Long-term performance and durability of masonry structures: degradation mechanisms, health monitoring and service life design 9780081021101, 0081021100

Citation preview

Long-term Performance and Durability of Masonry Structures

This page intentionally left blank

Woodhead Publishing Series in Civil and Structural Engineering

Long-term Performance and Durability of Masonry Structures Degradation Mechanisms, Health Monitoring and Service Life Design

Edited by

Bahman Ghiassi Paulo B. Lourenço

Woodhead Publishing is an imprint of Elsevier The Officers’ Mess Business Centre, Royston Road, Duxford, CB22 4QH, United Kingdom 50 Hampshire Street, 5th Floor, Cambridge, MA 02139, United States The Boulevard, Langford Lane, Kidlington, OX5 1GB, United Kingdom Copyright © 2019 Elsevier Ltd. All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Details on how to seek permission, further information about the Publisher’s permissions policies and our arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions. This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted herein). Notices Knowledge and best practice in this field are constantly changing. As new research and experience broaden our understanding, changes in research methods, professional practices, or medical treatment may become necessary. Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds, or experiments described herein. In using such information or methods they should be mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility. To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. Library of Congress Cataloging-in-Publication Data A catalog record for this book is available from the Library of Congress British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library ISBN: 978-0-08-102110-1 For information on all Woodhead publications visit our website at https://www.elsevier.com/books-and-journals

Publisher: Gwen Jones Acquisition Editor: Gwen Jones Editorial Project Manager: Ali Afzal-Khan Production Project Manager: Debasish Ghosh Designer: Christian J. Bilbow Typeset by TNQ Technologies

Contents

List of contributors Preface

Part One 1

2

3

4

Durability and degradation mechanisms of bare and strengthened masonry

Clay bricks Francisco M. Fernandes 1.1 Introduction 1.2 Origin and properties of clay bricks 1.3 Decay and degradation factors and mechanisms 1.4 Durability of clay bricks 1.5 Treatments to improve/extend brick’s durability References Concrete block Guilherme Parsekian, Humberto Ramos Roman, Cl audio Oliveira Silva and Marcio Santos Faria 2.1 Introduction 2.2 Block specifications 2.3 Common pathologies and causes 2.4 Wall performance and durability 2.5 Final remarks References

ix xi

1 3 3 4 10 15 16 17 21 21 27 30 35 54 55

Stone Ainara Zornoza-Indart and Paula Lopez-Arce 3.1 Introduction 3.2 Properties of stone materials and degradation mechanisms 3.3 Durability 3.4 Characterization analysis and standard tests 3.5 Conservation treatments References

59

Earth masonry Ioannis Ioannou and Rogiros Illampas 4.1 Introduction 4.2 Factors affecting the durability of earth masonry

89

59 60 73 74 80 81

89 90

vi

5

6

7

Contents

4.3 Assessment of earth masonry durability 4.4 Suggestions for future research directions 4.5 Conclusions References

112 117 118 118

Timber Christian Brischke 5.1 Introduction 5.2 Wood anatomy, structure, and chemical composition 5.3 Hazards: abiotic and biotic agents affecting the serviceability of timber 5.4 Wood durability 5.5 Wood preservation 5.6 Wood modification 5.7 Wood protection by design 5.8 Quantifying decay-influencing factors: contribution to performance prediction of timber structures References

129

Mortars Maria do Ros ario Veiga and Ant onio Santos Silva 6.1 Introduction 6.2 Theoretical background 6.3 Transport properties 6.4 Testing and modeling 6.5 Protective measures 6.6 State of codes and standards implementation 6.7 Conclusions References

169

FRP-strengthened masonry Maria Antonietta Aiello, Bahman Ghiassi and Paulo B. Lourenço 7.1 Introduction 7.2 Degradation mechanisms 7.3 Experimental background 7.4 Modeling of durability 7.5 State of codes and standards implementations References

Part Two 8

Health monitoring and testing

Dynamic identification of historic masonry structures Maria-Giovanna Masciotta and Luís F. Ramos 8.1 Introduction 8.2 Theoretical background

130 131 133 145 149 152 154 158 160

169 172 176 183 196 200 201 202 209 209 211 214 224 228 234

239 241 241 243

Contents

vii

8.3 Applications 8.4 Conclusions References 9

Laser scanning and its applications to damage detection and monitoring in masonry structures Ana S anchez Rodríguez, Belén Riveiro Rodríguez, Mario Soil an Rodríguez and Pedro Arias S anchez 9.1 Introduction 9.2 State of the art in the application of terrestrial laser scanning to masonry structures 9.3 Laser scanning technology 9.4 Successful applications 9.5 Conclusions References

10 Acoustic emission testing Els Verstrynge 10.1 Introduction into AE testing 10.2 AE for crack monitoring in masonry 10.3 AE for compressive creep monitoring in masonry 10.4 AE testing of FRP- and SRG-strengthened structures 10.5 Conclusions References

Part Three

Long-term performance and service life design

248 262 263 265 266 266 269 274 281 282 287 287 290 295 299 305 305

309

11 Service life design of timber structures Maxime Verbist, Lina Nunes, Dennis Jones and Jorge M. Branco 11.1 Introduction 11.2 Wood deterioration 11.3 Natural durability 11.4 Use class 11.5 Performance assessment 11.6 Durability models 11.7 Protection systems 11.8 Maintenance and monitoring 11.9 Conclusions References

311

12 Service life design of stone masonry structures Elsa Garavaglia, Giuliana Cardani and Anna Anzani 12.1 Introduction 12.2 The lifetime prediction of decayed stone masonry

337

312 312 317 319 321 326 329 331 333 333

337 338

viii

Contents

12.3 12.4 12.5 12.6

Fragility curve for service life prediction Laboratory aging tests on treated stone masonry On-site study of surface decay of treated stone masonry Conclusions Acknowledgments References Further reading

342 346 353 362 363 363 365

13 Probabilistic modeling of aging masonry Elsa Garavaglia, Giuliana Cardani and Anna Anzani 13.1 Introduction 13.2 Laboratory creep and pseudo-creep tests 13.3 A probabilistic approach to model the strain rate evolution 13.4 The probabilistic approach applied to the Bell Tower of Monza 13.5 Conclusions Acknowledgments References Further reading

367

Index

391

367 369 374 385 387 388 388 389

List of contributors

Maria Antonietta Aiello Salento, Lecce, Italy Anna Anzani

Department of Engineering for Innovation, University of

Department of Design, Politecnico di Milano, Milano, Italy

Pedro Arias S
anchez Department of Natural Resources and Environmental Engineering, University of Vigo, Vigo, Spain Jorge M. Branco Portugal

ISISE, University of Minho, Campus de Azurém, Guimar~aes,

Christian Brischke University of Goettingen, Faculty of Forest Sciences and Forest Ecology, Department of Wood Biology and Wood Products, Goettingen, Germany Giuliana Cardani Department of Civil and Environmental Engineering, Politecnico di Milano, Milano, Italy Maria do Ros
ario Veiga Portugal Marcio Santos Faria

National Laboratory for Civil Engineering, Lisbon,

Arq.EST Consultoria & Projetos Ltda, Juiz de Fora, Brazil

Francisco M. Fernandes Faculty of Engineering and Technologies, Universidade Lusíada - Norte, Vila Nova de Famalic~ao, Portugal Elsa Garavaglia Department of Civil and Environmental Engineering, Politecnico di Milano, Milano, Italy Bahman Ghiassi Centre for Structural Engineering and Informatics, Faculty of Engineering, University of Nottingham, Nottingham, United Kingdom Rogiros Illampas Department of Civil and Environment Engineering, University of Cyprus, Nicosia, Cyprus Ioannis Ioannou Department of Civil and Environment Engineering, University of Cyprus, Nicosia, Cyprus Dennis Jones DJ Timber Consultancy Ltd., Neath, United Kingdom; Wood Science and Engineering, Luleå University of Technology, Skellefteå, Sweden

x

List of contributors

Paula Lopez-Arce University College London (UCL), Institute for Environmental Design and Engineering (IEDE), The Bartlett, School of Environment, Energy and Resources, Faculty of the Built Environment, London, United Kingdom; Property Care Association (PCA), Huntingdon, United Kingdom Paulo B. Lourenço Portugal

ISISE, Department of Civil Engineering, University of Minho,

Maria-Giovanna Masciotta ISISE, University of Minho, Department of Civil Engineering, Guimar~aes, Portugal Lina Nunes LNEC, Structures Department, Lisboa, Portugal; cE3c, Centre for Ecology, Evolution and Environmental Changes, Azorean Biodiversity Group and University of the Azores, Angra do Heroísmo, Portugal Guilherme Parsekian

Universidade Federal de S~ao Carlos, Sao Paulo, Brazil

Luís F. Ramos ISISE, University of Minho, Department of Civil Engineering, Guimar~aes, Portugal Belén Riveiro Rodríguez Department of Materials Engineering, Applied Mechanics and Construction, University of Vigo, Vigo, Spain Humberto Ramos Roman Brazil

Universidade Federal de Santa Catarina, Santa Catarina,

Ana S
anchez Rodríguez Department of Materials Engineering, Applied Mechanics and Construction, University of Vigo, Vigo, Spain Ant
onio Santos Silva Cl
audio Oliveira Silva Brazil

National Laboratory for Civil Engineering, Lisbon, Portugal Associaç~ao Brasileira do Cimento Portland, Sao Paulo,

Mario Soil
an Rodríguez Department of Materials Engineering, Applied Mechanics and Construction, University of Vigo, Vigo, Spain Maxime Verbist Portugal

ISISE, University of Minho, Campus de Azurém, Guimar~aes,

Els Verstrynge Building Materials and Building Technology Division, Civil Engineering Department, KU Leuven, Leuven, Belgium Ainara Zornoza-Indart Department of Painting, Faculty of Fine Arts, University of the Basque Country (UPV/EHU), Lejona, Spain

Preface

Masonry is one of oldest and most used construction materials around the world. Masonry materials have been used since several thousand years ago and are still being used. Durability and sustainability of masonry are clearly noticeable from the large number of historical masonry structures that exist around the world. Though, depending on the construction method and materials used, environmental or mechanical deterioration are inevitable. Most of the deterioration in structures usually occurs in the exterior walls that are directly exposed to environmental conditions or in the base walls that are subjected to capillary rise. Construction practice has also significantly evolved in the last centuries. These changes, which have been mainly toward new production methods and more economical designs, have led to significant reduction of component sizes compared to that of historical structures, making them more vulnerable to deterioration. Masonry bridges are also under continuous deterioration and fatigue, due to combined mechanical and environmental loadings. Understanding these degradation mechanisms and taking protective measures against them is critical for sustainable design and use of masonry structures. The poor performance of masonry structures against tensile and cyclic loads (e.g., induced by earthquakes) has led to the development of different strengthening techniques. Durability of these systems in connection with masonry also needs to be well understood but has received insufficient attention. In this field, durability of FRP-strengthened masonry is among the few topics that have been addressed in the literature. There is therefore a significant need in future fundamental studies in this field. This book, as a step forward in collecting the current knowledge on durability of masonry and historical structures, covers three main topics arranged in different sections. The first section, which is also the most comprehensive one, presents an overview of durability of the main materials used in masonry structures (bricks, mortar, and timber). Durability of FRP-strengthened masonry is also presented. The main focus of these chapters is to provide an overview of the current knowledge on durability, degradation mechanisms, transport properties, testing and modeling methods, and protective measures for each material. The second section, which consists of three chapters, provides an overview of application of different techniques for detection and identification of damage and deterioration at the structural level. These include dynamic identification, laser scanning, and acoustic emission testing methods.

xii

Preface

The last section, also consisting of three chapters, is focused on service-life design and long-term performance modeling of existing masonry and historical structures. In particular, stone and timber masonry are addressed. We are convinced that the presented discussions can provide a first step in understanding durability issues for both theoreticians and experimentalists, and in bringing forward the knowledge gaps in the field. Finally, we would like to thank all the contributors for their valuable effort in finalizing this volume and all the reviewers for their comments and kind assistance to improve the book quality. Bahman Ghiassi Paulo B. Lourenço

Part One Durability and degradation mechanisms of bare and strengthened masonry

This page intentionally left blank

1

Clay bricks

Francisco M. Fernandes Faculty of Engineering and Technologies, Universidade Lusíada - Norte, Vila Nova de Famalic~ao, Portugal

Chapter Outline 1.1 Introduction 3 1.2 Origin and properties of clay bricks

4

1.2.1 Raw material characteristics 4 1.2.2 Production process 6 1.2.3 Properties of final clay bricks 8

1.3 Decay and degradation factors and mechanisms 1.3.1 1.3.2 1.3.3 1.3.4 1.3.5

1.4 Durability of clay bricks 15 1.5 Treatments to improve/extend brick’s durability References 17

1.1

10

Production/manufacture 11 Moisture and soluble salts 12 Temperature 14 Biological deterioration 14 Atmospheric contamination 15

16

Introduction

Since man began to establish himself and replace nomad hunting by local farming and agriculture, he sought to protect himself from the aggressions of the environment (snow, rain, cold, heat, etc.) and wild animals. The first protections were natural caves, and the first huts and houses were built with what nature provided, such as tree trunks, animal fur, straw, clay, etc. With the appearance of the first civilizations, around 9000e7000 BC, the construction techniques evolved, and stone, adobe, wood, and clay brick began to be used. The first vestiges of brick masonry buildings were found in the region of Israel (Mesopotamia) and dated from 9000 to 8000 BC. Masonryfortified walls in Jericho (7000 BC) and rectangular brick houses in C¸atal-H€uy€uk, Anatolia, (6500 BC) were also found. Clay brick masonry is, effectively, one of the finest and most durable construction techniques ever invented by man. Masonry consists of building stable bonded stacks of small pieces by hand (Vekey, 1998). Used since the time of the first villages and cities

Long-term Performance and Durability of Masonry Structures. https://doi.org/10.1016/B978-0-08-102110-1.00001-7 Copyright © 2019 Elsevier Ltd. All rights reserved.

4

Long-term Performance and Durability of Masonry Structures

built by man, masonry application has been growing and evolving to new uses all over the entire civilized world. It was a fundamental building material in Mesopotamian, Egyptian, and Roman periods. During Roman period, the use of clay brick increased and become specialized to maximize its benefits. Moreover, clay brick masonry was used as well during medieval and modern times. Despite several modifications of the clay brick use, shape, and manufacture along thousands of years of constant evolution, the simplicity that made its success remained. Numerous buildings built with clay bricks prevailed until the 21st century, which testifies to the strength and durability of this material through centuries of rainstorms, snow, thaw-freezing cycles, high temperatures (sunlight, fire), and human-induced deterioration. Durability, according to Cooper (1994), is the ability of a product to perform its required function over a lengthy period under normal conditions of use without excessive expenditure on maintenance or repair. Therefore, the main pillars that affect the durability of clay bricks are (1) the quality of the raw material; (2) the manufacture process and firing conditions and, consequently, (3) the properties of hardened bricks; (4) the usage conditions to which bricks were exposed since construction; and (v) the rate of maintenance and repair as well as the types of treatments carried out.

1.2 1.2.1

Origin and properties of clay bricks Raw material characteristics

In the natural state, raw clay is a complex and heterogeneous material, formed by a variable proportion of different minerals, known as clay minerals. A general definition  of raw clay is given in Alvarez de Buergo and Lim on (1994), where it is referred to as a material constituted by fine grain, with a size lower than 0.1 mm. These minerals are found to be in the range of 2e4 m soil depth, with earthen texture, and acquiring plasticity when mixed with water. Additionally, clay is a grain-sized term enclosing all the sediments where the dominant particles have an equivalent spherical diameter lower than 2 mm (Gomes, 1988). In geologic terms, clay is classified as a clastic sedimentary rock, originated by the mechanic accumulation of individual fragments of rocks and mainly constituted by clay minerals and quartz. The word sediment has its origin in Latin sedere (to accumulate), and it is a reference to the process of formation of clays, while the term clastic (from the Greek klastos, which means broken) describes the fragments of rocks and sediments that constitute raw clay. The formation process of sedimentary rocks is composed of four stages: alteration, transport, sedimentation, and lithification. These rocks derive from the consolidation of rock fragments and mineral grains (sediments), which were originated by erosion or weathering (physical alteration) and chemical decomposition of masses of preexisting rocky material (generally, acid aluminosilicate rocks such as feldspars, granites, and gneisses) by atmospheric agents (water, wind, and ice) and live animals. These rock fragments and sediments are then dragged by rainwater or transported by the wind. The third phase consists in the accumulation of the eroded material in regions of low elevation, such as sedimentary basins or in the

Clay bricks

5

ocean. The last stage of the sedimentary process is described as the process by which those rock fragments and sediments transform again into rocks. This process can be done by compaction (action of pressure), cementation (action of binding agents, particularly calcite, quartz, and iron oxide), and crystallization. Clays took millions of years to form. A long period of time was necessary to allow for the accumulation of sediments and for the lithification process to finish. Thus, every geologic period since Carboniferous (about 345 million years ago) had produced adequate clays for the manufacture of ceramic elements. Presently, the principal origin locations are the interface between the terrestrial surface and the atmosphere, where clays that come from the alteration of rocks are found, and fluvial systems, estuaries, lakes, oceans, and fluvial deltas. Clay minerals are complex crystalline elements that have particularities that provide clays with their characteristic plastic behavior. Clay minerals are included in the group of the phyllosilicates due to their fine, or foliate, shape. The plastic behavior showed by these minerals (Andrade et al., 2011) is due to three characteristics: the reduced size of their particles, the foliate shape of the particles, and the crystalline and chemical properties of the surface of the particles that allow them to interact with polar liquids, such as water. The crystals of these clay minerals typically have a thin, layered structure that enables them to absorb water and consequently to expand on wetting. For example, montmorillonite is particularly noted for this later characteristic and has a higher degree of plasticity than kaolinite. The most common clay minerals are kaolinite, montmorillonite, illite, talc, and pyrophyllite. Their complex chemical composition has as main components the silica dioxide or silica (SiO2) and the aluminium dioxide or alumina (Al2O3). Thus, clay minerals are mainly silicates, which are the largest, the most interesting, and the most complicated class of minerals by far. Approximately 30% of minerals are silicates, and some geologists estimate that 90% of the Earth’s crust is made up of silicates. With oxygen and silica, the two most abundant elements in the Earth’s crust, silicate, and thus clay, abundance is not surprising. Clay deposits are generally sedimentary materials of mixed composition, meaning that deposits with clay constituted by a major single clay mineral are quite rare. When found, such special clays are used to produce specific ceramic ware, such as porcelain, earthenware, refractory products, clearance pipes, etc. Generically, the fundamental element of clay is kaolinite, which is a complex silica and hydrated alumina compound (Al2Si2O5[OH]4). Besides kaolinite, other minerals and elements can usually be found such as iron oxides, sand, shales, marls, water, complex silicates, and even chalk. In most cases, these clays also contain organic matter and other materials associated with the deposit. The presence of other components such as sandy material, lime, and organic matter, which are not clay minerals, induces high heterogeneity, or unexpected results, that in most cases can be prejudicial to the final product. Historic clay brick specimens typically include diverse components besides clay minerals, such as ferric minerals (iron oxides and hydroxides), limestone, sandy material, plaster, lime, and organic matter. In fact, clays also possess in their composition a nonnegligible part that is not considered argillaceous material. Nowadays, the clays are carefully selected and cleaned before

6

Long-term Performance and Durability of Masonry Structures

their use. However, during the history of clay brick, those harmful elements were not always removed from the raw material. Iron oxide (Fe3O4) is often found in clay deposits. This compound gives the color to the fired bricks, and generally, the higher the proportion of this compound, the darker are the bricks. Additionally, this compound has the capability to lower the fusion point, which represents a beneficial aspect since old ovens or kilns could not reach the high temperatures encountered in modern ovens. Limestone and other calcareous rocks have the inconvenience of decomposing during the burning stage and transforming into calcium oxide, which suffers volume expansion when in contact with water. These elements can consequently cause the early cracking of the brick. Sands and sandy material are mainly constituted by silica and can be harmful according to their dimensions. If on one hand, fine sands reduced the shrinkage during drying and firing stages, on the other hand, sands with large grains are not convenient as the bricks tend to crack due to shrinkage in the drying process.

1.2.2

Production process

The manufacture of fired clay bricks can be divided into four stages (Límon and  Alvarez de Buergo, 1997) according to basic principles followed for thousands of years: (1) selection and preparation of the clays; (2) mixing and molding; (3) drying of the fresh material; and (4) firing of the clay units. Firstly, the extraction and preparation of the raw clay take place. As soon as the raw material is extracted, it is accumulated and moved to an open-air deposit, where it is left to putrefy for several days or weeks. During this period, the raw material should be rummaged to reduce soluble salts to a minimum, leading to a more homogeneous material. The analysis of the constituents of historic bricks showed that they were not always produced using treated clays. In some cases, bad quality clays were used. Marcus Vitruvius Pollio (Vitruvius and Morgan, 1960) stated that the choice of the raw material was very important to improve the performance and durability of the bricks. However, the selection of the raw material depended essentially on its availability in the construction location or  nearby (Alvarez de Buergo and Lim on, 1994). The raw material is further crushed and mixed with water, in an operation designated by tempering (Weaver, 1997). In early times, the mixing was carried out by hand, in a crude and often ineffective manner, but later, horse-driven heavy rollers, or wheels, in a ring-pit were used. The amount of water used depends on the type of element being produced and, usually, the finest the final piece, the greater the amount of water needed. The resultant mixing must be characterized by sufficient plasticity to facilitate the molding. For being readily molded or formed by hand, raw clays might also be “too plastic.” This characteristic leads to severe shrinkage during the drying phase, resulting in warping, twisting, or cracking. In this case, the plasticity of the clay must be reduced by adding sand, for example. Early brick makers often used a mix of about 30% of sand and 70% of plastic clay (Weaver, 1997; Vekey, 1998). Generally, the molds were bottomless wood molds placed down over the ground or over tables, which, usually, were covered with a thin film of sand to avoid the brick remaining

Clay bricks

7

attached to the bottom base during the drying process. The excess was removed with the aid of a rope, wooden ruler, or with bare hands. The clay elements still crude are removed from the mold and are dried in a protected space, which is generally a shelter made of scraps of wood and with straw thatch roofs, known as hovels, where it acquires its final shape. Although inexpensive, this primitive method required a lot of open free space and was severely conditioned by climacteric conditions. Generally, the drying of clay bricks lasts for a week or more, depending on the specific climacteric conditions. In hot temperature regions, drying is faster, but bricks should be protected from direct sunlight since they can undergo warping and  cracking (Alvarez de Buergo and Lim on, 1994). In colder regions, drying takes more time due to the low temperatures and moisture. The importance of the drying phase was already mentioned by Vitruvius, who wrote that “bricks should be made in Spring or Autumn, so that they may dry uniformly.” Additionally, too fast drying hardens the surface faster than the core, which remains crude for a longer time. Again, Vitruvius states that bricks “made in Summer are defective, because the fierce heat of the sun bakes their surface and makes the brick seem dry while inside it is not dry.” During this phase, the size reduction, or shrinkage, of the clay bricks occurs due to the elimination of the mixing water. If a too fast drying occurs or too much water is used, cracks will appear in the brick. Finally, the last stage is the hardening of the bricks. Bricks could be further dried at the sun, in the open air, designated by sun-dried bricks, or put in a kiln with temperatures on the order of 1000
C, where they were fired. Firing allowed the bricks to acquire much more resistance from both mechanical and chemical points of view. See Dalkılıç and Nabiko glu (2017) for more details. Early kilns used wood or straw as the combustible and took several days to finish combustion. Coal was not commonly used until the last quarter of the 19th century and was sometimes responsible for introducing sulfates into the bricks. During this phase, complex chemical reactions take place, creating diverse clay products, according to the firing temperature and the quality of the clay. The quality of the final clay bricks strongly affects strength and durability of the units, as well as buildings, and according to Vitruvius, sun-dried clay bricks need a minimum of 2 years to dry. Additionally, he gave the example of Utica where clay bricks used for building walls had to be 5 years old. The chemical reactions that take place during the firing of clay bricks are essentially related to what happens to the kaolinite. The most relevant chemical reactions were unknown for thousands of years, but now, they help to understand different clay brick properties. Generally, different chemical reactions are triggered with the temperature  increase (Alvarez de Buergo and Lim on, 1994). The first process, occurring at
100 C, is the elimination of hygroscopic water. At 200
C, the kaolinite loses the water attached to its surface, which is accompanied by shrinkage. Shrinkage should occur very slowly, so any early cracking is retarded, while the clay is still in a crude state. Between 350 and 650
C, the oxidation of organic matter takes place, and the constitutive water of the clay is freed. Between 600 and 700
C, the kaolinite loses two molecules of water, and a dehydrated compound is formed, called metakaolin (2SiO2Al2O3). At 650
C, the dissociation between silica and alumina starts to take

8

Long-term Performance and Durability of Masonry Structures

place, which enables other compounds to be formed with the increase of the temperature. Thus, between 850 and 950
C, the collapse of the carbonate structure occurs with the release of CaO and the formation of calcite and dolomite. At these temperatures, another considerable shrinkage takes place because of the formation of new compounds of silica and alumina. This additional shrinkage can reach 15% of the initial volume, and therefore, to reduce it, sand was added to the clay mass. Between 800 and 1000
C, the first signs of sintering and vitrification can be observed, and at 900
C, the gehlenite disappears totally. At temperatures higher than 1000
C (1200e1500
C), the fusion of the kaolinite with vitrification takes place, and the silicoaluminates are fused into glass (mullite Al4Si2O10 and cristobalite).

1.2.3

Properties of final clay bricks

The durability and resistance of clay bricks are intimately related to its diverse inner properties, but the following stand for the most important and fundamental ones: porosity and compressive strength. All the others generally derive from these two main ones. Additionally, chemical composition and color can inform about firing conditions, which relate also to porosity and compressive strength. The firing of clay bricks produces a series of mineralogic, textural, and physical changes that depend on many factors and influence the porosity (Lopez-Arce et al., 2003; Cultrone et al., 2004), resulting in a much porous material than natural stone (Mecha, 1998). Porosity is an important parameter concerning clay bricks due to its influence on properties such as chemical reactivity and mechanical strength, effectively affecting its quality and durability, which increases with decreasing values of porosity. Commonly, clay bricks exhibit high porosity values, ranging between 15% and more than 40% (Esbert et al., 1997; Livingston, 1993; Maierhofer et al., 1998). Clay bricks are porous materials, with pores constituting a large part of the brick volume. The dimension and distribution of the pores have a very strong impact on the durability of the bricks and are influenced by the quality of the raw clay, the amount of water, and the firing conditions and temperature. Mamillan (1979), Lopez-Arce et al. (2003) and Cultrone et al. (2000, 2004) all observed that if the firing temperature increases, the proportion of large pores (3e15 mm) increases, and the connectivity between pores is reduced, whereas the number of thin pores diminishes, promoting bricks with large, unconnected pores. It is a fact that large pores are less influenced by soluble salts and freeze/thaw cycles. Moreover, studies from Cultrone et al. (2004) and Elert et al. (2003) reported the formation of thin pores ( < ¼ 1:267  0:00267T > > : 0

(7.6.a)

0  C  T < 100 C 100 C  T < 475 C

(7.6.b)

T  475 C

Model 2 (Wang et al., 2011)

8 > > > > > > > > > >
fu 490 > > > > > > > > ðT  420Þ1:8 > : 0 :48  76000

22 C  T < 150 C 150 C  T < 420 C

(7.7)

420 C  T < 706 C

The difference between Eqs. (7.6.a) and (7.6.b) is that the first one refers to GFRP and the second one to CFRP materials.

7.5

State of codes and standards implementations

In the field of FRP-strengthening of civil buildings and infrastructures a wide number of national and international technical codes and design guidelines are available for concrete structures (ACI, 2017; FIB, 2001). Due to the impossibility to have a worldwide standardization of the masonry and historical construction, the efforts in producing guidelines for FRP-strengthening in this context produced only a small number of documents. The two guidelines that can be considered to have a framework for an international approach are CNR-DT 200 R1/2013 (CNR, 2013) and ACI 440.7R-10 (ACI, 2010). The first document is the result of a concerted effort by all the Italian scientists and practitioners involved in the field of FRP-strengthening of existing structures. The CNR-DT 200 R1/2013 document was produced in 2004 as its first edition (CNR, 2004) and afterward revised into the present form that was edited in 2013 (CNR, 2013). This document is focused on the typical constructions that represent the great part of existing buildings: concrete and masonry. Other new documents

FRP-strengthened masonry

229

were edited by CNR (Italian Research Council) for timber and steel structures in forms of preliminary studies (CNR, 2005a, 2005b). The CNR-DT 200 is focused on the use of externally bonded FRP; the first part of the document is addressed to reinforced concrete (RC) applications, while the second part illustrates the provisions for design, applications, and control of FRP strengthening for masonry construction. The document reports not only design equations but also material specifications that are related to the mechanical behavior in the short term and to the potential vulnerability in the long term. Durability safety coefficients are provided, even if FRP materials specifications in terms of durability are not distinguished for concrete or masonry substrate. The ACI 440.7R document is the result of a concerted effort by all the American and international scientists and practitioners involved in the Committee 440 of the American Concrete Institute, named 440 Fiber-Reinforced Polymer Reinforcement. This document provides guidelines for flexural and shear strengthening of masonry walls, confinement of masonry, and repair of cracked masonry with FRP systems. Provisions related to the long-term behavior are illustrated, as it regards the timedependent behavior of FRPs and their possible deterioration due to external environmental agents. Another document edited by the ACI is related to the applications of accelerated protocols that can be used to evaluate at a laboratory scale the potential degradation of FRP exposed to harsh environment. This recent document is titled ACI 440.9R15: Guide to Accelerated Conditioning Protocols for Durability Assessment of Internal and External FRP Reinforcement (ACI, 2015). This document was conceived for the use of FRPs in concrete structures, in forms of both internal reinforcement (bars) or EBR. Due to the extended validity of the principles and procedures that are provided herein, it is considered useful to recall it in the present discussion.

7.5.1

CNR-DT 200

In the CNR-DT 200 document, three types of FRP are counted: CFRP (carbon FRP), GFRP (glass FRP) and AFRP (aramid FRP); all other possible fibrous materials are not considered to be complying with the requirements of these guidelines. In the design principles that are presented in the first part of the document, it is reported that adequate performance of the FRPs strengthening systems should be assured, also with reference to their potential degradation. This level of reliability in the long term should be assured also by considering the adoption of adequate construction details and use of educated and skilled workers. It is also underlined that service conditions that could cause a degradation of the materials or interfaces in the long term should be identified and evaluated during the design process. And whenever the quantification of durability safety coefficients is not easily available from the guidelines, an estimation is required by considering the scientific literature or by leading experimental tests. Possible degradation in the mechanical properties of the FRPs, such as tensile strength, ultimate strain, and Young’s modulus, are considered with reference to specific aging conditions such as alkaline environment, moisture (water and chloride

230

Long-term Performance and Durability of Masonry Structures

solutions), extreme temperatures, thermal cycles, freeze and thaw cycles, and UV radiation. As it regards the chemical attack in alkaline environments, it is specified that this type of degradation is mainly due to the presence of pore water in concrete. In this type of degradation the role of the resin is recognized to be very important. Resins that are highly resistant, such as epoxy resins, are strongly recommended, while polyester resins are not considered a safe solution. The chemical vulnerability of the structural materials, due to the presence of alkaline ions in aqueous solutions, is also addressed to possible fractures in interfaces or at the matrix level. Even if it is not deeply specified in the document, it is well known in literature that E-glass fibers are prone to chemical degradation in the presence of alkaline species. This process is due to a combination of degradation mechanisms that may occur within the microstructure of the composite materials. They include an effect of the OH ions that produce corrosion of the fibers; a second effect due to the precipitation of hydration products, which may reduce the flexibility of the fibers and change the behavior at the interface with the inorganic matrix; and the presence of chemical products that cause a densification of the matrix at the interface level, which may produce a bending effect in the fibers (Bentur and Mindess, 2006; Majumdar and Laws, 1991). Some recent results on this topic may be found (in Micelli and Aiello, 2017). Also the presence of moisture is considered a potential threat for the FRP materials. The permeability of the resins and their possible degradation due to changes in glass transition temperature or strength reduction is pointed out to be the key for a durable solution. The effects of extreme temperatures and thermal cycles are recommended to be considered in relationship with thermal fatigue degradation of the FRP. These cycles may cause microcracks at resin or intralaminar interfaces, which can reduce the strength and stiffness of the composite material in the long term. Moreover, it is specified that the thermal cycles should not exceed the lower limit of glass transition temperature range of the polymeric matrix. This circumstance may cause a softening of the resin that may not be able to transfer the stress to the reinforcing fibers. When peaks of temperature may be close to the glass transition range, it is recommended to protect the surface of the FRP by avoiding the transmittance of the external heat to the composite surfaces. The case of thermal cycles that produce freeze-thaw effects should also be considered potentially detrimental for the resins and the fiber/resin interfaces that could be damaged, especially in the presence of high moisture levels. The effect of UV radiation is not considered to be an important detrimental factor, and the potential effects are considered to locate mainly at the fiber-matrix interface, especially when permeability of the resins is reduced. To quantify the possible reduction in mechanical properties, a set of safety coefficients, named environmental conversion factors, is provided. They are related to the possible degradation of the structural properties of the FRP materials in different environmental conditions. Two aspects mainly contribute to the quantification of these coefficients: the susceptibility of the fibrous material and the aggression potential of the environment.

FRP-strengthened masonry

231

Carbon fibers are considered to be the most resistant in terms of chemical or mechanical degradation, in all environments. Aramid fibers are considered in an intermediate position, while glass fibers (without specific differences among them) are considered the most vulnerable. Three environments are considered representative of external conditions: internal, external, and aggressive. The last case includes industrial environments in which specific emissions may produce chemical species that are very aggressive. Table 7.1 taken from the CNR-DT200 is a quantitative resume of the mentioned provisions. The use of the environmental conversion factors in design is compliant with the philosophy of the Eurocodes approach, since the generic design property Xd of the FRP material is expressed as: Xd ¼ ha

Xk gm

where ha, is the environmental conversion factor accounting for durability, Xk is the characteristic value of the FRP property being considered, and gm is the partial factor of the material that takes into account the type of application. Thus the environmental coefficient (ha < 1) is used as a reduction factor that takes into account the material susceptibility and the potential environmental threat. In bond-controlled applications the environmental conversion factor is used to reduce the strain of the FRP at the considered ultimate limit state (i.e., debonding, peeling etc.). This can be easily justified due to the elastic behavior of FRPs up to tensile failure.

Table 7.1 Environmental conversion factor provided by CNR- DT200 R1-2013 Exposure conditions

Type of fiber/resin

ha

Internal

Glass/epoxy

0.75

Aramid/epoxy

0.85

Carbon/epoxy

0.95

Glass/epoxy

0.65

Aramid/epoxy

0.75

Carbon/epoxy

0.85

Glass/epoxy

0.50

Aramid/epoxy

0.70

Carbon/epoxy

0.85

External

Aggressive environment

232

Long-term Performance and Durability of Masonry Structures

7.5.2

ACI 440.7R

The ACI 440.7 document is dedicated to the use of FRP composites in masonry construction, including those structural systems made by clay, or concrete masonry units, or natural stones, bonded together to create a well-defined stiff and resistant system. In this document the concept of durability is defined as “the ability of a material to resist weathering action, chemical attack, abrasion, and other conditions of service for an extended period of time.” The durability issues are mainly discussed in Chapter 3 (Constituent materials and properties) where important concepts are introduced. The problem of durability is considered from the first step of choosing the proper materials. In fact, in the list of the constituent materials a set of protective coatings is firstly introduced. The type of coating, which may include polymer coatings, acrylic coatings, cement-based systems, and intumescent coatings, depends on the specific agent or external factor that may lead to a reduction of the material properties in the structural system. The different harsh agents that make the use of specific coatings necessary are UV light exposure, vandalism, impact, abrasion, and wear, aesthetics degradation, chemical aggressive agents, and fire. To take into account the possible degradation of the mechanical properties in the long term, it is specified that the mechanical properties of the materials obtained from standard tests performed by the manufacturers are not representative of the long-term properties. In many FRP materials the exposure to environmental factors such as high temperature, humidity, and chemical exposure may strongly decrease the mechanical characteristics. This reduction depends also on the specific properties of the FRP systems. The type of application, the exposure times, and conditions and the properties of the fibers, matrix, and interfaces are strongly related to the degree of chemical/mechanical damage that influences the engineering properties. The design provisions that take into account quantitatively the potential degradation of the FRP material in the long term introduce an environmental reduction factor CE. This approach is in accordance to the design philosophy adopted by other ACI 440 documents. The environmental reduction factor is quantified by considering the type of fiber and the type of application. The first variable takes into account the different chemical susceptibility of the fibers. The second variable considers the type of environment, the duration of the exposure, the presence of prolonged harsh conditions, and the application of coatings and protective surface agents. It is admitted by the guidelines to consider an increased value of CE in aggressive environment whenever additional materials are used. This is allowed when the used coatings are proved to be able in reducing the susceptibility of the material in that specific environment. The values of the environmental reduction factor CE are reported in Table 7.2. They are applied in forms of reduction factors as follows: ffu ¼ CE f fu where ffu is the design ultimate strength of FRP, and f fu is the ultimate tensile strength of the FRP material as reported by the manufacturer.

FRP-strengthened masonry

233

Table 7.2 Environmental conversion factors provided by ACI.7R-10 Exposure conditions

Fiber type

Environmental reduction factor CE

Interior exposure

Carbon

0.95

Glass

0.75

Aramid

0.85

Carbon

0.85

Glass

0.65

Aramid

0.75

Carbon

0.85

Glass

0.50

Aramid

0.70

Exterior exposure

Aggressive environment

A particular aspect of the long-term behavior that is recalled in the document is also the possible occurrence of “creep rupture” which is a tensile failure under sustained loads having values lower than the short-term tensile strength. In particular, it is assessed how the creep rupture limit is influenced by the contemporary presence of harsh conditions that may be represented by high temperature, UV radiation, high alkalinity, wetting-drying cycles, or freezing-thawing cycles. As known from scientific literature, it is confirmed that carbon fibers are materials resistant to creep rupture, while aramid fibers are moderately susceptible, and glass fibers are susceptible. Experimental results recalled in the ACI document show that after an equivalent exposure time of 50 years the ultimate strength of the GFRP, AFRP, and CFRP bars was reduced to 30%, 47%, and 91%, respectively. Specific recommendations for design are provided for FRP materials that can be subjected to sustained or cyclic loads. In particular, it is recommended to maintain the stress levels within a range that cannot exceed 20% of the ultimate strength for GFRP, 30% of the ultimate strength for AFRP, or 55% of the ultimate strength for CFRP.

7.5.3

ACI 440.9R

The ACI 440.9R-15 document is devoted to providing standardized accelerated conditioning protocols that could be used in the future as standardized tests for qualification of FRPs used in concrete structures as internal reinforcement or external strengthening systems. Even if the document is not conceived for masonry structures, many aspects may be considered of general purpose. By editing this document the ACI 440 Committee wanted to fill a gap in the field of durability tests for FRP in construction, since there is a recognized lack of standard experimental procedures in this field. The guidelines provided herein are related to (1) the aging of FRP bars immersed in a concrete structural element under sustained flexural loads; and (2) the aging of bond properties between FRP sheets/fabrics and a concrete substrate.

234

Long-term Performance and Durability of Masonry Structures

To individuate the most effective protocols, it is specified that even the tensile failure of the fibers is considered the ultimate limit state of the FRP bars; often the damage at resin level due to alkalinity and moisture ingress is a primary concern. In this perspective a specific indicator of the vulnerability of the FRP bars in a moist environment is indicated to be the absorption rate and amount. Gravimetric tests are recommended to measure the increase in weight after accelerated exposure. An accelerated test protocol is provided for FRP bars immersed in alkaline solution at high temperatures (60 C) subjected to a sustained bending load. After exposure the tensile test is performed until failure, and the results are compared to those of control virgin specimens to provide a strength retention ratio. As it regards the EBR FRP to concrete a combined protocol is recommended to perform appropriate accelerated conditioning and mechanical tests. A concrete notched short beam is strengthened with FRP sheet/fabric. Moreover a sheet of FRP is also bonded on the end face of the beam. Then the specimen is immersed in a moist environment according to a specified accelerated aging protocol. After exposure the specimens are tested in a three-point bending configuration until complete debonding of the FRP sheet. Then a pull-off test is performed on the EBR FRP sheet placed on the end face of the beam. The same tests are performed on virgin control specimens, having the same materials and preparation protocol. The comparison between the results of the mechanical tests may provide the quantification of the bond retention factor. A further test is recommended to compare the tensile properties of control FRP specimens (sheets or fabric) to those of aged specimens subjected to an accelerated protocol. The protocol needs to be specified in terms of aqueous solution used and temperature level. This document is intended to be also a guide for future research and experimental studies in the field. In addition, it is clearly specified that the results of the accelerated tests cannot be used, at the moment, as design reference values since there is a lack of knowledge that needs to be filled in the near future by the scientific community.

References Abanilla, M.A., Li, Y., Karbhari, V.M., 2006. Durability characterization of wet layup graphite/ epoxy composites used in external strengthening. Composites Part B 37 (2e3), 200e212. https://doi.org/10.1016/j.compositesb.2005.05.016. ACI 440.7R-10, 2010. Guide for Design & Constr of Externally Bonded FRP Systems for Strengthening Unreinforced Masonry Structures, ACI Committee 440. American Concrete Institute. ACI 440.9R-15, 2015. Guide to Accelerated Conditioning Protocols for Durability Assessment of Internal and External Fiber-Reinforced Polymer (FRP) Reinforcement, ACI Committee 440. American Concrete Institute. ACI 440.2R-17, 2017. Guide for the Design and Construction of Externally Bonded FRP Systems for Strengthening Concrete Structures, ACI Committee 440. American Concrete Institute. ASTM D 2565-99, 1999. Standard Practice for Xenon-Arc Exposure of Plastics Intended for Outdoor Applications.

FRP-strengthened masonry

235

Bank, L.C., Gentry, T.R., Thompson, B.P., Russell, J.S., 2003. A model specification for FRP composites for civil engineering structures. Construction and Building Materials 17, 405e437. https://doi.org/10.1016/S0950-0618(03)00041-2. Barnes, B.A., 1990. Bond and Low Cycle Fatigue Behavior of Thermoset Composite Reinforcing for the Concrete Industry (MS thesis). Iowa State Univ., Ames. Bentur, A., Mindess, S., 2006. Fibre Reinforced Cementitious Composites, second ed. CRC Press. Binda, L., Tedeschi, C., Valluzzi, M.R., Garbin, E., Panizza, M., 2011. Salt crystallization tests on brick masonry reinforced by CFRP textiles. In: Proc. of XII Int Conf Durab Build Mater Components, XII DBMC, Porto, Portugal. Briccoli Bati, S., Rotunno, T., 2001. Environmental durability of the bond between the CFRP composite materials and masonry structures. In: Proc. of III International Seminar on Historical Constructions, pp. 1039e1046. Cardani, G., Valluzzi, M.R., Panizza, M., Girardello, P., Binda, L., 2015. Influence of salt crystallization on composites-to-masonry bond evaluated on site by pull-off tests. Key Engineering Materials 624, 338e345. https://doi.org/10.4028/www.scientific.net/ KEM.624.338. Cardani, G., Binda, L., Valluzzi, M.R., Girardello, P., Panizza, E., Garbin, M., Casadei, P., 2016. On site composites-to-masonry bond evaluation in presence of rising damp and salt crystallization. In: Modena, Da Porto, Valluzzi (Eds.), Brick Block Mason. e Trends, Innov. Challenges. Taylor & Francis Group, London, pp. 365e372. Chen, Y., Davalos, J.F., Ray, I., 2006. Durability prediction for GFRP reinforcing bars using short-term data of accelerated aging tests. Journal of Composites for Construction 10, 279e286. https://doi.org/10.1061/(ASCE)1090-0268(2006)10:4(279). CNR DT 200, 2004. Guide for the Design and Construction of Externally Bonded FRP Systems for Strengthening Existing Structures - Materials, RC and PC Structures, Masonry Structures. Italian National Research Council, Rome. CNR DT 201, 2005a. Guidelines for the Design and Construction of Externally Bonded FRP Systems for Strengthening Existing Structures - Timber Structures. Italian National Research Council, Rome. CNR DT 202, 2005b. Guidelines for the Design and Construction of Externally Bonded FRP Systems for Strengthening Existing Structures - Metallic Structures. Italian National Research Council, Rome. CNReAdvisory Committee on Technical Recommendations for Construction, 2013. Guide for the Design and Construction of Externally Bonded FRP Systems for Strengthening Existing Structures. Materials, RC and PC Structures, Masonry Structures. CNR-DT 200 R1/2013. Crank, J., 1975. Mathematics of Diffusion, second ed. Oxford University Press. Cultrone, G., Torre, M.J.D.E.L.A., Sebastian, E.M., Cazalla, O., 2000. Behavior of brick samples in aggressive environments. Water, Air, and Soil Pollution 119, 191e207. https:// doi.org/10.1023/A:1005142612180. Desiderio, P., Feo, L., 2005. Durability evaluation of EBR CFRP strengthened masonry structures. In: Proc. of the International Symposium on Bond Behaviour of FRP in Structures (BBFS 2005), pp. 481e488. Dejke, V., 2001. Durability of FRP reinforcement in concrete: literature review and experiments. In: Edizione 1 di Publikation/Chalmers tekniska h€ ogskola, vol. 1. Institutionen f€ or byggnadsmaterial, p. 211. ISSN 1104-893X. Elert, K., Cultrone, G., Navarro, C.R., Pardo, E.S., 2003. Durability of bricks used in the conservation of historic buildingsdinfluence of composition and microstructure. Journal of Cultural Heritage 4, 91e99. https://doi.org/10.1016/S1296-2074(03)00020-7.

236

Long-term Performance and Durability of Masonry Structures

Eslami, S., Taheri-Behrooz, F., Taheri, F., 2012. Effects of aging temperature on moisture absorption of perforated GFRP. Advances in Materials Science and Engineering. https:// doi.org/10.1155/2012/303014. FIB TG 9.6, 2001. Externally Bonded FRP Reinforcement for RC Structures. FIB International Federation for Structural Concrete. Technical Report 138. Frigione, M., Aiello, M.A., Naddeo, C., 2006. Water effects on the bond strength of concrete/ concrete adhesive joints. Construction and Building Materials 20 (10), 957e970. https:// doi.org/10.1016/j.conbuildmat.2005.06.015. Ghiassi, B., Marcari, G., Oliveira, D.V., Lourenço, P.B., 2013a. Water degrading effects on the bond behavior in FRP-strengthened masonry. Composites Part B: Engineering 54, 11e19. https://doi.org/10.1016/j.compositesb.2013.04.074. Ghiassi, B., Oliveira, D.V., Lourenço, P.B., 2013b. Recent developments in durability of Frpmasonry systems. In: Proc. of International Conference on Rehabilitation and Restoration of Structures, Madras, Chennai, India, pp. 107e116. Ghiassi, B., Silva, S.M., Oliveira, D.V., Lourenço, P.B., Bragança, L., 2013c. Assessment of the bond quality degradation in FRP-strengthened masonry using IR thermography technique. In: Proc. of 11th Int Symp Fiber Reinf Polym Reinf Concr Struct, pp. 1e9. Ghiassi, B., Oliveira, D.V., Lourenço, P.B., 2014a. Hygrothermal durability of bond in FRPstrengthened masonry. Materials and Structures/Materiaux et Constructions. https:// doi.org/10.1617/s11527-014-0375-7. Ghiassi, B., Silva, S.M., Oliveira, D.V., Lourenço, P.B., Bragança, L., 2014b. FRP-to-Masonry bond durability assessment with infrared thermography method. Journal of Nondestructive Evaluation 33, 427e437. https://doi.org/10.1007/s10921-014-0238-8. Ghiassi, B., Oliveira, D.V., Lourenço, P.B., 2015. Accelerated hygrothermal ageing of bond in FRP-masonry systems. Journal of Composites for Construction 19 (3). https://doi.org/ 10.1061/(ASCE)CC.1943-5614.0000506. Green, M.F., Dent, A.J.S., Bibsy, L.A., 2003. Effect of freeze-thaw cycling on the behavior of reinforced concrete beams strengthened in flexure with fiber reinforced polymer sheets. Canadian Journal of Civil Engineering 30, 1081e1088. Hollaway, L.C., 2010. A review of the present and future utilisation of FRP composites in the civil infrastructure with reference to their important in-service properties. Construction and Building Materials 24 (12), 2419e2445. https://doi.org/10.1016/j.conbuildmat.2010.04.062. Elsevier Ltd. Johnson, W.S., Masters, J.E., Wilson, D.W., Chin, J., Nguyen, T., Aouadi, K., 1997. Effects of environmental exposure on fiber-reinforced plastic (FRP) materials used in construction. Journal of Composites Technology and Research 205. https://doi.org/10.1520/CTR10120J. Karbhari, V.M., 2003a. Durability gap analysis for fiber-reinforced polymer composites in civil infrastructure. Journal of Composites for Construction 7 (3), 238e247. Karbhari, V.M., 2003b. Durability of FRP composites for civil infrastructure myth, mystery or reality. Advances in Structural Engineering 6 (3), 243e255. https://doi.org/10.1260/ 136943303322419250. Karbhari, V.M., Chin, W., Hunston, D., Benmokrane, B., Juska, T., Morgan, R., Lesko, J.J., Sorathia, U., Reynaud, D., 2003. Durability gap analysis for fiber-reinforced polymer composites in civil infrastructures. Journal of Composites for Construction 7 (3), 238e247. Khoshbakht, M., Lin, M.W., Berman, J.B., 2006. Analysis of moisture-induced stresses in an FRP composites reinforced masonry structure. Finite Elements in Analysis and Design 42, 414e429. https://doi.org/10.1016/j.finel.2004.12.013.

FRP-strengthened masonry

237

Khoshbakht, M., Lin, M.W., 2010. A finite element model for hygro-thermo-mechanical analysis of masonry walls with FRP reinforcement. Finite Elements in Analysis and Design 46, 783e791. https://doi.org/10.1016/j.finel.2010.04.002. Kinloch, A.J., Little, M.S.G., Watts, J.F., 2000. The role of the interphase in the environmental failure of adhesive joints. Acta Materialia 48 (18e19), 4543e4553. https://doi.org/ 10.1016/S1359-6454(00)00240-8. Kralj, B., Pande, G.N., Middleton, J., 1991. On the mechanics of frost damage to brick masonry. Computers & Structures 41 (1), 53e66. https://doi.org/10.1016/0045-7949(91)90155-F. Kumar, A., Gupta, R.K., 2003. Fundamentals of Polymer Engineering. Marcel Dekker, Inc., New York. Lin, M.W., Berman, J.B., Khoshbakht, M., Feickert, C.A., Abatan, A.O., 2006. Modeling of moisture migration in an FRP reinforced masonry structure. Building and Environment 41, 646e656. https://doi.org/10.1016/j.buildenv.2005.02.026. Litherland, K.L., Oakley, D.R., Proctor, B.A., 1981. The use of accelerated ageing procedures to predict the long term strength of GRC composites. Cement and Concrete Research 11, 455e466. https://doi.org/10.1016/0008-8846(81)90117-4. Litherland, K.L., Maguire, P., Proctort, B.A., 1984. A test method for the strength of glass fibres in cement. International Journal of Cement Composites and Lightweight Concrete 6 (1), 39e45. Majumdar, A.J., Laws, V., 1991. Glass Fibre Reinforced Cement. BSP Professional Books, Oxford. Maljaee, H., Ghiassi, B., Lourenço, P.B., Oliveira, D.V., 2016a. Moisture-induced degradation of interfacial bond in FRP-strengthened masonry. Composites Part B: Engineering 87, 47e58. https://doi.org/10.1016/j.compositesb.2015.10.022. Maljaee, H., Ghiassi, B., Lourenço, P.B., Oliveira, D.V., 2016b. FRPebrick masonry bond degradation under hygrothermal conditions. Composite Structures 147, 143e154. https:// doi.org/10.1016/j.compstruct.2016.03.037. Elsevier Ltd. Martin, R., 2008. In: Martin, R. (Ed.), Ageing of Composites. Woodhead Publishing, Cambridge. Micelli, F., Nanni, A., 2004. Durability of FRP rods for concrete structures. Construction and Building Materials 18, 491e503. https://doi.org/10.1016/j.conbuildmat.2004.04.012. Micelli, F., Aiello, M.A., 2017. Residual tensile strength of dry and impregnated reinforcement fibres after exposure to alkaline environments. Composites Part B: Engineering. https:// doi.org/10.1016/j.compositesb.2017.03.005. Elsevier. Nadjai, A., Talamona, D., Ali, F., 2005. Fire performance of concrete beams reinforced with FRP bars. In: Proc. of the International Symposium on Bond Behaviour of FRP in Structures (BBFS 2005), pp. 401e410. Nguyen, T.C., Bai, Y., Zhao, X.L., Al-Mahaidi, R., 2012. Durability of steel/CFRP double strap joints exposed to sea water, cyclic temperature and humidity. Composite Structures 94, 1834e1845. https://doi.org/10.1016/j.compstruct.2012.01.004. Ochsner, A., Silva, L.F.M., Altenbach, H., 2013. Design of adhesive joints under humid conditions. In: da Silva, L.F.M., Sato, C. (Eds.), Advanced Structured Materials. Springer Berlin Heidelberg, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-37614-6. Phani, K.K., Bose, N.R., 1986. Hydrothermal ageing of CSM - laminate during water immersion - an acousto-ultrasonic study. Journal of Materials Science 21, 3633e3637. Proctor, B.A., Oakley, D.R., Litherland, K.L., 1982. Developments in the assessment and performance of GRC over 10 years. Composites 13, 173e179. https://doi.org/10.1016/ 0010-4361(82)90056-8.

238

Long-term Performance and Durability of Masonry Structures

Purnell, P., Short, N.R., Page, C.L., 2001. A static fatigue model for the durability of glass fibre reinforced cement. Journal of Materials Science 36, 5385e5390. https://doi.org/10.1023/ A:1012496625210. Purnell, P., Beddows, J., 2005. Durability and simulated ageing of new matrix glass fibre reinforced concrete. Cement and Concrete Composites 27, 875e884. https://doi.org/ 10.1016/j.cemconcomp.2005.04.002. Ray, B.C., 2006. Temperature effect during humid ageing on interfaces of glass and carbon fibers reinforced epoxy composites. Journal of Colloid and Interface Science 298, 111e117. https://doi.org/10.1016/j.jcis.2005.12.023. RILEM MS-A.1, 1998. Determination of the Resistance of Wallettes Against Sulphates and Chlorides. https://doi.org/10.1007/BF02486406. Roylance, D., Roylance, M., 1976. Influence of outdoor weathering on dynamic mechanical properties of glass/epoxy laminate. In: Environmental Effects on Advanced Composite Materials. ASTM, Philadelphia, PA. Schutte, C.L., 1994. ‘Environmental durability of glass-fiber composites’. Materials Science and Engineering R 13 (7), 265e323. https://doi.org/10.1016/0927-796X(94)90002-7. Sciolti, M.S., Frigione, M., Aiello, M.A., 2010. Wet lay-up manufactured FRPs for concrete and masonry repair: influence of water on the properties of composites and on their epoxy components. Journal of Composites for Construction 14, 823e833. https://doi.org/ 10.1061/(ASCE)CC.1943-5614.0000132. Sciolti, M.S., Aiello, M.A., Frigione, M., 2012. Influence of water on bond behavior between CFRPsheet and natural calcareous stones. Composites Part B: Engineering 43, 3239e3250. https://doi.org/10.1016/j.compositesb.2012.03.002. Sciolti, M.S., Aiello, M.A., Frigione, M., 2015. Effect of thermo-hygrometric exposure onFRP, natural stone and their adhesive interface. Composites Part B: Engineering 80, 162e176. https://doi.org/10.1016/j.compositesb.2015.05.041. Soudki, K.A., Green, M.F., 1997. Freeze-thaw response of CFRP wrapped concrete. Concrete International 19 (8), 64e67. Tedeschi, C., Kwiecien, A., Valluzzi, M.R., Zając, B., Garbin, E., Binda, L., 2014. Effect of thermal ageing and salt decay on bond between FRP and masonry. Materials and Structures 47, 2051e2065. https://doi.org/10.1617/s11527-014-0448-7. Valluzzi, M.R., Garbin, E., Panizza, M., Binda, L., Tedeschi, C., 2011. Moisture and temperature influence on FRP masonry bonding. In: Proc. of XII Int Conf Durab Build Mater Components. XII DBMC, Porto, Portugal. Verghese, N., Haramis, J., Lesko, J.J., 1999. Freeze-thaw durability of polymer matrix composites in infrastructure. In: Proceedings of the Fourth International Conference on Durability Analysis of Composite Systems. Duracosys 99, Brussels, Belgium. Wang, K., Young, B., Smith, S.T., 2011. Mechanical properties of pultruded carbon fibrereinforced polymer (CFRP) plates at elevated temperatures. Engineering Structures 33, 2154e2161. https://doi.org/10.1016/j.engstruct.2011.03.006.

Part Two Health monitoring and testing

This page intentionally left blank

Dynamic identification of historic masonry structures

8

Maria-Giovanna Masciotta, Luís F. Ramos ISISE, University of Minho, Department of Civil Engineering, Guimar~aes, Portugal

Chapter Outline 8.1 Introduction 241 8.2 Theoretical background 8.2.1 8.2.2 8.2.3 8.2.4

243

Testing procedures for modal identification of structural systems 243 Output-only modal identification 244 Input-output modal identification 246 General remarks 247

8.3 Applications

248

8.3.1 Mogadouro Clock Tower 248 8.3.2 Qutb Minar tower 252 8.3.3 Saint Torcato church 257

8.4 Conclusions References 263

8.1

262

Introduction

Dynamic identification is the characterization of the main dynamic properties of structural systems based on the analysis of their vibration responses to an input force. Such properties, referred to as modal parameters (i.e., frequencies, mode shapes, and damping ratios), define the inherent characteristics or “fingerprints” of the system and provide useful information about its state. Modal parameters are related to the physical and mechanical properties of the analyzed structure, like mass, stiffness, and energy dissipation, thereby allowing for their characterization even in the absence of viable experimental testing procedures, through the solution of an inverse problem. This relationship also implies that any structural changes a system may undergo over time will be reflected by changes in its modal properties, hence the importance of tracking the dynamic response of structures also for damage identification purposes. Circumstances show that built environment is continuously exposed to the risk of damage, whether due to exogenous or endogenous causes. If not detected in due time, such an adverse condition can compromise the structural integrity and ultimately jeopardize users’ safety. Long-term Performance and Durability of Masonry Structures. https://doi.org/10.1016/B978-0-08-102110-1.00008-X Copyright © 2019 Elsevier Ltd. All rights reserved.

242

Long-term Performance and Durability of Masonry Structures

Within this context, dynamic identification plays a leading role, as it offers real-time punctual checkups of the structural fitness, driving the application of supervised vibration-based damage identification methods to spot the weakest and most vulnerable areas of a structure in due time. Furthermore, by following the evolution of modal parameters through the consecutive repetition of dynamic measurements over time, one can obtain on a systematic basis nearly real-time information about the health status of the monitored system and timely detect anomalies if the structure does not behave as expected. This continuous acquisition process and analysis of data from the structure is referred to as structural health monitoring. Either performed in a continuous or intermittent way, dynamic testing can be considered a kind of global nondestructive health monitoring tool since it enables one to estimate the modal features by only embedding an array of sensors in the structure and recording the corresponding response processes, without resorting to any invasive technique. This aspect, which definitely represents one of the major strengths of dynamic testing, gains further importance when dealing with cultural heritage assets, where the need to respect the historical value of the constructions often limits the range of applicable techniques for the system characterization. Another benefit that is worthwhile to mention consists of the possibility to exploit dynamic testing for the evaluation of the system response before and after structural interventions, allowing one to control and appraise the effectiveness of the adopted remedial solutions. All these considerations throw light on the remarkable increase that dynamic vibration testing has seen in the last decades as a preferred tool for the assessment of the global health conditions of civil and monumental structures, as well as for the development of realistic behavioral models of complex engineering systems. In detail, modal analysis has been widely implemented in aerospace, mechanical, and civil engineering applications for vibration trouble shooting, structural dynamics modification, analytical model updating, optimal dynamic design, vibration control, and vibration-based structural health monitoring (Zhang et al., 2005). Well-established methods and system identification algorithms for the extraction of the most relevant modal parameters from the measurements of the dynamic response have also been developed (Maia and Silva, 1997; Reynders, 2012). However, the transfer of input devices typically used in mechanical engineering to the civil engineering field is not always feasible, as exciting large civil structures in a controlled manner can often be impractical. Notwithstanding, the ceaseless technologic progress in transducers and analog-to-digital converters has allowed us to overcome this limitation by making operational modal analysis possible. Unlike traditional experimental modal analysis (EMA) where both excitation and structural response are measured, and modal parameters are deterministically estimated from input-output data, operational modal analysis (OMA) only requires records of the structural response to freely available natural excitation sources, such as traffic, wind, streams, and microtremors. As this type of excitation is random in nature and cannot be measured exactly, the main assumption on which OMA relies is the consideration of ambient excitation as a stationary Gaussian white noise stochastic process, viz. a broadband random signal with flat power spectrum in the frequency range of interest for the structure. In this case,

Dynamic identification of historic masonry structures

243

with the input load being unknown, modal parameters are identified from output-only data by applying suitable stochastic modal identification techniques. The main scope of this chapter is to provide a brief overview of both forced and ambient vibration testing procedures along with their relevant input-output and output-only modal identification techniques, focusing on their application to historic masonry buildings. As widely known, the uniqueness and complexity of ancient structures make the understanding of their behavior a true challenge. Full-scale in situ experimentation represents the only way to shed light upon the actual performance of these constructed systems and to create a thorough knowledge about built heritage. Hereafter, the results from full-scale dynamic testing of three historic masonry buildings are presented and discussed. The Qutb Minar tower in India, one of the tallest stone masonry towers in the world, is among the investigated monuments. For each of the presented case studies, the application of vibration tests and modal analysis procedures for the characterization of the system’s dynamic behavior is fully described, showing how the information obtained enlightens as to the actual response of the structure and may be exploited for structural assessment purposes.

8.2 8.2.1

Theoretical background Testing procedures for modal identification of structural systems

Conventional modal testing procedures are performed by exciting the structure with a known input force and capturing its response by a set of sensors deployed at selected locations along the structure, trying to operate with high enough spatial density and frequency resolution. The identification of the modal parameters is then obtained by estimating the frequency response functions (FRFs) or the impulse response functions (IRFs) from input-output data. Before performing experimental tests, it is recommended to carry out preliminary finite element (FE) modal analyses to drive the selection of measurement points (number and location), time duration, and sampling rate of output signals. The dynamic response of the structure can be measured by any kind of device able to convert physical quantities such as displacements, velocities, accelerations, strains, etc., into proportional electrical signals, ready to be processed by the data acquisition system (DAQ). Displacements and velocities transducers are all suited for this purpose, but usually equipment based on accelerometers (piezoelectric, piezoresistive, capacitive, or force balance) are preferred, because of their relatively low cost and high sensitivity at the same time (Fig. 8.1). However, the so-recorded response signals are rather low and must be amplified by conditioning units provided with both noise and antialiasing filters. Then, to be processed, the measured continuous analog signals are converted to discrete digital signals through an analog-to-digital converter (Masciotta, 2015). Once acquired, the digital raw signals have to be preliminarily analyzed and processed. This means: (1) check the data for clipping, drop-out voltage, and spikes;

244

Long-term Performance and Durability of Masonry Structures

Figure 8.1 Examples of high-sensitivity piezoelectric accelerometers for vibration monitoring of civil structures (www.pbc.com): (a) Model 393B12, (b) Model 393B04, (c) Model 393C.

(2) de-trend the signals by removing any possible trend due to a slowly varying mean value; (3) filter the signals to get rid of undesired frequency components through the application of different types of window functions in the frequency domain, i.e., high-pass, low-pass, or band-pass; (4) down-sample (decimate) the signals to reduce the number of values in the time histories and speed up the processing time for the subsequent modal identification; and (5) reduce leakage errors caused by differences between sampling time and signal period through appropriate time windowing, such as Hanning, cosine-taper, and the like (Fig. 8.2). After preprocessing the data, it is possible to proceed with the modal identification of the structure either by frequency domain (FD) or by time domain (TD) approaches. Obviously, the use of two or more identification methods will lead to major confidence in the results.

8.2.2

Output-only modal identification

Modal identification methods capable of estimating the modal parameters of a structure from unknown natural ambient excitation are named output-only identification techniques, also referred to as OMA (Brincker and Kirkegaard, 2001; Peeters and De Roeck, 2001; Reynders, 2012). The main assumption on which OMA relies is the consideration of the excitation as a stationary Gaussian white noise stochastic process. Although this does not reflect the reality, it is anyway a good approximation since the excitation can be seen as the response to a linear filter excited with white noise input. Compared to traditional EMA, OMA results are much more attractive and reliable as a dynamic identification tool. The main benefits characterizing output-only identification techniques are (1) the possibility of measuring the response of the structure using freely available environmental excitations, such as wind, traffic, microtremors, and human walking (the more the input randomness, the better the modal results); (2) the low cost of the tests since no heavy and expensive equipment is necessary;

Dynamic identification of historic masonry structures

245 T S = 9s

2.0

2.00

1.0

1.00

0.0 0.0 -1.0

2.0

4.0

6.0

8.0

10.0

x (t )w (t )

-2.0

2.0

4.0

6.0

8.0

10.0

-2.00

Time(s)

(a)

Time(s)

0.25 0.20 0.15 0.10 0.05 0.00

0.20

Xn

0.15 0.10

2.56

2.11

2.33

1.89

1.44

1.67

1.22

1.00

High-pass Amplitude

Band-pass Amplitude

Amplitude

Low-pass

Frequency

0.78

ω (Hz)

ω (Hz)

(b)

0.56

0.00

0.11

2.56

2.11

2.33

1.89

1.67

1.44

1.22

1.00

0.78

0.56

0.33

0.05

0.11

Xn

0.00 -1.000.0

0.33

x (t )

T S = 9s

Frequency

Frequency

Figure 8.2 Digital processing. (a) Hanning window application: signal before and after windowing and corresponding discrete Fourier spectra; and (b) main frequency filtering windows.

and (3) the possibility of carrying out real in situ testing, based on true boundary conditions and on real in-operation conditions, without interrupting the normal use of the structure. Additionally, since ambient excitation provides multiples inputs, OMA is clearly based on multiple-input/multiple-output (MIMO) techniques, thus even closely spaced modes can be estimated. The only shortcoming might arise in the presence of a low level of ambient excitation, a factor that can hamper the identification of highfrequency modes. Two main groups can be distinguished within output-only modal identification methods: nonparametric methods developed in FD and parametric methods developed in TD. The first group is based on the estimation of modal parameters from the power spectral densities of the measured output signals after the application of the fast fourier transform (FFT) process. The second group is based on the identification of modal parameters by fitting the response correlation functions (obtained from FFT algorithm or random decrement method) of each measurement point to a mathematical model representative of the dynamic behavior of the structure. It is worth noting that FD methods are simpler and faster in comparison to TD methods, but they are limited by the frequency resolution of the spectral density estimates that, if low, can lead to heavily biased modal estimates.

246

Long-term Performance and Durability of Masonry Structures

Depending on the number of steps involved, both FD and TD methods can be further classified in two-stage or one-stage approaches. The two-stage approach estimates correlation functions (TD) or power spectral densities (FD) in the first stage, and then extracts modal parameters from these estimates. The one-stage approach directly extracts modal parameters from the discrete time histories (TD) or the Fourier transforms (FD) of the output signals. Many dynamic identification algorithms have been developed since the 1990s, from the basic frequency domain technique, namely the peak picking (PP), to the improved frequency domain decomposition (FDD) and the more refined enhanced frequency domain decomposition (EFDD) and frequency-spatial domain decomposition techniques, from stochastic subspace identification (SSI) methodsdeither covariance or data-drivendto procedures that originate from traditional EMA techniques, such as Ibrahim time domain (ITD), auto regressive moving average (ARMA), complex exponential (CE), least square complex exponential (LSCE), polyreference complex exponential (PRCE), and eigensystem realization algorithm (ERA).

8.2.3

Input-output modal identification

Modal identification methods based on the use of both input (excitation) and output (response) measurements to estimate the modal parameters are called input-output dynamic identification techniques and belong to the field of traditional EMA (Maia and Silva, 1997; Ewins, 2000; Cunha and Caetano, 2006). Since the 1960s, EMA has obtained substantial progress and numerous modal identification algorithms have been developed, from single-input/single-output techniques to single-input/multi-output and MIMO techniques, either in the TD or FD. The goal of these identification methods is to extract as many modal information as possible by properly exciting the structure during the experimental testing. As far as civil structures are concerned, forced-vibration tests can be carried out using different excitation mechanisms, such as impact hammers, drop weight systems, or shakers, which have to be adequately chosen with respect to the size of the structure to be tested. For instance, impact hammers can work satisfactorily in the case of small and medium-size structures, but in the case of large structures, a greater amount of energy is needed to excite all the relevant vibration modes, thus electro-dynamic or electrohydraulic shakers as well as eccentric mass vibrators may be more suitable for attaining higher frequency resolution. Input-output modal identification algorithms rely on deterministic estimates of FRFs in the FD or IRFs in the TD. Such functions describe the response of a linear time-invariant system for all frequencies. In the FD, modal identification techniques can range from simpler SDOF formulationsdlike PP, circle-fit and inverse method (IM)dto more sophisticated MDOF formulations, like rational fraction polynomial, complex exponential frequency domain, and polyreference frequency domain. On the other hand, in the TD, either direct methods, such as ARMA, and indirect methodsdlike CE, LSCE, PRCE, ITD, and ERAdcan be employed. Given the limitations on the resolution of FD methods and the possible presence of leakage errors in

Dynamic identification of historic masonry structures

247

the estimates, TD methods are usually preferred when a large frequency range or a large number of modes exist in the data (Cunha and Caetano, 2006). Although there is a wide field of application, traditional EMA presents some drawbacks, such as the difficulty in artificially exciting large and complex structures with sufficient energy and in a controlled manner, the high cost of the equipment required for testing, and the impossibility to adequately simulate real operation conditions in a lab environment. These limitations led the civil engineering community to focus on modal identification techniques based on response measurements only, a great alternative to traditional EMA.

8.2.4

General remarks

Although there are outstanding advantages, output-only modal identification techniques present several shortcomings. One of the main drawbacks is the impossibility of mode shapes scaling due to lack of input information. Mass normalized mode shapes are needed in applications such as structural health monitoring or vibrationbased damage identification, so the knowledge of the scaling factors of the mode shapes is important. To overcome this issue, some methods based on repeated testing introducing mass changes have already been proposed (at first by Parloo et al., 2001; Brincker and Andersen, 2003). Another approach recently addressed (Aenlle and Brincker, 2013) is to update a finite element model of the structure using modal parameters estimated by OMA and, if a good correlation is present, to scale the experimental mode shapes using the mass matrix of the finite element model. The structural mode sorting is another drawback of OMA: in many practical cases, in addition to random loads, background noise and/or harmonic excitations due to rotating machinery and/or fluctuating forces are also present. Due to that, the distinction between structural modes and noise or spurious modes can become very difficult and lead to an inaccurate modal identification. This issue mostly concerns TD OMA methods, since FD methods perform much better in structural determination: for instance, FDD techniques can almost eliminate spurious mode problems thanks to a statistical indicator (Brincker et al., 2000). A step forward has been done in TD methods as well; for example, Mohanty and Rixen (2004) proposed a modification of the LSCE method to account for harmonic components in the response. Altogether, all the major issues in OMA have been troubleshot little by little, contributing to the refinement of the relevant modal identification algorithms. Nevertheless, some shortcoming still persists, such as the inaccuracy of modal parameters estimates in FD if the signal power spectral densities (PSDs) are low in resolution, the difficulty in performing repeated tests with mass changes to get scaled modes in case of large and complex structures, the ill-conditioning of outputs measurements when dealing with broad-banded background noise during testing, or even the limitations in modal parameters identification if the frequency content of ambient forces is quite narrow-banded. In the light of these considerations, it is clear that issues related to sensing equipment, type of modal analysis technique to adopt, data acquisition system, data processing, and removal of noise and environmental effects must be carefully handled to

248

Long-term Performance and Durability of Masonry Structures

succeed in the dynamic characterization of structural systems. Furthermore, a thorough test planning tailored to the specificity of the structure to characterize and to the objectives to achieve should always be made beforehand. The applications described in the next section will give a better insight into these aspects when dealing with historical constructions.

8.3

Applications

Ambient vibration tests (AVTs) supported by output-only modal identification techniques are definitely the best nondestructive tools available to characterize the dynamic behavior of ancient masonry structures. Although still limited in number, several applications of OMA techniques to historic buildings can be found in literature (e.g., Gentile and Saisi, 2007; Ramos et al., 2010; Saisi and Gentile, 2015; Ubertini et al., 2016; Masciotta et al., 2016, 2017). Hereafter, three case study applications on full-scale monumental structures are presented with the aim of illustrating the main steps involved in the system characterization process of historic masonry structures and highlighting the important role that dynamic identification plays for a complete comprehension of the actual structural behavior of such complex constructions.

8.3.1

Mogadouro Clock Tower

Located within the castle perimeter of the homonymous town in the Northeast of Portugal, the Mogadouro Clock Tower is a historic masonry structure built after 1559 to serve the nearby church as a bell tower (Fig. 8.3). The fabric is 20.4 m high and features a rectangular cross-section of 4.7
4.5 m2 with masonry walls of about 1 m thickness. Large granite stones are used at the corners, whereas the central part of the walls mainly consists of rubble stones with thick lime mortar joints. At the top level, eight masonry columns support the roof body.

Figure 8.3 The clock tower and the Mogadouro castle.

Dynamic identification of historic masonry structures

249

Due to the lack of maintenance, the tower appeared in very poor condition, showing out-of-plane displacements, large cracks, material deterioration, and biologic growth throughout. The parts mainly affected were the east and west façades, where deep cracks were splitting the box cross-section of the tower into two separate U bodies, thereby compromising the structural integrity (Fig. 8.4). Conservation works were carried out to reinstate the tower safety, including lime grout injections for the consolidation of the walls, replacement of material with high level of degradation, filling of voids and losses, and installation of tie-rods at two levels. To evaluate the global response of the tower before and after the works, AVTs were conducted making use of wind and traffic as excitation sources (Ramos et al., 2010). The response of the tower was acquired in 54 selected points by means of uniaxial piezoelectric accelerometers, deployed according to the layout shown in Fig. 8.5. Based on the results of a preliminary FE dynamic analysis, the nodal processes were recorded with a sampling frequency of 256 Hz for a duration of about 11 min to ensure an acquisition time window 2000 times larger than the fundamental period of the structure. The same test planning and measurement points were adopted before and after the works. For either structural condition, the dynamic parameters were estimated by comparing the results from two well-known and complementary OMA techniques: the EFDD (Brincker et al., 2001) and the SSI (Peeters and De Roeck, 1999) methods, (a)

Loss of material

(b)

(c)

Level 5 Level 4 Loss of material

Level 3

Level 2

(d) Loss of material

Out of plane displacements

Main cracks

Loss of material Main cracks

Biological growth

Main crack

Level 1

z x

(e)

y

x

y

(f)

z

z

z

(g)

(h)

Figure 8.4 Damage survey in the tower: (a) south, (b) east, (c) north, and (d) west façades; cracks on the (e) east and (f) west fronts; (g) inner crack in the west façade; and (h) example of material loss.

250

Long-term Performance and Durability of Masonry Structures

(a)

(b)

S7 S8 S9

S4E5S5

E10

(c)

E11 E13 E14 E12

E5 E6

S6

E8

N11

E9

N4

E7 S1

S2 S3

E1 E2

z

(d)

N12 N14 N15 N13 N6

W13 W15 W16 W12 W14

N9 N10

W7

N5 N7 N8

N1 N2 N3

E4 E3

W1

W11 W10

W3 W5 W6

W2 W4

z

z x

W9 W8

y

x

z y

Figure 8.5 Sensor layout adopted in the AVTs of the Mogadouro tower: (a) south, (b) east, (c) north, and (d) west façades (reference sensors are indicated inside a grey box).

implemented in the commercial software ARTeMIS (SVS, 2006). The application of both techniques allowed the identification of seven vibration modes in the frequency ranges 2e9 Hz and 2e17 Hz for the damaged and undamaged conditions, respectively. The estimated natural frequencies and damping ratios are summarized in Table 8.1, whereas the corresponding mode shapes and modal assurance criterion (MAC) values are displayed in Fig. 8.6. For the sake of brevity, only the modal features identified by the SSI are shown. The global parameter results relevant to the two structural conditions reveal an average increase of 50% in terms of frequency values, while damping decreases around 40%. Concerning the experimental mode shapes, similar configurations are observed for the first five vibration modes, distinguishing four dominant bending modes in the two main planes of the tower (modes 1, 2, 4, and 5) and one torsional mode (mode 3), whereas modes 6 and 7 appear swapped. However, despite the apparent resemblance, the MAC, i.e., a statistical indicator of the degree of similarity between two mode shape vectors, denotes a weak correlation between comparable mode shapes, reading values lower than 0.65. This result is clearly due to the local protuberances affecting the mode shape configurations of the damaged tower both in the upper part and in the areas close to the cracks because of the presence of local damage mechanisms before the works. On the contrary, a monolithic behavior characterizes the global response of the tower after the conservation works. The analysis of the results allows one to conclude that the presence of damage changed significantly the dynamic behavior of the tower with respect to the possible original configuration, but the structural intervention enabled the reduction of the nonlinear phenomena effects, leading to a stiffer system. Hence, the strengthening works were considered efficient. Finally, a 3D numeric model was built, and an FE model updating analysis was performed to better assess the dynamic response of the tower before and after the retrofitting (Ramos, 2007). The numeric models simulating the two structural conditions

Before

Before

After

After

Mode

f (Hz)

CVf (%)

f (Hz)

CVf (%)

Df (%)

x (%)

CVx (%)

x (%)

CVx (%)

Dx (%)

1st

2.15

1.85

2.56

0.21

þ19.28

2.68

219.51

1.25

0.13

53.26

2nd

2.58

1.05

2.76

0.30

þ6.70

1.71

94.02

1.35

0.17

21.00

3rd

4.98

0.69

7.15

0.27

þ43.67

2.05

65.33

1.20

0.14

41.32

4th

5.74

1.56

8.86

0.47

þ54.37

2.40

24.27

1.31

0.13

45.72

5th

6.76

1.13

9.21

0.21

þ36.13

2.14

31.74

1.16

0.12

45.65

6th

7.69

2.94

15.21

2.24

þ97.87

2.33

55.98

2.54

0.24

þ9.11

7th

8.98

1.21

16.91

1.40

þ88.27

2.30

46.39

1.49

0.23

35.07

Avg

e

1.49

e

0.73

þ49.47

2.23

76.75

1.47

0.17

40.34

Dynamic identification of historic masonry structures

Table 8.1 Dynamic response of Mogadouro Tower before and after consolidation works

*Average value of damping calculated only for negative differences.

251

252

Long-term Performance and Durability of Masonry Structures

Before rehabilitation

1st mode 2nd mode 3rd mode 4th mode

x

After rehabilitation

6th mode

7th mode

y

z

2.15 Hz

2.58 Hz

4.98 Hz

5.74 Hz

6.76 Hz

7.69 Hz

8.98 Hz

2.56 Hz 0.48

2.76 Hz 0.65

7.15 Hz 0.45

8.86 Hz 0.16

9.21 Hz 0.42

15.21Hz 20%

Beetles Termites Blue stain fungi Brown rot White rot

Predominantly or permanently wet MC > 20%

Beetles Termites Blue stain fungi Brown rot White rot Soft rot

UC 5

In salt water

Permanently wet MC > 20%

Marine borers (or UC 4 for elements outside the water)

As an illustration to the application of those guidelines, poplar (Populus canescens) features a very low durability against biologic deteriorations (Tables 11.1e11.4); yet it has to be used indoors with protection measures, sheltered from weather exposition (UC 1 or UC 2) to keep the wood dry with MC  20%. On the other hand, European oak (Quercus robur) features high durability against fungal decay and good resistance to insect attack and can be used for every exposure condition. However, some treatments should ideally be applied on wood surfaces when timber elements made of European oak belong to UC 4 and UC 5, since the respective exposure conditions are severe. Nevertheless, checking the UC requirements does not enable one to predict the long-term performances of timber structures over time. To this end, empirical durability models must be established with respect to the natural durability of wood and climate exposure conditions.

Service life design of timber structures

11.5

321

Performance assessment

11.5.1 Definition and context Performance assessment over time is required to predict the service life of timber structures for new and existing buildings. Before proceeding, several specific terms introduced in the preceding sentence have to be explained. The service life is the period of time after installation during which a building or its parts meets or exceeds the performance requirements (ISO 15686-1, 2000). On the other hand, the performance stands for the behavior of the product (e.g., wood material, timber structure) in terms of its effectiveness in tests (EN 1001-2, 2005). The term “performance” also applies to the effectiveness of the product (e.g., mechanical strength, wood durability, aesthetic aspect) in practice against the individual or collective effects of particular biologic agents through its service life (EN 1001-2, 2005). Furthermore, performance requirements consist of the minimum acceptable level of a critical property of the product, being defined as a limit state between acceptable and nonacceptable performance (Viitanen et al., 2010). The performance assessment of timber structures may be necessary at a certain moment of time in the building life, under different reasons (van de Kuilen, 2007): • • • •

damage of timber structures, standing for biologic, physical, or chemical wood deteriorations (e.g., fungal decay, insect attacks, cracks, etc.), resulting in a reduction of mechanical and aesthetical performances; reuse of timber coming from former structures demolished after their service life, for which the adequacy of their performance in new buildings has to be checked; change of use, alterations, and changes in timber structures for which the performances may not meet with current configuration of the building; maintenance planning.

Without going further in detail, (ISO 15686-7, 2017) determines the main phases for the performance survey: defining the task, planning, examining, evaluating, and reporting. Furthermore, different types of inspection on site can be sorted out into four levels when assessing the performance of timber structures (ISO 15686-7, 2017): •

preliminary: introductory inspection through visual observation and basic measurements to provide a very rough overview; • regular: examination of supporting data (e.g., drawings, requirements, documentation) at regular intervals through extensive measurements; •maintenance-driven: performance registration of the existing conditions at the time of a loss function (i.e., unexpected minor failure); • specific/detailed: performance registration of a special loss function (i.e., unexpected major failure), implying accurate measurements or test methods through research work.

11.5.2 Evaluation of performance-influencing factors Before considering any durability model, several factors that influence the long-term performance of timber structures must be evaluated over time: climate exposure, cracks, biologic wood deteriorations, and MC. Based on these factors, the long-term

322

Long-term Performance and Durability of Masonry Structures

performance of timber structures can thereafter be evaluated through different aspects: optical, aesthetical, durability, moisture, mechanical, and functional performance. Note that the insect attacks are not mentioned as performance-influencing factors since their assessment over time is challenging in respect with environmental storage conditions of timber structures.

11.5.2.1 Climate exposure When considering the climate exposure of outdoor and indoor timber structures, it is crucial to firstly define the different investigation scales. In the literature, three scales of climate exposure can then be evaluated (Viitanen et al., 2010): • • •

macroclimate: regional climate in the area where the building and/or timber structure is situated, supported by meteorological data (e.g., temperature, air relative humidity, solar radiation, rain, wind); mesoclimate: climate conditions in the vicinity of the building and/or timber structure, which are still not disturbed either by the wood properties nor by the geometric features of the building and/or timber structure; microclimate: climate close to the wood material, featured by the wood MC and temperature, which highly depend on the geometric features of building and/or timber structures, on the wood properties, and on the environmental exposure conditions.

Before assessing the long-term performance of timber elements, their exposure conditions to macro-, meso-, and microclimates must be defined through collecting data, according to the geographic location and the method of test sites (Humar et al., 2015): shading or free exposure, underground, above with or without contact with soil, etc. As the climatic data of timber elements need to be recorded over large periods of time, extra monitoring devices must be used to guarantee the long-term safety of the test sites against human actions, storms, or other climatic catastrophes. After determining the exposure conditions of test sites, general information about the respective exposure site has to be collected through available documentation (Humar et al., 2015): geographic position, height above sea level, and photographs showing the test site and its surrounding in terms of geographic orientation. On the other hand, detailed information about the exposure test site is required through long-term measurement of data: precipitation, fluctuation (i.e., maximum and minimum) and average of air temperature, average relative humidity, wind speed and direction, and global irradiance. Note that these data can be obtained by using the closest weather station to the test sites. It should also be mentioned that some climate dataset systems, such as Scheffer Climate Index or ERA-40, for example, can be used as a database to predict the risk of mold or fungal decay in timber structures through empirical durability models, with respect to their environmental storage conditions (Viitanen et al., 2010).

11.5.2.2 Cracks Because wood material is hygroscopic, climate variables such as the temperature and air relative humidity realize a fluctuation of wood MC under the fiber saturation point

Service life design of timber structures

323

(i.e., MC  26%e30%) (Shupe et al., 2008), resulting in the variation of timber element dimensions mostly in the tangential and radial directions through the swelling and shrinking process. It can then be obvious to relate the development of cracks with the MC variation. The total number of drying and wetting cycles as well as the time of an individual shrinkage process may condition the development of cracks inside timber elements. Furthermore, the appearance and dimensions of cracks strongly depend on wood species (i.e., density and hygroscopic behavior), on outdoor or indoor climate variables (e.g., temperature, air relative humidity, rain, wind, solar radiation, etc.), on the MC oscillation, and on the annual ring orientation on the wood surfaces (Sandberg and S€oderstr€ om, 2006). Cracks become more apparent for dry timber elements, causing minor aesthetic damage. On the other hand, the wood durability and mechanical properties are reduced since the development of cracks increases the risk of biologic decay occurring beyond the surface of the wood. The development of cracks can be determined through visual assessment on the external surfaces of timber elements. To this end, several geometric parameters of cracks have to be quantified (Humar et al., 2015): the total crack length, number of cracks, and mean maximum crack width. Any cracks present having total lengths less than 5 mm can be disregarded since their respective impact on the performance of timber elements is very low. In addition to this, complementary visual analysis by a tomographic X-ray scanner can be performed in practice on each external surface of timber element to classify the cracks into several categories (Mergny et al., 2016): shakes, loosened grains, checks, and splits. So far, it has been challenging to monitor cracks based on visual assessment only, as well as to establish any durability model for the prediction of growing cracks with respect to the MC fluctuations over time. Nonetheless, extended knowledge on the moisture states and gradients close to the external surfaces of timber elements is crucial to identify zones subject to moisture-induced stresses, thereby triggering the cracking. This objective could be achieved, for example, by simulating the hygrothermal response of timber structures under natural climates through a numeric approach (Fortino et al., 2013).

11.5.2.3 Fungal decay Since wood material is subject to biologic deteriorations such as fungal decay, performance loss (i.e., mechanical properties, wood durability, etc.) may occur in affected timber structures in their service life. To predict the performance loss through durability models, the onset and development of fungal decay must firstly be evaluated over time with respect to wood and fungus species, indoor and outdoor climate exposure conditions for timber structures, wood MC, and temperature. Fungal decay can usually be assessed by visual inspection combined with a picktest using a pointed knife, which has been pricked into the wooden specimens beforehand. This method has been used at different European tests sites for which the related climate variables have already been measured (Humar et al., 2015). The knife is pulled out and pricked again to evaluate the surface strength over time, the fracture depth, and

324

Long-term Performance and Durability of Masonry Structures

the splinter characteristics of any rotted wood with respect to the decay rating scale from EN 252 (2012) given in Table 11.5. Based on data recorded from the visual assessment, wooden specimens can be sorted out into five classes of decay rating (EN 252, 2012): no attack, slight attack, moderate attack, severe attack, and failure. The decay rating scale provides a good understanding for estimating the performance loss of wooden zones damaged by fungal decay. Meanwhile, reliable durability models are still required to predict the development of fungal decay and the performance loss over time per the climate variables. In the last decade, experimental research (Brischke and Rapp, 2008; Brischke and Meyer-Veltrup, 2016; Meyer-Veltrup et al., 2017a) on the development of fungal decay in wooden specimens has been carried out for different test methods for in- and aboveground exposure conditions, at 23 European sites over a period of several years. The results showed that the wood MC, the wetting time, and the decay development highly depended on the test method, the related climate variables (e.g., temperature and air relative humidity), the types of rot induced by fungus species, and on the wood species featured by their natural durability. Furthermore, fluctuations of MC and temperature, due to repetitive wetting and drying cycles with respect to climate exposure, have a significant impact on the fungal growth and survival in timber structures (Pasanen et al., 2000).

11.5.2.4 Wood MC In addition to the temperature, wood MC strongly influences the long-term performance of timber structures (e.g., mechanical, durability, hygroscopic, aesthetical, etc.), mainly through the conditioning of wood deformations, the onset and Table 11.5 Rating scale for the extent of fungal decay according to EN 252 (EN 252, 2012) Rating

Classification

Definition

0

No attack

No change perceptible by the means at the disposal of the inspector in the field; if only a change of color is observed, it shall be rated 0

1

Slight attack

Perceptible changes, but very limited in their intensity and their position or distribution: changes that only reveal themselves externally by superficial degradation, softening of the wood being the most common symptom

2

Moderate attack

Clear changes: softening of the wood to a depth of a least 2 mm over a wide surface (covering at least 10 cm2) or by softening to a depth of a last 5 mm over a limited surface area (covered less than 1 cm2)

3

Severe attack

Severe changes: marked decay in the wood to a depth of a least 3 mm over a wider surface (covering at least 25 cm2) or by softening to a depth of at least 10 mm over a more limited surface area

4

Failure

Impact failure of the stake in the field

Service life design of timber structures

325

development of cracks, fungal decay, or insect attacks. The wood MC is the most important factor affecting the use classification and the durability of timber structures in their service life. Therefore, the determination of MC values in timber structures has to be carried out over time with respect to their environmental storage conditions. Overall MC values of timber elements can be estimated by using a database of wood hygrometric equilibrium curves according to the ambient temperature, air relative humidity, and wood species (Cruz et al., 2015). However, the mere determination of wood EMC may be not enough when individual zones of timber elements are subject to dampness within the building envelope, resulting in local wood deterioration. In that case, experimental measurement systems (e.g., gravimetric, electrical resistance, or capacitive methods) and numeric methods are required to assess the MC distribution inside wood and on the external surfaces to identify the damp regions where decay is more likely. Because the MC oscillation may occur close to the external wood surface, indoor and outdoor climate variables (i.e., ambient temperature and air relative humidity) can accurately be approximated as harmonic cycles with different periods (Svensson et al., 2011). Several possible MC states near the external wood surface are then obtained by keeping the same sinusoid cycle of relative humidity. As a result, the MC response to the pure harmonic case can be characterized by four parameters, illustrated in Fig. 11.2 (Svensson et al., 2011): • • • •

EMC of wood, with respect to ambient temperature and air relative humidity; MC amplitude A, featured by a minimal (maximal) MC value, respectively noted MCmin (MCmax ) at the external wood surface during the drying (wetting) process; MC penetration, distance from the external surface to inner wood along which the MC values differ from the EMC; MC gradient close to the external surface of wood.

The development of moisture-induced stresses triggering cracking may occur due to the presence of significant moisture states and gradients near the external surfaces.

Moisture content Mc (%)

17

Mcmax

16 A = Mcmax – Mcmin

15 14 13

C

12 11

Mcmin 0

2

4

B 8 Length L (mm)

10

12

14

Figure 11.2 Moisture states and parameters at the timber member end initially in balance with the EMC (Svensson et al., 2011).

326

Long-term Performance and Durability of Masonry Structures Tangential direction Δ7.7M%

Symmetry

22 20

Δ2.0M%

0 day 30 days 60 days 90 days

Passive zone

180 days 270 days 360 days

Passive zone Active zone

Active zone

18

14 12 10

0

Δ3.6M%

16 Δ12M%

Moisture content (M%)

24

25

R 50 75 100 125 Width of cross section (mm)

L T

Figure 11.3 Experimental results of MC distribution (Franke et al., 2016) in tangential direction during the absorption process (left), with the distinction between the passive and active zones in the timber element cross-section (right).

To assess the moisture-induced stresses, Fortino et al. (2013) proposed a numeric approach to simulate the MC distribution in the cross-section of elements from a timber bridge, based on the air relative humidity and temperature histories from a nearby meteorological station. The authors (Fortino et al., 2013) showed that high moisture gradients can be caused by daily variations of air relative humidity in different seasonal periods. Meanwhile, a protective coating can strongly reduce the MC peaks and moisture gradients in the timber element cross-section, resulting in preventing the development of moisture-induced stresses. The inner MC values may be measured at several points in the radial and tangential directions of timber samples through an electrical resistance method (Franke et al., 2016). When comparing with experimental results, numeric programs and analytic equations were used to estimate the MC distribution. As shown in Fig. 11.3, the active and passive zones in the timber sample cross-section were then defined. From the use classification (Table 11.4), the passive zone belongs to UC 1, since the maximum value of MC recorded is less than 20%. On the other hand, the active zone featured by high risk of biologic wood deteriorations belongs to UC 2 or 3, as the maximum value of MC measured is greater than 20%.

11.6

Durability models

After assessing every performance-influencing factor (i.e., climate exposure, cracks, fungal decay, wood MC), empirical durability models may be used to predict the occurrence risk of wood deteriorations and the performance loss of timber structures in their service life. For example, durability models can be used when designing timber structures at the early stage of construction or when assessing their performance in existing buildings. However, they should be used very carefully, taking into account the use classification, the natural durability and permeability of the wood species, and last but not least, the presence of preventive measures.

Service life design of timber structures

327

11.6.1 General models In the last 2 decades, several Australian research projects (Leicester, 2001; Foliente et al., 2002) have developed general durability models for timber engineering, by considering biologic and chemical damage (i.e., fungal decay, attacks of termites or marine borers, and corrosion). In their work, different environmental conditions were investigated, including timber elements in ground, in sea water, above ground, exposed outdoors, and within the building envelope. Based on the same concepts used to predict fungal decay development inside wood, idealized durability models establish a linear evolution of the wood deterioration depth and of the related rate in the timber element cross-section over time (Nguyen et al., 2008; MacKenzie et al., 2013). As illustrated in Figs. 11.4 and 11.5 for the fungal decay, both depth and rate of decay depend on the natural durability of wood, the location of damaged regions within the wood specimen, the potential presence of preventive measures, and on the climate exposure conditions. A lag time is noticed between the emergence time of fungal decay and the starting time of significant performance loss of timber structures in relation to the increase of decay depth in the element cross-section (Fig. 11.4). Furthermore, the application of antifungal treatments may introduce additional lag time on the wood decay process, although periodical maintenance is necessary to ensure their long-term effectiveness. The prediction of depth and rate of wood degradations is crucial to estimate the cross-section reduction of damaged timber elements over time, resulting in checking their structural safety and performance loss. Nonetheless, the relevance of these durability models from previous work (Nguyen et al., 2008; MacKenzie et al., 2013) is under scrutiny because most of the input data requested (e.g., initial storage conditions and climate exposure before the emergence of wood deteriorations, wood age and species, maintenance history, etc.) are rarely accessible. Besides, they may not be reliable for some case studies of timber structures (e.g., roof or floor timber beam ends in contact with moist masonry walls) because the climate variables are more complex and need to be calibrated on a case by case level (Sousa et al., 2014). Through a failure analysis, other durability models (van de Kuilen, 2007) can estimate the residual lifetime of timber structures damaged as a result of fungal decay to determine the necessity of reinforcement, repair, or replacement. Base decay progress

Decay progress when there is maintenance by an external treatment

Depth of decay (mm) Rate

Extra lag due to an external treatment

Basic lag

Application time of an external treatment

Time (years)

Figure 11.4 Idealized model for the progression of decay in timber elements with possibility of external treatments (Nguyen et al., 2008).

328

Long-term Performance and Durability of Masonry Structures Sapwood

Rate of decay

ds, sapwood thickness

For untreated sapwood Inner heartwood or core wood

Outer heartwood

D

Sapwood d

For treated sapwood

Pith

Outer heartwood

d/4 d/4

Core wood or inner heartwood

Distance from pith

Figure 11.5 Definitions of different wood parts in the log cross-section and their relative durability in terms of rate of decay (Nguyen et al., 2008).

11.6.2

Modeling the fungal decay risk

To predict the fungal decay risk and to estimate the performance loss for damaged timber structures, reliable durability models must be proposed through establishing empirical relationships between the incipient spreading of fungal decay and the moisture/ temperature-induced exposure dose (Brischke and Rapp, 2008; Brischke and Meyer-Veltrup, 2016). Based on experimental data, Brischke et al. (Brischke and Rapp, 2008) succeeded in defining the total daily dose inducing fungal decay inside wood as the product of two factors: the daily temperature-induced dose dT and the daily MC-induced dose dMC . As shown in Fig. 11.6, the daily development of fungal decay is optimal (d y 1) for temperatures between 20 and 40 C and wood MC between 35% and 70%. On the other hand, the growth of fungal decay may stop (d y 0) for temperature below 0 C and for wood MC levels below fiber saturation point (i.e., MC  26%e30%). 1,0

dMC dT

Dose

0,8 0,6 0,4 0,2 0,0

0

10

20

30

40 50 60 MC (%), Tav (ºC)

70

80

90

100

Figure 11.6 Relationships between the moisture content (MC) and daily moisture contentinduced dose dMC , and between the average wood temperature Tav and daily temperatureinduced dT for Pinus sylvestris wood species (Brischke and Rapp, 2008).

Service life design of timber structures

329

As illustrated in Fig. 11.7, Brischke et al. (Brischke and Rapp, 2008) established an empirical relationship between the total dose and the mean decay rating (Table 11.5) for small wooden specimens tested under different climatic exposure conditions and over different periods of time. The definition of total and daily doses significantly depended on the fungus and wood species, since they also condition decay growth as well as the hygrothermal behavior of wood. Other durability models against fungal decay directly compare the exposure daily dose (Brischke and Rapp, 2008) with the resistance dose against wood deterioration fungi (Meyer-Veltrup et al., 2017b), by taking into account the wetting ability of wood and its inherent natural durability. Through the mapping based on the Scheffer’s Climate Index and durability models, the risk of fungal decay can be estimated for different European countries, per their respective climate variability (Brischke et al., 2011). Over a 10-year period (1961e70), the growth of fungal decay and the related mass loss were evaluated through empirical durability models on several pine sapwood specimens, in different European test sites, protected or not from rain exposure (Viitanen et al., 2010). It was shown that the higher the mass loss, the higher the mechanical performances loss for timber structures as a result of damage due to fungal decay. Through mapping based on ERA-40 dataset system, the evaluation of mass loss was modeled in different European countries, per their respective climate exposure conditions.

11.7

Protection systems

Protection systems such as preventive actions or treatments have to be applied at the design stage of timber structures to preserve their long-term performance by reducing the risk of wood deteriorations in their service life. Before thinking about any

Mean decay rating (0–4)

4 3 2 1 y = 3.8392*EXP*(–EXP(1.6141–(0.0045*x))) 2 R = 0.9389

0

0

100

200

300

400

500 Dose

600

700

800

900

1000

Figure 11.7 Empirical relationship between the total dose and the mean decay rating from EN 252 (EN 252, 2012) for Pinus sylvestris wood species (Brischke and Rapp, 2008).

330

Long-term Performance and Durability of Masonry Structures

preventive system, the inspection of the building envelope and of timber structures must firstly be undertaken to establish cause-effect relationships for different structural problems or other performances loss noticed on site. To this end, a wide diagnosis of the building has to carried out (Teles and Do Valle, 2001; Freas, 1982) to determine the following inventory: • • • • • •

building history, including structural modifications and changes in use; classification, geometry, wood species, and use conditions of each timber element; environment of timber structures in service (e.g., temperature, MC, acidity, proximity to soil or contact with damp regions, etc.); previous intervention techniques related to the maintenance of timber structures (e.g., repair, replacement, reinforcement, protection measures, etc.); wood deterioration, structural damage, or defects taking place and their extent; areas of timber structures exposed to high moisture or to other severe environmental storage conditions.

Wood deterioration usually occurs in regions of timber structures under damp conditions, resulting in defects or performance loss within the building envelope, Therefore, protection systems mainly aim at decreasing the MC, at creating inspection accesses for easy maintenance over time, and at isolating wood material from its deterioration agents. From the nonexhaustive list, intervention measures may stand for (Cruz et al., 2015; Teles and Do Valle, 2001): • • • • • • • • •

ensuring the required performances of timber structures in service; using surface treatments (e.g., paint, stain, coats) that protect timber elements against moisture and wood-deteriorating agents; applying chemical treatment on timber elements in permanent contact with dampness to decrease the hygroscopic behavior and permeability of wood; reducing the risk of leaking water through periodic maintenance and repair of functional components within the building envelope; ensuring proper ventilation through air circulation spaces or openings in the building envelope to decrease wood MC; promoting passages to otherwise inaccessible places for the maintenance of building; improving contact areas between timber elements and the moist wall support through humidity barrier (e.g., metal sheets, bitumen paper, etc.); detecting/identifying wood deteriorations and stopping their development through monitoring exposure conditions of timber structures (e.g., MC); establishing repair and/or reinforcement strategies for damaged regions of timber elements and joints that threaten the stability of the whole structure.

In addition to the intervention measures stated, other protection techniques (not detailed in the present chapter) can be found in the literature (Jones and Brischke, 2017), for which prevention actions and treatments are more specific with respect to the different wood-degrading agents encountered on site: dry rot (Digest 299, 1993; Carey and Grant, 1999), insect attacks (Leary, 2002; Nunes, 2008; Nobre and Nunes, 2007; Hutton, 2008; Demaus, 1995; Digest 327, 1993), and fire (Hutton, 2012).

Service life design of timber structures

11.8

331

Maintenance and monitoring

11.8.1 Context and definitions In terms of biologic wood deterioration, treatments with chemicals against insect attacks and fungal decay may be expensive, inconvenient, operationally hazardous, environmentally unacceptable, and in some cases, unnecessary (Carey and Grant, 1999). Conversely, maintenance and monitoring can provide less destructive solutions by preserving the performance of timber structures over time in their service life. Although they have to be carried out over long time periods to provide meaningful data, maintenance and monitoring are preferred among different intervention measures to detect the regions at risk and prevent the development of biologic wood degradation, while using the durability models at disposal. Maintenance consists of continuously evaluating the timber structure health and overall building envelope over time, by seeking out the potential presence of defects, malfunctions, or wood deteriorations. Once identified, they have to be corrected through promoting suitable intervention techniques that will preserve or improve the long-term performance of timber structures. Furthermore, attention should be paid to zones of timber elements exposed to occasional or frequent moisture, which will realize a higher occurrence risk of wood biologic degradations. In that context, the climate variables and performances of timber structures will also need to be assessed over time through monitoring systems. The experimental methodology aims at designing more efficient and suitable assessment techniques on site such as non- and semidestructive tests, whereas the purpose of the numeric methodology is to evaluate theories and numeric techniques used to simulate the realistic behavior of wood under changing environmental conditions (Tannert et al., 2011). For example, the numeric tool proposed by Fortino et al. (2013) allows the monitoring of MC distribution in the cross-section of timber elements in terms of temperature and air relative humidity response.

11.8.2 Assessment techniques Assessment techniques such as non- and semidestructive tests enable the monitoring of long-term performance of timber structures, the climate variables, and the performance-influencing factors over time. In the following sections, two categories of monitoring are introduced as a result of overviewing several assessment techniques: MC monitoring and health assessment of timber structures.

11.8.2.1 Moisture content monitoring While ambient temperature and air relative humidity can be recorded with data loggers near the external surfaces of timber elements, the wood MC can be evaluated through two methodologies (Tannert et al., 2011): •

direct methodology: which gives direct MC values conforming to standard methods, through gravimetric (also called oven-drying or kiln-dry), distillation, and extraction methods;

332

•

Long-term Performance and Durability of Masonry Structures

indirect methodology: which infers MC values based on the assessment of different wood physical properties, through the capacitive, electrical resistance, microwave, radiometric, spectrometric, and color reaction methods.

Although the direct methodology provides accurate MC values, the different related methods are referred to as destructive tests since the extraction and segmentation of small samples from timber structures is required (Dietsch et al., 2015). Therefore, the direct methodology cannot be applied to monitor the wood MC in timber structures on site. Indirect methodology standing for non- or semidestructive testing is more suitable in that use. As low destructive tests featured by their easy use, the capacitive and electrical resistance methods have stood out from other indirect methodologies, since they have been the object of extensive research in these last 2 decades. For example, Brischke et al. (2008) developed an accurate and automated system for long-term MC measurements in weathered conditions through the electrical resistance method, by fixing insulated electrodes embedded with conductive glues inside timber elements. The indirect methodology can also be applied to monitor the fungal growth in moist regions of timber elements. On the other hand, the MC monitoring is not enough to control the insect infestation because added factors influencing their development inside wood have to be considered.

11.8.2.2 Health assessment of timber structures When evaluating the health of timber structures, visual assessment provides quick identification of different pathogens encountered on site, their state of progress with respect to environmental conditions, and their extent within timber elements. Visual assessment can be aided through several nondestructive techniques (Riggio et al., 2015; Kasal and Tannert, 2010): photogrammetry, IR thermography, radiography, stress wave measurement, microwave scanning, and penetrating radar. Every visual technique is characterized by its accuracy degree of data measurement, types of data recorded and interpretation, investigation domain scale, execution time, and feasibility. To guarantee a good interpretation of visual parameters, the adoption of multivariate and complementary analyses is always advisable. To estimate the extent of wood degradation in timber elements and their impact on the wood mechanical performances, several non- and semidestructive methods can be used (Sousa et al., 2014; Riggio et al., 2015): ultrasonic pulse velocity, impact penetration test, drilling resistance test, screw withdrawal test, and core drilling, to name several examples. In addition to these, destructive tests (e.g., hardness test, tension test, compression test, bending test) can be applied on small samples previously extracted from specific regions within existing timber structures to accurately determine the wood mechanical properties investigated. Since mechanical performances differ with wood species and aging, their identification is then crucial in existing timber structures, and it should be performed through dendrochronological dating.

Service life design of timber structures

11.9

333

Conclusions

Wood material can be subject to different categories of deterioration, with respect to suitable climate exposure conditions, resulting in threatening the long-term performances of timber structures in their service life. In that sense, the present chapter has gathered notions, guidelines, and current research that can help engineers and architects when designing the service life of timber structures. It provides basic knowledge about different wood-deteriorating agents, the natural durability of wood, and the use classes of timber structures. Furthermore, some durability models have been introduced for estimating the risk of wood deteriorations and the performance loss of timber structures in their service life. On the other hand, overall protection systems and other intervention measures have been proposed, while the maintenance and monitoring of timber structures through various assessment techniques have been highlighted. Despite this background knowledge, further work still has to be undertaken to propose reliable models to design properly the service life of timber structures and the occurrence risk of wood deteriorations, in relation to the climate exposure conditions. Since several international research groups are very well advanced in the field of fungal decay; future contributions should then promote further investigations on the onset and development of other wood deteriorations (e.g., cracks, insect attacks, etc.) with respect to the climate variables, wood MC, and other added parameters. Through on-site monitoring of damaged timber structures by using suitable and efficient assessment techniques over large time periods in different geographic sites, reliable durability models could be established that extend or complement the existing ones.

References Bravery, A.F., Berry, R.W., Carey, J.K., Cooper, D.E., 1993. Recognising Wood Rot and Insect Damage in Buildings. BRE e Building Research Establishment, Watford, UK, ISBN 0 85125 535 3. Brischke, C., Meyer-Veltrup, L., 2016. Modelling timber decay caused by brown rot fungi. Materials and Structures 49 (8), 3281e3291. Brischke, C., Rapp, A.O., 2008. Dose-response relationships between wood moisture content, wood temperature and fungal decay determined for 23 European field test sites. Wood Science and Technology 42, 507e518. Brischke, C., Rapp, A.O., Bayerbach, R., 2008. Measurement system for long-term recording of wood moisture content with internal conductively glued electrodes. Building and Environment 43 (10), 1566e1574. Brischke, C., Fr€uhwald Hansson, E., Kavurmaci, D., Thelandersson, S., 2011. Decay hazard mapping for Europe. In: IRG/WP 11-20463: The International Research Group on Wood Protection e Section 2: Test Methodology and Assessment. 42nd Annual Meeting, Queenstown, New Zealand. Carey, J., Grant, C., 1999. The Treatment of Dry Rot in Historic Buildings. The Building Conservation Directory. http://www.buildingconservation.com/articles/rot/rot.htm.

334

Long-term Performance and Durability of Masonry Structures

Cruz, H., Jones, D., Nunes, L., 2015. Chapter 12-wood. Materials for construction and civil engineering. In: Clara, G.M., Margarido, F. (Eds.), Science, Processing and Design. Springer, Switzerland, pp. 557e583. https://doi.org/10.1007/978-3-319-08236-3_12. Demaus, R., 1995. Precision Treatment of Death Watch Beetle Attack. The Conservation and Repair of Ecclesiastical Buildings. http://www.buildingconservation.com/articles/beetle/ beetle.html. Dietsch, P., Franke, S., Franke, B., Gamper, A., Winter, S., 2015. Methods to determine wood moisture content and their applicability in monitoring concepts. Journal of Civil Structural Health Monitoring 5 (2), 115e127. Digest 299, 1993. Dry Rot: Its Recognition and Control. BRE - Building Research Establishment e Concise Reviews of Building Technology, Watford, UK. ISBN 0 81525 348 2. Digest 307, 1992. Identifying Damage by Wood-boring Insects. BRE e Building Research Establishment e Concise Reviews of Building Technology, Watford, UK. ISBN 0 85125 531 0. Digest 327, 1993. Insecticidal Treatments against Wood-Boring Insects. BRE e Building Research Establishment e Concise Reviews of Building Technology, Watford, UK. ISBN 0 85125 278 8. Eaton, R.A., Hale, M.D.C., 1984. In: Wood e Decay, Pests and Protection. Chapman & Hall, London, UK. EN 1001-2, 2005. Durability of Wood and Wood-based Products. Terminology, Vocabulary. CEN, European Standardisation Institute, Brussels, Belgium. EN 1995-1-2, 2004. Eurocode 5. Design of Timber Structures. Part 1-2: GeneraldStructural Fire Design. CEN, European Standardisation Institute, Brussels, Belgium. EN 252, 2012. Field Test Method for Determining the Relative Protective Effectiveness of a Preservative in Ground Contact. CEN, European Standardisation Institute, Brussels, Belgium. EN 335, 2013. Durability of Wood and Wood-Based Products. Use Classes: Definitions, Application to Solid Wood and Wood-based Products. CEN, European Standardisation Institute, Brussels, Belgium. EN 350, 2016. Durability of Wood and Wood-Based Products. Testing and Classification of the Durability to Biological Agents of Wood and Wood-Based Materials. CEN, European Standardisation Institute, Brussels, Belgium. Foliente, G.C., Leicester, R.H., Wang, C.-H., Mackenzie, C., Cole, I., 2002. Durability design for wood construction. Forest Products Journal 52 (1), 10e19. Fortino, S., Genoese, Al., Genoese, A., Lina, N., Palma, P., 2013. Numerical modelling of the hygro-thermal response of timber bridges during their service life: a monitoring case-study. Construction and Building Materials 47, 1225e1234. Franke, B., Franke, S., Schiere, M., M€uller, A., 2016. Moisture diffusion in wood e experimental and numerical investigations. In: WCTE 2016 e World Conference on Timber Engineering. Vienna, Austria. Freas, A., 1982. Evaluation, Maintenance and Upgrading of Wood Structures e A Guide and Commentary. American Society of Civil Engineers (ASCE), New York, USA. Humar, M., Brischke, C., Meyer, L., Lesar, B., Thaler, N., Jones, D., Bardage, S., Belloncle, C., Van den Bulcke, J., Abascal, J.M., Alfredsen, G., Baisch, D., Brunnhuber, B., Cofta, G., Grodas, E., Fr€uhwald Hansson, E., Irle, M., Kallakas, H., Kers, J., Klamer, M., Larsson Brelid, P., Maider, A.B., Mahnert, K.-C., Melcher, E., Moeller, R., Noël, M., Nunes, L., Ormondroyd, G.A., Palanti, S., Pfabigan, N., Pilgard, A., Rapp, A.O., Schumacher, P., Suttie, E., Teppand, T., Touza, M., Van Acker, J., 2015. Introduction of the COST FP 1303 cooperative performance test e section 2: test methodology and assessment. In: Conference

Service life design of timber structures

335

paper for the 46th IRG/WP e International Research Group on Wood Protection, Vi~ na del Mar, Chile. Hutton, T., 2008. Woodworm e Anobium punctatum. The Building Conservation Directory. http://www.buildingconservation.com/articles/woodworm/woodworm.htm. Hutton, T., 2012. After the Fire. The Building Conservation Directory. http://www. buildingconservation.com/articles/fire-damage/fire-damage.htm. ISO 15686-1, 2000. Building and Construction Assets d Service Life Planning e Part 1: General Principles and Framework. International Organization for Standardization, Geneva, Switzerland. ISO 15686-7, 2017. Buildings and Constructed Assets e Service Life Planning e Part 7: Performance Evaluation for Feedback of Service Life Data from Practice. International Organization for Standardization, Geneva, Switzerland. Jones, D., Brischke, C. (Eds.), 2017. Performance of Biobased Building Materials. Woodhead Publishing, Duxford, UK. ISBN: 9780081009826. Kasal, B., Tannert, T. (Eds.), 2010. Situ Assessment of Structural Timber e RILEM State of the Art Reports, vol. 7. https://doi.org/10.1007/978-94-007-0560-9. Kranitz, K., Sonderegger, W., Bues, C.-T., Niemz, P., 2016. Effects of aging on wood: a literature review. Wood Science and Technology 50 (1), 7e22. Lamb, F., 1992. Splits and Cracks in Wood (Reno, United States, s.n). Leary, P., 2002. The Eradication of Insect Pests in Buildings. The Building Conservation Directory. http://www.buildingconservation.com/articles/eradication/eradication.htm. Leicester, R.H.. Engineered durability for timber construction. John Wiley & Sons, Ltd. Volume 3, Pages 216-227. CSIRO, Highett, Australia Progress in Structural Engineering and Materials MacKenzie, C.E., Wang, C., Leicester, R.H., Foliente, G.C., Nguyen, M.N., 2013. #05 Timber Service Life Design, Revised Version e Design Guide for Durability e Technical Design Guide. Wood Solutions. Forest and Wood Products Australia, CSIRO, Australia. Mergny, E., Mateo, R., Esteban, M., Descamps, T., Latteur, P., 2016. Influence of cracks on the stiffness of timber structural elements. In: WCTE 2016 e World Conference on Timber Engineering. Vienna, Austria. Meyer-Veltrup, L., Brischke, C., Alfredsen, G., Humar, M., Flaete, P.-O., Isaksson, T., Brelid, P.L., Westin, M., Jermer, J., 2017a. The combined effect of wetting ability and durability on outdoor performance of wood: development and verification of a new prediction approach. Wood Science and Technology 51 (3), 615e637. https://doi.org/ 10.1007/s00226-017-0893-x. Meyer-Veltrup, L., Brischke, C., K€allander, B., 2017b. Testing the durability of timber above ground: evaluation of different test methods. European Journal of Wood and Wood Products 75 (3), 291e304. Milton, F.T., 1986. The Preservation of Wood e A Self Study Manual for Wood Treaters. University of Minnesota, College of Natural Resources, Minnesota, USA. Nguyen, M.N., Leicester, R.H., Wang, C., Foliente, G.C., 2008. A Proposal for AS1720.5 e Timber Service Life Design Code e Durability of Structural Timber Members. Forest and Wood Products Australia, CSIRO, Australia. Nilsson, T., Daniel, G., 1990. Structure and the aging process of dry archaeological wood. In: Rowell, R.M., Barbour, R.J. (Eds.), Archaeological Wood: Properties, Chemistry, and Preservation. Advances in Chemistry Series 225. American Chemical Society, Washington, USA, pp. 67e86. Nobre, T., Nunes, L., 2007. Non-traditional approaches to subterranean termite control in buildings. Wood Material Science and Engineering 2 (3e4), 147e156.

336

Long-term Performance and Durability of Masonry Structures

Nunes, L., 2008. Termite Infestation Risk in Portuguese Historic Buildings. Cost Action IE0601-Wood Science for Conservation of Cultural Heritage, Braga, Portugal. https:// www.researchgate.net/publication/267778562_Termite_infestation_risk_in_Portuguese_ historic_buildings. Papp, G., Barta, E., Preklet, E., Tolvaj, L., Berkesi, O., Nagy, T., Szatmari, S., 2005. Changes in DRIFT spectra of wood irradiated by UV laser as a function of energy. Journal of Photochemistry and Photobiology A 173, 137e142. Pasanen, A.-L., Kasanen, J.-P., Rautiala, S., Ik€aheimo, M., Rantam€aki, J., K€a€ari€ainen, H., Kalliokoski, P., 2000. Fungal growth and survival in building materials under fluctuating moisture and temperature conditions. International Biodeterioration and Biodegradation 46 (2), 117e217. Riggio, M., Sandak, J., Franke, S., 2015. Application of imaging techniques for detection of defects, damage and decay in timber structures on-site. Construction and Building Materials 101 (Part 2), 1241e1252. Sandberg, D., S€oderstr€om, O., 2006. Crack formation due to weathering of radial and tangential sections of pine and spruce. Wood Material Science and Engineering 1 (1), 12e20. https:// doi.org/10.1080/17480270600644407. Shupe, T., Lebow, S., Ring, D., 2008. Causes and Control of Wood Decay, Degradation and Stain. LSU AgCenter e Research & Extension. Louisiana State University Agriculture Center, Baton Rouge, LA. Publication 2703. 27 Pages. https://www.fpl.fs.fed.us/documnts/ pdf2008/fpl_2008_shupe001.pdf. Sousa, H.S., Branco, J.M., Lourenço, P.B., 2014. Characterization of cross-sections from old chestnut beams weakened by decay. International Journal of Architectural Heritage e Conservation, Analysis, and Restoration 8 (3), 436e451. Svensson, S., Turk, G., Hozjan, T., 2011. Predicting moisture state of timber members in a continuously varying climate. Engineering Structures 33 (11), 3064e3070. Tannert, T., Berger, R., Mueller, A., Vogel, M., 2011. Remote moisture monitoring of timber bridges: a case study. In: SHMII-5: Proceedings of the 5th International Conference on Structural Health Monitoring of Intelligent Infrastructure. Canc un, México. Teles, C.D.M., Do Valle, A., 2001. In: Lourenço, P.B., Roca, P. (Eds.), Wood Structures: Acting before Deterioration. Historical Constructions, Guimar~aes, Portugal, pp. 857e866. van de Kuilen, J.-W.G., 2007. Service life modelling of timber structures. Materials and Structures 40 (1), 151e161. https://doi.org/10.1617/s11527-006-9158-0. Viitanen, H., Ritschkoff, A.-C., Paajanen, L., 2002. The durability of different wood products. In: European Timber Buildings as an Expression of Technological and Technical Cultures, pp. 173e186. Viitanen, H., Toratti, T., Makkonen, L., Peuhkuri, R., Ojanen, T., Ruokolainen, L., R€ais€anen, J., 2010. Towards modelling of decay risk of wooden materials. European Journal of Wood and Wood Products 68 (3), 303e313.

Service life design of stone masonry structures

12

Elsa Garavaglia 1 , Giuliana Cardani 1 , Anna Anzani 2 1 Department of Civil and Environmental Engineering, Politecnico di Milano, Milano, Italy; 2 Department of Design, Politecnico di Milano, Milano, Italy

Chapter Outline 12.1 12.2 12.3 12.4

Introduction 337 The lifetime prediction of decayed stone masonry 338 Fragility curve for service life prediction 342 Laboratory aging tests on treated stone masonry 346 12.4.1 Monitoring and damage quantification 347 12.4.2 Results of crystallization test on masonry wallettes 349 12.4.3 The probabilistic approach applied to laboratory salts crystallization tests 351 12.4.3.1 Stochastic modeling of a deterioration process 352 12.4.3.2 The choice of the significant damage threshold 353

12.5

On-site study of surface decay of treated stone masonry 353 12.5.1 Decay observations and results 355 12.5.2 The probabilistic approach 358 12.5.3 Fragility curves proposed for in situ damage modeling 359

12.6 Conclusions 362 Acknowledgments 363 References 363 Further reading 365

12.1

Introduction

During their service life, masonry structures can be subjected to decay due to environmental attacks, aging, and/or damage due to long-term heavy loads, as well as the application of incompatible materials used to repair and protect their surface. One of the most frequent causes of damage to masonry walls in many environments is salts crystallization behind the surface that induces exfoliation, delamination, and crumbling of the masonry components. The research on the causes and remedies of masonry surfaces decay has studied very deeply the microstructure and its influence on the durability of stones. Remedies were set up so that substitution of the original materials could be avoided and the stone Long-term Performance and Durability of Masonry Structures. https://doi.org/10.1016/B978-0-08-102110-1.00012-1 Copyright © 2019 Elsevier Ltd. All rights reserved.

338

Long-term Performance and Durability of Masonry Structures

surfaces could be consolidated and protected against further deterioration. Nevertheless, the attention was mostly dedicated to single materials, either stones or bricks, rather than to masonry as a whole. Furthermore, lower attention was given to the influence of the construction technique on the durability of the treated stone surfaces. The following questions still look relevant: • • • •

How deep does decay penetrate into the wall from the external surface? What is the mutual influence of mortar and stones in the decay processes of stone masonry? What is the influence of the surface decay on the structural behavior of masonry walls? How successful can a water-repellent and/or a strengthening product be?

Research on the service life of building materials requires in general developing and/or enhancing methods to generate data on the lifetime of materials and buildings (Masters and Brandt, 1987). In particular, the following aspects need to be developed: collect reliable data on the effective on-site behavior of the masonry structure, increase the knowledge on the mechanisms of failure, develop methods and procedures for the decay measurement, improve the knowledge on the aggressive agents and on their quantity and influence, implement methods to simulate or take into account the synergetic effects of aggressive agents, and implement mathematical models to describe the material behavior in specific aggressive environments. The great randomness associated in each decay phenomena suggests studying these processes from a probabilistic point of view. The deterioration of masonry materials has therefore been investigated through a simple probabilistic approach aimed to the prediction of the time needed to reach a given level of damage, which is an important issue in planning strategies for the maintenance and repair of existing buildings. The model has been applied to interpret sets of experimental data collected on samples of masonry (wallettes) subjected to salt crystallization and tested in the laboratory through accelerated aging tests. The problem of lifetime prediction is described in Section 12.2; the probabilistic approach in its theoretical form is presented in Section 12.3; laboratory aging tests and the application of the model to material deterioration are presented in Section 12.4; finally, in Section 12.5 the first application of this procedure to a full-scale model is proposed.

12.2

The lifetime prediction of decayed stone masonry

The durability of stones depends both on their intrinsic properties (their petrographic composition and internal structure) and on the environment to which they are exposed. In this sense, stones used as construction materials are subjected to the influence of a greater number of factors than those naturally acting on the rocks in a quarry, starting from the effects of extraction and working up to the action of pollution and weathering. They in turn depend on the kind of environmental placement of the material (city, rural, marine, etc.). In particular, the action of water is particularly dangerous and is always involved in the deterioration processes of stones: erosion is due to a physical

Service life design of stone masonry structures

339

action of rain and hail; freezing and thawing produce a mechanical damage; in the presence of water chemical reactions between the mineral components of stones and air pollutants can also take place. The following major categories of stones deterioration can be highlighted: effects of finishing and laying process, salt crystallization, freeze and thaw cycles, thermal shocks, biologic alteration, and air pollution in particular combined with the action of meteorologic phenomena (rain, fog, wind). Salts can be present in masonry for several reasons. They can be contained in the original materials, come from the soil, or be deposited on the masonry surface. If water is present, they can be easily solubilized and can migrate within the masonry. A cyclic process may take place during which an evaporation phase follows, and the salts are transported toward the external material surface and can crystallize. If new water is fed into the masonry, a new salts solubilization is produced; then a new movement toward the surface takes place and crystallization is caused again. This phenomenon due to fatigue effects can continue for many thermal-hygrometric cycles, causing material deterioration (Binda et al., 1985). As it is explained in (Binda and Baronio, 1987), the related decay is a continuous material delamination from the external surface caused by wet-dry cycles, which is especially harmful in the presence of soluble sulphates inside the material. The thickness of the laminated layer strongly depends on the microenvironmental conditions (temperature and R.H.), the porosity and the mechanical strength of the material (Lewin, 1981), and the duration of wet-dry cycles (Binda and Baronio, 1987). As far as the penetration of the decay is concerned, considering several samples of stones cored on site, in most cases the decay penetrates a few millimeters below the badly damaged external layer (Binda et al., 1992). The decay in a wall can therefore be considered a loss of material from the exterior toward the interior, actually a one-side decay (see Section 12.2). The case of a wall is certainly different from the case of a slender element, which is attacked by the decay from all sides: in this case, the decay penetration into the core of the element through continuous delamination can be very quick. However, whatever the thickness of the layer is, the material properties are modified only in a narrow millimetric portion underneath the detached layer. This fact has been observed several times and can be explained according to the mechanism of decay (Larsen and Nielsen, 1990; Binda et al., 1992). Fig. 12.1(a,b) shows the decay penetration in the case of two specimens, respectively a soft mud fired brick and a sandstone. The two specimens were cut normally to the deteriorated surface after being subjected to a severe crystallization test, which had badly destroyed the outer part of the specimens; the depth of the decay is no more than 1e1.5 mm, the material below this layer being practically undamaged. The durability assessment, in view of designing suitable protection interventions, requires carrying out reliable and replicable aging testing procedures, which are representative of the real environment in which the service life needs to be evaluated. Since stone masonry is a composite material made by stone/brick and mortar, its decay is highly influenced not only by the environmental conditions but also by the combination of its components (Fig. 12.2(a) and (b)). In fact, the durability of stonework masonry toward salt crystallization decay is largely influenced by the mortar, according to the porosity characteristics of the two materials. Therefore, when the prediction of

340

Long-term Performance and Durability of Masonry Structures

Figure 12.1 Cut portion of masonry units badly damaged due to salt crystallization test (top surface): (a) soft mud brick and (b) sandstone.

masonry durability is required, long-term and hence costly durability tests on single masonry components, brick, stone, or mortar are of very limited utility (Binda and Baronio, 1985; Binda et al., 1985; Van Hees et al., 1996) because of their continuous synergic interaction according to their chemical-physical-mechanical properties. Of course, the characteristics of the macro- and microenvironment should be known, and the presence and movements of water inside the masonry should be detected. The types of damage occurring to the masonry are well described in (Franke et al., 1998). A systematic approach dealing with masonry as a composite material was set up by Binda and Baronio (1985, 1987) and Binda et al. (1985) after a rather long research experience on the masonry components. In the Mediterranean countries, even if

Service life design of stone masonry structures

341

Figure 12.2 Exfoliation and erosion of stone masonry walls due to environmental actions and salt crystallization: (a) sandstone masonry, (b) limestone masonry.

frost-defrost action is still an important factor of decay, the most influencing factor on masonry deterioration is with no doubt salt crystallization. Efflorescence and cryptoefflorescence are nearly always present on the decayed or decaying masonry, distributed according to the combination of their masonry components and technique of construction (one-leaf or multiple-leaf masonry, brick or stone masonry, with or without mortar, rendered masonry, etc.). It is very important to consider that when the delamination/crumbling reaches a certain threshold [, the characteristics of a porous material and hence the reliability of the system could be seriously impaired. When is this threshold reached? And how can damage be quantified? The problem of reproducing realistic and complex environmental conditions needs to be balanced with the need of isolating single causes of damage (capillary rise, salt crystallization, etc.) and controlling their boundary conditions. In addition, the

342

Long-term Performance and Durability of Masonry Structures

definition of significant parameters to be measured during the test should be dealt with, to record the evolution of damage in time. The problem of quantifying the surface damage with nondestructive testing and of monitoring its evolution in time without excessively long tests is an open question. The objective of this research has been to develop a procedure that can predict, in probabilistic terms, the evolution of surface decay in time, without a large set of experimental data. In fact, considering that the great complexity of the environmental processes causing material deterioration cannot be completely reproduced in the laboratory, the adoption of appropriate probabilistic modeling for a better interpretation of the phenomena seems to be a good approach (Binda and Molina, 1990b; Cranmer and Richerson, 1998; Bekker, 1999).

12.3

Fragility curve for service life prediction

In view of the preservation of the cultural heritage toward physical and mechanical damage due to different causes, the probability that a system will reach, or exceed, a given damage threshold over time often requires to be evaluated. According to the specific damage process that in turn is analyzed, this threshold may correspond, for instance, to the thickness of the lost surface material due to salt decay (see Section 12.4 and 12.6) or other significant parameters. Therefore, a parameter [ defined as the loss of performance reached by the system at the time t* has to be chosen to describe the deterioration process of porous materials. The high randomness associated with the occurrence of the decay of a structural system in the natural environment justifies the consideration of [ as a random variable (r.v.) with a certain probability distribution (Binda et al. 1999b). Since the deterioration depends on the instant t* in which it is recorded, at each time t* the loss [ can be modeled with a probability density function (p.d.f.) f[ (L) that depends not only on the r.v. [, but also on the constant parameter t* (Fig. 12.3(a)). From this point of view, the deterioration process can be considered a stochastic process of the r.v. [. To model f[ (L), a theoretical distribution needs to be chosen at every time t*. Obviously, the distribution, modeling a given phenomenon, must be chosen on the basis of the physical aspects of the phenomenon itself and of the characteristics of the distribution function in its “tail”, where experimental results are scarce. The most suitable distribution function can be chosen analyzing at every time t* the behavior of the failure rate function f[ ðLÞ connected to the distribution function itself (details are given in Garavaglia et al., 2004): f[ ðLÞdL ¼ PrfL < [  L þ dLj[ > Lg ¼

f[ ðLÞ 1  F[ ðLÞ

c t

(12.1)

where, as said earlier, Pr denotes the conditioned probability of failure, f[ (L) is the probability density function, and F[ (L) the cumulative distribution of the r.v.

Service life design of stone masonry structures

343

(a) L (loss in %)

Noto- stone consolidant

5.0 4.5

Salt Na2So4

4.0

Salt concentration 2.5%

3.5 3.0 2.5

– l

2.0 1.5 1.0

fl(L)

0.5 0.0

0

1

2

3

4

5

6

7

8

9

t (months)

t*

(b) F–T(t)

1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0

0

1

2

3

4

5

6

7

8

9

10 11 12 13 14 15

t (months) Figure 12.3 Probability to cross the threshold [: (a) lognormal p.d.f. modeling damage at each t*; (b) experimental (⬥) and Weibull theoretical (d) fragility curves.

The behavior of f[ ðLÞ defines the immediate failure rate dL; if the event has not yet happened at [ ¼ L, then three possibilities are given: •

A constant f[ ðLÞ means that the probability of an immediate occurrence of the next failure dL is always the same and is not affected by the reached level of loss L; this is the typical

344

•

•

Long-term Performance and Durability of Masonry Structures

Poisson process and can describe the loss of performance due to the aging when no other factors play any particular role. A distribution characterized by this f[ ðLÞ is the exponential. A decreasing f[ ðLÞ means that the probability of an immediate occurrence of the next failure dL decreases when the reached level of loss L increases; this seems to happen in the delamination phenomenon where the thickness of the detached layers is always lower than characteristic values. A distribution characterized by this f[ ðLÞ is the lognormal distribution. An increasing f[ ðLÞ means that if the material has already reached a given loss of performance L and the loss [ > L has not yet happened, the probability to have a failure in the next dL increases if the reached level of loss L increases. At L tending to infinite, f[ ðLÞ can increase either tending to infinite or tending to an asymptotic value; both situations are typical of the extreme value distributions. By experimental evidence, the strain rate versus stress phenomenon (see Section 12.4) seems to obey to this model; in particular, it seems to tend to infinite when the stress increases above a certain level.

According to Eq. (12.1) and to the studied physical problems, a lognormal distribution was adopted in the case of salt crystallization process. Considering a significant damage [ and the variable time needed to exceed it, a physical or mechanical deterioration process can be treated as a reliability problem (Garavaglia et al., 2002b). The loss of the surface material experimentally recorded cycle by cycle is initially modeled through a probabilistic approach. The physical model suggests that the layers of surface material lost for salt crystallization can be classified according to their thickness into a given interval of size. The immediate detachment of layers with thickness larger than that usually recorded is an extremely rare event, and again the larger the detachment expected the smaller the probability that it could happen in the analyzed cycle. On the basis of the experimental evidence, the probability density function (p.d.f.) chosen to model the material loss is the lognormal p.d.f. having the following mathematical form: " # 1 ðln L  mÞ2 f[ ðLÞ ¼ pffiffiffiffiffiffi exp  2s2 L 2ps

(12.2)

where m and s are the mean and the standard deviation of distribution. The immediate failure rate function, f[ ðLÞ, of lognormal distribution is a function that decreases when the loss value L increases, so it seems good to interpret the behavior of the loss at each cycle. The next step is then the modeling of the deterioration process as a reliability process. Reliability R(t) is related to the performance of a system over time, and it is defined as the probability that the system does not fail by time t (Evans, 1992). This definition is extended here, denoting by RðtÞ the probability that a system will exceed a given significant damage threshold [ by time t. The random variable used to quantify

Service life design of stone masonry structures

345

reliability is T, which is simply the time taken to exceed the damage [. Thus, from this point of view, the reliability function is given by the following:   RðtÞ ¼ Pr T > t ¼ 1  FT ðtÞ

(12.3)

where FT ðtÞ is the distribution function for T. Assuming that the density function fT ðtÞ exists for the r.v. T, the failure rate function fT ðtÞ is given by (Evans, 1992): fT ðtÞ ¼

fT ðtÞ RðtÞ

(12.4)

Computing FT ðtÞ for different damage levels [ allows us to build the fragility curve for each [. A fragility curve describes the probability of reaching or exceeding a given damage [ over time (Singhal and Kerimidjian, 1996). At a given time t*, the probability that a particular damage level [ will be reached can be seen as the area below the threshold [, and the probability that it is exceeded can be seen as the area above the threshold [ (Fig. 12.3(a)) (Garavaglia et al., 2002b, 2004). The latter area can be calculated by using the survive function: J[ ðLÞ ¼ 1  F[ ðLÞ

c t

(12.5)

where F[ (L) is the cumulative distribution function of the probability density function p.d.f. chosen to model the parameter [ at every t*. The computation of J[ ðLÞ at every t* is performed by means of numeric integration. For each chosen damage threshold [, the areas computed provide the experimental fragility curves FT  ðtÞ at different t* (Fig. 12.3(b)). To model numerically the experimental fragility curves, a function has also to be chosen that should provide a good interpretation of the physical phenomenon: in all the damage cases studied here, a Weibull distribution has been chosen (Cranmer and Richerson, 1998, Bekker, 1999; Garavaglia et al., 2002b, 2004): FT ðtÞ ¼ 1  exp½ðrtÞa 

(12.6)  G

1 1þ a



with a > 0 indicating the shape parameter of the distribution, r ¼ > 0 where m m is the mean value of the distribution, and G is the Eulero’s Gamma function. Of course, the choice has been based on a physical interpretation of the decay process induced in the porous building materials by environmental aggressive agents over time. The decay process can be interpreted like a dynamic process that, being the material under attack, constantly evolves describing how its performance be compromised. If some significant damage levels [ are defined, the decay process shows how the material reaches and exceeds each of them over time. If the cause of damage endures, the damage advances; therefore, if a given damage level [ has not yet been reached at

346

Long-term Performance and Durability of Masonry Structures

the time t, the probability that it can happen in the next immediate interval dt increases; so, the probability increases if t tends to infinite. The behavior here described seems to be well modeled by a distribution having an increasing hazard rate fT ðtÞ. Weibull distributions present exactly this type of hazard rate.

12.4

Laboratory aging tests on treated stone masonry

Durability tests are generally aimed to reproducing site situations and evaluating the resistance of the materials to deterioration, to design damage prevention and quality control. Applying surface treatments is an easy and common, but costly, practice, which is thought to avoid or reduce masonry damage and hence delay the ordinary maintenance. Research carried out in the past (Binda et al., 1992) concluded that waterproof and strengthening surface treatments of porous material in the presence of soluble salts induce greater detachment of thicker material layers (the whole treated parts) due to the possible formation of cryptoefflorescence than in the case of untreated material. Within an EC contract (De Witte, 2001) an attempt was made to individuate, in the presence of water, the maximum allowable salt content in stone masonry, below which the surface protection treatments do not fail. Crystallization tests according to RILEM TC 127 MS recommendation were carried out at the Material Tests Laboratory of Politecnico di Milano on treated and untreated stone masonry specimens (Cardani et al., 2002) (Fig. 12.4). A large number of tests had previously been carried out on the single units used for the masonry specimens (Garavaglia et al. 2002a). According to the aims of the EC contract, the salt concentration compatible with surface treatments depends on the capillary moisture content, which is calculated on the basis of the water absorbed by capillary rise in 48 h. Therefore, since the goal of the experimental Lost area (li) Foam Cross section

Wallette Gravel

Plan view

Figure 12.4 Box scheme for the crystallization tests, following RILEM TC 127 MS recommendation.

Service life design of stone masonry structures

347

campaign was to define a limit salt content, a saturated salt solution as indicated by the RILEM recommendation was not used. Salt solutions with two low concentrations of sodium sulphate (1% and 2.5% of the measured capillary moisture content CapMC, referred to the percent weight of the dry specimens) were introduced in masonry wallettes treated with a water-based water repellent (solventless silicone microemulsion concentrate based on silanes and siloxanes) and with a strengthening product (reactive silicic acid ethyl ester compounds). The aging test was carried out in environmental conditions of 20 C and 50% R.H. on masonry wallettes (250  200  120 mm) built with natural stones: Serena sandstone (widely used in central Italy for architectural heritage) and Noto limestone from Italy and Savonni
ere limestone from France. The wallettes were all prepared with bedding joints, 15 mm thick, made with a putty lime-based mortar and a calcareous sand (Cardani et al., 2002). The wallettes were treated by immersing their upper surface respectively for 10 s into the water repellent and for 30 s into the consolidant. One wallette of each material was left blank to be used as a reference (Table 12.1).

12.4.1 Monitoring and damage quantification The chosen testing procedure is useful to study the performance of stone/mortar systems in the laboratory, with and without surface treatments. To gain significant results from crystallization tests, the monitoring time needs to be sufficiently long (6 months or more), and the number of cycles, corresponding to a new water supply after complete evaporation, must be not less than three. Salt crystallization produces high stresses inside the material and results in a continuous crumbling and delamination of the external surface of the wall, the material underneath remaining unaltered. The loss of material as a function of time is a good parameter that can be adopted to measure damage in the case of salt decay of porous surfaces (Binda et al., 1992). Therefore, the variation of subsequent surface profiles has been assumed as a measurement parameter of damage, and a laser profilometer was used for its monitoring in time (Fig. 12.5), as described in (Binda et al., 1999b; Berra et al., 1993; Baronio et al., 1993; Cardani et al., 2002). Table 12.1 Physical mechanical characteristics of the stones subjected to the testing procedure Bulk density (kg/m3)

Porosity (% vol.)

Capillary coeff. (g/cm2 h0,5)

CapMC (%)

st dry (N/mm2)

st wet (N/mm2)

Noto limestone

1856

29

0.75

12.87

2.82

1.32

Savonni
ere limestone

1682

18.8

0.34

9.47

1.35

0.95

Serena sandstone

2600

5.2

0.25

0.57

6.83

3.54

Sample

348

Long-term Performance and Durability of Masonry Structures

Figure 12.5 Laser profilometer device during measurements.

The device allows one to draw plots of the surface profile by interpolation of their readings in chosen positions k. Subsequent measurements show how the profile changes due to any superficial decay and allows one to measure the material loss in time. Fig. 12.6(a) shows four example surface profiles obtained in four different measurements. Since bulging (or swelling) occurs before delamination, it can be considered the onset of damage. The experimental measurements need to be converted into new deterioration diagrams where bulging has been eliminated; this allows one to quantify correctly the swelled material as a “loss.” The procedure compares the vertical coordinates of two subsequent diagrams: the current plot n and the previous plot (n  1). The coordinates of the diagram n are usually lower than those of the diagram (n  1), except for the points affected by swelling. In these points the computer code continues the procedure, comparing the coordinates of plot n with those of the subsequent plots (n þ 1), (n þ 2), (n þ 3) . until a plot m ¼ (n þ i) having lower vertical values than that of profile n is found. The coordinates of diagram m, corresponding to the points affected by bulging, become the new reference coordinates of profile n when redrawn.

(a)

(b)

mm

mm –250 –230 –210 –190 –170 –150 –130 –110 –90 –70 –50 –30 –10

–250 –230 –210 –190 –170 –150 –130 –110 –90 –70 –50 –30 –10

5 4

5 4 3

0 months

3

0 months

1 months

2

1 months

2

1

2 months

1

0

6 months

2 months 6 months

Bulging

0

–1

–1

–2

–2

–3 mm

–3 mm

–4

–4 –5

–5

i th profile Softmud brick untreated Na2So4 2.5%

–6 –7 –8 –9 –10

loss li

i th profile

Softmud brick untreated Na2So4 2.5%

–6 –7 –8 –9 –10

Figure 12.6 Example of deterioration measurements over time on a unit of masonry specimen: (a) raw readings and (b) after the swelling has been subtracted from raw readings.

Service life design of stone masonry structures

349

As a result, a clean plot of the evolution of the surface damage as a function of time and space is obtained (Fig. 12.6(b)) (Garavaglia et al. 2002a). The loss of material can be quantified using the new diagrams of Fig. 12.6(b). Here, for each profile i, the loss [i of the cross-sectional area (in mm2) of the wallette, calculated at every measurement time t* (t* ¼ 0, 1, 2, 6 months), has been assumed as the damage parameter. To quantify [i, the area included between two consecutive diagrams is automatically calculated at every time t* by the computer code studied to eliminate bulging by means of the composite trapezoidal rule (Garavaglia et al., 2002a).

12.4.2 Results of crystallization test on masonry wallettes As far as the specimens built with Noto limestone are concerned, according to the results achieved with the crystallization test on single masonry stone units (Garavaglia et al. 2001, 2002a), damage was always observed for the highest salt concentrations adopted (5% or 7.5% of the CapMC), in many cases for the lower concentrations (2.5% or 5% of CapMC), never for the minimum (1% of CapMC), the latter being consequently considered a threshold value. In the case of untreated wallettes, efflorescence started from the stone/mortar interface, the damage being very poor for the lowest salt concentration (1%) in the first 3 months (Fig. 12.7(a)). With the highest concentration (2.5%) the efflorescence was more evident in the stone/mortar interface, and a greater powdering of the stone took place; the stone damage was serious from the beginning, showing spalling of layers of about 1.4 mm. In the case of specimens treated with water repellent, at the lowest salt concentration (1%) the damage started slowly, and no particular surface decay was visible on wallettes after the first 3 months. After 6 months the damage became serious, showing exfoliation and spalling of stones (Fig. 12.7(b)). Mortar joints showed damage with both concentrations.

(a)

(b)

Untreated

(c)

Water repellent

Consolidant

Figure 12.7 Noto stone wallettes with Na2SO4 after the aging test: (a) in the untreated specimen, (b) in the presence of water repellent, and (c) in the presence of consolidant.

350

Long-term Performance and Durability of Masonry Structures

In the case of specimens treated with consolidant treatment, in a period of 1 month the stones, also with the lowest concentration of salt, presented detachment of a thin layer (about 0.65 mm) corresponding to a level of damage of about 0.8% (Fig. 12.7(c)). After detachment of the layers the stones started powdering uniformly. The mortar showed damage after 3 months but only with the highest concentration. Measuring the detached layers thickness and the stone wettability of specimens treated both with consolidant and with water repellent, the depth reached by the surface treatment seems to be respectively around 0.65 mm in the case of the consolidant, and of 1.4 mm (corresponding to a damage level of 1.2%) in the case of the water repellent. The difference in penetration depth could explain the delay in the exfoliation and the more irregular damage observed in the case of water repellent treated surfaces with the lowest salt concentration. Since the material loss in untreated wallettes after the same period of time (8 months) is less or equal to that of the treated ones (Fig. 12.7(b)), treatments on Noto limestone wallettes seem to be unnecessary to prevent salt crystallization decay. In Fig. 12.8 the presence of salts below the water-based water repellent (WBWR) treated layer was evident in all the treated wallettes, also where the surface damage was still not visible, as for the Savonni
ere stone wallettes (Fig. 12.8(a) and (b)). The damage in the Serena sandstone wallettes was localized at the interface between mortar and stone (Fig. 12.8(c)). The data coming from the wallettes tests could be reliably used for the elaboration of the probabilistic model; in fact, for each wallette, there were at least four units to be measured and therefore a sufficient number of results. The elaboration of the p.d.f. here described applies to the units combined with mortar tested with a Na2SO4 salt solution. In the case of the Noto stones, a visible damage was observed also at the lowest salt concentration (1%). The experimental data elaboration, as in Fig. 12.9, is reported in (Cardani et al., 2002; Garavaglia et al., 2002a). Due to the low damage of the Serena and Savonni
ere stone wallettes, only the data on Noto stone wallettes are here presented as an example of the used approach.

Figure 12.8 Presence of salts below the WBWR treated layer: (a) Noto limestone, (b) Savonni
ere limestone, and (c) Serena sandstone.

Service life design of stone masonry structures

351

(a)

(b)

L (loss in %)

1.0

Noto-stone consolidant Salt: Na2SO4

5.0 4.5 4.0

FT–(t)

0.9 0.8

Salt concentration 2.5%

0.7

3.5

0.6

3.0

l

2.5

0.5 0.4

1.5

0.3

1.0

0.2

L(t*,l)

2.0

0.5 0.0 0

1

2

– l = 1.2 % – l = 1.6 % – l = 2.0 % – l = 2.4 %

0.1

t* 3

5 4 t (months)

6

0.0

7

8

9

0

2 4

6

8 10 12 14 16 18 20 22 24 t (months)

Figure 12.9 (a) Probability to cross a threshold [: interpolation of the loss diagrams (*) and LogN p.d.f; (b) Experimental (•, :, ♦, *) and theoretical (Weibull) fragility curves (d).

12.4.3 The probabilistic approach applied to laboratory salts crystallization tests In each statistical analysis an important issue is the statistical sampling significance. A too small number of samples could compromise the statistical significance of the results obtained. Since the experimental campaigns are often expensive, and consequently limited in time and number of samples analyzed, their results must usually be considered symptoms of a possible behavior, but they cannot be assumed as the statistical truth. To bypass this problem, some numeric simulation techniques, like Monte Carlo simulation technique, could be adopted; however, the reliability of such methods strongly depends on the goodness of the available information: the more the characteristics of the phenomenon are known, the better the simulation results will be. The number of samples cannot be less than four to six, and the time interval must be long enough. A too-short testing time gives information that can be modeled only by the lower distribution tail. As a consequence, the distribution parameters are evaluated on the base of these data; therefore the fitting can suffer from unreliability. Of course, this depends also on the investigated damage. If the damage is progressive and increases with a “monotonic” behavior and if the damage level considered is small ([ < 0.2%e0.4%), the time to reach it is short, so the time test during which 4e5 measurements are done can be short (4e5 months). If the damage investigated is serious ([ > 2.0%e3.0%) the time test must be longer (10e12 months). In this case, it will be possible to make prevision of the damage evolution for a long period of time (more than 30 months). In each analysis here proposed, the fitting of the experimental fragility curve is satisfactory, though a small number of samples was used. Anyway, to interpret these results, great caution is needed.

352

Long-term Performance and Durability of Masonry Structures

12.4.3.1 Stochastic modeling of a deterioration process The parameter [, describing the surface deterioration process of porous materials, can be defined as the loss of performance reached by the system at the time t*. To compare all the experimental results obtained during the period, the damage has been plotted in percentage: [i ðkÞ ¼

Ai ðkÞ AT

(12.7)

where Ai(k) (lost area) is the area included between two consequent profiles in position k (Fig. 12.6(b), dashed area) recorded at the ti, and AT is the transversal section area. A simple linear interpolation of the experimental data provided a quite readable trend of the behavior of the loss [ðkÞ over time, where [ðkÞ is assumed to be the sum of the area lost till the instant t* by profile in position k: [ ðkÞ ¼

n X

[i ðkÞ

(12.8)

i¼1

In Fig. 12.9(a), where the readings at 8 months have been reported in addition to the data of Fig. 12.6, a simple interpolation of the experimental points permits one to better read the loss trend [ of every profile over time (linear slices). As mentioned before, at each time t* the loss [ can be modeled with a probability density function (p.d.f.) f[ ðLÞ depending on [ and the constant parameter t* ¼ 0, 1, 2, 6, 8 months. To model f[ ðLÞ, a theoretical distribution needs to be chosen at every time t*. Usually, the recorded experimental data collected by monitoring deteriorating structures show dispersion around the average value of [. However, the loss of performance seems to be contained within a certain range of values. Therefore, it seems correct to assume that, at a given time t*, the probability of a loss (L < [  L þ dL) decreases as the value L (magnitude of the loss) increases. This assumption can be made as a gooddalthough not uniquedphysical interpretation of the decay process. Analyzing the failure rate function expressed with Eq. (12.1) and considering the experimental evidence, a lognormal p.d.f. looks like the most suitable choice to model [ for each t* (Fig. 12.8(a)), according to Eq. (12.2). The parameters m and s of the distribution in Eq. (12.2) have been estimated through a computer code using the maximum likelihood method and the Rosenbrock’s optimization method (IMSL Fortran Library). As suggested in (Maybeck, 1979), to estimate the parameters of a density distribution (Eq. 12.2), this method is preferable to the least squares method (more details in Maybeck, 1979 and Augusti et al., 1984). On the contrary, to estimate the parameters of a cumulative distribution (Eq. 12.6) the least squares method and the Rosenbrock’s optimization method have been preferred. This estimation approach has been used, with good results, to evaluate the parameters of all the distribution used here.

Service life design of stone masonry structures

353

Once some significant damage levels [ are chosen, the computer code evaluates for each threshold the area over the threshold and under the p.d.f. curve (Fig. 12.6(b), dashed area). That area represents an experimental reading of the probability of reaching or exceeding each specific damage level [ over time (Fig. 12.9(b), symbols). Through Eq. (12.6) the building of the theoretical fragility curves FT ðtÞ is now possible, as shown in Fig. 12.9(b). The fragility curves represent the probability for the system to reach or exceed the damage level [ at each instant t of their life. The results of the salt crystallization tests suggested that, for the treated single substrates, the lowest concentrations chosen (1% and 2.5% of CapMC) can be considered compatibility threshold values of the salt concentration. Nevertheless, in the equivalent untreated substrates, after 7 months, also the lowest salt concentrations were able to produce small efflorescence.

12.4.3.2 The choice of the significant damage threshold The experimental data recorded on Noto wallettes through the laser profilometer and modeled with the lognormal p.d.f. have put in evidence some aspects, as shown in Fig. 12.10: • • •

In the case of untreated stones, 1% of Na2SO4 could be considered a compatibility threshold of salt concentration; the prevision shows that the probability to reach a serious damage [ ¼ 1.6 is in 2 years. Also in the case of samples treated with a consolidant (Fig. 12.10(a)), the damage appears very soon at 1% salt concentration; the chosen thresholds levels of loss are [ ¼ 0.8 and [ ¼ 1.2. In the case of samples treated with a water repellent (Fig. 12.10(b) and (c)), significant levels of damage appear at 2.5% salt concentration; the chosen thresholds levels of loss are [ ¼ 0.4%, 0.8%, 1.2%, 1.6%, 2.0%, and 2.4%, 2.5%.

12.5

On-site study of surface decay of treated stone masonry

To investigate the interaction between mortar and stone as well as the durability of stonework masonry treated with different surface products and to study more realistic situations than those created in the laboratory, three full-scale models were built in Milan, a highly polluted urban area. The constructions were single-story buildings with oriented modular walls, one made with stonework, one with brickwork, and one with mixed masonry (Baronio et al., 1992). Thanks to a layer of bentonite underneath the foundations, water could be continuously supplied to allow capillary rise (Fig. 12.11). A sodium sulphate solution was also fed into the walls at their corners to study the effects of salt crystallization on the masonry and on the surface treatments. The water was fed naturally by rain or artificially for short periods. The presence of

354

Long-term Performance and Durability of Masonry Structures

(a)

(b) F T– (t)

1.0

F T– (t)

1.0

0.9

0.9

0.8

0.8

0.7

0.7

0.6

0.6

0.5

0.5

0.4

0.4

0.3

0.3

–

l = 1.2%

–

0.2

–

l = 1.6%

0.2

–

l = 0.8%

0.1

–

l = 1.2%

0.0 0

2

4

6

8

(c)

10 12 14 16 18 20 22 24 t (months)

–

l = 2.4%

0.0 0

4

6

8

10 12 14 16 18 20 22 24 t (months)

F T– (t)

1.0

0.9

2

(d)

F T– (t)

1.0

0.9

0.8

0.8

0.7

0.7

0.6

0.6

0.5

0.5

0.4

–

l = 0.8%

0.4

–

0.3

–

0.3

–

0.2

l = 0.4%

0.2

l = 0.8%

0.1

–

l = 1.2%

0.0 0

2

4

6

(e) 1.0

l = 2.0%

0.1

8 10 12 14 16 18 20 22 24 t (months)

F T– (t)

l = 1.2% –

l = 1.6% –

0.1

l = 2.0%

0.0

l = 2.4%

–

0

2

4

6

8

(f) 1.0

0.9

0.9

0.8

0.8

0.7

0.7

0.6

0.6

0.5

0.5

0.4

0.4

10 12 14 16 18 20 22 24 t (months)

F T– (t)

–

l = 0.4% –

0.3

–

l = 0.8%

0.2

–

l = 1.2% 0.1

l = 0.8%

0.3

–

l = 1.2%

0.2

–

l = 1.6%

0.1

–

–

l = 1.6%

0.0 0

2

4

6

8

10 12 14 16 18 20 22 24 t (months)

l = 2.0%

0.0 0

2

4

6

8

10 12 14 16 18 20 22 24 t (months)

Figure 12.10 Experimental (•, :, ♦, *) and theoretical (Weibull) fragility curves, respectively, at 1% and 2% salt concentration, in the case of specimens treated with a consolidant (a) and (b) with a water repellent (c) and (d), and in the case of reference wallettes (e) and (f).

Service life design of stone masonry structures

355

(a) Plan of the model

(b) Section A-A

1740

150

Stone model

B1

A4

B2

B3

A5

30

0

B5

B4

38 120

Mixed model

C2 A

30

C3

38

0

C4

30

Brick

25

Stone

Brick

17

C1

Bentonite

33 17

0

A

80 25

(c) Details of the walls

15

Brick model

High porous concrete slab

Salt container 64

46

50

A3

312

A2

385

A1

61

Figure 12.11 Full-scale models in Milan: plans, section, and details of walls and foundations.

water in the subsoil was controlled by five piezometers. The stone used was a sandstone coming from a quarry in Firenzuola, very similar to the Serena sandstone. A continuous monitoring of temperature and R.H. of the air inside the models and immediately out of them was carried out, with the aim to know the environmental actions affecting the masonry surfaces. As an example, the total number of frost-defrost cycles within 2 years turned out to be approximately 60, while a much higher number of possible crystallization cycles took place in the same period (more than 400). It could be observed that the conditions for the formation of thenardite and mirabilite, the two most stable phases of sodium sulfate, take place in Milan every month of the year, several times a month with an average duration of 48 h. The capillary rise was visually surveyed and controlled by gravimetric measurement and radar detection (Forde et al., 1993). The environmental effects were then compared to the ones obtained in the laboratory through accelerated tests, and probabilistic models were eventually implemented. The different behavior of treated and untreated walls and the interaction between mortar and stone could be clearly analyzed. Since 1990, no other signs of decay were found apart from the deterioration due to salt crystallization. A first conclusion is that there is an influence of the mortar on the stone behavior in the masonry and that the treated walls behave in a different, but not exactly better, way than the untreated ones.

12.5.1 Decay observations and results In the model made with stonework, a high influence of vertical and horizontal mortar joints on the moisture rise in the stones was detected (Fig. 12.12). Stones were sampled from the bottom of the walls, and their moisture content was measured. The results

356

Long-term Performance and Durability of Masonry Structures

Figure 12.12 Water and salt distribution in a stonework wall.

were compared to the absorption data of the materials previously obtained, and the stones resulted nearly saturated. The presence of Na2SO4 in solution accelerated the rise of water. The observed decay was mainly exfoliation and powdering of the stone masonry surface (Figs. 12.2 and 12.12). Like the mechanism of failure that took place in the laboratory, this is clearly a long-term mechanism occurring under repeated cycles, so it can be considered a deterioration under fatigue phenomenon. The external surface gradually delaminated when salts crystallization took place underneath the masonry surface. Nevertheless, the depth of deterioration corresponds to the mere thickness of the detached layer, the inner material remaining unaltered till a new salt crystallization cycle occurs. From the structural point of view, the surface decay can be well represented as a decrease of the cross-sectional area of a masonry element (load bearing wall, column, pier, arch, etc.), which after many years may even lead to a reduction of the load-carrying capacity of the element itself. Like in the laboratory testing, also in the case of on-site buildings, the damage that occurred to the masonry was measured by determining vertical profiles of the wall in chosen positions through a laser profilometer and by calculating the material loss as the difference between two subsequent measurements (Fig. 12.13(a) and (b)). Subsequent measurements show how the profile changes in time due to any superficial decay caused by a combination of freeze-thaw, salt crystallization, air pollution, etc. Also in this case, it was possible to measure the material loss in time, assumed as a parameter of damage for the wall. Fig. 12.14(a) shows an example of profiles corresponding to six different measurements carried out in 6 years. The damage appeared in correspondence to the highest level of the capillary rise surveyed by visual inspection. It was characterized, in the first and the second stone courses, by an initial phase of relatively fast material loss (between 0 and 22 months), followed by a steady state (between 22 and 44 months) and a subsequent renewal of the decay process, which started after 44 months. The damage was not deep but

Service life design of stone masonry structures

357

(a) Laser unit controller

Data acquis system

Plotter

Feeder

Floppy disk recorder

Motor unit controller

Frame Trolley

M

Step by step motor

Laser sensor

Surface

(b)

4th course

3rd course

2nd course

Profile 31 Profile 33 Profile 35

Profile 26 Profile 28

Profile 16 Profile 19 Profile 20 Profile 23

Profile 14

Profile 11

Profile 2 Profile 5 Profile 7

1st course

Figure 12.13 (a) Scheme of the laser profilometer equipment; (b) profiles positions on the wall.

extended over a large surface till the second horizontal stone course was influenced by the stone’s disposition in the wall (stretchers or headers). In fact, the capillary rise in a stonework masonry takes place mostly through the joints due to the higher capillary rise coefficient of the mortar versus the stone: in the case studied, it was 0.27 (kg/m2)/s0.5 for the mortar and 0.006e0.007 (kg/m2)/s0.5

358

Long-term Performance and Durability of Masonry Structures

(a)

(b) 600

First measurement Last measurement

550

First measurement Last measurement

500

4th course 500

Joint

Joint

450 450

3rd course

350 Joint 300 250

2nd course

200

Joint

150 External surface 100

0 48

Bulging

400

External surface

3rd course

350

1st course Joint

50

Profile length (mm)

Profile length (mm)

400 Bulging

52 56 (mm)

60

Joint 300 48

52 56 (mm)

60

Figure 12.14 (a) Raw deterioration measurement over time (before subtraction of swelling data) (profile 7); (b) raw deterioration measurement over time (before subtraction of swelling data (profile n.7, detail).

for the stone with 5.2% vol. of porosity. Consequently, the most damaged areas were those adjacent to the joints, and the header stones showed an apparent larger damage than the stretcher stones, expressed as a percentage to their area.

12.5.2

The probabilistic approach

On the basis of the laser profilometer plots, the difference between the coordinates of the same points, detected in subsequent measurements, allowed the calculation of the material loss. Like in the laboratory measurements, bulging of the profile at a certain measurement indicates the initial detachment of a surface layer, which in turn is followed by a sudden profile reduction when the layer drops off (Fig. 12.14(b)). Again, the presence of swelling phenomena could compromise the damage measurements. Nevertheless, since bulging is an initial deterioration step before detachment, it is possible to consider it as the starting point of damage. Under this consideration, a simple computer code was studied to convert the raw experimental diagrams (Fig. 12.14) into modified diagrams where bulging was subtracted (Fig.12.15) (Binda et al., 1999a).

Service life design of stone masonry structures

359

The obtained clean plot of the evolution of the surface damage, as a function of time and space, is shown in (Fig. 12.15). The damage of the wall was identified as the loss of its cross-section in time, quantified as the area DAi between two consecutive diagrams (DAi ¼ the dashed area in Fig. 12.15). This area was automatically calculated within the assumed computer code (Binda et al., 1999a). In this case, DAi has the same meaning of [i in Section 12.4. To quantify the evolution of the loss of cross-sectional area over time, for each profile shown in Fig. 12.16, the areas DAi, with i ¼ 1, .,5, were evaluated. The measurements were made at times t1 ¼ 6, t2 ¼ 18, t3 ¼ 22, t4 ¼ 44 and t5 ¼ 60 months. Starting from the evaluated DAi and through a simple data by data linear interpolation, a first evolution process over time was obtained, as shown in Fig. 12.16.

12.5.3 Fragility curves proposed for in situ damage modeling The proposed probabilistic approach based on the construction of fragility curves is able to model the deterioration in terms of probability to reach or exceed a given 600

First measurement Last measurement

550 4th course 500 Joint

Profile length (mm)

450 400

3rd course

350 Joint 300 2nd course

250 200 150

Joint External surface

∆Ai

100 1st course

50

Joint 0 48

52

56

60

(mm)

Figure 12.15 Deterioration measurement over time (after subtraction of the swelling data) (profile n. 7).

360

Long-term Performance and Durability of Masonry Structures

1200

L (mm2)

∆Ai

1000

800

600

400

200 t (months) 0 0

10

20

30

40

50

60

Figure 12.16 Diagram describing the material loss over time on site real wall.

damage threshold L over time (here L has the same meaning of [ in the previous sections) (Figs. 12.17 and 12.18). The modeling has been obtained through a computer code involving the maximum likelihood method. The logarithm of the maximum likelihood function and the values of the other statistical test performed, associated to the L (mm2) 1200

1000

l 800 600

400

L(t*,l)

200

t*

0 0

5

10 15 20 25 30 35 40 45 50 55 60 t (months)

Figure 12.17 Deterioration process modeled as a process depending on the r.v. l.

Service life design of stone masonry structures

1000

361

L(mm2)

900 800 700 600

–L

500 400 300

100 0

L(t*,l)

200

t (months) 0

6

17 22

44

60

Figure 12.18 Survive function JL ðt; [Þ for damage level L.

physical knowledge on the deterioration, seem to well support the choice made. The assumption of a lognormal distribution to model the experimental data has pointed out that the deterioration trend changed over time, showing an increased scattering (Fig. 12.17). This behavior is probably due to the randomness connected with the greater complexity due to the reproduction of the decay process in a real environment and to the characteristics of the model structures (i.e., presence of mortar joints, prevailing of headers along the profile, profile position close to the mortar-stone interface, etc.). The experimental fragility curves, obtained by computing Eq. (12.6), are plotted in Fig. 12.18. The fitting is obtained through a computer code involving the least squares method. Also in this case the values given by the least squares method and the values of the other statistical test performed, associated to the physical knowledge of the deterioration process and the statistical knowledge, support the choice made. From Fig. 12.19, it is evident that the probability of exceeding a given damage L in a short time is lower if L > 600 mm2 and higher if L < 400 mm2. In fact, the plot shows a high probability that a small delamination (L ¼ 200e400 mm2) happens in t ¼ 20 months from the initial time. On the other hand, the probability to have a higher loss (L ¼ 800 mm2) increases only for t [ 120 months after the initial time. Also in this case, the proposed approach can point out the higher or lower probability of failure occurrence in the given time t. The fitting shown by Fig. 12.19 is satisfactory. However, given the small number in the sample tested, much caution is needed to interpret these results. This means that, to have significant results, the time of monitoring and the data recorded have to be very long.

362

Long-term Performance and Durability of Masonry Structures

F – (t)

– L = 200mm2

T

1.00 – L = 400mm2

0.90

– L = 600mm2

0.80 0.70 0.60

– L = 800mm2

0.50 0.40 0.30 0.20 0.10

t (months) 0.00 0

10

20 30

40

50

60

70

80 90 100 110 120

Figure 12.19 Fragility curves for different L.

12.6

Conclusions

The study of the delamination effects of salt crystallization has suggested a nondestructive way of measuring the decay accumulation over time, through a laser profilometer. Given the positive results of the measurements and the difficulty in interpreting the phenomena through deterministic approaches, a probabilistic model looked more suitable. The model was used to interpret the decay mechanisms during accelerated aging tests and to detect the occurrence of decay particularly when using surface treatments and/or material substitution. The results can help in the future choice of appropriate repair and protection techniques for external surfaces of historic buildings. As far as the laboratory aging tests are concerned, the proposed modeling procedure was able to evaluate the effectiveness of the surface treatment on different materials and in the presence of different salts concentrations. The presence of salts, shown after cutting the specimens at the end of the tests, was evident below the treated layer. The results showed how this approach was able to predict, in probabilistic terms, the magnitude of the expected damage over time and the occurrence time for a given damage level. Therefore, the use of this approach allows one to evaluate the treatment effectiveness on different building materials. Great caution is needed in the data collection in order for the results to be statistically significant: the number of samples cannot be less than four, and the time interval between measurements must be long enough. Of course, this depends also on the

Service life design of stone masonry structures

363

investigated damage. If the damage investigated is serious ([ > 2.0%e3.0%) the testing time must be long (10e12 months). In this case, it will be useful trying to use the model to make a prevision of the damage evolution over a long period of time (more than 30 months) without prolonging too much the test duration. In conclusion the application of this approach is simple, but to have significant results the time of monitoring should be very long and a great number of data has to be recorded. The proposed approach was applied also to interpret data collected on walls exposed to environmental attacks and accelerated salt crystallization tests. Also in this case, the approach was able to predict the occurrence probability of a given damage over time. The chosen method, which seemed to appropriately achieve also lifetime prediction of masonry subjected to long-term heavy loads (Anzani et al. 2003, 2005), proved to be flexible in fact; although a small sample was used, satisfactory results were obtained that allow one to appropriately plan maintenance and repair actions of the wall surface. The success of each probabilistic approach depends on the level of knowledge of the physical aspects of the analyzed phenomenon and on the number of samples tested. Even a simple approach like the one here proposed can give significant results if the time of monitoring and the data recorded is long enough to make its application reliable.

Acknowledgments The paper has been written in memory of Luigia Binda and Chiara Molina, who devoted a great deal of their research and guidance to this topic. Giulia Baronio is gratefully acknowledged for her participation in the investigation activity, as well as Cristina Tedeschi and Barbara Lubelli for the experimental work. Many thanks to the Laboratory of Material, Structures and Construction Testing (LPMSC) of Politecnico di Milano for the technical support and to all those who have contributed to the development of this research over the years.

References Anzani, A., Binda, L., Garavaglia, E., 2003. The vulnerability of ancient buildings under permanent loading: a probabilistic approach. Kuopio, Finland. In: Proc. of 2nd Int. Symp. ILCDES 2003, Integrated Life-Cycle Design of Materials and Structures, vol. I. RIL/VTT, Helsinki, Finland, pp. 263e268. UE. Anzani, A., Garavaglia, E., Binda, L., 2005. A probabilistic approach for the interpretation of long term damage of historic masonry. In: Lissen, S., Benz, C., Hagel, M., Yuen, C., Shrive, N. (Eds.), Proc. of 10th Canadian Masonry Symposium, Banff, Alberta, Canada, June 8e12, vol. I. Printed by Dept. of Civil Eng. The University of Calgary, Canada, pp. 664e673. Augusti, G., Baratta, S., Casciati, F., 1984. Probabilistic Methods in Structural Engineering. Chapman and Hall, London, UK. Baronio, G., Binda, L., Cantoni, F., Rocca, P., 1992. Durability of preservative treatments of masonry surfaces: experimental study on outdoor physical models. Lisbona. In: 7th Int. Congr. of Deterioration and Conservation of Stone, vol. 3, pp. 1083e1092.

364

Long-term Performance and Durability of Masonry Structures

Baronio, G., Binda, L., Cantoni, F., Rocca, P., 1993. Durability of stone and Brick-work protectives and consolidants: experimental research on full-scale models. Omiya, Japan. In: 6th Int. Conf. on Durability of Building Materials and Components, vol. 2, pp. 824e833. Bekker, P.C.F., 1999. Durability testing of masonry: statistical models and methods. Masonry International 13 (1), 32e38. Berra, M., Fatticcioni, A., Binda, L., Squarcina, T., 1993. Laboratory and in-situ measurements procedure of the decay of masonry surfaces. Omiya, Japan. In: 6th Int. Conf. on Durability of Building Materials and Components, vol. 2, pp. 834e843. Binda, L., Baronio, G., 1987. Mechanisms of masonry decay due to salt crystallization. In: J. Durability of Buildings Materials, N.4. Elsevier, Amsterdam, pp. 227e240. Binda, L., Baronio, G., 1985. Alteration of the Mechanical Properties of Masonry Prisms Due to Aging, vol. 1. 7th IBMaC, Melbourne, Australia, pp. 605e616. Binda, L., Garavaglia, E., Molina, C., 1999a. Physical and mathematical modeling of masonry deterioration due to salt crystallization. In: Lacasse, M.A., Vanier, D.J. (Eds.), Proc. of 8DCMC, 8th International Conference on Durability of Building Materials and Components, vol. I. NRC-CNRC, Ottawa, Vancouver, Canada, pp. 527e537. ISBN 0 660 17737 4. Binda, L., Baronio, G., Lubelli, B., Rocca, P., 1999b. Effectiveness of surface treatments of stone and brick masonry: proposal and calibration of on site control techniques. In: Lacasse, M.A., Vanier, D.J. (Eds.), Proc. of 8DCMC, 8th International Conference on Durability of Building Materials and Components, vol. 1. NRC-CNRC, Ottawa, Canada, Vancouver, Canada, pp. 538e549. Binda, L., Baronio, G., Squarcina, T., 1992. Evaluation of the durability of bricks and stones and of preservation treatments. Lisbon, Portugal. In: 7th Int. Congr. of Deterioration and Conservation of Stone, vol. 2, pp. 753e761. Binda, L., Charola, A.E., Baronio, G., 1985. Deterioration of porous materials due to salt crystallization under different thermohygrometric conditions. In: 5th Int. Conf. on Deterioration and Conservation of Stone, pp. 279e288. Lausanne, Suisse, I. Binda, L., Molina, C., 1990. Building materials durability semi-markov approach. Journal of Materials in Civil Engineering 2 (4), 223e239. ASCE, USA. Cardani, G., Tedeschi, C., Binda, L., Baronio, G., 2002. Crystallization Test on Treated Brick/ Stone Masonry Specimens for Damage Evaluation. Int. Conf. 9th Durability of Building Materials, 17-20 March, 2002, Brisbane, Queensland, Australia. Paper n. 039, CD-ROM. Cranmer, D.C., Richerson, D.W., 1998. Mechanical Testing Methodology for Ceramic Design and Reliability. Marcel Dekker, New York, NJ, USA. De Witte, E., 2001. Salt Compatibility of Surface Treatments (SCOST): Final Report of the European Contract ENV4-CT98-0710. KIK-Irpa, Brussels. Evans, D.H., 1992. Probability and Its Applications for Engineers. Marcel Dekker Inc., New York, NJ, USA. Forde, M.C., McCavitt, N., Binda, L., Colla, C., 1993. Identification of moisture capillarity in masonry using digital impulse radar. In: Proceedings of V Int. Conf. on Structural Faults and Repair, Edimburgh, pp. 397e403. Franke, L., Schumann, I., van Hees, R.P.J., van der Klugt, L.J.A.R., Naldini, S., Binda, L., Baronio, G., van Balen, K., Mateus, J.M., 1998. Damage Atlas Classification and Analysis of Damage Patterns Found in Brick Masonry. Fraunhofer IRB Verlag, Stuttgart. ISBN 38167-4702-7. Garavaglia, E., Cardani, G., Binda, L., 2001. A probabilistic model to predict the durability of surface treatments. In: Wittmann, F.H., Charola, A.E. (Eds.), Proc. of Hydrophobe III e Third Int. Conf. on Surface Technology with Water Repellent Agents, vol. 1. Aedificatio Publishers, Hannover, Germany, pp. 61e78.

Service life design of stone masonry structures

365

Garavaglia, E., Cardani, G., Binda, L., 2002a. A probabilistic model to predict the durability of surface treatments. International Journal of Restoration of Buildings and Monuments/ Internationale Zeitschrift f€ur Bauinstandsetzen und Baudenkmalpflege, Ed. Aedificatio Verlag, Freiburg 8 (2/3), 223e254. Garavaglia, E., Lubelli, B., Binda, L., 2002b. Two different stochastic approaches modeling the deterioration process of masonry wall over time. Materials and Structures 35 (248), 246e256. RILEM Pubblications s.a.r.l., EU. Garavaglia, E., Gianni, A., Molina, C., 2004. Reliability of porous materials: two stochastic approaches. Journal of Materials in Civil Engineering 16 (5), 419e426. ASCE. Larsen, E.S., Nielsen, C.B., 1990. Decay of bricks due to salt. Materials and Structures 23, 16e25. Lewin, S.Z., 1981. The Mechanism of Masonry Decay through Crystallization, Conservation of Historic Stone Buildings and Monuments, Washington. Nat. Academy of Sciences, Washington, DC. Masters, L.W., Brandt, E., 1987. Prediction of Service Life of Building Materials and Components, vol. 20. RILEM Technical Committees CIB W80/RILEM 71-PSL$Final Report, Materials and Structures, pp. 55e77, 115. Maybeck, P.S., 1979. Stochastic Models, Estimation, and Control. In: Series of Mathematic in Science and Engineering, vol. 1. Academic Press, New York, NJ, USA. Singhal, A., Kerimidjian, A.S., 1996. Method for probabilistic evaluation of seismic structural damage. Journal of Structural Engineering 122 (12), 1459e1467. ASCE, USA. Van Hees, R.P.J., Koek, J.A.G., De Clercq, H., De Witte, E., Binda, L., 1996. Evaluation of the performance of surface treatments for the conservation of brick masonry. In: 8th Int. Congr. on Deterioration and Conservation of Stone, vol. 3, pp. 1695e1715. Berlin, Germany.

Further reading RILEM, M.S.A.1, 1998. Determination of the resistence of wallettes against sulphate and chloride. Materials and Structures 31, 2e19.

This page intentionally left blank

13

Probabilistic modeling of aging masonry

Elsa Garavaglia 1 , Giuliana Cardani 1 , Anna Anzani 2 1 Department of Civil and Environmental Engineering, Politecnico di Milano, Milano, Italy; 2 Department of Design, Politecnico di Milano, Milano, Italy

Chapter Outline 13.1 13.2

Introduction 367 Laboratory creep and pseudo-creep tests

369

13.2.1 Results of creep tests 370 13.2.2 Results of pseudo-creep tests 371

13.3

A probabilistic approach to model the strain rate evolution 13.3.1 13.3.2 13.3.3 13.3.4

374

The strain rate versus stress history 375 The strain rate evolution as a reliability problem 378 Application of the probabilistic approach to creep tests 380 Fragility curve ε_ versus S applied to pseudo-creep tests 381

13.4 The probabilistic approach applied to the Bell Tower of Monza 13.5 Conclusions 387 Acknowledgments 388 References 388 Further reading 389

13.1

385

Introduction

The collapse of monumental buildings, which has occurred during the last 30 years, enforces the structural analysis for the safety assessment of ancient constructions to take account of specific factors that were not traditionally considered relevant (Fig.13.1) (Fradeletto, 1912). In addition to some aspects that generally characterize masonry as a material, like its being not continuous, homogeneous, and isotropic, there are some that are peculiar of ancient structures and require special attention: (1) the masonry texture (presence of different layers characterized by different stiffness values, different materials ratio, and different constructive techniques) strongly influences the stress distribution; (2) the stress state due to the self-weight has been acting with very high values for centuries, and (3) the dimensions of the structure are often considerable. In particular, creep behavior and the creep-fatigue interaction have shown to strongly influence the Long-term Performance and Durability of Masonry Structures. https://doi.org/10.1016/B978-0-08-102110-1.00013-3 Copyright © 2019 Elsevier Ltd. All rights reserved.

368

Long-term Performance and Durability of Masonry Structures

(a)

(c)

(b)

(d)

Figure 13.1 (a) San Marco Bell Tower in Venice, Italy, collapsed in 1902; (b) Civic Tower of Pavia, Italy, collapsed in 1989; (c) Zichem Tower; and (d) Meldert Tower, collapsed in Belgium in 2006.

mechanical behavior of historic constructions (Abrams et al., 1985), and a continuous mechanical damage appeared to be caused by persistent heavy loads. The influence of time on the mechanical behavior of soft porous materials becomes evident in both uniaxial and triaxial compression tests at different loading rates and in compression tests at vertical constant load. In particular, if a constant load is applied, an increase of deformation can be observed, which is commonly subdivided into three phases (Fig. 13.2): the so-called primary, secondary, and tertiary creep (Jaeger and Cook, 1976). They respectively correspond to a viscoelastic branch at decreasing strain rate (decreasing slope of the tangent to the curve) and reversible strain; a viscoplastic branch at a constant strain rate (constant slope); and a final highly unstable branch, characterized by strain developing at increasing rate and ending with the specimen failure. The appearance of one or more of these phases and the strain rate value of the secondary creep phase depend on the stress level. Whereas the creep behavior of new-built brick masonry has been widely studied (Lenczner and Warren, 1982; Shrive et al., 1997), its relevance to ancient structures became evident only after the collapse of the Civic Tower of Pavia (Binda et al.,1990;

369

Tertiary creep

Secondary creep

Primary creep

εv

Probabilistic modeling of aging masonry

Time

Figure 13.2 Primary, secondary, and tertiary creep phases of vertical strain at constant uniaxial compressive stress.

Carpinteri et al., 1996; Ferretti et al., 1998). Rheologic models to interpret this damage has been applied by Papa et al. (2000) and by Papa and Taliercio (2003). Evaluating the vulnerability of historic buildings toward the effects of permanent loads and making a reliable estimate of their lifetime on a deterministic basis are often very complex. Besides, given the difficulties of setting up significant testing procedures for studying creep behavior, an attempt has been made by applying a probabilistic approach. Long-term damage has been modeled to obtain fragility curves corresponding to threshold values of the strain-rate. Since the strain rate is related to the residual life of the material, fragility curves could be adopted, when suitable data coming from the monitoring of ancient buildings were available, for the safety evaluation of existing structures. In the chapter, the results of creep and pseudo-creep tests carried out on the masonry of ancient structures are presented, and their interpretation by means of a probabilistic model are proposed, based on the individuation of an aleatory variable (the strain rate) as a significant vulnerability index, and on the solution of a classic problem of reliability in stochastic conditions.

13.2

Laboratory creep and pseudo-creep tests

After the sudden collapse of the Civic Tower of Pavia (built in different phases between the 11th and 16th centuries, Fig. 13.3) in 1989, many prisms of different dimensions were obtained from the large blocks recovered for testing from the 7000 m3 of ruins. The prisms, subjected to mechanical tests, had mainly been obtained from the rubble masonry forming the 2800 mm-thick internal part of the three-leaf walls constituting the medieval trunk of the structure. Only a few samples came from the fairly regular external leaves made of brickwork masonry of thickness varying between 150 and 490 mm (Fig. 13.4(a)e(c)). No specimens were sampled, at that time, from the plain brick masonry constituting the 16th-century belfry, covered by granite stone elements, and not involved in the initiation of the collapse.

370

Long-term Performance and Durability of Masonry Structures

Plain masonry XVI century

Three-leaf masonry XI-XII century

Figure 13.3 Civic Tower of Pavia before failure.

(b)

(a)

(c)

67

Figure 13.4 (a) Section of the 2800-mm-thick medieval masonry; (b) prism coming from the inner rubble masonry subjected to creep test; (c) prism coming from the outer regular brick masonry subjected to pseudo-creep test.

Different kinds of uniaxial compressive tests, including monotonic tests, fatigue tests to simulate the effects of the wind, and unloading/reloading cycle tests were carried out before focusing on creep and pseudo-creep tests.

13.2.1

Results of creep tests

Six prisms of dimensions 300
300
510 mm coming from the inner rubble masonry of the Tower of Pavia were tested in compression under controlled conditions of 20 C and 50% RH, using oleo-dynamic machines able to keep a constant load for years and with a capacity of 1000 kN. The dimensions adopted for the prisms were the maximum compatible with the dimensions of the testing machine.

Probabilistic modeling of aging masonry

371

The load was increased stepwise and kept constant until either the creep strain reached a constant value or the strain-versus-time diagram took a relatively low constant slope. The first stress level was chosen between 40% and 50% of the compressive static peak stress of the prisms, estimated by sonic tests thanks to a relationship between sonic velocity and compressive strength obtained by previous testing (Anzani et al., 2000) and similar to that shown in Fig. 13.7. All the test results are reported in Fig. 13.5(a) and (b). From the experimental data, the following aspects may be pointed out: 1. Primary creep strain develops at low stress level, corresponding to a viscoelastic phase, characterized by reversible strain at decreasing rate. 2. Secondary creepdnamely a steady-state viscoplastic phase at constant strain ratedshows at 41% of the estimated material peak stress. 3. Tertiary creepdcorresponding to a highly unstable behavior, characterized by strain developing at increasing rate and ending with the specimen failuredshows at about 70% of the estimated material peak stress. 4. Material dilation under severe compressive stress, corresponding to high values of the horizontal strain, develops before failure. 5. Cracks propagate slowly for a long time until failure.

In Fig. 13.6 the crack pattern of a prism at the end of a creep test is shown: vertical cracks responsible for material dilation are visible, mainly passing at the interface between mortar and resistant elements (bricks and stone fragments).

13.2.2 Results of pseudo-creep tests In addition to long-term tests, which require constant thermo-hygrometric conditions and specially designed testing apparatus, a more rapid and therefore more convenient testing procedure that proved reasonably suitable to study creep behavior was adopted. Four prisms were tested, applying the load by subsequent steps corresponding to 0.3 MPa kept constant for 28,800 s. Again, the dimensions 200
200
350 mm adopted for the prisms were the maximum compatible with the testing machine. The peak stress obtained by pseudo-creep tests on the masonry of Pavia and of the crypt of Monza (16th century) previously tested (Anzani et al., 2005) indicated an interesting direct relationship with the results of sonic tests (Fig. 13.7). Stress-versus-strain and strain-versus-time diagrams for both vertical and horizontal directions are reported as an example in Fig. 13.8 for one of the prisms tested. As appears from the diagrams, the test procedure allowed obtaining very regular data. As an initial qualitative consideration, it can be observed that the first branch of the stressversus-strain diagrams may be considered within the elastic range. The corresponding strain-versus-time diagrams at every load step only show initial elastic strain (at time ¼ 0) and primary creep strain. Therefore, the slope of the linear branch of the diagrams, corresponding to the strain rate, tends to 0. Subsequently, after reaching a stress level of about 1.5 MPa, the stress-versus-strain diagram departs from a pseudo-linearity and shows plastic deformations, which is particularly evident in the case of horizontal strain. The corresponding strain-versus-time diagrams start to show

372

Long-term Performance and Durability of Masonry Structures

(a)

4

Secondary creep Tertiary creep

3

εv (×103)

Primary creep 2 1 0 –1

εh (×103)

–2 –3 Dilation –4 –5 –6 –7 –8 0

200

400

600

800

1000

1200

400 600 800 Time (days)

1000

1200

Time (days)

(b) 3.0

εv (N/mm2)

2.5 2.0 1.5 1.0 0.5 0.0 0

200

Figure 13.5 (a) Strain-versus-time diagrams from creep tests; (b) stress-versus-time diagrams from creep tests.

Probabilistic modeling of aging masonry

373

Figure 13.6 Pavia prism 300
300
510 mm of inner rubble masonry after creep test: crack pattern.

also secondary creep strain. Again, this is more pronounced in the case of horizontal strain components: at increasing the stress level, the slope of the linear branch of the diagrams gets progressively higher than 0. Though it is not clearly visible in this case, at the last load step, tertiary creep strain takes place. 10

Monza Pavia

Peak stress (MPa)

8 6 4 2 0 0

500

1000

1500

2000

2500

3000

Sonic velocity (m/s)

Figure 13.7 Compressive peak stress versus sonic velocity of masonry prisms subjected to monotonic test.

374

Long-term Performance and Durability of Masonry Structures

6

σv (N/mm2)

5 4 3 2

1 0

t (s)

20000 40000 60000 80000

εh

–20

εv

–15

–10

–5

0

(μm/mm)

5

10

Figure 13.8 Results of pseudo-creep tests on Pavia medieval masonry.

Fig. 13.9 shows the final crack pattern of a prism subjected to pseudo-creep test: like in Fig. 13.6, vertical and subvertical cracks, responsible for material dilation, are evident, mainly passing at the interface between mortar and the resistant elements (bricks and stone fragments). Again, the horizontal secondary creep strain rate obtained for each specimen at each load step was considered a significant parameter in view of lifetime prediction. Considering the last pseudo-creep load step for each specimen, its duration was regarded as the residual life of the material; to investigate its relationship with the secondary creep rate before collapse, the two values were calculated for every tested specimen and plotted as shown in Fig. 13.10. Comparing the masonry of Pavia with that taken from the crypt of Monza, though the number of test results is not particularly large, an interesting inverse relationship can be found, which seems to apply to the two materials considered together. A similar relationship applies to other brittle materials, such as concrete when subjected to creep and fatigue tests (Taliercio and Gobbi, 1998). In view of the preservation of historic buildings from major damage or even failure, it would be very useful to detect a critical value of the strain rate under which the residual life of the building is greater than the required service life.

13.3

A probabilistic approach to model the strain rate evolution

To apply a probabilistic model to the interpretation of the mentioned experimental data, vertical and horizontal secondary creep strain rate (_εv and ε_ h ), corresponding

Probabilistic modeling of aging masonry

375

Figure 13.9 The 200
200
350 mm prism of Pavia inner rubble masonry after pseudo-creep test: crack pattern.

to the strain-versus-time linear branch slope, have been calculated for all specimens at the application of any load step. The horizontal secondary creep strain rate, chosen as the most significant parameter in terms of lifetime prediction, was then plotted for each test versus the stress value corresponding to any load step (Fig 13.11) (Anzani et al., 2003). By experimental evidence, as reported in Section 13.2, the strain evolution connected with a given stress history of a viscous material like a historic masonry can be described through the parameters ε_ v and ε_ h , defined as the vertical and horizontal strain rate, respectively, taken on the linear branches of the strain-versus-time diagrams corresponding to the applied constant stress level sv .

13.3.1 The strain rate versus stress history For each sv the high randomness connected with the changing of strain rate, due to the high nonhomogeneity of the masonry, allows one to treat ε_ v and ε_ h as random variables(r.v.) with a certain distribution of values (Fig. 13.12). In this perspective, the deformation process can be interpreted as a stochastic process of the considered r.v. The strain rate also depends on the corresponding stress level sv . Therefore, for each stress level sv , the strain
rate  (measured in ε/s) can be modeled with a probability density function (p.d.f.), fε_ E_ , which is dependent on the sv and on the strain

376

Long-term Performance and Durability of Masonry Structures 1E+001 Monza Pavia

Δεv/Δt (×106/s)

1E+000

1E–001

1E–002

1E–003 100

1000

10000

Time (sec)

100000

Figure 13.10 Secondary creep rate before failure versus duration of the last load step.


 rate. To model fε_ E_ , a family of theoretical distributions must be chosen at every stress level. Obviously, the distribution modeling a given phenomenon must be found according to the physical aspects of the phenomenon itself, and to the characteristics of the distribution function in its “tail,” where often no experimental data can be collected. This latter
 aspect can be investigated by analyzing the behavior of the failure rate function 4ε_ E_ at every stress level sv connected with the chosen distribution function.
   4ε_ E_ dE_ ¼ Pr E_ < ε_  E_ þ dE_ j ε_ >E_ csv ;

ε_ ¼ ε_ h or

ε_ ¼ ε_ v

(13.1)

More details on this subject are given in Binda and Molina (1990), Garavaglia et al. (2004), Anzani et al. (2005), and Garavaglia et al. (2008). The recorded experimental data can show a large dispersion around the average value of ε_ v and ε_ h . This is probably due to the randomness connected with the high nonhomogeneity of the masonry. Here a conventional value of ε_ may be assumed as a critical value indicating a safety limit. Consequently, for a given stress level sv the probability that the critical strain rate connected with the secondary creep safety limit shows increases if the strain rate increases. It seems therefore correct to assume that, at a given stress level sv ,  theprobability that the secondary creep strain rate falls in the interval E_ < ε_  E_ þ d E_ increases as the strain rate increases. This hypothesis, assumed as a satisfying (but not unique) physical interpretation of the decay process, leads to model the evolution of the strain rate at the stress level sv with a Weibull p.d.f. as follows: h
a1 1  a1 i

exp  r1 E_ fε_ E_ ¼ a1 r1 r1 E_

c sv

(13.2)

This family of distributions presents an immediate occurrence rate function (13.1) _ that increases if the value of E_ increases and tends to N if E/N; this fact seems to respect the physical interpretation of the strain rate behavior previously proposed.

Probabilistic modeling of aging masonry

(a)

377

0.005 0.0045 0.004

0.003 0.0025 0.002

h

h

. 3 ε ((ε × 10 ) / s)

0.0035

0.0015 0.001 0.0005 0 1

1.2

1.4

1.6

1.8

2

2.2

2.4

2.6

2.8

3

σv (N/mm2)

(b) 5E–005

3E–005

2E–005

h

h

. 3 ε ((ε × 10 ) / s)

4E–005

1E–005

0 0

1

2

3

4

5

6

7

σv (N/mm2)

Figure 13.11 Interpolation of the secondary creep horizontal strain rate versus applied stress: (a) pseudo-creep tests, (b) creep tests.

378

Long-term Performance and Durability of Masonry Structures

. . Ε( ε )

0.005 0.0045 0.004

0.003 0.0025

.

f . (ε)

.

⎯ε =0.002 h

Ε

. εh((εh × 103) / s)

0.0035

0.002 0.0015

.

⎯ε =0.001 h

0.001

.

⎯ε =0.0005 h

0.0005 0 1

1.2

1.4

1.6

σ*v

1.8

2

2.2

2.4

2.6

2.8

3

σv (N/mm2)

Figure 13.12 Creep tests: interpolation of the horizontal secondary creep strain rate versus applied stress (* * *) modeled with a Weibull p.d.f fE_ ð_εÞ (dd).

13.3.2

The strain rate evolution as a reliability problem

It is furthermore interesting to evaluate the probability that the system will reach or exceed a given strain rate level ε_ h or ε_ v over a stress history. Considering a significant strain rate level ε_ h or ε_ v and the variable stress needed to exceed it, the strain rate evolution can be treated as a reliability problem (Garavaglia et al., 2002, 2004). Reliability R(t) is related to the performance of a system over time, and it is defined as the probability that the system does not fail by time t (Evans, 1992). This definition is extended here, denoting by Rðsv Þ the probability that a system will not exceed a given significant strain rate level by stress sv. The random variable used to quantify reliability is s, which is simply the stress corresponding to which damage ε_ h or ε_ v is exceeded. Thus, from this point of view, the reliability function is given by (Evans, 1992; Aven and Jensen, 1999): Rðsv Þ ¼ Prðs > sv Þ ¼ 1  Fs ðsv Þ

(13.3)

where Fs ðsv Þ is the distribution function for s. Computing Fs ðsv Þ for different damage levels ε_ h (or ε_ v ) allows us to build the corresponding fragility curves describing the probability of exceeding a given strain rate versus stress (Singhal and Kerimidjian, 1996). Fragility curves are cumulative distributions, Fs ðsv Þ, and they could be modeled through experimental data. This procedure involves the construction of experimental

Probabilistic modeling of aging masonry

379

fragility curves Fs ðsv Þ useful as a basis for the probabilistic modeling of the curves themselves. In our case an experimental fragility curve is a discrete set of points, each of which represents the probability that a defined level of strain, ε_ h or ε_ v , could be exceeded when a certain level of stress, sv , is reached. From a computational point of view, this experimental probability is the area under the curve fE_ ð_εÞ and above the chosen strain rate threshold (Fig. 13.13) (Garavaglia et al., 2002, 2004). It can be calculated by using the survive function:

 Jε_ E_ ¼ 1  Fε_ E_ c sv

(13.4)


 function of the fE_ ð_εÞ at every sv . where Fε_ E_ is the cumulative
 distribution  _ The computation of Jε_ E at every sv is possible with the use of any kind of computer code for numeric integration. The areas computed over different strain rate thresholds provide the experimental fragility curves F s ðsv Þ. Therefore, the evaluation for different sv of the exceedance probability for each strain rate level results in an experimental fragility curve for each chosen ε_ h (or ε_ v ). To model the experimental fragility curves, a Weibull distribution has been chosen (Cranmer and Richerson, 1998; Bekker, 1999; Garavaglia et al., 2002, 2004) that seems to provide a good interpretation of the physical phenomenon (more details in Chapter 18).

. . Ε( ε )

0.005 0.0045 0.004

0.003 0.0025

h

h

. 3 ε ((ε × 10 ) / s)

0.0035

0.002

Ε

.

f . (ε)

0.0015 0.001

.

⎯ε =0.0005 h

0.0005 0 1

1.2

1.4

σ*v

1.6

1.8

2

2.2

2.4

2.6

2.8

3 2

σv (N/mm )

Figure 13.13 Exceedance probability to cross the threshold ε_ referred to creep test.

380

Long-term Performance and Durability of Masonry Structures

Eqs. (13.2) and (13.4) have been applied to model vertical and horizontal strain rate of both creep and pseudo-creep results; the experimental fragility curves obtained by the application of Eq. (13.4) will be presented in the following paragraphs.

13.3.3

Application of the probabilistic approach to creep tests

The identification of significant thresholds of critical strain rate is crucial for modeling fE_ ð_ε; sÞ. Considering the relationship between the secondary creep strain rate and the residual life of the material obtained with creep tests (Taliercio and Gobbi, 1996; Anzani et al., 2000), three conventional values of the horizontal critical strain rate ε_ h have been identified, corresponding to the initiation and the subsequent development of the secondary creep phase for most of the prisms, as shown in Fig. 13.12. In Fig. 13.14 the experimental and theoretical fragility curves related with these horizontal strain thresholds are reported. They describe the probability to exceed the critical thresholds as a function of the reached stress level sv. The same thresholds related to the initiation of the secondary creep may be indicated also for the vertical strain rate (Anzani et al., 2003). In Fig. 13.15 the experimental and theoretical fragility curves related with the thresholds previously adopted for the horizontal strain rate are shown. In Table 13.1 a comparison between vertical and horizontal strain rate is reported. The optimized estimate of the distributions parameters was obtained by fitting the experimental fragility curves with an error included within the interval 9.5e03 and 9.9e02. Though the error is greater than hoped, since the distribution assumed to model the phenomenon is supported by the physic of the phenomenon itself, it is however acceptable. The choice of the 3% values of the exceedance probability shown in Table 13.1 are respectively related to the limits of an interval of possible previsions (10%e90%) and to the value (63%) of the born-in method used to test the significance of an experimental test. In fact, for a Weibull modeling, 63% is the limit describing a reliable prevision because most of the experimental data fall below this limit. Comparing these results, it can be observed that the exceedance of the chosen threshold strain rate is always performed by horizontal strain at a lower stress level than by vertical. This is in good agreement with the dilatant behavior of ancient masonry when approaching failure, as shown by Fig. 13.5(b), where the horizontal strain appears to be higher and developing at a higher rate than the vertical ones. This behavior is typical of a damaged material presenting a response beyond the elastic limit. This is also confirmed by the crack pattern of prisms at the end of creep tests, as shown in Fig. 13.6: having loaded the specimen vertically in compression, the cracks follow a mainly vertical path, therefore giving an apparent horizontal dilation. However, from a probabilistic point of view, caution must be offered to the tails of the distribution where usually not much data are present and where the prediction becomes critical.

Probabilistic modeling of aging masonry

1

381

F (σv) ⎯σ

0.9 .

⎯εh=0.0005

0.8 0.7 0.6

.

⎯ε

=0.001

h

0.5 0.4 0.3

.

⎯ε

h

=0.002

0.2 0.1 0 1

1.2

1.4

1.6

1.8

2

2.2

2.4

2.6

2.8

3

2

σv (N/mm )

Figure 13.14 Horizontal strain rate: experimental (•, :, >) and theoretical (d) fragility curves from creep tests.

13.3.4 Fragility curve ε_ versus S applied to pseudo-creep tests Also in the case of pseudo-creep tests, threshold strain rate values related to the initiation and the evolution of the secondary creep phase were selected. The experimental and theoretical fragility curves corresponding to the thresholds ε_ ¼ 5:0e  005, ε_ ¼ 1:0e  004, and ε_ ¼ 2:0e  004 are reported in Fig. 13.16 and Table 13.2 for both vertical (Fig. 13.16(a)) and horizontal (Fig. 13.16(b)) strain rates. The optimized estimate of the distributions parameters was obtained by fitting the experimental fragility curves with a maximum error of 0.11 for the threshold ε_ ¼ 2:0e  004 of Fig. 13.15(a). It has to be noted that, due to the different load histories, different strain rate thresholds were found in this case. In fact, pseudo-creep tests were carried out applying the load stepwise starting from zero; therefore the strain rate evolution started from lower values than in the case of creep tests. In the latter case the first load step was applied at 40%e50% of the peak estimated stress; therefore higher strain rate values were obtained. Comparing these results with those reported in Table 13.1, despite a certain scatter in the results, a similar trend to that shown by the creep test is confirmed, but here some additional comments have to be made. In the 10% of the cases, the exceedance of the chosen threshold strain rate is performed at a lower stress level in the case of

382

Long-term Performance and Durability of Masonry Structures 1

F σ (σv) ⎯

0.9 0.8 0.7 0.6 0.5 .

⎯εv =0.001

0.4 0.3

.

⎯εv =0.002

⎯εv =0.0005

.

0.2 0.1 0 1

1.2

1.4

1.6

1.8

2

2.2

2.4

2

σv (N/mm )

Figure 13.15 Vertical strain rate: experimental (•, :, A) and theoretical (d) fragility curves from creep tests.

Table 13.1 Probability to exceed ε_ v and ε_ h for different sv ε_ v [ 0:0005 2

ε_ h [ 0:0005

Exceedance prob. of ε_ v

sv (N/mm )

sv (N/mm2)

10%

1.40

1.38

63%

1.81

1.75

90%

2.00

1.90

vertical strain. On the contrary, at 63% and 90% of the cases the exceedance of the chosen threshold strain rate is performed at a lower stress level by the horizontal strain, exactly like in the creep test. This indicates that the prisms tested with pseudo-creep presented a viscoelastic nondilatant response at low stress level, whereas at higher stress level, their behavior was beyond the elastic limit, and a dilation became evident. The influence of the testing procedure is evident. In fact, in the case of creep tests the first load step corresponded to the 40%e50% of the estimated static peak stress; therefore the elastic nondilatant behavior related to low stress value could not be

Probabilistic modeling of aging masonry

(a)

1

383

F σ (σv) ⎯

0.9 0.8 .

⎯εv =5.0e–005

⎯ε =2.0e–004 v

.

0.7 0.6 0.5 0.4 0.3

.

⎯εv =1.0e–004

0.2 0.1 0

0

1

2

3

4

5

2

σv (N/mm )

(b)

1

F σ (σv) ⎯

0.9 0.8 .

⎯ε =5.0e–005 h

0.7 0.6 0.5

.

⎯ε =1.0e–004 h

0.4 0.3 0.2

.

⎯ε =2.0e–004 h

0.1 0 0

1

2

3

4

5

2

σv (N/mm )

Figure 13.16 Experimental (: - • - >) and theoretical (d) fragility curves obtained from pseudo-creep tests: (a) vertical strain; (b) horizontal strain.

384

Table 13.2 Probability to exceed ε_ v and ε_ h for different sv 2

ε_ v [ 1:0e L004 2

ε_ v [ 2:0e L004 2

ε_ h [ 5:0e L005 2

ε_ h [ 1:0e L004 2

ε_ h [ 2:0e L004

Exceedance prob.

sv (N/mm )

sv (N/mm )

sv (N/mm )

sv (N/mm )

sv (N/mm )

sv (N/mm2)

10%

0.01

0.49

1.03

0.13

0.15

0.46

63%

0.09

1.38

1.99

0.64

0.79

1.28

90%

0.80

2.02

2.58

1.18

1.48

1.87

Long-term Performance and Durability of Masonry Structures

ε_ v [ 5:0e L005

Probabilistic modeling of aging masonry

385

expressed, the material performing a response beyond the elastic limit. On the contrary, in the case of pseudo-creep tests the load was applied through small steps, so at the initial stage, the material still had to be within the elastic limit, whereas at higher stress level, it performed a dilatant response.

13.4

The probabilistic approach applied to the Bell Tower of Monza

The proposed probabilistic approach has been applied for the first time to evaluate the results of monitoring of a massive historic building subjected to persistent vertical loading, mainly originating from self-weight. The Bell Tower of Monza, a 16thcentury structure built in solid brickwork masonry, had suffered major and diffused cracks due to high compression (Fig. 13.17). After the constitution of a technical committee in 1976, the building, together with the Cathedral, was subjected to a systematic control, setting up 31 fixed bases, seven of which on the tower corresponded to the major cracks. The measuring basis had a length of about 400 mm, and the measurements were taken, starting in January 1978, every month during the first 3 years and every 3 months subsequently. The instrument used was a millesimal deformometer. After the recorded increase of the crack opening on the tower and an anomalous geometry recorded on the Cathedral, in 1992 a new committee was constituted by the DIS (now DICA) department of Politecnico di

(a) (b)

West front

(c)

Est front

Figure 13.17 The Bell Tower of Monza: (a) crack pattern on the prospects, (b) crack pattern on the wall cross-section, (c) sampling of a brick to inspect the crack depth.

386

Long-term Performance and Durability of Masonry Structures

Milano, who installed a static control system, which included the continuation of the geometric evaluation of the cracks. Each of the previous bases was substituted by a couple of new bases placed above and below the other ones, having a length of 200 mm, the readings being taken at the same periodicity (every 3 months). Considering the data collected until 1999 (Fig. 13.18), the influence of thermal variation on the crack opening can clearly be observed. Nevertheless, tracing a regression line through the data, a net increase of the crack opening in time is visible; in the case of the base shown in Fig. 13.18, a rate of crack opening of 6.48 micron/year can be measured until 1986, and a higher rate of 24.94 micron/year can be measured subsequently. This clear worsening of the tower static conditions could not be attributed to any external reason, so an intrinsic damage may only be assumed. After the static survey, a consolidation intervention on the tower was carried out (Modena et al., 2002). To compare the rate of crack opening of the tower with the strain rate measured in the laboratory, the average strain rate was calculated based on the monitoring data. This calculated rate was related with the values of the vertical stress, locally measured by flat jack tests at the same height of the crack monitoring. Flat jack tests gave respectively a vertical stress of 0.98 N/mm2 at a height of about 12 m and of 1.67 N/mm2 at a height of about 5 m, which were in good agreement with the FE elastic analysis considering the self-weight (Binda et al., 1998). The fragility curves shown in Fig. 13.19 were built for the tower cracks for two different thresholds ε_ h of horizontal strain rate. As expected, in this case the parameter optimization gives very small errors (7.70e11 and 3.41e09, respectively) because only two points were fitted. The recorded strain rate and stress level, and consequently the chosen threshold strain rate values, are lower than those obtained by creep and pseudo-creep laboratory tests:

400 300

Aperture variation (μm)

200 100 0 –100 –200 6.48 μm/year

–300

24.94 μm/year

19 7 19 8 79 19 8 19 0 8 19 1 8 19 2 8 19 3 84 19 8 19 5 86 19 8 19 7 88 19 8 19 9 9 19 0 91 19 9 19 2 9 19 3 9 19 4 95 19 9 19 6 9 19 7 98 19 9 20 9 00

–400

Years

Figure 13.18 Monitoring of the Bell Tower of Monza: opening variation of a crack versus time.

Probabilistic modeling of aging masonry

1

387

F (σo) ⎯σ ⎯ε =1e–9 h

.

0.9 0.8 0.7

.

⎯ε =2e–9 h

0.6 0.5 0.4 0.3 0.2 0.1

0 0

1

2 3 σo (N/mm2)

4

5

Figure 13.19 Horizontal strain rate, experimental (• :) and theoretical (d) fragility curves.

in fact, the latter were aimed to take the prisms to failure, whereas the tower of Monza is still performing its service life. Here the modeling appears hazardous, but on the basis of the results obtained by laboratory tests, it is possible to suppose the same distribution (Weibull distribution) for modeling the probability of exceedance of the ε_ h thresholds chosen. Although two data are not sufficient to investigate the Weibull shape, the results obtained look like an interesting example of possible application of the procedure to real cases, where more suitable data are available. The decision of strengthening the tower was, of course, taken independently from the present research.

13.5

Conclusions

The unexpected collapse of some tall buildings like historic masonry towers (San Marco Bell Tower of Venice, Civic Tower of Pavia, and others) led to understanding the creep behavior of ancient structures when subjected to high constant loads for a long time. The recognition of the signs showed by the masonry over time and their interpretation make us assert that their collapse cannot be defined as sudden any longer. The safety assessment of monumental buildings in view of their protection requires to take account of different aspects that act in a synergetic way. A probabilistic model has been applied to interpret experimental results aimed at studying the time-dependent

388

Long-term Performance and Durability of Masonry Structures

behavior of historic masonry by means of creep and pseudo-creep tests. An estimate of the exceedance of critical thresholds in vertical and horizontal strain rate that are related to the residual service life was possible. The results obtained through the application of the proposed probabilistic approach indicate an interesting research direction toward the interpretation of data collected through the monitoring of monumental buildings detecting, for instance, whether the recorded creep strain indicates a critical condition. The choice of significant geometric parameters itself is another important research issue: if in the case of clearly visible crack pattern, the evolution of crack opening rate would be suitable, this would not be in the case of multiple leaf masonry, characterized by cracks only diffused in the inner leaf and not visible on the facade. In similar cases, an alternative geometric parameter indicating damage could be, for instance, the wall thickness increase rate at the base of the building, which indicates transversal material dilation, associated with the measurement of a corresponding vertical compressive stress by flat jack test. Of course, the precocious recognition of a critical state will allow the design of a strengthening intervention to prevent total or partial failure of the construction.

Acknowledgments Prof. L. Binda is gratefully acknowledged for her extraordinary vision in the experimental understanding of ancient masonry and breakthrough interpretation of its long-term behavior.

References Abrams, D.P., Noland, J.L., Atkinson, R.H., 1985. Response of clay-unit masonry to repeated compressive forces. In: Proc. of the 7th IBMaC, Melbourne, Australia, vol. I, pp. 565e576. Anzani, A., Garavaglia, E., Binda, L., 2005. A probabilistic approach for the interpretation of long term damage of historic masonry. In: Lissen, S., Benz, C., Hagel, M., Yuen, C., Shrive, N. (Eds.), Proc. of 10th Canadian Masonry Symposium, Banff, Alberta, Canada, June 8e12. Dept. of Civil Eng. The University of Calgary, Canada, I, pp. 664e673. Anzani, A., Binda, L., Garavaglia, E., 2003. The vulnerability of ancient buildings under permanent loading: a probabilistic approach. Helsinki, Finland, UE, I. In: Proc. of 2nd Int. Symp. ILCDES 2003, Integrated Life-cycle Design of Materials and Structures, Kuopio, Finland, RIL/VTT, pp. 263e268. Anzani, A., Binda, L., Mirabella Roberti, G., 2000. The effect of heavy persistent actions into the behaviour of ancient masonry. Materials and Structures 33, 251e261. Aven, T., Jensen, U., 1999. Stochastic Models in Reliability. Springer, New York, NY, USA. Bekker, P.C.F., 1999. Durability testing of masonry: statistical models and methods. Masonry International 13 (1), 32e38. Binda, L., Gatti, G., Mangano, G., Poggi, C., Sacchi Landriani, G., 1990. La Torre Civica di Pavia: indagini sui materiali e sulla struttura. L’Edilizia e L’Industrializzazione 11, 713e735. Binda, L., Molina, C., 1990. Building materials durability semi-markov approach. Journal of Materials in Civil Engineering, ASCE, USA 2 (4), 223e239.

Probabilistic modeling of aging masonry

389

Carpinteri, A., Cerioni, R., Iori, I., 1996. Alcuni pensieri in merito al crollo della Torre Civica di Pavia, Studi e Ricerche. Politecnico di Milano 17, 491e511. Cranmer, D.C., Richerson, D.W., 1998. Mechanical Testing Methodology for Ceramic Design and Reliability. Marcel Dekker, New York, NJ, USA. Evans, D.H., 1992. Probability and Its Applications for Engineers. Marcel Dekker Inc, New York, NJ, USA. Ferretti, D., Iori, I., Riva, R., 1998. Un approccio allo studio della stabilita delle antiche torri: il crollo della torre civica di Pavia, Studi e Ricerche. Politecnico di Milano 19, 169e191. Fradeletto, A. (Ed.), 1912. Il campanile di San Marco riedificato: studi, ricerche, relazioni, Venezia, Carlo Ferrari. Garavaglia, E., Anzani, A., Binda, L., Cardani, G., 2008. Fragility curve probabilistic model applied to durability and long term mechanical damages of masonry. Materials and Structures 41, 733e749. Garavaglia, E., Gianni, A., Molina, C., 2004. Reliability of porous materials: two stochastic approaches. Journal of Materials in Civil Engineering, ASCE 16 (5), 419e426. Garavaglia, E., Lubelli, B., Binda, L., 2002. Two different stochastic approaches modeling the deterioration process of masonry wall over time. Materials and Structures 35, 246e256. Jaeger, J.C., Cook, N.G., 1976. Fundamentals of Rock Mechanics, second ed. Chapman & Hall, London. Lenczner, D., Warren, N., 1982. In situ measurement of long-term movements in a brick masonry tower block. In: Proc. of the 6th IBMaC, Rome, vol. I, pp. 1467e1477. Modena, C., Valluzzi, M.R., Tongini Folli, R., Binda, L., 2002. Design choices and intervention techniques for repairing and strengthening of the Monza cathedral bell-tower. Construction and Building Materials 16 (7), 385e395. Oct. Elsevier Science Ltd. Papa, E., Taliercio, A., 2003. A theoretical model for the description of static and creep-induced damage in brittle materials under nonproportional loading. In: Owen, D.R.J., O~ nate, E., Suarez, B. (Eds.), Computational Plasticity VII e Fundamentals and Applications. CIMNE, Barcelona (CD ROM). Papa, E., Taliercio, A., Mirabella Roberti, G., 2000. A damage model to predict the behaviour of masonry under sustained loading. In: Proc. 12th IBMaC, Madrid, III, vol. I, pp. 1777e1790. Shrive, N.G., Sayed-Ahmed, E.Y., Tileman, D., 1997. Creep analysis of clay masonry assemblages. Canadian Journal of Civil Engineering 24, 367e379. Singhal, A., Kerimidjian, A.S., 1996. Method for probabilistic evaluation of seismic structural damage. Journal of Structural Engineering, ASCE, USA 122 (12), 1459e1467. Taliercio, A.L.F., Gobbi, E., 1996. Experimental investigation on the triaxial fatigue behaviour of plain concrete. Magazine of Concrete Research 48 (176), 157e172. Taliercio, A., Gobbi, E., 1998. Fatigue life and change in mechanical properties of plain concrete under triaxial deviatoric cyclic stresses. Magazine of Concrete Research 50 (3), 247e256.

Further reading Larsen, E.S., Nielsen, C.B., 1990. Decay of bricks due to salt. Materials and Structures 23 (1990), 16e25. Lewin, S.Z., 1981. The Mechanism of Masonry Decay through Crystallization, Conservation of Historic Stone Buildings and Monuments. Washington, Nat. Academy of Sciences, Washington, D.C.

390

Long-term Performance and Durability of Masonry Structures

Masters, L.W., Brandt, E., 1987. Prediction of service life of building materials and components, RILEM technical committees CIB W80/RILEM 71-PSL$final report. Materials and Structures 20 (1987), 55e77. Maybeck, P.S., 1979. Stochastic models, estimation, and control. In: Series of Mathematic in Science and Engineering, vol. 1. Academic Press, New York, NJ, USA.

Index ‘Note: Page numbers followed by “f ” indicate figures and “t” indicate tables.’ A Accelerated aging tests, 78, 79f acidic fog test, 80 environmental conditions, 77e78 freezing-thaw cycles test, 79 humidity-dryness cycles test, 79 light stability test, 78 repair mortars, 191, 192te195t saline fog test, 79 salt crystallization test, 78e79, 80f tests types, 78 thermal shock cycles test, 79e80 ACI 440.7R, 228e229, 232e233, 233t ACI 440.9R, 233e234 Acidic fog test, 80 Acoustic emission (AE) testing compressive creep monitoring. See Compressive creep monitoring crack monitoring. See Cracks; monitoring debonding detection, 299e300 features, 288e289, 288f fracture process, 289 frequency spectrum, 288 glass fiberereinforced polymer/ steel-reinforced grout (GFRP/SRG)strengthened bricks aging and moist environments, 301 composite materials, 300e301 cumulative AE energy, 303e304, 304f debonding mechanism, 300 emissions in, 301, 302f failure modes, 303e304 long-term bond quality monitoring, 304 one-directional medium-density steel fiber, 300e301 preamplifier gain, 301 single-lap shear bond tests, 301e302, 303f solid clay bricks, 300e301

high-frequency elastic waves, 287e288 longitudinal (P) waves, 287e288 metallic pressure vessels, 287 nondestructive technique, 287 on-site monitoring, 289e290 parameter-based approach, 289 Rayleigh (R) waves, 287e288 transversal/shear (S) waves, 287e288 waveform-based analysis, 289 Acquisition process, 242 Activation energy, 225e226 Aging, probabilistic modeling Bell Tower of Monza, 385e387, 385fe387f creep-fatigue interaction, 367e368 laboratory creep and pseudo-creep tests brickwork masonry, 369, 370f Civic Tower, 369, 370f compressive peak stress vs. sonic velocity, 371, 373f crack pattern, 374, 375f experimental data, 371 Pavia prism, 371, 373f secondary creep rate, 374, 376f strain-versus-time diagrams, 370e373, 372f, 374f testing machine, 370e371 uniaxial compressive tests, 370 long-term damage, 369 phases, 368, 369f safety assessment, 367 San Marco Bell Tower, 367, 368f strain rate evolution, probabilistic model application, 380, 381fe382f, 382t exceedance probability, 379, 379f experimental fragility curves, 379e380 fragility curves, 378e379, 381e385, 383f, 384t

392

Aging, probabilistic modeling (Continued) horizontal secondary creep strain rate, 374e375, 377f reliability function, 378 strain rate vs. stress history, 375e376, 378f survive function, 379 uniaxial and triaxial compression tests, 368 Air conditioning systems, 17 Air/natural hydraulic lime, 172 Alkaline compounds, 67e68 Ambient vibration tests (AVTs), 248 Arrhenius equation, 227 Arrhenius model, 224e225 Artificial weathering cycles, 191 Atmospheric agents atmospheric pollution, 67e68 chemical hydration process, 65 hydrolysis process, 65e66 physical hydration, 65 silicates and aluminosilicate minerals, 65e66 stone pore system, 65 temperature and insolation, 66 wind and saline fog-marine spray, 66e67 Australian research projects, 327 B Bacteria, 141, 141f Bell Tower of Monza, 385e387, 385fe387f Bio-based fibers, 136 Biocidal Product Regulation (BPR), 150 Biocidal Products Directive (BPD), 150 Biodeterioration, 32 earth masonry, 91e92 stone, 71e72 Building information models (BIM), 268e269 C Calcitic air lime mortars, 172 Carbonation process, 67e68 Carbon FRP (CFRP), 216 Carboniferous, 5 Chemical hydration process, 65 Civic Tower, 369, 370f Clastic sedimentary rock, 4e5 Clay bricks, 222, 223f

Index

clastic sedimentary rock, 4e5 clay minerals, 4 components, 5e6 deterioration factors and mechanisms atmospheric contamination, 15 biologic agents, 14e15 capillarity ascension, 12 causes, 10e11 crypto-efflorescence, 13 crystallization mechanism, 13 environmental factors, 11 external factors, 10 freezing/thawing cycles, 12 humidity conditions, 13 moisture transport properties, 11 pore size distribution, 14 porous mortar, 14 production/manufacture, 11e12 salt crystals, 13e14 saturation of, 12 temperature, 14 types, 11, 11t untreated clays, 13 durability air conditioning systems, 17 compressive strength, 16 definition, 4 lime-based/gypsum-based mortars, 16e17 permeability test, 16 porosity level, 15 water/moisture prevention, 16e17 waterproof membrane, 17 evolution of, 3 illite, 5 iron oxide (Fe3O4), 6 kaolinite, 5 limestone, 6 modifications, 3e4 montmorillonite, 5 phyllosilicates, 5 plastic behavior, 5 production process calcite and dolomite, 7e8 chemical reactions, 7 drying, 6e7 firing, 6e7 hygroscopic water elimination, 7e8 kaolinite, 7e8

Index

metakaolin, 7e8 mixing and molding, 6e7 quality, 7 selection and preparation, 6 sintering and vitrification, 7e8 stages, 6 properties chemical elements, 9 colors range, 9e10 compressive strength, 8 expansion/shrinkage phenomena, 9 mineralogic composition, 9 porosity, 8 suction rate, 8e9 temperature and operating conditions, 10 water absorption, 8e9 pyrophyllite, 5 raw clay, definition, 4 rock fragments and mineral grains, 4e5 sands and sandy material, 6 sedimentary rocks, formation process, 4e5 talc, 5 Climate and global warming, 130 Climatic cycles, 191 CNR-DT 200, 228e231, 231t Compressed earth blocks (CEBs), 91 Compressive creep monitoring creep curve and the cumulative AE events, 295, 295f creep failure, 298e299, 299f periodic measurements, 296e298, 296fe297f time-dependent damage progress, 295e296 Concrete block abrasion, 30 absorption, 27 acoustic performance components and building elements, 42 international codes, 39 National Mansory Concrete Association TEK 13-1C manual, 41 natural frequency, 41 noise control, 39 plaster/stucco, 41e42 reverberation time, 42 sound absorption, 39 sound behavior, solid object, 41, 41f sound transmission class (STC), 41 sound transmission loss (STL), 41

393

surface porosity, 41e42 test results, 42, 43t admixtures, 22 advantages, 21 applications, 21 biodeterioration, 32 block machine, 22, 22f chemically aggressive environments, 30 compression failure, 27e28, 28f compressive strength, 27e28 cracks. See Cracks; concrete block curing process, 22e23, 23f drying/carbonation, 29 efflorescence block molds, 32, 36f soluble salts migration, 31 on wall painting, 32, 35f fire resistance, concrete block, Fire resistance freeze and thaw resistance, 30 functions, 21 hard body impact testing, 36e38, 38t industrialized process, 26 lightweight blocks, 29 long-term performance, 26 nonreinforced cementitious components, 29 performance evaluations, 35 pigmented facade units, 30 pigments, 22 plan design, 23e25, 25f Portland cement, 29 pozzolanic materials, 22 production process, 22 properties, 27 quality control, 23e25 regular concrete block, 23, 24f shapes, 23, 24f shrinkage testing, 27f, 29 soft body impact testing, 36, 37t, 38f split and colored-face, 23, 24f split tensile failure, 28e29, 29f standard test method, durability, 53e54, 54t steam curing process, 29 steel bars corrosion, 30 structural performance, 36, 55 sulfate attacks, 30 suspended point load resistance

394

Concrete block (Continued) criteria, 39, 39t testing, 39, 39f test results, 39, 40t technical limitations, 26 tensile strength, 28e29, 28f thermal performance air infiltration/exfiltration mitigation, 45 climate zone number, 42, 44t equivalent U factors, 45, 45t heat capacity (HC), 42, 45 International Energy Conservation Code, 45 multilayer walls, 45 Technological Research Institute (IPT), 42 thermal resistance (R), 42 thermal transmittance (U), 42, 44t training programs, 26 UV-resistant pigment, 30 wall elevation, 23e25, 26f water permeability. See Water permeability, concrete block Conservation treatments, stone, 80e81 Cracks concrete block design error, 31, 31fe32f failure modes, 31 lintel and windowsill, design detailing, 31, 33f mortar compressive and tensile strengths, 30 movement joints, lintel cracks, 31, 34fe35f windowsill, wrong detailing, 31, 33fe34f formation, 135, 136f monitoring applied force and cumulative AE energy, 293, 294f brick masonry wall, three-point bending test, 292e293, 292f brick-mortar interfaces, 291 cyclic testing, bricks, 291e292, 291f detection range analysis, 291 elastic strains, 293 Kaiser and Felicity ratios, 291e292 LOAD ratios, 293e294, 294f

Index

PZT-type AE sensors, 293 wave velocities and source-sensor distances, 290e291, 290t Creosote-treated glulam, traffic bridge, 150, 150f Crypto-efflorescence, 13 Crystallization mechanism, 13 Crystallization tests, 221 Curing process, 22e23, 23f D Darcy law, 180 Data acquisition system (DAQ), 243 Decay fungi, wood, 313 brown rot fungi, 139e140, 139f characteristics, 138e139 colonization and degradation, 138e139 discoloring and destroying fungi, 138e139, 138f factors, 140 lignocellulose, 140 mold and stain fungi, 138e139 soft rot fungi, 139e140, 139f types, 139 waterlogged wood, 140 white rot fungi, 139e140, 139f Diffuse-porous hardwoods, 131e132 Discoloring fungi, 140e141, 140f Drying process, 6e7, 29 Dynamic identification acquisition process, 242 ambient vibration tests (AVTs), 248 data acquisition system (DAQ), 243 data analysis, 242 digital raw signals, 243e244 experimental testing procedures, 241 finite element (FE) modal analyses, 243 frequency domain (FD) approach, 244 frequency filtering window, 243e244, 245f frequency response functions (FRFs), 243 full-scale in situ experimentation, 243 full-scale monumental structures, 248 Hanning window application, 243e244, 245f high-sensitivity piezoelectric accelerometers, 243, 244f impulse response functions (IRFs), 243 input-output modal identification, 246e247 modal parameters, 241

Index

Mogadouro Clock Tower. See Mogadouro Clock Tower noise and environmental effects, 247e248 nondestructive health monitoring tool, 242 operational modal analysis (OMA), 242e243 output-only modal identification benefits, 244e245 correlation functions, 246 disadvantages, 247 enhanced frequency domain decomposition (EFDD), 246 fast fourier transform (FFT) process, 245 frequency domain decomposition (FDD), 246 linear filter, 244 nonparametric methods, 245 power spectral densities, 246 stochastic subspace identification (SSI) methods, 246 Qutb Minar tower. See Qutb Minar tower Saint Torcato church. See Saint Torcato church stationary Gaussian white noise stochastic process, 242e243 structural integrity, 241 time domain (TD) approach, 244 E Earth masonry adobe bricks, 91 air voids, 92e93 assessment capillary water absorption, 113, 114f climatic conditions, 117 compressive strength, 112 cyclic in-plane lateral loading tests, 113 dry compressive strength approach, 116 flexural testing, 112 freeze-thaw tests, 116 loading tests, 112e113 mechanical strength testing, 113 particle size distribution, 114 spray and drip erosion tests, 115 stabilized bricks, 115 water-based organic silicone emulsion, 116 weight loss, 115

395

biodeterioration, 91e92 clay, 90e91 cohesive soils stabilization, 92 compressed earth blocks (CEBs), 91 compressive strength, 91 disadvantages, 90 drainage system, 94e95 durability characteristics, 90e91 external structural loading damage ballistic tests, 100e101 and boundary conditions, 98e99 causes, 96 crack propagation, 97e98, 97f cyclic loading shift failure, 98 differential ground settlement collapse, 99, 99f earthquakes, 96e97 geometry and configuration, 96 geometry and mechanical characteristics, 98 kinematic rocking mechanism, 98 masonry spandrels crack, 99e100, 100f monolithic rammed earth constructions, 97 out-of-plane failure mechanism, 97e98 seismic damage patterns, 97e98 seismic forces, 98 soilestructure interaction, 99 structural integrity, 98e99 thermal movements, 100 wind velocities, 98e99 fabrication procedure, 93 factors, 90e91 fractal analyses, 92e93 freeze-thaw cycles, 91e92 grading and plasticity characteristics, 91 incompatible and ineffective interventions Cementitious coatings, 107e108 clay-based grouts, 106e107 coating surface collapse, 107e108, 109f defective structural behavior, 107 elastic resistance, 111 fragility of, 111 grout injections, 106e107 mechanical connectors, 107 reconstruction works, 107, 108f rehabilitation and maintenance schemes, 107e108 renovation/remodelling projects, 110

396

Earth masonry (Continued) restoration projects, 106e107 seismic strengthening strategies, 111 shear forces, 110e111 surface erosion, 109e110 tensile capacity, 107 waterproof membrane, 110 interconnection structures, 94, 95f laboratory tests, 91e92 load-bearing walls, 93 long-term field exposure, 92 moisture-induced deterioration, 94e95, 96f mud walls, 89 non-homogeneous composition, 92 normative test procedures, 117 performance levels, 117e118 physico-mechanical properties, 89 restoration, 89 rigidity, 93 semi-solid state, 92e93 shaping/compaction method, 92e93 shrinkage cracks, 90e91 smectite and montmorillonite groups, 90e91 soil mixtures, 90e91 stiffness of, 91e92 thermo-mechanical loading, 92 timber sub-structures, 94 unit-mortar interfaces, 93e94 vernacular earth masonry structures, 89, 90f water absorption coefficients, 92e93 water penetration, 94e95 water-soluble salts, 91 weathering and moisture-driven damage. See Weathering and moisture-driven damage X-ray computed tomography, 92e93 Efflorescence block molds, 32, 36f soluble salts migration, 31 on wall painting, 32, 35f Eigenentropy, 273 Elastic modulus, 216 Electric conductivity, 75 Energy dispersive x-ray spectroscopy, 75 Enhanced frequency domain decomposition (EFDD), 246 Equilibrium moisture content (EMC), 133e134, 135t

Index

Eurocode 5, 134e135 Externally bonded reinforcement (EBR), 209e210, 214 F Fast Fourier transform (FFT) process, 245 Federal US Highway Administration, 226 Fiber-reinforced polymers (FRPs) activation energy, 225e226 aging treatments, 215 Arrhenius equation, 227 Arrhenius model, 224e225 bond durability adhesive-substrate interface, 219 bond decay, 218 bond fracture energy, 220 brick-primer interface, 219 clay bricks, 222, 223f conditioning method, 222 crystallization tests, 221 cyclical procedures, 218 debonding load, 220, 223, 225f degradation phenomenon, 219e220 environmental cycles, 220 epoxy resin, 222, 222f failure mode, 219e220 freeze and thaw cycles, 221 humidity variation, 221 infrared (IR) thermography technique, 220 Lecce stone, 218e219, 222, 223f on-site tests, 221e222 porosity, 223, 224f pull-off tests, 219e220 single-lap shear bond tests, 219e220 stress-slip curves, 218e219 substrate-reinforcements interface, 221e222 surface preparation, 219e220 tensile strength reduction, 222, 224f thermal aging, 221 thermal expansion coefficient, 220 codes and standards implementations 440 Fiber-Reinforced Polymer Reinforcement, 229 ACI 440.7R, 228e229, 232e233, 233t ACI 440.9R, 233e234 CNR-DT 200, 228e231, 231t guidelines, 228e229

Index

reinforced concrete (RC) applications, 228e229 degradation mechanisms environmental conditions, 211 moisture, 211e213 structural components, 211 synergistic effects, 211 temperature, 213e214 ultraviolet (UV) exposure, 214 detrimental mechanism, 225e226 environmental key factors, 224e225 evaluation of, 224e225 exponential degradation model, 226 externally bonded reinforcement (EBR), 209e210, 214 Federal US Highway Administration, 226 fiber-resin interface, 210 Fick’s law, 227 finite element method (FEM), 227 industrial-made composites, 210 long-term delamination load/bond strength, 210 long-term exposure, 209e210 materials durability carbon FRP (CFRP), 216 detrimental effect, 217e218 elastic modulus, 216 firing temperature, 216 hydrolysis and plasticization, 216e217 hygrothermal durability, 217 realization phase, 216 tensile strength, 216 water on, 215e216 wet lay-up procedure, 216 mechanical properties, 210 moisture diffusion, 227 quantification, 210 reinforced elements, 215 strengthening materials and performance, 224e225 structural elements, 209e210 substrates types, 210 tensile strength, 227e228 time shift factor, 226 treatments, 210 Fiber saturation point (FSP), 133e134 Fick’s law, 227 Finite element models (FEM), 227, 243, 268e269

397

Fire resistance, concrete block characteristics, 46 equivalent thickness, 46, 47f National Concrete Mansory Association TEK 7-1C, 46, 48t noncombustible material, 46 testing, 46, 49te50t Firing process, 6e7 Freezing/thawing cycles, 12, 79 Frequency domain (FD) approach, 244 Frequency domain decomposition (FDD), 246 Frequency response functions (FRFs), 243 Fungal infestation risk, 136, 137f G Glass fiber-reinforced polymer (GFRP)- strengthened bricks aging and moist environments, 301 composite materials, 300e301 cumulative AE energy, 303e304, 304f debonding mechanism, 300 emissions in, 301, 302f failure modes, 303e304 long-term bond quality monitoring, 304 one-directional medium-density steel fiber, 300e301 preamplifier gain, 301 single-lap shear bond tests, 301e302, 303f solid clay bricks, 300e301 Gypsum-based mortars, 16e17 H Hanning window application, 243e244, 245f Hard body impact testing, 36e38, 38t Heartwood, 131, 318e319, 318t Heat capacity (HC), 42, 45 Humidity-dryness cycles test, 79 Hydrolysis process, 65e66 I Illite, 5 Impulse response functions (IRFs), 243 Infrared (IR) thermography technique, 220 Insects, wood, 314e315, 314f beetles, 142e143, 142t termites, 143, 143f Interfacial transition zone (ITZ), 177

398

International Energy Conservation Code, 45 Ion chromatography, 75 Iterative Closest Point technique, 274 K Kaiser and Felicity ratios, 291e292 Kaolinite, 5, 7e8 Kinematic rocking mechanism, 98 L Laboratory aging tests cryptoefflorescence, 346 crystallization test results, 349e350, 349fe351f damage threshold, 353, 354f environmental conditions, 347 fragility curve, 351 monitoring and damage quantification, 347e349, 348f “monotonic” behavior, 351 physical mechanical characteristics, 347, 347t RILEM TC 127 MS, 346e347, 346f statistical analysis, 351 stochastic modeling, deterioration process, 352e353 surface treatments, 346 Laboratory-made mortars, 185 Laser scanning advantage, 266e267 data processing, automation, 268e269 engineering evaluation, historical constructions, 266 experience-based protocols, 266 geomatic technologies, 266e267 health monitoring, 266 light detection and ranging technology (LiDAR), 267e268. See also Light detection and ranging technology (LiDAR) limitations, 266 mobile laser technology, 266e267 operational mode, 269 photogrammetric technologies, 266e267 point cloud processing automation algorithms, 273 preprocessing, 274 registration, 273e274

Index

remote sensing, 266e267 service life, 266 static sensors, 266e267 structural analysis, 268 structural damage detection Guimar~aes wall, Portugal, 278e281, 280fe281f Segura roman bridge segmentation, 277e278, 278f superficial pathologies characteristics, 274 clusterization techniques, 276e277, 277f contactless techniques, 274e275 efflorescence and humidity, 275 environmental and geometric parameters, 275 intensity data, 275e276, 276f transmission efficiency, 275 Light detection and ranging technology (LiDAR), 267e268 data acquisition, 270e271 distance measurement device, 270 mobile laser scanner (MLS), 271e272 scanner types, 270 terrestrial laser scanner (TLS), 270e271, 270f Light stability test, 78 Lightweight blocks, 29 Lime-based mortars, 16e17 air/natural hydraulic lime, 172 calcitic air lime mortars, 172 degradation mechanisms, 171 chemical, 173e174 mechanical, 176 physical, 174e176 properties, 172 stone masonry, 170, 170fe171f transport properties bimodal pore size distribution, 179 calcium carbonate, 179 capillary coefficients, 179 capillary pressure, 180e181 carbonation kinetics, 176e177 Darcy law, 180 drying curves, air lime mortars, 181e182, 182f interfacial transition zone (ITZ), 177 liquid water absorption, 180e181

Index

mercury intrusion porosimetry (MIP), 179, 179fe180f permeability, 182 pore network, 176e177 pores, classification, 177, 178t salt crystallization, 179 scanning electron microscope (SEM), 180 suction, 182 water absorption curves, 180e181, 181f water and salt solutions, 176e177 Limestone, 6 LOAD ratios, 293e294, 294f M Mercury intrusion porosimetry (MIP), 75, 179, 179fe180f Metakaolin, 7e8 Mixing and molding process, 6e7 Mobile laser scanner (MLS), 271e272 Mogadouro Clock Tower, 248f damage survey, 249, 249f experimental mode shapes, 249e250, 252f FEMU, 250e252, 253f, 253t granite stones, 248 modal assurance criterion (MAC), 249e250, 252f natural frequencies and damping ratios, 249e250, 251t 3D numeric model, 250e252 sensor layout, 249, 250f uniaxial piezoelectric accelerometers, 249 vibration modes, 250 Montmorillonite, 5 Mortars anomalies and impact, 184 characterization testing, 183 chemical-mineralogic characteristics, 185 codes and standards implementation, 200e201 compressive strength, 185 degradation mechanisms, 171 diagnostic test techniques, 185, 186te190t durability and performance, 170 global analysis test methods, 183e184 hydric characteristics, 185 inorganic and organic natural materials, 170e171

399

in situ wide-spectrum nondestructive techniques, 183 laboratory-made mortars, 185 lime-based mortars. See Lime-based mortars maintenance, 171 mechanical characteristics, 185 mechanical strength, 172 microscopic characteristics, 185 microstructural characteristics, 185 nondamaged mortars, 185 Portland cement, 172 protective measures atmospheric pollutants, 196e197 capillary coefficient, 199 capillary rising effects, 196, 198f chemical barriers, 198 conservation strategy, 198e199 degradation effects, 196 ethyl silicate and nanolimes, 199 factors, 198e199 human action, 197 multicoat technique, 196 Roman mortars, 198e199, 198f salts, 197 water repellents, 198 water vapor coefficient, 199 repair mortars accelerated aging tests, 191, 192te195t artificial weathering cycles, 191 climatic cycles, 191 factors, 185 salt crystallization cycles, 191 roles, 169 self-healing properties, 170 shore hardness measurement, 184, 184f thermography, 183e184, 183f ultrasound pulse test, 184, 184f visual analysis, 183 N Nanolime, 67e68 National Mansory Concrete Association TEK 13-1C manual, 41 Nitrification process, 67e68 Noise control, 39 Nuclear magnetic resonance, 75e76

400

P Phyllosilicates, 5 Point cloud processing automation algorithms, 273 preprocessing, 274 registration, 273e274 4-Points Congruent Sets (4PCS) algorithm, 273 Polarized light optical microscopy, 74 Portland cement, 29 Power spectral densities, 246 Principal components analysis (PCA) algorithm, 273 Pyrophyllite, 5 Q Qutb Minar tower, 252e253, 254f bending modes, 256, 257f experimental results, 256, 256t geometric survey, 253e255, 254f output-only modal identification techniques, 255 sensor layout, 255, 255f SSI-driven implementations, 255 structure, 252e253 vibration modes, 256, 257f R Reverberation time, 42 Ring-porous hardwoods, 131e132 S Saint Torcato church experimental mode shapes, 260, 261f FEMU process, 261e262, 261t, 262f general view and plan, 257e258, 258f geotechnical survey, 259 operational modal analysis (OMA) tests, 260, 260f, 260t static nonlinear analysis, 261e262 structural damage, 259, 259f vibration modes, 260e261 Saline fog test, 79 Salt crystallization, 78e79, 80f, 191 deliquescence moisture, 69 destructive agents, 70 environmental RH, 69 hydration phases, 68e69 ionic compound, 69

Index

magnesium sulfate salts, 68e69 restoration mortars, 68e69 saline solution absorption, 69, 70f salty soils, 68e69 solubility products, 69, 69t San Marco Bell Tower, 367, 368f Sapwood, 131, 318e319, 318t Scanning electron microscopy, 75 Shannon entropy, 273 Silicates and aluminosilicate minerals, 65e66 Soft body impact testing, 36, 37t, 38f Softwoods, 131e132 Sound transmission class (STC), 41 Sound transmission loss (STL), 41 Spectrophotometry, 77 Stationary Gaussian white noise stochastic process, 242e243 Steam curing process, 29 Steel-reinforced grout (SRG)-strengthened bricks aging and moist environments, 301 composite materials, 300e301 cumulative AE energy, 303e304, 304f debonding mechanism, 300 emissions in, 301, 302f failure modes, 303e304 long-term bond quality monitoring, 304 one-directional medium-density steel fiber, 300e301 preamplifier gain, 301 single-lap shear bond tests, 301e302, 303f solid clay bricks, 300e301 Stochastic subspace identification (SSI) methods, 246 Stone accelerated aging cycles, 78f. See also accelerated aging cycles aesthetic properties, 77 anthropic agents, 72e73 artificial materials, 60 biodeterioration, 71e72 carving, 59e60 conservation treatments, 80e81 construction technique, 337e338 decorative element, 60 deterioration patterns, 60, 61f durability, 73 external environmental conditions, 59e60

Index

extrinsic degradation factors, 60 geologic resource, 59e60 hydric properties, 77 impact resistance, smith hammer tool, 75 instrumental techniques and test methods, 74, 76f electric conductivity, 75 energy dispersive x-ray spectroscopy, 75 ion chromatography, 75 polarized light optical microscopy, 74 scanning electron microscopy, 75 X-ray diffraction, 75 intrinsic and external factors, 60 laboratory aging tests cryptoefflorescence, 346 crystallization test results, 349e350, 349fe351f damage threshold, 353, 354f environmental conditions, 347 fragility curve, 351 monitoring and damage quantification, 347e349, 348f “monotonic” behavior, 351 physical mechanical characteristics, 347, 347t RILEM TC 127 MS, 346e347, 346f statistical analysis, 351 stochastic modeling, deterioration process, 352e353 surface treatments, 346 lifetime prediction aging testing procedures, 339e340 characteristics, 340 delamination/crumbling, 341 deterioration process, 338e339 exfoliation and erosion, 339e340, 341f fatigue effects, 339 fragility curve, 342e346, 343f intrinsic properties, 338e339 mechanical strength, 339 salt crystallization test, 339, 340f surface damage, 342 water chemical reactions, 338e339 masonry surfaces decay, 337e338 mercury intrusion porosimetry, 75 nuclear magnetic resonance, 75e76 on-site behavior, surface decay, 338 capillary rise, 353e355, 355f environmental effects, 355

401

frost-defrost cycles, 355 in situ damage modeling, 359e361, 360fe362f laser profilometer equipment, 356, 357f porosity, 357e358 probabilistic approach, 358e359, 359fe360f raw deterioration measurement, 356, 358f second horizontal stone course, 356e357 water and salt distribution, 355e356, 356f physical-chemical properties, 62f chemical and mineralogical composition, 62 condensation and capillary forces, 63 conservation and restoration treatments, 64, 64f low surface roughness, 63e64 mechanical strength, 63e64 permeability, 63 petrophysical characteristics, 62 porosity, 63 shape and tortuosity, 63 surface area, 63 water retention, 63e64 planning strategies, 338 salt crystallization, 337. See also Salt crystallization standard tests, 74 structural construction material, 60 surface adhesion test, 75 surface roughness, 77 ultrasonic velocity propagation, 75 weathering process, 60e61 agents, 64, 65f atmospheric agents. See Atmospheric agents X-ray computed tomography, 76e77 Stone pore system, 65 Structural damage detection Guimar~aes wall, Portugal, 278e281, 280fe281f Segura roman bridge segmentation, 277e278, 278f Sulfate attacks, 30 Sulfating process, 67e68 Surface adhesion test, 75

402

T Talc, 5 Technological Research Institute (IPT), 42 Terrestrial laser scanner (TLS), 270e271, 270f Textile-reinforced mortar (TRM). See Steelreinforced grout (SRG)-strengthened bricks Thenardite, 67e68 Thermal expansion coefficient, 220 Thermal resistance (R), 42 Thermal shock cycles test, 79e80 Thermal transmittance (U), 42, 44t Timber. See Wood Time domain (TD) approach, 244 Time shift factor, 226 U Ultrasonic velocity propagation, 75 Ultrasound pulse test, 184, 184f UV-resistant pigment, 30 W Water absorption, 92e93 under atmospheric pressure, 77 contact angle determination, 77 desorption under atmospheric pressure, 77 Karsten tube penetration test, 77 under vacuum, 77 water vapor permeability, 77 Water-based organic silicone emulsion, 116 Water permeability, concrete block capillarity, 51 cement-lime stucco, 47 control joints and horizontal reinforcement, 52 faceshell bedding, 47 moisture sources, 47e51, 51f mortar and grout shape, 52, 53f multileaf wall solution, 47 paint coatings, 47 physical properties, 52 rainfall precipitation, 47 rain with wind pressure, 51 source of, 47 steam retarders, 53 surface treatments, 52 wall waterproofing testing, 51, 52f water-repellent additives, 52 water vapor, 51

Index

Waveform-based analysis, 289 Weathering and moisture-driven damage bitumen-based waterproof courses, 104e106 capillary absorption, 101e102, 101t cement and lime, 104 damp and salt crystallization, 102, 103f drying shrinkage, 104 freeze-thaw damage, 102e103 hydrophobic organic polysiloxane, 103e104 kinetic energy, 102 moisture penetration, 101e102 New Mexico Building Code, 104 straw-reinforced clay-rich plaster, 104, 105f water-based silane/siloxane emulsion, 103e104 water resistance, 106 wetting-drying cycles, 101e102 Weathering process, 60e61 agents, 64, 65f atmospheric agents. See Atmospheric agents Wind and saline fog-marine spray, 66e67 Wood, 131f aging, 317 applications, 130 assessment techniques health assessment, 332 moisture content (MC) monitoring, 331e332 non- and semidestructive tests, 331 bacteria, 141, 141f benefits, 130 biochemicals, 131 biodegradability, 130 biologic degradation, 312 cell types and configuration, 131 cellulose, 132 cell wall regions, 132 chemical and physical agents, 312e313 environmental conditions, 315e316 fire, 316 weathering, 316 classification schemes, 145, 146t climate and global warming, 130 decay fungi, 313. See also Decay fungi, wood decay-influencing factors

Index

dosimeter model, 159 European research projects, 158 exposure and resistance dose, 159, 160f factorization, 158e159 in hazard classes, 159, 159t design protection coatings, 155 construction phase, 154e155 maintenance, 157, 158f moisture-induced risk, 155 moisture protection, 155, 156fe157f physical barriers, 155e156, 157f physical properties, 154e155 planning phase, 154e155 deterioration agents, 312 diffuse-porous hardwoods, 131e132 discoloring fungi, 140e141, 140f durability, 147 antifungal treatments, 327 Australian research projects, 327 classification, 147e148, 148t, 318e319, 319t depth and rate, fungal decay, 327, 327fe328f European wood species, 317e318, 318t "extractives", 147 fungal decay risk, 328e329, 328f laboratory and field test methods, 148e149, 149f long-term performances, 317 moisture dynamics, 147 natural resistance, 317 nontarget organisms, 148e149 permeability, 326 sapwood and heartwood treatability, 318e319, 318t test results, 147e148 wood-destroying organisms, 145 extractive components, 132e133 hardwoods structure, 131 heartwood, 131 hemicelluloses, 132 insects, 314e315, 314f beetles, 142e143, 142t termites, 143, 143f intervention techniques, 331 lignin, 132

403

load-carrying function, 131 marine borers, 144e145, 144f modification biocidal substances, 152 chemical, 152e154 with oils and waxes, 152e153 properties, 152 thermal, 152e153 water repellents, 152 moisture adsorption/capillary suction, 133 allergic reactions, 137e138 bio-based fibers, 136 biodegradability, 137e138 building components, 133 cracks formation, 135, 136f equilibrium moisture content (EMC), 133e134, 135t erosion, 135e136, 137f Eurocode 5, 134e135 fiber saturation point (FSP), 133e134 fungal infestation risk, 136, 137f lignocellulose strength, 135 L-joint tests, 137 load-bearing structures, 134e135 moisture content (MC), 133 mold growth, 137e138 relative humidity (RH), 133 service classes, 134e135, 136t sorption isotherm, 133, 134f structural integrity, 135e136 thermal conductivity, 136 transport, 134 organic solvents, 132e133 performance-influencing factors climate exposure, 322 cracks, 322e323 fungal decay, 323e324, 324t long-term performance, 321e322 moisture content (MC), 324e326, 325fe326f phases, performance survey, 321 polymeric/nonpolymeric compounds, 132 preservation techniques, 151, 151f preservatives, 149e150, 150f primary structural compounds, 132 products, 130

404

Wood (Continued) protection systems, 312, 329e330 ring-porous hardwoods, 131e132 sapwood, 131 service life, 130, 321 softwoods, 131e132 technologic properties, 130 tree rings, 131 use class (UC)

Index

four-class system, 319, 320t long-term performances, 320 X X-ray computed tomography, 76e77 X-ray diffraction, 75