however,  the  interfacial  performance  of  beams  strengthened  with  prestressed  PFS  is  the  same  as  those  of  beams  strengthened with no­prestressed PFS.  (c)  The  load­carrying  characteristics  in terms  of the yield  load  and  ultimate  load  increase  due  to  the  additional  cracking  occurring  under  fatigue  loading. Moreover, it alters some other characteristics  such as  stiffness. 

REFERENCES 1.  Wight  R.G.,  and  Erki  M.A., Prestressed CFRP Sheets for Strengthening Concrete Slabs in Fatigue,  International  Conference  on FRP  Composites  in Civil Engineering, Hong Kong, pp.  1093­1100, 2001.  2.  T.C.  Triantafillou  and  Deskovic  N., Innovative prestressing with FRP sheets, Mechanics of short-term behavior,  Journal  of  Engineering  Mechanics, ASCE, Vol. 117, No.7, pp. 1653­1672, 1991.  3.  EI­Hacha  R.,  Wight G., and  Green.  M.F., Strengthening Concrete Beams with Prestressed Fiber Reinforced Polymer Sheets:Behavior at Room and Low, Fourth International Symposium Fiber Reinforced Polymer  Reinforcement  for Reinforced Concrete Structures, ACI, SP­188, pp.737­749,2000.  4.  Wu Z.S., Matsuzaki  T. Fukuzawa  K.and  Kanda T, Strengthening Effects on RC Beams with Externally Prestressed Carbon Fiber Sheets,  Journal  of Material, Concrete  Structures and Pavements, JSCE, pp. 153­165, 2000.  5.  Wu  Z.S., Iwashita  K.,  Hayashi  K.,  Higuchi  T,  Murakami  S.,  Koseki  Y,  Strengthening PC structures with externally prestressed PBO fiber sheets, International  Conference  on  FRP  Composites  in  Civil  Engineering,  Hong Kong, pp.  1085­1092, 2001.  6.  Wu  Z.S.,  Iwashita  K.,  Hayashi  K.,  Higuchi  T,  Murakami  S.,  Koseki  Y.,  Strengthening Method for RC structures with externally prestressed PBO fiber sheets,  Journal  of  The  Japan  Society  for  Composite  Materials,  Vol.28, No.4, pp.146­155, 2002. 

FRPRCS-6,  Singapore, 8­10  July 2003  Edited by Kiang Hwee Tan  ©World Scientific  Publishing Company 

PRESTRESSED  CFRP SHEETS FOR  STRENGTHENING  REINFORCED  CONCRETE  STRUCTURES IN FATIGUE  R. EL­HACHA,  R.G. WIGHT, P.J. HEFFERNAN  AND  M.A.  ERKI  Department of Civil Engineering, Royal Military College of Canada P.O. Box 17,000 Station Forces, Kingston, ON, K7K 7B4 CANADA Strengthening  with  externally  bonded  prestressed  carbon  fibre  reinforced  polymer  (CFRP)  sheets  combines  the  benefits  of  excellent  durability  and  structural  improvement  in  terms  of  serviceability  and  ultimate  conditions.  The  addition  of  CFRP  sheets  to  the  tension  surface  of  a  beam  or  slab  reduces  the  stresses  in  the  steel  and  this  can  dramatically  improve  the  fatigue  life  of  the  specimens.  By  prestressing  the  sheets,  stresses  are  further  reduced  and  can  result  in  even  greater  increases  in  fatigue  life.  During  the  initial  stages  of this  research  program,  three  one­way  slabs  and  three  large­scale  reinforced  concrete  T­beams  were  constructed  and  tested  to  investigate  the  effectiveness  of  using  prestressed  CFRP  sheets  to  strengthen  concrete  beams  subjected  to  fatigue  loading.  The  loading  applied  was very  severe with  an  amplitude  of  loading  from  a small  preload  to  a  load  from  90­100%  of  the  yield  load  of  the  steel  reinforcement.  For  both  the  beams  and  the  slabs,  one  specimen  was  unstrengthened  and  used  as  a  control,  the  second  specimen  was  strengthened  with  non­prestressed  CFRP  sheets,  and  the  third  specimen  was  strengthened  with  prestressed  CFRP  sheets  tensioned  to  approximately  30%  of  their  ultimate  tensile  strength.  An  anchorage  system  was  developed  to  directly  prestress  the  CFRP  sheets  against  anchors mounted  of the strengthened  structure  itself.  The  testing  program  for  the slabs confirmed  the benefits  of prestressing  the  CFRP  sheets  and  the  practicality  of  the  prestressing  technique  used.  The  prestressed  sheets  bonded  to  the  lower  surface  of  the  beam  relieved  the  stresses  present  in  the  internal  reinforcing  steel  and  were  much  more  effective  at extending the fatigue  life,  increasing  the strength  and  improving  the  serviceability  of  the  beams  than  non­prestressed  sheets.  For  the  T­ beams  however,  the  details  of the  anchorage  were  such  that  a failure  at  the  anchor  location  caused  a premature  failure  of the  strengthening  system  and  the  benefits  of  prestressing  were  less  significant.  Similar  anchors  had  performed  very  well  under  static  loading.  It  is  clear  that  caution  must  be  exercised  when  designing  prestressing  anchors  if  the  structure  will  be  subjected  to very severe  loading. 

896 FRPRCS-6: Sustained and Fatigue Loads

INTRODUCTION Non­prestressed  FRP sheets bonded to the tension face of an under­strength  or  structurally  damaged  concrete  member  supplement  the  flexural  reinforcement  of the deficient  member. Studies have shown that the flexural  strength under service load was improved slightly and the ultimate strength  was significantly  increased. However,  only  a portion of the strength  of the  FRP  sheets  is  effective  in  non­prestressed  strengthening  applications.  To  improve  the  efficiency  of this  strengthening  technique,  the  sheets  may be  prestressed  prior  to  bonding.  This  strengthening  technique  offers  the  benefits  of  both  a  prestressed  system  that  contributes  to  load  carrying  capacity  even  before  further  deformation  occurs  in  the  structure,  and  a  bonded  system  that  sustains  a significant  portion  of the load under  further  deformations.  Prestressed  FRP sheets can improve the serviceability  of the  beam by limiting deflection  and providing excellent  control  of cracks, and  can  restore  prestress  to  a  system  that  has  suffered  a  loss  of  internal  prestressing.  The benefits  and advantages  associated  with prestressed  FRP  sheets have been discussed by El­Hacha et al., (2001).  There  is very  little research  into the use of prestressed  FRP sheets  for  strengthening  beams  and  slabs.  Generally,  researchers  found  that  failure  was  due  to  fracture  of  the  internal  reinforcing  steel,  the  fatigue  life  was  increased and the deflections were decreased2'3,4,5'6.  Research  into  the  fatigue  behaviour  of  members  strengthened  with  prestressed  FRP  sheets  is  also  limited.  A  fatigue  test  was  carried  out  at  EMPA  with  the  CFRP  plate  prestressed  to  50%  of  its  strength.  Thirty  million cycles were performed without any evidence of damage to either the  concrete  or  the  CFRP  plate7.  Wight  and  Erki  (2001)  found  that  the  prestressed sheets were much more effective  at extending the fatigue life of  the reinforced  concrete slabs than non­prestressed sheets.  OBJECTIVES This investigation is part of a large experimental/analytical  project studying  the fatigue behaviour of concrete beams and slabs strengthened with bonded  non­prestressed  and  prestressed  CFRP  sheets.  Only  the  effectiveness  of  CFRP  sheets  when  used  to  strengthen,  in  flexure,  simply  supported  reinforced  concrete T­beams  is reported in this paper and will be compared  with the results of the one­way reinforced  concrete slabs. A new mechanical  anchorage  system  has  been  developed  as  part  of  this  study  and  used  to  directly  prestress the CFRP sheets by jacking  and reacting against  anchors  mounted on the web of the strengthened beam itself. 

Prestressed CFRP Sheets for RC Concrete Structures 897

EXPERIMENTAL PROGRAM Specimen Details, Test Set-Up and Strengthening Materials The experimental  program consisted  of testing, simply supported, three 4.0  m  reinforced  concrete  T­beams,  and  three  3.0  m  one­way  reinforced  concrete  slabs (90mm  x  1000mm) under four­point  loading. In each group,  one specimen was unstrengthened and used as control, the second specimen  was strengthened with non­prestressed CFRP sheets, and the third specimen  was strengthened with prestressed CFRP sheets. The specimens were tested  to failure 7 days after strengthening. A summary of these specimens is given  in  Table  1. The  28­day  compressive  strength  of the  concrete  was 40MPa.  The  test  set­up  of  a  T­beam  specimen,  reinforcement  details  and  strengthening schemes are shown in Figure  1. The total load was applied to  the  specimen  using a 500kN  capacity  actuator through an MTS  controller­ testing  machine  operating  under  load­control  mode.  The  specimens  were  initially  loaded to 5kN to ensure stability before  starting the cyclic loading.  All  fatigue  loads  were  applied  at  a rate  of 2Hz, and  data  was  recorded  at  every  1000th cycle and  10,000th cycle. The beams were subjected to a cyclic  load  from  5 to  65kN,  and  the  slabs  to  a cyclic  load  from  5 to  55kN.  All  specimens were fully  instrumented.  Table  1. Experimental  Program 

Beam # TBeamC  TBeamNP  TbeamP2  SF  SF­N  SF­P 

Type of Strengthening Unstrengthened Control Beam  Four layers of Non­Prestressed CFRP Sheets  Four layers of Prestressed CFRP Sheets  Unstrengthened Control Slab  Two Strips (two layers each) of Non­Prestressed CFRP Sheets  Two Strips (two layers each) of Prestressed CFRP Sheets 

Strengthening System The FRP system used for  strengthening the T­beams and slabs was the wet  lay­up  composite  system  consists  of  dry unidirectional  high tensile  carbon  fiber  sheet9.  According  to  the  manufacturer  the  sheet  had  a  nominal  thickness of 0.165mm/ply, and ultimate tensile strength and tensile modulus  of  elasticity  of  the  sheets  were  3800MPa  and  227GPa,  respectively.  The  design  strength  per  unit  width  was  627N/mm.  Details  of  T­beam  strengthening  are  given  in  the  following  sections  and  details  of  slab  strengthening are given in Wight and Erki (2001). 

898 FRPRCS-6: Sustained and Fatigue Loads 1.0 m

LVDT

-1.55m

1.55m

«-'• - —0.8 m —H^~"

L  2000000 

No  evident  deformation 

N„16 

54 

60%F 0 7 . 

150% Fo.7 

49 

140 

266037 

Concrete  shear  failure 

N 0 23 

58 

20% F0.7 

112% F0.7 

52 

180/260 

1040582  (stop)  (rebar  640002) 

1 rebar  fracture 

N024 

58 

12%Fo,7 

120%F0,7 

63 

207/274 

371571  (rebar  370003) 

2rebars  fracture 

Beams  strengthened  with  carbon  sheet 

Precracked RC Beams with CFRP Sheets 921

Figure  11.  Deflections of strengthened specimens 

0

100000 200000 300000 400000 500000 600000 700000 800000 cyclas

Figure 12.  Propagation of crack widths in specimen N023  CONCLUSIONS This test study has shown  that:  (a)  The  external  bonding  of  CFRP  sheets  on  initially  crack­damaged  RC  structures  is  an  effective  method  to  restore  and  improve  the  structural  strength  and ultimate fatigue  life.  (b)  For  beams  with  a  small  shear  span/beam  depth  ratio,  besides  strengthening  on  tensile  surface,  additional  strengthening  in  shear  is  usually necessary  under static and  fatigue  loads.  (c)  Under  static  and  fatigue  loads,  CFRP  sheet  reduces  the  propagation  of  cracks  because  of  the  bridging  effect.  On  the  other  hand,  the  stress  redistribution  between  CFRP  sheet  and  steel  reduce  reduce  the  stress  range in the steel rebars.  (d)  The  development  of  strains  in  steel  rebar  under  fatigue  load  in  beams  strengthened  with  carbon  cloth  is  different  from  that  in  unstrengthened  beams.  (e)  The  fracture  of  steel  rebar  is  the  main  reason  for  fatigue  failure  of  RC  beams strengthened  with CFRP  sheet. This kind  of failure  is not  sudden: 

922 FRPRCS-6: Sustained and Fatigue Loads

after  the  fracture  of  the  steel  rebars,  the  specimen  does  not  fail  completely,  (f)  The  stress  range  is  the  principal  reason  for  the  fatigue  fracture  of  steel  rebars.  Morever,  the  magnitude  of  crack  width  affects  this  fracture  significantly.  ACKNOWLEDGMENTS This study was supported by DRAST(MEMR)  and the  CFRP sheets  were  TFC®, provided by Freyssinet  International. 

REFERENCES 1.  Clement,  J.L.,  Dumas,  C.  and  Belhoul,  M.,  Numercial  simulations  of  RC  beams  strengthened  by  carbon  cloth.  Computational  modelling  of  concrete  structures  (EURO­C1998),  de  Brost,  Bicanic,  Mang  and  Meschke(eds), Balkema, Rotterdam, ISBN  90 54  10 946, pp.741­748.  2.  Hamelin,  P. and Ferrier,  E., Etude bibliographique  sur  les  renforcements  par  materiaux  composite  de  structures  du  genie  civil.  Rapport  N°  LCPC/01/du  27/04/01.  3.  Ferrier,  E.,  Naseri,  H.  and  Hamelin,  P.,  Fatigue  Behavior  of  Composite  Reinforcement  for  Concrete  Structure.  Fourth  International  Symposium  Fiber  Reinforced  Polymer  Reinforcement  for  Reinforced  Concrete  Structures. ACI  special publication  SP­188, pp. 535­545,  1999.  4.  Shahawy,  M.  and  Beitelman,  T.E.,  Fatigue  Performance  of  RC  Beams  Strengthened by CFRP Laminates. CDCC 98, pp.  169­178, 1998.  5.  Barnes, R.A.  and  Mays,  G.C.,  Fatigue  Performance  of  Concrete  Beams  Strengthened  with  CFRP  Plates.  Journal  of  composite  for  construction  may,  1999 pp.63­72.  6.  Chollaway,  L.  and  Bleeming,  M.  Strengthening  of  Reinforced  Concrete  Structure  ­  Using  Externally  Bonded  FRP  Composites  in  Structure  and  Civil Engineering  .  Woodhead  publishing  england and  CRC press  USA  2001.  7.  Wu,  Z.Y.,  Clement,  J.L.,  Tailhan,  J.L.,  Boulay,  C.  and  Fakhri,  P.,  Fatigue  Test  on  Damaged  reinforced  Concrete  Specimens  Strengthened  by  Carbon  Cloth.  Proceedings  of  HPSC2002,  Seville  Spain  2002,  pp.  347­355. 

FRPRCS-6, Singapore, 8­10  July 2003  Edited by Kiang Hwee Tan  ©World Scientific Publishing Company 

FATIGUE INVESTIGATION OF CONCRETE BRIDGE DECK SLAB REINFORCED WITH GFRP AND STEEL STRAP

A. H. MEMON, A. A. MUFTI AND B. BAKHT  ISIS Canada, University of Manitoba #A250 Agricultural and Civil Engineering, Winnipeg, MB, R3T2N2, Canada This  paper  describes  the  behavior  of  one  segment  of the cast  in  situ  full­ scale  model  of  bridge  deck  slab  reinforced  with  glass  fibre  reinforced  polymer  (GFRP)  internally  and  externally  with  steel  strap.  Due  to  this  hybrid  GFRP and steel strap, corrosion  can be eliminated  completely  from  the  deck  slab,  leading  to  an  economical  durable  bridge  deck  system. The  model  of  the  deck  slab was tested  under  cyclic  loading to  investigate  the  fatigue  behavior.  The cyclic load was simulating  the effect  of truck wheel  load. From the test results, it was observed that at 25 tones load level, the  model  deck  slab completed  1,000,000 cycles without any damage and this  satisfied  the serviceability  and  lifetime  number of axles that a bridge deck  would experience. The deflection  behavior,  crack width, strain distribution  in GFRP bars and steel straps with the number of cycles are reported.  INTRODUCTION The  decline  in North  American  infrastructure  has never been more  prevalent  than  it  is today.  In particular,  the  highway  system  and  its  bridges  have  been  adversely  affected  by  age  and  weathering  over  the  past  two  decades.  Surveys  show  that  the  majority  of  the  highway  bridges  have  reinforced  concrete  decks  supported  on steel  or concrete girders. Concrete has been  the  choice  for  highway  bridge decks for  a long time. Over the years, the  weather  has taken  its toll  on these reinforced  concrete  decks. Rainwater  and  de­icing  chemicals  applied  to  roadway  surfaces  during  the  winter  months  have  seeped through  many concrete  decks and caused  corrosion  of the  reinforcing  steel. To  avoid the corrosion  of steel, Mufti  et al.1 proposed the concept  of a  steel­free  deck  slab  entirely  free  of  any  internal  steel  reinforcement.  In  a  steel­free  deck slab, steel  straps  are connected  to the top  flange  of the  girder  providing  lateral  restraint,  which  develops  compressive  membrane  force  in  the  deck  slab.  In  operation,  as  the  bridge  deck  withstands  the  stress  of  the  traffic  loads,  the  deck  will  deform  and  stresses  develop.  Eventually  the 

924 FRPRCS-6: Sustained and Fatigue Loads

loads  will  reach  a  magnitude  where  the  tensile  stresses  will  cause  the  concrete to crack. Once the bridge deck cracks it resists traffic  loads through  arching  action.  Such  arching  action  is  characterized  by  compressive  membrane action and failure by punching shear.  This  paper  describes  a  study  to  replace  the  steel  reinforcement  completely from  the deck slab with hybrid internal GFRP and external steel  strap  as an alternative  solution  to  increase the  service  life  of bridges. This  study investigates the fatigue behavior of a cast in situ full­scale  model of a  bridge  deck  slab. This full­scale  model  of a deck slab is divided  into three  segments  (A, B and C). Segment A is designed according to the Canadian Highway Bridge Design Code  Section  8,  while  segment  B  and  C  are  designed  according to  Section  16 and make use of the principle of a steel­ free  deck  slab  that  is  confined  transversely  by  steel  straps.  In  segment B,  carbon  fibre  reinforced  polymers  (CFRP)  are  used,  while  in  segment  C,  GFRP  are  used  to  reduce  the  chances  of  development  of  cracks.  The  test  results of segment C, GFRP panel is reported in this paper.  During  its  lifetime,  a  bridge  deck  slab  is  subjected  to  a  very  large  number  of  wheels  of  different  magnitudes.  By  contrast,  the  laboratory  investigation  of  the  fatigue  resistance  of  a  bridge  deck  slab  is  usually  conducted  under  wheel  loads  of  constant  magnitude.  The  design  codes  (AASHTO  2  and  CHBDC  3  )  are  not  explicit  with  respect  to  the  design  fatigue  loads on deck slabs. An analytical  method was developed  by  Mufti  et  al.4  for  establishing  the  equivalence  between  fatigue  test  loads  and  a  given population of wheel loads. While the method is general  enough to be  applicable  to  all  deck  slabs  of  concrete  construction,  it  is  developed  especially  for  steel­free  deck  slabs  1,s  which  are relatively  new and  do not  have a long track record of field performance. In this paper, fatigue behavior  of segment C, model of bridge deck slab is reported.  EXPERIMENTAL PROGRAM Details of Bridge Deck Slab The full­scale  model  of bridge deck slab was tested under cyclic loading to  investigate fatigue behavior. The model of deck slab consisted of 3.0 m each  with  overall  dimensions  of  9.0  x  3.0  mm  and  a thickness  of  175  mm,  as  shown  schematically  and before  casting in Figures  1(a) and  1(b). The deck  slab  was  cast  in  situ  compositely  on two steel  girders  at a center­to­center  spacing  of  2.0  m through  the  use  of  shear  connectors  and  had  a  500 mm  long  cantilever  overhanging  beyond  the  center  of  the  each  girder.  In the 

Concrete Bridge Deck Slab with GFRP and Steel Strap 925

longitudinal direction, (i.e.: parallel to the girders), the deck slab was 9.0 m  long.  Although cast monolithically, the slab was conceptually  divided  into  three  segments, as shown in Figure  1(a).  Segment C of the deck slab was  reinforced  with  hybrid  GFRP  internally  and  externally  with  steel  strap.  GFRP  bars  with  a  modulus  of  elasticity  40.8  GPa  and  ultimate  tensile  strength of 690 MPa were used.  One GFRP bar with a diameter of 12.7 mm  (#4) is spaced at  150 mm in the transverse direction, providing a reinforcing  ratio of 0.48% (0.16% per m), and one 12.7 mm (#4) is spaced at 200 mm in  the longitudinal direction, providing a reinforcing ratio of 036% (0.12% per  m). The steel strap having dimensions 25.4 x 38.1 mm (1" x  1.5")  is spaced  at  1000 mm in the transverse direction only, providing the reinforcing  ratio  of 0.55% (0.18% per m).  A steel free  concrete  deck slab of girder  bridge  derives  its  strength  from  an  arching  action,  which  is  harnessed  by  both  longitudinal,  and  transverse  confinement  system,  the  latter  may  comprise  transverse  steel  strap connected  to the top flange  of the girder.  To control  the temperature and thermal cracking, concrete is mixed with 0.3% chopped  polypropylene  fibers.  The average concrete  compressive  strength  was 57.7  MPa.  H-lfflttttlilhTiirtliTfl _L 1 I 1

•

1 M

1 1

y

if

1 InMiHlri-J|J-1J_M^IM

00

-4rnTrTOTtr MM JL N i l hj' M1 Segment A

4

Segment B

00 • iTtfrf l 1 , ,., iJ

Segment C

r

m (12.7 mm) cSa Asian 190 QFRP rsbar @ spacing ef 15C mm In Transvsrs* and 200

Figure  I (a), Reinforcement  detail or the deck slab 

Figure 1(b). Picture of the deck slab showing GFRP 

926 FRPRCS-6: Sustained and Fatigue Loads

Instrumentation To Investigate the performance  under cyclic loading, the model of the deck  slab  was  monitored  through  a  number  of  sensors,  which  included  linear  variable  displacement  transducers  (LVDTs),  strain  gauges5  and  pi­gauges.  Vertical  deflection  of  the  model  was  measured  by  LVDTs.  In  order  to  measure deflection  of the deck slab with respect to girders, the displacement  transducers  (5 LVDTs) were attached to steel beams resting directly above  the center of the girders, as shown in Figure 2. 

Figure 2. Test Setup and Instrumentation 

The displacement transducers were located  at mid­span and on all  four  sides in both the longitudinal and transverse centerline of the model, so that  profile  of  the  deflection  can  be  obtained  with  respect  to  the  increasing  number of cycles. In order to monitor the performance of the steel straps, an  electrical resistance  strain gauge was mounted on the middle of each strap.  To  monitor  internal  performance  of  the  deck  slab through  GFRP bars, an  electrical  resistance  strain gauge was mounted  on five transverse and three  longitudinal bars at the center and away from the center as shown in Figure  1(b).  The  purpose  of  doing this  was  to  get  the  distribution  of  strain  with  respect to the number of cycles. The ultimate goal is to monitor the yielding  of the  steel  strap and rapture of the  GFRP bar. Pi­gauge instruments  were  used  to  measure  the  crack  width.  To  monitor  the  crack  width  of  the  longitudinal  crack of the model deck slab, three pi­gauges were mounted at  the  bottom  surface  of  deck  slab  at  the  center  (under  the  load)  and  1.0 m towards West end (near the edge beam)­ and 1.5 m towards East, as shown in  Figure 3.  Testing Procedure To understand the fatigue behavior of cast in situ full­scale  model of bridge  deck  slab,  cyclic  test  was  conducted  at  25  t  load.  The  load  was  applied 

Concrete Bridge Deck Slab with GFRP and Steel Strap 927

through a hydraulic actuator. The load cell has a capacity of  1000 kN. The  load  was controlled  by an MTS controller.  For this  dynamic test, the  load  control was selected at 1  Hz frequency. The data was recorded through Data  Acquisition  System.  The  deck  slab  was tested  under  a central  rectangular  patch load measuring 610 x 305 mm, with the later dimension being in the  longitudinal direction of the deck slab. The test setup is shown in Figure 2. 

Figure 3. Bottom View,  steel straps and  Instrumentation 

WHEEL LOADS DATA  Commentary  Clause  C3.6.1.4.2  of  the  AASHTO  2  Specifications  (1998)  notes that the Average Daily Traffic  (ADT) in a lane is physically limited to  20,000 vehicles, and the maximum fraction of tracks in traffic  is 0.20. Thus,  the  maximum  number  of trucks per  day in one direction (ADTT) is 4,000.  When  two  lanes  are available  to trucks, the number  of trucks per  day in a  single lane, averaged over the design life, (ADTTSL) is found by multiplying  ADTT by 0.85, giving 3,400. It is assumed that the average number of axles  per truck is four  (a conservative assumption), and that the life of a bridge is  75  years.  The  maximum  number  of  axles  that  a  bridge  deck  would  experience on one lane during its lifetime is 372 million.  Matsui  and Tei 6  described the maximum axle load observed  in Japan  as 32 t, or 314 kN. The close correspondence between the expected annual  maximum  axle weight  in Canada4 and the maximum observed  axle load in  Japan  indicates  a  similarity  between  the  axle  loads  in  the  two  countries.  Matsui and Tei 6 also provided a histogram of axle weights observed on 12  bridges in Japan. In the absence of data on Canadian tracks, this histogram  was used to construct the wheel load statistics, as shown in Table 1. 

928 FRPRCS-6: Sustained and Fatigue Loads Table  1. Statistics of wheel  loads  Wheel Weight (Tons)

1

2

3

Percentage of Total (%)

21.25

32.06

21.61

12.60 6.48 3.24 1.44

0.54

No. of Wheels (million)

79.05

119.3

80.39

46.87 24.1 12.1 5.37

2.01

Wheel Weight (Tons)

9

10

11

12

Percentage of Total (%)

0.32

0.18

0.11

0.07

0.04 0.02 0.01 0.003

No. of Wheels (million)

1.19

0.67

0.41

0.26

0.15 0.07 0.04

4

5

13

6

14

7

15

8

16

0.01

This table also includes the numbers of wheels of various magnitudes,  corresponding to a total  of 372 million wheel  passes. Any fatigue  test load  on  a  bridge  deck  slab  should  cause  the  same  damage  in  the  slab  as  the  damage caused  by all the wheel  loads included  in this or any similar table.  An  analytical  method  was  developed  by  Mufti  et  al.,4  according  to  this  approach:  N2=Nl*e(K>-*>>'3° 

(1) 

RX=P,/PS 

(2) 

R2=P2/PS 

(3) 

with  and 

Pi and  P2 are  two  different  wheel  loads; Ni  and N2  are the  corresponding  number  of  passes  of  Pi  and  P2  respectively;  while  Ps  is  the  static  failure  load.  The ultimate capacity of the deck slab can be predicted by using Punch  Program developed by Newhook and Mufti7.  The reliability of this program  is  discussed  in  Mufti  and  Newhook  8.  Therefore  according  to  the  Punch  Program,  the  ultimate  capacity  of the  segment  C model  deck  slab is about  827 kN (84.41).  Consider  that  model  deck  slab  has  a static  failure  load  (Ps)  of  84.4  t;  while Pi, P2 and Ni are  11, 161 and 79.1 million, respectively. By using the  equation (1), N2 will be 0.38 million, as shown in Table 2. From this table 2,  it  is  clear  that  the  maximum  number  of  axles  that  a  bridge  deck  would  experience in one lane, during its lifetime,  is 372 million. This includes the  number of wheels of various magnitudes from  1  t to 161. 

Concrete Bridge Deck Slab with GFRP and Steel Strap 929 Table 2. Lifetime number of cycles, equivalent to 161 wheel load  Static Load (Ps)

Load (Pi)

No. of Cycles (n1)

Load (P2)

No. of Cycles (n2)

(Tons)

(Tons)

(million)

(Tons)

(million)

84.4

1

79.05

16

0.38

84.4 84.4

2 3 4

119.26 80.39 46.87

16 16 16

0.82 0.79 0.66

84.4

5

24.11

16

0.48

84.4

6

12.05

16

0.34

84.4 84.4

7 8

5.37 2.01

16 16

0.22 0.12

84.4 84.4

9 10

1.19 0.67

16 16

0.10 0.08

84.4 84.4 84.4 84.4

11 12 13 14

0.41 0.26 0.15 0.07

16 16 16 16

0.07 0.06 0.05 0.03

84.4

15

0.04

16

0.03

84.4

16

0.01

16

84.4

Total No. of Cycles @ 16 Tons

0.01

4.25

RESULTS AND DISCUSSION Test  results  show  that  at  25  t  load  level;  model  deck  slab  completed  1,000,000  cycles,  without  any  damage.  As  mentioned  earlier,  during  its  lifetime,  a bridge deck slab  is subjected  to 372 million  cycles of  different  magnitudes. By using an analytical  approach,  which was explained  earlier,  372  million  cycles  are  equivalent  to  4.25  million  cycles  of  16 t  load,  as  shown  in Table 2. From the experimental  results, the deck  slab completed  1,000,000  cycles  at  25  t,  which  is  equivalent  to  4.25  million  cycles  at  a  20.93 t. This can increase the carrying capacity to approximately 30% more.  Vertical  deflection  of  the  deck  slab  was  measured  by  displacement  transducers.  It was  observed  that  maximum  deflection  was  obtained  at the  center  of  the  deck  slab,  as  shown  in  Figure  4.  From  Figure  4,  it  is  clear  deflection  was  increased  with  increasing  number  of  cycles.  It  was  also  found that the maximum deflection is less than the permissible limit.  In order to measure the internal  and external response of the deck slab,  strain  gauges  were  mounted  on  every  strap,  GFRP  bars  in transverse  and  longitudinal  and maximum response was measured under the applied cyclic  load,  as shown  in Figure  5. From Figure  5, it is clear that maximum  strain  was achieved by GFRP bar in transverse direction, followed  by GFRP bar in  longitudinal  direction  and  steel  strap.  All  values  are  lower  than  the  serviceability limit. 

930 FRPRCS-6: Sustained and Fatigue Loads

2.S 2.0 -j

I,.. 5

',

s

1.0 LVDT2@Center

0.5

0

200000

400000 600000 800000 No. of Cycles (Nos.)

1000000

1200000

Figure 4. Deflection­No. of cycles near the applied load 

1000 O X Z.

800 600

|

400.

.*-

-'

.

..- ""

200

-

-

-

"

•

"

- - - -Steel Strap — — GFRP-Longitudinal

0

200000

400000 600000 800000 No. of Cycles (Nos.)

1000000

1200000

Figure 5. Strain­No. of Cycles under the applied load  A longitudinal crack was observed along the centerline of the deck slab  at  20,000  cycles.  Few  cracks  were  noticed  when the  deck  slab  completed  1,000,000 cycles. To monitor the propagation of the centerline  longitudinal  crack, a pi­gauge was mounted at the bottom of the deck slab under the load.  The crack width with the increasing number of cycles is shown in Figure 6.  From  Figure  6,  it  is  clear  that  crack  width  increased  with  the  increasing  number  of  cycles  and  maximum  crack  width  was  found  less  than  the  permissible limit.  From these  results  it  is clear  that the  maximum  deflection  of the  deck  slab,  strain  in  the  steel  strap,  GFRP  transverse  and  longitudinal  bars  and  crack width in the deck slab are within the permissible limits. 

Concrete Bridge Deck Slab with GFRP and Steel Strap 931

0.4 -,

I" ack

3

1o

PG2@Under Load

(

200000

400000

600000

800000

1000000

1200000

No. of Cycles (Nos.)

Figure 6. Crack width­No. of cycles under the applied  load 

CONCLUSIONS  Based  on the analysis  and findings  of this  investigation,  the  following  conclusions can be drawn:  (a)  The  presence  of GFRP  bars in the  deck  slab  reduced  the  chances of  development of cracks.  (b)  The  concrete  deck  slab  reinforced  with  hybrid  GFRP  and  steel  strap  completely eliminate the corrosion of the deck slab.  (c)  Test  results  show  that  GFRP  bar in  transverse  direction  reached  maximum  strain  under  load  of about  0.12%,  which  is  60%  of the  service strain, hence area of the GFRP bar  in transverse  direction can  be reduced up to 40%.  (d)  Experimental  results  show  that  fatigue  damage  induced  at 25 t,  are  within permissible limits.  (e)  Analytical  results  show  that  during  the bridge  deck's  lifetime,  it is  subjected  to 372  million  cycles  of different  magnitudes,  which are  equivalent to about 4.25 million cycles of a 161 wheel load.  (f)  Experimental  results show that,  1,000,000 cycles at 25 t, equivalent to  4.25  million  cycles of a 20.93 t, can increase the carrying capacity to  approximately 30% more during the lifetime.  (g)  The deck slab satisfied  the serviceability  and lifetime  number of axles  that a concrete bridge deck slab would experience. 

932 FRPRCS-6: Sustained and Fatigue Loads

ACKNOWLEDGEMENTS The financial  assistance provided by ISIS Canada, A Network of Centres of  Excellence,  and  the  Cement  Association  of  Canada  are  gratefully  acknowledged. The authors gratefully  acknowledge the GFRP bars provided  by  Hughes  Brothers,  Inc.  USA.  Special  thanks  to  Moray  Mcvey,  Grant  Whiteside and Liting Han for their assistance during fabrication  and testing  of the deck slab and the administrative support from ISIS Canada.  REFERENCES 1.  Mufti,  A.A.; Jaeger,  L.G.; Bakht, B.; and Wegner, L.D., "Experimental  Investigation  of  FRC  Slabs  Without  Internal  Steel  Reinforcement,"  Canadian Journal of Civil Engineering, V. 20 No.3,  1993, pp. 398­406.  2.  AASHTO, LRFD Bridge Design Specifications,  American  Association  of State Highway and Transportation Officials,  Washington, D.C., 1998.  3.  CHBDC,  "Canadian  Highway  Bridge  Design  Code,  "Canadian  Standards Association International, Toronto, 2000.  4.  Mufti,  A.A.,  Memon,  A.H.,  Bakht,  B.,  and  Banthia,  N.,  "Fatigue  Investigation  of  the  Steel­Free  Bridge  Deck  Slabs," ACI  International  SP­ 206, American Concrete Institute, 2002, pp. 61­70.  5.  Bakht,  B.  and  Mufti,  A.A.,  "Five  Steel­Free  Bridge  Deck  Slabs  in  Canada," Journal of the International Association for Bridge and Structural Engineering (IABSE), V. 8, No. 3, 1998, pp. 196­200.  6.  Matsui,  S.  and  Tei,  K.,  "Researches  and  Japanese  Developments  on  Highway Bridge Slabs and Contribution of Wheel Running Machines,"  Proceedings, Third International Conference on Concrete under Severe Conditions, Vancouver, June  18­20, 2001, V. 1, pp. 992­1008.  7.  Newhook,  J.P. and Mufti,  A.A., "Punch Program User Manual  ", Nova Scotia CAD/CAM Centre Dalhousie University,  Hallifax,  Nova  Scotia,  September, 1998.  8.  Mufti, A.A. and Newhook, J.P., "Punching Shear Strength of Restrained  Concrete  Bridge Deck slabs", ACI Structural Journal, 95(4),  1998, pp.  375­381. 

Prestressed FRP Reinforcement  and  Tendons 

This page is intentionally left blank

FRPRCS-6, Singapore, 8­10 July 2003  Edited by Kiang Hwee Tan  ©World Scientific Publishing Company 

FATIGUE OF HIGH STRENGTH CONCRETE BEAMS PRETENSIONED WITH CFRP TENDONS B. B. AGYEI AND J. M. LEES  Dept. of Engineering, Univ. of Cambridge, Trumpington St., UK G. P. TERRASI  SACAC Schleuderbetonwerk AG, Lenzburg, Switzerland The  research  investigates the  fatigue  resistance  of  high  strength  concrete  beams  prestressed  with  CFRP tendons.  The  fatigue  performance  of  bare  CFRP tendons has been found  to be very good.  However, the influence of  stress  concentrations  at concrete  crack  locations  on the  fatigue  behaviour  requires  further  verification.  In  the  main  experiments,  concrete  beams  pretensioned  with  CFRP  tendons  were  subjected  to  cyclic  loading.  The  maximum and minimum  loads were chosen so that the beams were cycled  above and below the cracking load.  Thus, as the cracks opened and closed  the concrete crack faces  rubbed against each other and potentially  against  the  tendon.  It  is  concluded  that  the  fatigue  resistance  of  CFRP  tendons  does  not appear to be affected  by  stress concentrations  at crack  locations  since debonding occurs and reduces the strain concentrations. 

INTRODUCTION The  first  CFRP  prestressed  high  strength  concrete  pylon  for  transmitting  electricity was produced  in Switzerland, in September 2000'. The 27m pole  was  manufactured  using  a  centrifugally­cast  high  strength  concrete  containing  silica­fume­blended  cement  and  polypropylene  chopped  fibre  reinforcement.  The  CFRP  prestressed  pole  had  a wall  thickness  of  only  40mm  and  the  weight  was  40%  less  than  that  of  a  conventional  steel  prestressed concrete pole design.  Having  successfully  used  CFRP  tendons  in  pylons,  the  potential  of  extending  this  novel  idea  to  the  design  of  wind  turbine  towers  was  identified.  Wind  turbines  are  very  efficient  in  onshore  and  offshore  environments.  However,  these  environments  are  very  aggressive  causing  corrosion  of  steel  or  steel­prestressed  concrete  towers.  It  is  thus  an  ideal  application  for  combining  concrete  with  a non­corrodible,  lightweight  and  high strength tendon material such as CFRP. Although the initial cost of the  structure  is  expected  to  be  higher,  lower  maintenance,  transportation  and  installation  costs  are  expected  to  make  CFRP­prestressed  poles  cost 

936 FRPRCS-6: Prestressed FRP Reinforcement and Tendons

effective.  The  most  severe  loading  condition  for  wind  turbine  towers  is  fatigue  in a corrosive  environment. Hence, the  initial  focus  of the research  was to study the fatigue of CFRP tendons embedded in concrete.  FATIGUE REVIEW The goal was to carry out a series of experiments to investigate the  fatigue  behaviour  of high  strength  concrete  prestressed  with  CFRP.  In  particular,  the  required  stress  range  and  mean  stress  to  be  used  in  the  experiments  needed  to  be  determined.  As  a  basis  for  the  design  of  the  experiments,  various  codes  of  practice  and  relevant  publications  were  studied.  The  findings  are shown in Tables  1 and 2.  Table  1 summarises various  fatigue  tests  on  bare  CFRP  tendons  (not  in  concrete)  undertaken  by  others.  The  tests  were  conducted  under  varying  stress  ranges  and  frequencies  but  on  average for about 3000000 cycles.  All the tests shown in the table survived  the specified  number of cycles.  However, in Saadatmanesh et al's2 tests, the  tendons failed  for a stress range of 200 MPa when the minimum stress was  90%  of  the  ultimate  tendon  strength  and  also  for  a  stress  range  of  400  MPa  when  the  minimum  stress  was  60%  of  the  ultimate  tendon  strength.  In  the  case  of  Uomoto's3  experiments,  tendons  subjected  to  maximum stresses greater than 87.5% failed during testing.  Table  1.  Summary  of CFRP  fatigue  tests. 

Ref Uomoto* Saadat­ manesh 2 UomotoJ  Adim?

Freq. (Hz)

Stress (MPa)

Range (MPa)

Min. no. of cycles

-

mean  85%  m/«30­90%  min  30­60%  min  30%  max 87.5%  min/max=0.1 

100  100  200  400  100­1000  98.8­988 

2000000  3000000  3000000  3000000  4000000  4000000 

3­5  3­5  3­5  1­10  4 

Ultimate Strength (MPa) 1295  2000  2000  2000  1390  2600 

Table 2 shows the specifications  of various design codes for the  fatigue  performance  of bare steel tendons and also for  steel embedded  in concrete.  The table shows that most of the codes specify  a mean stress range of about  180 MPa for 2000000 cycles. As the CFRP tendons are expected to be used  instead  of  steel, these  results  would  represent  a minimum  baseline  fatigue  performance. 

High Strength Concrete Beams with CFRP Tendons 937 Table 2.  Summary of code recommendations for the fatigue resistance of steel.  Ref/Steel status CEB-fip "/bare CEB-flp '/bare Mallet*/bare Mallet1'Ibare Mallei/bare SiA^/in cone. SiA^lin cone.

Stress (MPa) max 70%  ­

Range (MPa) 170­180  80­280  80­280  180­26 • 0  50­220  100­150  10­12% of Strength

Cycles (N) 2000000  2000000  1000000  2000000  10000000  4000000 

* In the absence of fretting  effects. 

An alternative approach to designing fatigue experiments is to specify a  load range above and below the beam cracking load.  Hence, an example of  a possible practical stress range that could be used in the experiments is that  which corresponds to a minimum load (Pmi„) below the cracking load and a  maximum  load (Pmsx) above the cracking load as indicated  in Figure  110.  It  is  expected  that  fatigue  failures  would  almost  always  occur  at  crack  locations as a result of stress concentrations. These loads (Pmax and Pmi„) can  be used to calculate the tendon stress range Afp .  Using the S­N  curve  for  the tendon, Figure  lb, the corresponding number of cycles until failure  (N)  can  then  be  deduced  from  the  S­N  plot.  This  approach  was  used  in  the  design  of  the  experiments.  A  typical  S­N  curve  (as  shown  in  Figure  1)  predicts the  fatigue  life  of a material  at given  stress ranges or stress ratios  for  a  given  frequency  of  tensile  oscillation  and  a  given  temperature.  The  fatigue life is usually predicted within a 90/95% confidence limit.  A 00

d2(j>

dy

ox 

d2d> '

dxdy

Hence, the stresses in both the x and y directions can be determined by  selecting  an  appropriate  stress  function  which  satisfies  all  boundary  conditions.  There  are  numerous  methods  for  obtaining  a suitable  form  for  the  stress  function,  but  here  dynamic  relaxation  (DR)  incorporated  into  a  finite  difference  grid  is  used10.  Using DR, the  form  of the  stress  function  (x, y)  can be found for the various loading conditions, essentially by using  numerical  partial  differentiation,  so  that  the  stresses  in  both  the  x  and  y  directions can be calculated and hence the cracking load predicted. Figure 6  shows  a  typical  stress  plot  for  a  specimen  loaded  under  two  distant  concentrated anchorage locations. 

Figure 6.  Elastic stress distribution in secondary anchorage zones 

Anchorage Zones for FRP-Prestressed Concrete 961 Figure 7 shows correlation between predicted and actual cracking loads  for all secondary zone specimens. Agreement is reasonably good.  Correlation between predicted and actual cracking loads ouu

z"

.* -~

•

/ •

500-

n o -a) 400-

*

C

o 300ra

y»/

•

•

\s* * * S • l

=5 1000. n. 0

U

100

i

i

i

i

i

200

300

400

500

600

700

Actual cracking load (kN)

Figure 7.  Cracking load predictions for all secondary zone specimens  Plastic Model In order to predict the ultimate failure load of the specimens, it was decided  to use plasticity theory once again, as had been used for the primary zones,  as  it  was  noticed  during  testing  that the  final  collapse  state  in  every  case  always  involved  wedging  beneath  the  load  plates. The  level  at  which this  wedging  occurred  was  found  to  be  related  to  the  quantity  of  secondary  reinforcement.  Further,  in  every  case,  the  secondary  reinforcement  was  inspected  following  failure  of the overall specimen, and it seemed that rather than the  FRP being mainly stretched (as one might expect from  steel bars crossing a  shear  discontinuity),  there  was  substantial  transverse  shear  damage  where  the wedge action had attempted to guillotine the FRP bars.  Therefore,  the  model  shown  in  Figure  8  has  been  adopted  for  the  ultimate  capacity  analysis  of  the  secondary  zones.  In  tandem  with  this  model,  the  same  plastic  assumptions  as  were  made  for  the  primary  zone  situation are made here, with the only exception being that stretching of the  FRP  bars  does not  occur.  Instead, the  guillotine  shear  strength  of the  bars 

962 FRPRCS-6: Prestressed FRP Reinforcement and Tendons was  measured  and  used  in the  analysis. Figure  9 shows  correlation  between  the  actual  ultimate  collapse  capacities  and  the  predicted  values  using  this  plasticity­based  approach.  Correlation  is  reasonably  good  for  specimens  in  which the  anchorage  loads are  spaced  far  apart, but  less good  for  specimens  in which the anchorage  loads are closer  together. 

Figure 8.  Isometric and elevation of the assumed failure planes 

Correlation between predicted and actual test loads _

16001

z £. 1400   X v\

1000

800

-200

^ _  AvgPAF 

1800

*k 

-100 -50 Load(kN)

Figure 6.  Average strain in PAF and steel 

Development of PAF strain with different PAF ratios 10 

d/dy  Figure 7. Maximum strain development in PAF with different  amount of PAF 

Figure 7 shows the effect  of amount of PAF on the maximum strain of PAF  (of both  inner and outer PAF). The strain development  at the first yielding  displacement is greater in SI­2 where there is smaller amount of PAF. Also,  after  the flexural  yielding, the strain development is greater in the pier with  smaller  amount  of  PAF  and  it  is  approximately  linear.  The  strain 

Simple Continuous System of Shear Reinforcement

973

development in steel is relatively smaller at the initial phase until it yields in  both the piers.  Effect of repeated cyclic displacement on PAF strain The increment  in maximum strain of PAF with increment  in repeated  load  reversals  at  constant  displacement  in  addition  to  the  non  repeated  displacement reversals is shown in Figure 8.  The effect  is more pronounced  when  the  amount  of  PAF  is  smaller.  The  increment  in  strain  reduces  decrease  as the  number  of  repeated  cycles  at the  same displacement  level  increases. There is no specific  trend  in increment in strain with increase in  displacement  level. The mean value of available non­negative and non­zero  data from Figure 8 shows 0.78 and 0.56 times yielding strain of steel stirrup  for  the  first  repeated  cycles  of pier  SI­2  and  SI­3  respectively.  The mean  values are 0.29 and 0.17 times yielding strain of steel stirrup for the second  repeated  cycles  of the piers  SI­2  and  SI­3  respectively.  Conservatively,  it  can be concluded that the increment in strain in PAF for the second repeated  cycle, excluding non repeated cycle, is half that of the first repeated cycle.  1.5 

1.5  Sl-2

1.2 

DCycle3

• Cycle2

Sl-3

1.2 

DCycle3

• Cycle2

1

^0.9 

«0.9 

i

i

W

T ?

i

T

Scheme (a) Scheme (b) Scheme (c) Figure 2. Schematic of the computation of stresses in the groove­filling material in  the linearly­elastic range.  where dg  is the  size  of  the  (square)  groove,  and ux and uy are  displacements  along x  and y,  respectively.  Once the  system  is  solved  for  the  unknowns X, and X2,  it  is  possible  to  compute  the  stresses  in  the  epoxy.  Based  on  phenomenology,  stresses  whose  computation  is  more  relevant  are  the  "hoop"  stresses  on  the x  and y  axes.  Particularly  important  is  to  determine  the  "hoop" tensile  stress  on  the  positive y  axis,  i.e.,  in  the  cover  of  the  bar  where  initiation  of  longitudinal  cracking  in the epoxy can thus be  predicted.  The  maximum  in­plane  principal  stresses  on  the  two  locations  are  as  follows:  ai(0,y) = -kW

) = 

9y + 1 + + kjiv-) 4^— 3 + ~ ~l0do 4

l-{*l("')-J*2("') + {. -ki(.v') + -k2(v')-

(3a)  /

2xl

+ 9kj(v')-

k2-\

(3b)  2x  1  , 1 2 * .  valid  for  ­1