UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

In-plane shake-table testing of unreinforced masonry walls strengthened with fibre reinforced-plastics Turek, Martin Edward 2002

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2002-0588.pdf [ 26.19MB ]
Metadata
JSON: 831-1.0063871.json
JSON-LD: 831-1.0063871-ld.json
RDF/XML (Pretty): 831-1.0063871-rdf.xml
RDF/JSON: 831-1.0063871-rdf.json
Turtle: 831-1.0063871-turtle.txt
N-Triples: 831-1.0063871-rdf-ntriples.txt
Original Record: 831-1.0063871-source.json
Full Text
831-1.0063871-fulltext.txt
Citation
831-1.0063871.ris

Full Text

In-Plane Shake-Table Testing of Unreinforced Masonry Walls Strengthened with Fibre Reinforced-Plastics by MARTIN EDWARD TUREK B.E.Sc, The University of Western Ontario, 2000 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE in ^ THE F A C U L T Y OF GRADUATE STUDIES Department of Civil Engineering We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA October, 2002 © Martin E. Turek, 2002 UBC Rare Books and Special Collections - Thesis Authorization Form In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of C\Vu_ SAJ<§/A/t££//J{j The University of British Columbia Vancouver, Canada Date Oc£Oi?£-*L ta / ^® Abstract Unreinforced-masonry structures have typically performed poorly during earthquakes. Many structures that were built with this material still have a continued economic and social benefit, even though they may be structurally unfit. For this reason, many seismic retrofit schemes have been developed, one being the use of composite materials. Various testing has been performed on these materials, and on some of their structural applications in seismic retrofits. Limited full-scale shake table testing has been performed, and a research collaboration between The University of British Columbia and Public Works and Government Services Canada, Pacific Region was established to explore this type of testing on the application of these materials. A series of in-plane shake-table tests were performed on a set of unreinforced and FRP strip reinforced concrete-masonry walls. The unreinforced walls were used as a benchmark for the study, and five different configurations of FRP strip reinforcing were tested. The walls were subjected to code-level and near-code level to determine their behaviour at design levels. Then the walls were subjected to extreme-level records until failure, to determine the failure modes and behaviour of the various reinforcing schemes. It was observed from the testing that all of the strengthened specimens, regardless of reinforcing configuration, performed well during the application of the code- and near-code level records. Four of the five reinforcing configurations also performed well during application of the extreme-level record. It was concluded from these tests that the use of vertical FRP strips is an adequate reinforcing configuration to improve the in-plane performance of U R M walls. The behaviour of vertical strips was comparable to that of horizontal strips, which were also tested and found to be very effective. The vertical strips were found to be effective in repairing damaged walls. They were also found to help control the failure modes of the specimens, and prevent collapse even after severe damage had occurred. This can be a strong contributor to improving life-safety during a severe event. ii Table of Contents Abstract 1 1 Table of Contents iii List of Figures v i i List of Tables xiv Acknowledgements x v Project Funding xvii Glossary of Terms xviii Chapter 1 - Introduction 1.1 General 1 1.2 Seismic Performance of U R M 2 1.2.1 Historic Performance of U R M 3 1.2.2 Structural Performance of U R M 7 1.2.3 Guidelines for Behaviour of U R M 11 1.3 Use of Fibre-Reinforced-Plastics in U R M Reinforcement 13 1.4 Objectives of Thesis 16 1.5 Scope of Thesis 17 Chapter 2 - Literature Review 2.1 General 18 2.2 Cyclic and Static Testing 18 2.3 Pseudo-Dynamic Testing 21 2.4 Shake-Table Testing 21 2.5 Out-of-Plane Testing 22 Chapter 3 - Shake Table Testing 3.1 General 25 3.2 Specimen Design 25 3.2.1 Description of Specimen 25 3.2.2 Design Details 26 3.2.3 Construction 27 3.2.4 FRP Materials and Layout 29 3.3 Shake Table 40 3.3.1 Description 40 3.3.2 Table Modifications 43 3.4 Ground Motions 47 3.4.1 Kobe, Japan 1995 (Kobe) 48 3.4.2 LlayLlay, Chile 1985 (Chile) 50 3.4.3 Telcordia VERTEQII Synthetic Record (VERTEQII) 52 3.5 Experimental Procedure and Instrumentation 54 3.5.1 Test Setup 54 3.5.2 Shake-Table Tests 55 3.5.3 Instrumentation 59 3.5.4 Data Acquisition System 62 Chapter 4 - Results and Observations 4.1 General 63 4.2 Observations 63 4.2.1 U R M Specimens 63 4.2.2 FRP Reinforced Specimens 70 4.2.3 Test Summary 82 4.3 Displacement Results 82 4.4 Acceleration Results 87 4.5 Force Results 95 4.6 Drift Results 95 Chapter 5 - Analysis of Results and Interpretation 5.1 General 100 5.2 Failure Modes and Behaviour of U R M Specimens 100 5.2.1 Specimen UI 100 iv 5.2.2 Specimen U2 101 5.2.3 Specimen U3 102 5.2.4 Comparison of the U R M Specimens 103 5.3 Failure Modes and Behaviour of FRP Reinforced Specimens 104 5.3.1 Specimen RI 104 5.3.2 Specimen R2 106 5.3.3 Specimen R3 107 5.3.4 Specimen R4 108 5.3.5 Specimen R5 110 5.4 In-Plane Stiffness Analysis 112 5.5 FRP Configuration Effects 119 5.5.1 Effect of Vertical Strips 119 5.5.2 Effect of Spacing on Vertical Strips 120 5.5.3 Effect of Anchoring of Vertical Strips 120 5.5.4 Effect of Vertical Strips on Damaged Wall 122 5.5.5 Effect of Horizontal Strips 122 5.5.6 Effect of X-Pattern Strips 123 5.5.7 Effect of One-Sided Application Only 123 Chapter 6 - Conclusions and Recommendations 6.1 Summary 124 6.2 Conclusions 125 6.2.1 Code and Near-Code Level Performance 125 6.2.2 Extreme Level Performance 126 6.3 Recommendations 129 6.3.1 Recommendations for Further Study 129 6.3.2 Recommendations for Design and Application 131 References 133 v Appendix A - Test Setup Details 138 Appendix B - Additional Earthquake Record Details 147 Appendix C - Complete Displacement Results 151 Appendix D - Complete Acceleration Results 164 Appendix E - Complete Force Results 183 vi List of Figures Figure 1.1: Loma Prieta Earthquake U R M Damage [Courtesy of NISEE] 5 Figure 1.2: Northridge Earthquake U R M Damage [Courtesy of NISEE] 6 Figure 1.3: Nisqually Earthquake U R M Failures [Courtesy of EQE] 7 Figure 1.4: In-Plane Behaviours of U R M Walls 10 Figure 1.5: Typical In-Plane Forces and Stresses 10 Figure 1.6: Tensile Forces in Mortar Joints 11 Figure 3.1: Wall Design Details 26 Figure 3.2: Basebeam Construction 27 Figure 3.3: Wall Construction 28 Figure 3.4: Fibre Wrapping 29 Figure 3.5: Specimen RI FRP Layout 30 Figure 3.6: Photo of Specimen RI 31 Figure 3.7: Tyfo Fibre Anchor Installation 32 Figure 3.8: Specimen R2 FRP Layout 33 Figure 3.9: Photo of Specimen R2 33 Figure 3.10: Specimen R3 FRP Layout 34 Figure 3.11: Photo of Specimen R3 35 Figure 3.12: Specimen R4 FRP Layout 36 Figure 3.13: Photo of Specimen R4 36 Figure 3.14: Anchor #1 Details (Specimen R4) 37 Figure 3.15: Specimen R5 FRP Layout (Backside Only) 38 Figure 3.16: Photo of Specimen R5 38 Figure 3.17: Anchor #2 Details (Specimen R5) 39 Figure 3.18: Shake-Table 40 Figure 3.19: Shake-Table Actuator 41 Figure 3.20: Shake-Table Control and Data Acquisition System 42 Figure 3.21: Base Saddle 43 Figure 3.22: Bottom Roller 45 vii Figure 3.23: Top Roller 46 Figure 3.24: Centre HSS Beam Roller Layout : 46 Figure 3.25: Kobe Acceleration Time History 48 Figure 3.26: Kobe Displacement Time History 49 Figure 3.27: Response Spectrum of the Kobe Record 49 Figure 3.28: Chile Acceleration Time History 50 Figure 3.29: Chile Displacement Time History 51 Figure 3.30: Response Spectrum of the Chile Record 51 Figure 3.31: VERTEQII Acceleration Time History 52 Figure 3.32: VERTEQII Displacement Time History 53 Figure 3.33: Response Spectrum of the VERTEQII Record [After Telcordia] 53 Figure 3.34: Test Setup 55 Figure 3.35: Instrumentation Layout 59 Figure 4.1: Test UI .S Crack Patterns 64 Figure 4.2: Test U2.5 Crack Patterns 66 Figure 4.3: Test U3.1 Initial Crack Patterns 68 Figure 4.4: Test U3.1 - Initial Cracks and Sliding Behaviour 69 Figure 4.5: Test U3.1 - Southside X-Pattern Begins 69 Figure 4.6: Test U3.1 - Collapse Mechanism 70 Figure 4.7: Test RI .6 Crack Patterns 72 Figure 4.8: Test R1.6 Anchor Failure 72 Figure 4.9: Test R2.4 Failure Mode 74 Figure 4.10: Test R3.4 Crack Patterns 75 Figure 4.11: Test R3.4 Anchor Failure 76 Figure 4.12: Test R4.3 Crack Patterns 78 Figure 4.13: Test R4.3 Anchor Failure 78 Figure 4.14: Test R4.3 FRP Strip Failure 79 Figure 4.15: Test R5.3 Anchor Failure 80 Figure 4.16: Specimen R5 - East View 81 Figure 4.17: Specimen R5 - North View 81 viii Figure 4.18: Test Ul .S Table Displacement 83 Figure 4.19: Kobe Table Displacement 84 Figure 4.20: Chile Table Displacement 84 Figure 4.21: VERTEQII Table Displacement 84 Figure 4.22: Test Ul .S Top Displacement 85 Figure 4.23: Test U2.5 Top Displacement 85 Figure 4.24: Test U3.1 Top Displacement (Partial Record) 85 Figure 4.25: Test R1.6 Top Displacement (Partial Record) 86 Figure 4.26: Test R2.4 Top Displacement 86 Figure 4.27: Test R3.4 Top Displacement 86 Figure 4.28: Test R4.3 Top Displacement 87 Figure 4.29: Test R5.3 Top Displacement 87 Figure 4.30: Test Ul .S Table Acceleration 89 Figure 4.31: Kobe Table Acceleration 89 Figure 4.32: Chile Table Acceleration 90 Figure 4.33: VERTEQII Table Acceleration 90 Figure 4.34: Test Ul .S Top Acceleration 91 Figure 4.35: Test U2.5 Top Acceleration 91 Figure 4.36: Test U3.1 Top Acceleration (Partial Record) 92 Figure 4.37: Test R1.6 Top Acceleration (Partial Record) 92 Figure 4.38: Test R2.4 Top Acceleration 93 Figure 4.39: Test R3.4 Top Acceleration 93 Figure 4.40: Test R4.3 Top Acceleration 94 Figure 4.41: Test R5.3 Top Acceleration 94 Figure 4.42: Test U2.3 Drift Time History 96 Figure 4.43: Test U2.5 Drift Time History 96 Figure 4.44: Test R2.3 Drift Time History 97 Figure 4.45: Test R2.4 Drift Time History 97 Figure 4.46: Test R3.3 Drift Time History '.. 98 Figure 4.47: Test R3.4 Drift Time History 98 Figure 4.48: Test R4.3 Drift Time History 99 ix Figure 4.49: Test R5.3 Drift Time History 99 Figure 5.1: Test U2.3 Force-Deformation Diagram 114 Figure 5.2: Test U2.5 Force-Deformation Diagram 115 Figure 5.2: Test R1.5 Force-Deformation Diagram 115 Figure 5.2: Test R2.3 Force-Deformation Diagram.. 116 Figure 5.2: Test R2.4 Force-Deformation Diagram 116 Figure 5.2: Test R3.3 Force-Deformation Diagram 117 Figure 5.2: Test R3.4 Force-Deformation Diagram 117 Figure 5.2: Test R4.3 Force-Deformation Diagram 118 Figure 5.2: Test R5.3 Force-Deformation Diagram 118 Figure A . l : Test Setup Details 139 Figure A.2: Test Setup Expanded View 140 Figure A. 3: Shake Table Details (1) 141 Figure A.4: Shake Table Details (2) 142 Figure A.5: Shake Table Details (3) 143 Figure A.6: Basebeam Details 144 Figure A.7: Base Saddle Details 145 Figure A.8: Top Beam and Cable Details 146 Figure B . l : Landers Displacement Time History 147 Figure B.2: Landers Acceleration Time History 148 Figure B.3: Response Spectrum of Landers Record 148 Figure B.4: Sherman Oaks Displacement Time History 149 x Figure B.5: Sherman Oaks Acceleration Time History 149 Figure B.6: Response Spectrum of Sherman Oaks Record 150 Figure C . l : Test Ul .S Top Displacment 153 Figure C.2: Test U l . l Top Displacment 153 Figure C.3 Test U1.2 Top Displacment 153 Figure C.4 Test U2.1 Top Displacment 154 Figure C.5 Test U2.3 Top Displacment 154 Figure C.6 Test U2.4 Top Displacment 154 Figure C.7 Test U2.5 Top Displacment 155 Figure C.8 Test U2.7 Top Displacment 155 Figure C.9 Test U2.8 Top Displacment 155 Figure CIO: Test U2.10 Top Displacment . 156 Figure C. l 1: Test U3.1 Top Displacment (Partial Record) 156 Figure C.12: Test R l . l Top Displacment 156 Figure C.l3: Test R1.2 Top Displacment 157 Figure C.14: Test R1.3 Top Displacment 157 Figure C.15: Test R1.4 Top Displacment 157 Figure C.l6: Test R1.5 Top Displacment (Partial Record) 158 Figure C.17: Test R1.6 Top Displacment (Partial Record) 158 Figure C.l8: Test R2.1 Top Displacment 158 Figure C.l9: Test R2.2 Top Displacment 159 Figure C.20: Test R2.3 Top Displacment 159 Figure C.21: Test R2.4 Top Displacment 159 Figure C.22: Test R3.1 Top Displacment 160 Figure C.23: Test R3.2 Top Displacment 160 Figure C.24: Test R3.3 Top Displacment 160 Figure C.25: Test R3.4 Top Displacment 161 Figure C.26: Test R4.1 Top Displacment 161 Figure C.27: Test R4.2 Top Displacment 161 Figure C.28: Test R4.3 Top Displacment 162 xi Figure C.29: Test R5.1 Top Displacment 162 Figure C.30: Test R5.2 Top Displacment . . 162 Figure C.31: Test R5.3 Top Displacment 163 Figure C.32: Test R5.4 Top Displacment (Partial Record) 163 Figure D . l : Test Ul .S Top Acceleration 166 Figure D.2: Test U l . l Top Acceleration 166 Figure D.3: Test U1.2 Top Acceleration 167 Figure D.4: Test U2.1 Top Acceleration 167 Figure D.5: Test U2.3 Top Acceleration 168 Figure D.6: Test U2.4 Top Acceleration 168 Figure D.7: Test U2.5 Top Acceleration 169 Figure D.8: Test U2.7 Top Acceleration 169 Figure D.9: Test U2.8 Top Acceleration 170 Figure D.10: Test U2.9 Top Acceleration 170 Figure D . l 1: Test U2.10 Top Acceleration 171 Figure D.12: Test U3.1 Top Acceleration (Partial Record) 171 Figure D.l3: Test R l . l Top Acceleration 172 Figure D.14: Test R1.2 Top Acceleration 172 Figure D.l5: Test R1.3 Top Acceleration 173 Figure D.l6: Test R1.4 Top Acceleration 173 Figure D.l7: Test R1.5 Top Acceleration (Partial Record) 174 Figure D.l8: Test R1.6 Top Acceleration (Partial Record) 174 Figure D.l9: Test R2.1 Top Acceleration 175 Figure D.20: Test R2.2 Top Acceleration 175 Figure D.21: Test R2.3 Top Acceleration 176 Figure D.22: Test R2.4 Top Acceleration 176 Figure D.23: Test R3.1 Top Acceleration 177 Figure D.24: Test R3.2 Top Acceleration 177 Figure D.25: Test R3.3 Top Acceleration 178 Figure D.26: Test R3.4 Top Acceleration..... 178 xii Figure D.27: Test R4.1 Top Acceleration 179 Figure D.28: Test R4.2 Top Acceleration 179 Figure D.29: Test R4.3 Top Acceleration 180 Figure D.30: Test R5.1 Top Acceleration 180 Figure D.31: Test R5.2 Top Acceleration 181 Figure D.32: Test R5.3 Top Acceleration 181 Figure D.33: Test R5.4 Top Acceleration (Partial Record) 182 xiii List of Tables Table 3.1: Shake-Table Modification Summary 44 Table 3.2: Test Program: Input and Test Codes 57 Table 3.3: Test Program: Simulated Earthquake Levels 58 Table 3.4: Instrumentation Layouts 60 Table 4.1: Peak Displacement Results for the Failure Tests 83 Table 4.2: Peak Acceleration Results for the Failure Tests 88 Table 4.3: Force Results for the Failure Tests 95 Table 5.1: Specimen In-Plane Stiffness Values 113 Table A. 1: Drawing List 138 Table C . l : Complete Displacement Results of U R M Specimens 151 Table C.2: Complete Displacement Results of Reinforced Specimens 152 Table D. l : Complete Acceleration Results of U R M Specimens 164 Table D.2: Complete Acceleration Results of Reinforced Specimens 165 Table E . l : Complete Force Results of U R M Specimens 183 Table E.2: Complete Force Results of Reinforced Specimens 184 xiv Acknowledgements I would first like to acknowledge the contributions to this thesis, and to all of my work at UBC, of my supervisor Dr. Carlos E. Ventura. He encouraged me to take on my work from many different levels and taught me new ways to learn. But most of all, he showed me that the most important thing is to enjoy my work. Secondly, I would like to acknowledge Dr. Steven Kuan, of Public Works and Government Services Canada, Pacific Region. His contributions to the development of this thesis were invaluable. Many thanks. Bill McEwen, of the Masonry Institute of British Columbia, was instrumental in providing the wall specimens for this study. He was the key contributor to their design and construction. J.P. LeBerg, also of MIBC, was also very helpful with the construction of the walls. Ed Fyfe, of Fyfe Co., was responsible for providing and installing the fibre-material for the project, and some aspects of the specimen construction. Pete Milligan, also of Fyfe Co., was helpful in the design of the reinforcing configurations. Scott Jackson, Harald Schrempp, Doug Smith and John Wong, all technicians at UBC, were extremely helpful with test preparations throughout the course of this work. I would like to specially thank Doug Hudniuk, also a technician at UBC, for his assistance throughout the project. His contributions to the experimental part of this project were very important and greatly appreciated. Terry Davies, Carlos Perdomo and Mark Forsyth, all civil engineering students from UBC, deserve special credit for their contributions to this project. Their time and dedication to the experimental work were greatly appreciated. XV Additional lab assistance was provided by my fellow graduate students. Thanks to Rozlyn Bubela, Canisius Chan, Houman Ghalibafian, Mehdi Kharrazi and Jean-Francois Lord. I would also like to acknowledge my parents, Mike and Kathy, and my two sisters, Lisa and Vicki, all of whom gave endless support through this whole project. Lastly, I would like to acknowledge Kristopher Chandroo, without whom, I couldn't have made it this far. And that is an understatement. At the end of the project, looking back, I realize how much work it has been, and how many people put in so much work with me. To everyone, thanks. xvi Project Funding The primary funding for this project was provided by Public Works and Government Services Canada, Pacific Region. Fyfe Co., of San Diego was responsible for provision and installation of the FRP materials, and for provision and construction of the reinforced concrete basebeams used in the tests. The Masonry Institute of British Columbia and The Masonry Contractors Association of British Columbia were responsible for provision and construction of the wall specimens. . Lower Mainland Steel donated the reinforcing steel for the basebeams. xvii Glossary of Terms Bed-Joint - The horizontal joints in the mortar. Bondbeam - A course of CMU's with notched-out webs, which allows for the placement of horizontal reinforcement that is grouted in place. Code-Level Record - An earthquake record that has a response spectrum with frequencies that are close to those of the design spectrum and has an acceleration magnitude scaled close to the code level. Concrete-Masonry-Unit (CMU) - Hollow concrete blocks fabricated with a standard shape, sizes and strengths. Cycle - A discreet portion of an earthquake record defined by two peaks, one in either direction. The cycle begins, passes through, and returns to the zero position of the shake-table. These cycles can have any frequency and amplitude of peaks. Extreme-Level Record - An earthquake record that has a response spectrum with frequencies that are higher (~2x) than those of the design spectrum and has an acceleration magnitude that is scaled higher (~4x) than the code level. Failure - The word "failure" in this thesis is defined as any breach of structural integrity, which may include cracking, debonding of the composite material or a complete collapse. Fibre-Reinforced-Plastic (FRP) - A composite material consisting of glass-fibre strands in an epoxy matrix. These materials are applied externally to U R M walls as reinforcement. Head-Joint - The vertical joints in the mortar. In-Plane - Any action or physical property that is in the long horizontal direction of the walls. Near-Code-Level Record - An earthquake record that has a response spectrum with frequencies that are close to those of the design spectrum and has an acceleration magnitude that is scaled higher (~3x) than the code level. Out-of-Plane - Any action or physical property that is in the short horizontal direction of the walls. Running Bond - A method of laying CMU's that staggers the courses, with no two head-joints directly connected. xviii Sliding Failure - Failure of a mortar joint in a horizontal plane. Stair-Stepping Cracks - Typical shear cracking in masonry that follows the head-joints and bed-joints along a stepped path. Surcharge Load - Load applied to the top of the wall that simulates upper story walls and a portion of the floor slab. Unreinforced Masonry (URM) - In Chapters 1 and 2 of this thesis, U R M is defined as any type of masonry construction without reinforcing steel or with a negligible amount of reinforcing steel. In Chapters 3, 4, 5 and 6 of this thesis, U R M is defined as unreinforced concrete-masonry that has no reinforcing steel. Wythe - A thickness of bricks or blocks; eg. A double-wythe wall is 2 bricks or blocks thick; typically mortared together. xix Chapter I Chapter 1 - Introduction Introduction 1.1 General The economical and effective seismic upgrade of structures is an important issue for civil engineers. There are many structures that can have a continued economic benefit, but are reaching the end of their service life or are no longer safe. There are also structures of historic significance, which have a social rather than economic benefit. Many of these structures are built using unreinforced masonry. Masonry is the oldest building material, and its simple and effective use has made it widespread and popular. Masonry is typically either clay-brick or concrete block. The behaviour of unreinforced masonry during earthquakes is typically poor, and makes it a good candidate for structural upgrading. Because of a large building stock made with these materials and their continued public use, it is desirable to examine simple upgrade schemes that can prolong the life of these structures. One such scheme is the use of Fibre-Reinforced-Plastics (FRP's) in seismic strengthening and repair [Saadatmanesh, 1997]. Public Works and Government Services Canada (PWGSC) is committed to employing state-of-the-art technologies for upgrading their structures. They are responsible for an aging building stock, many of which are located in seismically active regions. This creates a distinct opportunity to explore the use of new technologies [Cheung et al, 1999] to upgrade these buildings, and to achieve two things in the process: 1. Solve a practical problem with the future use of many buildings. 2. Provide a basis for Universities and Industry to push the state-of the-art. It was from these initiatives that a research collaboration was established between PWGSC, Pacific Region and The University of British Columbia. Key sponsors were The Masonry Institute of British Columbia and Fyfe Co. of San Diego, C A , a manufacturer of FRP products. The research is to examine the use of FRP's for the retrofitting and in-plane strengthening of unreinforced-concrete-block-masonry walls. 1 Chapter 1 1.2 Seismic Performance of Unreinforced Masonry Walls Introduction The performance of U R M during earthquakes is variable, and although there have been many recorded observations during past earthquakes, some testing done, and a few basic guidelines set, it is difficult to predict that performance. This is due to many factors, both structural and economic, and this section w i l l provide a brief introduction. This w i l l include a historical review, a summary of structural theory and some of the available guidelines. When looking at the behaviour of this material, it is important to take into account the type of masonry used, and its application. The most common type of U R M found is the solid clay-brick unit. It is used in many different applications, with different importance and hazard levels. A common hazardous application that U R M is found is in the parapets of buildings. It forms relatively high, unreinforced cantilevers of brick that can be located high above street level. There are several other types of U R M , including concrete masonry units ( C M U ) . These units come in many different sizes, strengths, and are used in many different applications and construction techniques. The strength of a C M U system is largely determined by the mortar properties. These mortar types and strengths can vary significantly. The performance of a U R M wall during an earthquake is also determined by its use and construction. A wall can have one of many different functions, such as a load-bearing wall , or an in-f i l l wall between load-bearing columns. U R M walls can also be a single wall , multi-wythe or walls with an attached brick-veneer. These variations can have an effect on the loading on the wall , which wi l l in turn affect the behaviour during an earthquake. For example, it has been observed that in-fil l walls attracted more load than were anticipated and as a result were heavily damaged. 2 Chapter 1 1.2.1 Historic Performance of U R M Introduction Although historically the behaviour of U R M during earthquakes has been variable, it is generally expected and has been observed that it performs poorly. The most common failures reported are to clay-brick masonry, and in the out-of-plane direction. Extensive in-plane damage has been reported, however complete collapses typically occur from out-of-plane motion. The consequences of these failures will depend on the use of the URM. In a clay-brick building that uses concrete columns for its gravity load support, the consequences of an out-of-plane failure of the facade would be in terms of property damage or casualties due to falling debris. In a U R M building that uses the masonry as a load-bearing system, this type of failure can be catastrophic. Studying past earthquakes displays the variability of the performance of the U R M . After the Chilean earthquake of May 22, 1960, a study of the damage was made and documented [Arya, 1996]. Of the 6000 buildings that were studied, 50% of the U R M buildings were damaged, most of them severely. Of the reinforced masonry buildings that were studied, less than 1% were damaged. In contrast, the Armenian earthquake of December 7, 1988 showed that in the most heavily damaged areas, the U R M buildings performed better than the reinforced buildings [Langenbach, 1989]. A majority of casualties occurred due to the failure of those reinforced structures. A common use of U R M is the in-fill wall. It is used in many countries, in varying ways. For instance, the October 9, 1995 Manzanillo, Mexico earthquake showed several completely collapsed in-fill walls, mostly in industrial operations, which reduced the public risk [EQE, 1995]. The June 26, 1998 Adana-Ceyhan, Turkey earthquake showed an example of partial-height U R M in-fills causing pounding damage to the steel gravity-load columns [EQE, 1998]. There was also damage to stone-masonry reported and several chimney collapses. The August 17, 1999 Izmit, Turkey earthquake showed another behaviour of in-fill walls [EQE, 1999]. Due to poor construction quality of reinforced concrete many complete collapses of structures occurred, and little damage was done due to collapse of in-fill walls only was reported. The September 7, 1999 3 Chapter 1 Introduction Athens, Greece earthquake reported many losses of in-fill panels, even though the peak ground acceleration of 0.2g was relatively low [Anastasiadis et al., 1999]. The September 21, 1999 ChiChi, Taiwan earthquake showed structures similar to Izmit, but in this case the concrete frames performed better while the in-fill walls failed [EQE, 1999]. Izmit and ChiChi had similar magnitude, but Izmit showed predominantly large horizontal displacements, while ChiChi was mostly vertical. The January 26, 2001 Buhj, India earthquake again had similar construction and behaviour [Goel, 2001]. The concrete load-bearing columns performed well, while the in-fills collapsed. This was seen in 7 and 8 story buildings, and there was a severe threat of damage to property and persons below. Both of the two significant California earthquakes in recent history, Loma Prieta and Northridge, illustrated the performance of U R M , and the importance of retrofit. The October 17, 1989 Loma Prieta, C A earthquake caused much more damage to U R M buildings [EQE, 1989]. This is mostly due to the type of construction and age of the San Francisco - Oakland region of California. There are many U R M load-bearing buildings, and several of these suffered complete collapses. A common failure of U R M buildings is caused by the disconnection of the floor/roof diaphragm from the walls, which precipitates the collapse. In cases where complete collapses did not occur there was extensive diagonal cracking in the piers. Also it was observed that some buildings that were damaged but did not collapse likely would in future earthquakes. There were also in-fill wall structures that performed poorly. Several photos showing failures of U R M are shown in Figure 1.1. 4 Chapter I Introduction Figure 1.1: Loma Prieta Earthquake U R M Damage [Courtesy of NISEE] The January 17, 1994 Northridge, CA earthquake showed different behaviours with less damage, primarily because there aren't as many URM buildings in the Los Angeles region, and that many of those U R M buildings were built before 1940 and have since been dealt with structurally [EQE, 1994]. In areas surrounding Los Angeles, there are more U R M buildings, but the intensity of shaking was less. Los Angeles is one of the leading cities in the world with respect to seismic retrofit, however there were still some U R M buildings that were severely damaged and collapsed. Some of the collapses that occurred were reported from as much as 40km from the epicentre. An important observation from the Northridge earthquake is that there was severe damage to some retrofitted structures. A common retrofit was to bolt through the masonry, anchoring entire walls at certain points to the rest of the structure. During the earthquake, the walls would buckle in between anchors, and the entire wall would collapse, leaving the anchors embedded in the structure. Other retrofitted structures performed very well. This shows that although retrofitting is important, proper application is equally important. Several photos showing failures of U R M are shown in Figure 1.2. 5 Chapter 1 Introduction Figure 1.2: Northridge Earthquake U R M Damage [Courtesy of NISEE] There were two earthquakes in the Pacific Northwestern United States that have occurred recently. These moderate earthquakes had a more significant effect on the public perception of safety in these regions than it did to the structures. The July 2, 1999 Western Washington earthquakes did not have significant damage, but there were some cases reported [EQE, 1999]. In one of these, a U R M column supporting the roof of a firestation showed spalling, and exposed the anchor bolts. This jeopardizes the roof to wall connection, and further shaking could have collapsed the roof. This is especially significant because it was an emergency facility. The February 28, 2001 Nisqually, Washington earthquake is an example of a local earthquake of moderate intensity, Magnitude 6.8, causing damage to U R M components of structures [MacRae and Lehman, 2001]. Most of the damage observed after the earthquake was to clay-brick masonry and in the out-of-plane direction. Many parapet failures occurred, with brick material collapsing onto the street below. Although there were no casualties caused by the falling debris, it posed a significant life-safety hazard. There were instances where walls failed more completely, and damaged structures to an extent that they were braced for stability of the exterior facade. Diagonal cracks were also observed in some structures, and one example is in the Governor's mansion in Olympia, Washington. Several photos showing in-plane failures of U R M are shown in Figure 1.3. 6 Chapter 1 Introduction Figure 1.3: Nisqually Earthquake U R M Damage [courtesy of EQE] The real significance of the Nisqually earthquake is that it showed the poor performance of U R M during a moderate earthquake with a deep focal point. While there were recorded ground accelerations of 0.3 lg, it did not exceed 0.1 Og in most places. This region is susceptible to stronger ground motions, as is the Southwest British Columbia region. The damage observed could be an indication of what could happen in a much larger event. 1.2.2 Structural Performance of U R M This section introduces some of the key structural aspects that can influence the in-plane behaviour of U R M walls that has been observed in past earthquakes. The structural design, and construction quality are both major factors in U R M seismic response. With regard to the design of the walls, there are several variables that can affect the behaviour. One of these is the aspect ratio of the wall, which determines the primary mode of behaviour. This behaviour could be rocking, flexure, shear or a combination. Another is the block and mortar type, which can influence behaviour by their strength, ductility and durability. Another is the intended use of the wall, such as load-bearing walls or in-fill walls, which can influence behaviour in several different ways. As mentioned in Section 1.2.1, it has been seen in past earthquakes that in-fill walls took 7 Introduction Chapter I more load than they intended to and as a result were heavily damaged. This had been observed with partition walls also. The construction quality variables have a higher uncertainty and create difficulty in predicting the behaviour of U R M walls. This is especially true with older structures. The construction quality of a building comes into effect during both the fabrication and the installation of the materials. Mortar fabrication and block fabrication have different degrees of quality control, since mortar is manufactured on site, and blocks are typically pre-fabricated. Although most modern mortar types will have a minimum strength of 10 MPa and is tested frequently on-site, this is not always the case, and has often been not historically. The installation of these materials also affects the behaviour. There are various construction techniques performed, as well as different skill levels of tradesperson. The Canadian Masonry Design Code, S304.1 1994, is presented with a design handbook as part of the text Engineered Masonry Design [Glanville et al., 1996]. The remainder of this section summarizes the theory of in-plane behaviour of U R M , using excerpts and concepts from the handbook: The handbook describes the in-plane resistance of U R M according to two separate components, flexure and shear. The bending resistance of the wall is determined by the applied axial load, P. The flexural forces will create a tensile force in part of the wall, and the dead load must be large enough to prevent cracking in this region. Whether a wall is load-bearing or not, the effect of in-plane bending is to create a bearing condition. The result is that the resistance is dependent on its axial resistance, in the following way: Pf/L-6Mf/L2 >= 0 (1.1) This assumes a U R M wall where there is no tensile resistance created by the mortar strength. 8 Chapter I Introduction In the case of pure shear, the consideration for the bearing forces caused by bending are disregarded. The dead load, in compression, still plays the key role in determining the shear resistance. Clause 11.5.3 of S304.1 1994 considers the shear resistance to be: Vr = q>m(vmbd+0.25P)yg (1.2) A second term would be applied to the above equation for reinforced masonry to account for the reinforcing steel. The shear strength of masonry, rjm, is determined with the following expression: vm = 0.16 [2 - Mf/ (Vfd)] (f'J1/2 (1.3) As described in Section 1.2.3, there are characteristic modes of failure for these different loadings. In shear, there is a typical diagonal tension crack pattern. As in concrete behaviour, the shear load is largest in a 45° plane across the material. The associated tensile forces that develop will instigate the cracking. Unlike the 45° cracks that appear in concrete, masonry tends to show a stair-stepped crack along the mortar joints. This is due to the crack following the path of least resistance. This diagonal stair-stepping is shown in Figure 1.4(a). If the loading is not pure shear, a sliding failure in the bed-joints may happen at the onset of the diagonal cracking. In this case, the shear-sliding resistance of the wall becomes important. This is stated in the masonry handbook as: Vr = <pmfiC (1.4) Where C is the compressive resistance of the wall and p is a factor between 0.7 and 1.0 depending on the material contact surface. The behaviour of a diagonal crack can be exaggerated when the load is purely transverse, and sliding effects are present. The crack shown in Figure 1.4(a) shows the exaggeration caused by sliding. Flexural cracking is typically observed as a horizontal crack along part of the wall. It can be mixed with a diagonal crack, because the tensile force caused by bending will vary across the width of the wall, and the compression force varies across the height of the wall. The wall cracks at the locations where the tensile force overcomes the resistance. This type of cracking can be confused with shear, but it is typically characterized by long horizontal cracks, with smaller diagonal sections. An example of this is shown in Figure 9 Chapter 1 Introduction 1.4(b). The figure is shown with a single step, to emphasize the variation of resistance across the width of the wall, and as a result varying the height of the flexural cracking. The rotation of the wall is exaggerated for clarity. length —> (a) Stairstepping (b) Flexural Cracking (c) Forces in the Wall Figure 1.4: In-plane Behaviours of U R M Walls The forces in the wall caused by bending and compression are shown in Figure 1.4(c). The tension due to the applied moment M varies along the length linearly, and that compression due to axial load P varies linearly with height. These stresses are shown by the lines on the figure. The wall can crack anywhere the tensile force becomes larger than the available compressive resistance. In a theoretical case, it would have to be assumed that the bending forces are equal throughout the height of the wall. This would not be the case when shear is also applied. From the handbook, the distribution of forces and the resulting stress in a purely in-plane loaded wall are as shown in Figure 1.5. Forces Stresses Figure 1.5: Typical In-Plane Forces and Stresses 10 Chapter 1 Introduction The in-plane response of U R M is also dependent on the characteristics of the loading. Although a U R M wall is strongest in-plane, the loads do not usually act only in this direction. Vertical loads pulling upwards can put tension on the mortar along the entire bed-joint (Figure 1.6a); out-of-plane loads put bending forces on the wall, which puts tension on part of the bed-joint (Figure 1.6b) and in-plane loads can create bending or shear, which can put tension on head-joints and bed-joints, and shear' in the mortar (Figure 1.6c). M V TTT TTT TTT TTT hi (a) Tension due to Vertical'Loads (b) Tension due to Out-of Plane Loads (c) Tension due to In-Plane Loads Figure 1.6: Tensile Forces in Mortar Joints 1.2.3 Guidelines for Behaviour of U R M Based on field observations, such as those mentioned above, and on the theoretical information available, several guidelines for the behaviour of U R M have been developed. These guidelines deal with many issues, including the assessment of U R M and its potential hazard, and with its strengthening/rehabilitation. Both of these issues are addressed in two documents by ABK, which were produced for the National Science Foundation (USA). These documents are part of a set of guidelines entitled Methodology for Mitigation of Seismic Hazards in Existing Unreinforced Masonry Buildings. They are split into two sections, Categorization of Buildings [ABK, 1981] and The Methodology [ABK, 1984]. Much of the work that followed has been based on these guidelines. Recent work has been done by the International Conference of Building Officials (ICBO), which collaborates to develop building codes, and helps individual governments 11 Chapter 1 Introduction adopt these codes. They have developed a set of Guidelines for the Seismic Retrofit of Existing Buildings [ICBO, 2000]. The first chapter of these guidelines deals with seismic strengthening of U R M bearing wall buildings. The provisions set forth in the chapter are a "minimum standard for structural resistance". The intention is not to mitigate economic loss but to reduce the risk of life loss or injury. These guidelines set forth a set of definitions, similar to the A B K documents, which include the following: • Definition of the observations of existing masonry • Specifications for testing of existing masonry • Requirements for construction • Analysis and design procedures • Detailed system requirements, including anchorage, non-structural masonry and adjacent buildings These guidelines are designed to be adopted by individual municipalities and are not a part of the Uniform Building Code. The Federal Emergency Management Agency (FEMA) in the US has developed extensive guidelines, including NEHRP Guidelines for the Seismic Evaluation of Existing Buildings [FEMA, 1992] and NEHRP Guidelines for the Seismic Rehabilitation of Existing Buildings [FEMA, 1996]. More specifically, they have developed guidelines in categorizing the different U R M types and behaviours. These are available in F E M A 306 [ATC, 1998], 307 [ATC, 1998] and 308 [ATC, 1998]. Due to the frequency of earthquakes in California, F E M A has had much experience in observing the behaviour of U R M components to seismic events. F E M A 306, Evaluation of Earthquake Damaged Concrete and Masonry Wall Buildings, defines different wall components and behaviours in discreet units. It categorizes walls according to five types, as well as three different categories of ductility. The behaviours described are dependent on the ductility of the system and are based on field observations and experimental research. While U R M walls are the weakest in the out-of-plane 12 Chapter I Introduction direction, there are several failure modes shown in a wall that is loaded in-plane. Rocking of the wall, partially or completely is a common behavioural mode. Depending on the severity and nature of the rocking, this can have varying levels of contribution to the wall damage. If the rocking is continued for several cycles of loading, toe crushing can result, which would form a more severe level of damage in a rocking mode. Another common mode is bed-joint sliding. This type of behaviour exhibits sliding along the bed-joints, either completely in a horizontal plane, or along diagonals with the head-joints cracking and opening. Bed-joint sliding can also occur along a single course. This mode is dependent on the fixity of the wall to the base, and on the weight of the wall and the resulting friction forces at the base. Both of these sliding modes can result in significant damage and failure of the wall. Flexural behaviour constitutes another set of modes in walls that are loaded in-plane. This type of behaviour can result in several different combinations of behaviours, such as toe crushing, diagonal tension cracking and bed-joint sliding. This behaviour is related to the rocking mode, except that rocking does not occur and the forces instead are distributed through the toe into the wall. This type of cracking can result in triangular sections of the wall coming loose from the rest of the wall, and these can collapse. Other useful guidelines include a set created recently by PWGSC. They have developed several sets of similar guidelines including: Procedure for Seismic Assessment of Existing Buildings [PWGSC, 1998], Guidelines for the Seismic Upgrading of Stone-Masonry Structures [PWGSC, 2000] and Guidelines for the Seismic Assessment of Stone-Masonry Structures [PWGSC, 2000]. 1.3 Use of Fibre-Reinforced-Plastics for URM Reinforcement To improve this poor behaviour of the U R M walls, many retrofit techniques have been applied in the last several decades. One common method is to physically insert or apply externally steel reinforcing. Another technique is to brace entire walls or components externally. It is also possible to cover the face of a wall with shotcrete or a similar 13 Chapter 1 Introduction material. These methods tend to be expensive and labour intensive, and the effectiveness of these methods has been questionable. The use of composite material in external retrofit applications is a more modern technique, and it is introduced in this section. The use of FRP's for U R M reinforcement is becoming a popular, effective and relatively economical method of seismic upgrading. It has been used in different forms and applications for over 40 years [Triantafillou, 1998]. It has been applied to many structures already, and some extensive studies of the material and some of its applications have been performed. A review of some of the previous testing involving FRP reinforced U R M walls is presented in Chapter 2. FRP's are a composite material that is made up of a fibre/epoxy matrix. The fibre sheets are composed of either glass, aramid or carbon fibres. These fibres are weaved into sheets and the structure of the sheets varies depending on their intended use. The fibre can be uni- or multi-directional. The material is also available in different types of sheets, sizes and thickness. This material is useful based on its defining characteristics, such as the strength. The fibres have a very high tensile strength, and can be used in high-load structural applications. Another important characteristic is the ductility. The modulus of elasticity in different types of FRP material varies, but in general it can experience significant deformation before failure. This is useful in reinforcement of brittle materials, where a sudden high-displacement may occur. Another important characteristic is the durability of the material. Two aspects of the durability are the fire protection and the resistance to environmental degradation. The material has a well-developed fire protection standard [Fyfe, 2000], and this varies among different products. It has excellent resistance to corrosion and environmental degradation [Fyfe, 1998]. This makes it a useful reinforcement material for exterior applications. The application of the material can be accomplished in different ways. Depending on the clearance available the material can be applied with a machine, or it can be applied by 14 Chapter 1 . Introduction hand [Fyfe, 1999]. In both cases, the surface of the U R M must be prepared for application. If wall is damaged, the cracks and openings must be injected with epoxy. In all other locations, the mortar joints need to be filled to provide a smooth surface. The block faces are saturated with epoxy for a short period of time prior to application. The fibre must also be saturated with epoxy. This is done either by hand, or by using a saturation machine. In applications where a large amount of FRP is to be applied, the machine is necessary to properly saturate the material. Once the preparation is complete, the material can be placed on the wall. It has a curing time of between 3 to 7 days, depending on the ambient temperature. The practical aspect of the material application is very important. Since access is often restricted due to building services, such as plumbing, electrical or mechanical, the placement of the material in the optimal structural location is not always possible. Also openings such as doorways and windows can cause discontinuities in the FRP/wall system. This must be taken into consideration during the design process. The concept of reinforcing U R M walls with FRP's is similar to reinforcing with steel. Since mortar, and hence U R M , is weak in tension, the FRP's are used to take the tensile forces. The difference between the two is that the FRP's are applied externally, on the exterior of the material. This can have both a positive and negative effect. The material would provide better containment; hence better ductility, as in the case of concrete columns. Aesthetically however, especially with clay-brick masonry, an external application can be undesirable. It is important to note that the application of FRP reinforcement on U R M components creates a structural system. Although the FRP's have a high tensile strength, they have negligible resistance to compression. Therefore in order for the system to work, the U R M must be able to take the compression forces. The FRP's are bonded to the external surface of the wall. They act to resist loads in tension, although they can do this in different ways. When the load is primarily out-of-plane, the wall experiences flexural forces. Part of the blocks, on the side opposite to the direction of loading, go into compression. The rest of the blocks go into tension. Since 15 Chapter I Introduction they will not take the tension without cracking, the load is transferred into the FRP's. This has the effect of holding the wall down. When the load is primarily in-plane, the wall experiences shear, sliding and overturning forces. Some of the blocks will tend to lift from the bed-joint, and some will crack at the head-joint and then slide along the bed-joint. In all cases where the mortar would go into tension, the FRP's take the load. In this case either a multi-directional fibre is used, or multiple layers of FRP's with the fibres oriented in different ways. These two methods of reinforcement are common, and the application of FRP's will differ between them. The use of FRP's for reinforcement has been performed mainly in the out-of-plane direction. It is also useful as a means of reinforcement in the in-plane direction of the walls. 1.4 Objectives of Thesis The purpose of this research project, in general, was to examine the performance of FRP strips applied as reinforcement to U R M walls when subjected to in-plane seismic loading. The specific objectives were: 1) To examine the behaviour of these walls subjected to code and near-code level simulated earthquake motions. It was expected that the reinforced walls would perform well during these simulations, and the purpose of this experiment was to add further proof of the benefits of using FRP's for seismic retrofit. 2) To examine the behaviour of these walls when subjected to a much higher extreme-level loading. The purpose of this experiment was to examine the failure modes and behaviour of various FRP reinforcing configurations and to examine the life-safety issues that are involved with U R M and certain retrofitting techniques. 16 Chapter I 1.5 Scope of Thesis Introduction This thesis was limited to the in-plane testing of eight full-scale unreinforced concrete-masonry walls, subjected to simulated earthquake ground motions generated by a shake-table. Each of the walls had identical dimensions, with the only variable between them being the reinforcing configuration. Three of the walls did not have FRP reinforcement, and were used as the benchmark for the U R M behaviour. One of the walls was subjected to an impulse load, one was subjected to the code-level record, and one was subjected to the extreme-level record. Five of the walls were reinforced using five different FRP configurations. The reinforced walls were subjected to a series of at least three earthquakes. The first two were the code level near-field record and the stronger subduction-type record. The walls were then subjected to the extreme-level record, which has a much higher peak acceleration value. This record was applied repeatedly until failure was observed. 17 Chapter 2 Chapter 2 - Literature Review Literature Review 2.1 General The use of FRP's for structural upgrades has been in use for many decades. Over this time the material has evolved significantly, and is now produced by many different companies. To further develop the material and to increase awareness of its use, these companies have sponsored various testing projects. Typically the testing is performed at universities and is in collaboration with governments or structural engineering organizations. Many of the tests present results on the evaluation of specific products, but some of them present results on the use of FRP's in general. This chapter will review some of these tests. From the literature review performed in this thesis, it has been found that much of the testing has been done on clay-brick masonry, and in the out-of-plane direction. This corresponds to the fact that most of the damage during earthquakes has been observed from these situations. There have been some tests performed in the in-plane direction, but most of these have been on clay-brick masonry walls. There are also several different types of tests performed, such as slow cyclic tests, which form the majority, pseudo-dynamic tests, and shake-table tests. Although the focus of this thesis is on in-plane testing, a short summary of some out-of-plane tests has been included in this chapter. This is due to the direct correlation between the in-plane and out-of-plane effects during an earthquake. 2.2 Cyclic and Static Testing An in-plane cyclic/push-over test was performed on two clay-brick specimens [Reinhorn and Madan, 1995] at the University of New York at Buffalo. This test was a product evaluation of the Tyfo W Fibrwrap system. Of the two specimens, one was unwrapped and used as a benchmark, and the other was wrapped, with a different configuration on 18 Literature Review Chapter 2 each side. One side used two vertical strips, and the other was coated with a continuous fibre web. The specimens consisted of a double-wythe of standard clay-bricks, and had dimensions of 1800mm wide x 1750mm high. The test setup had a specimen mounted on a steel beam, with bearing plates at either end for shear transfer. It also had a similar steel beam at the top, held down at either end by pin-ended rigid links, to keep the top beam level and transfer shear more effectively. The top beam was connected to the hydraulic actuator. Axial load was applied to the wall using two dywidag bars, one on either side of the wall. The wall was instrumented with a series of displacement transducers, measuring its in-plane and out-of-plane movements. The applied force of the actuator was measured by a load cell. The specimens were first subjected to small displacements in a sinusoidal pattern to determine the cracking loads and formation. The specimens were then subjected to a push-over test to determine the in-plane resistance in one direction, and then the loading was reversed to determine the crack widths. Both of the specimens showed a dominant shear failure mode. The unreinforced specimen failed initially along its diagonals. The cracking increased until a large diagonal crack opened and the resistance dropped suddenly. In the case of the wrapped specimen, the failure occurred first on the full wrapped side, then on the banded side. The failure was initiated with a debonding of the material near the diagonals, up to a maximum width of 400mm. The fibre then began to tear in a stepping pattern, along the upper edge of the debonded region. The failure of the strips was not as dramatic, but did have a concentration of debonding near the diagonals of the specimen. The improvement of the ultimate strength of the walls with FRP added was 120%. The improvement in deformations was 150%. The wrapping was observed to not completely prevent cracking, but it did control the cracking, and prevented total shear failure. It was also observed that the material failed in tension, and did not completely debond. The debonding reduces the effectiveness of the FRP reinforcing. It also worth noting that the 19 Chapter 2 Literature Review full wrapped side failed first. This is presumably because it takes more load across its entire area. A series of in-plane cyclic tests were performed on several masonry wall specimens [Weng et al., 2001], to examine the lateral load resisting and deformation capacities. The walls tested were made of clay brick and strengths were chosen to reflect common construction in the Shanghai region. The study was carried out in comparison to the specifications of the Chinese Design Code, and found that the maximum capacity of the walls was about 2 to 4 times the calculated capacity according to that code. As a part of this study, the walls were pushed until cracking was observed, and then they were repaired and loaded again. The results showed that the repaired walls could recover the original capacity. Additional "static" testing has been performed. One study was on GFRP reinforced brick specimens [Almusallam et al., 2001]. This study observed a significant increase in strength in all of the reinforced specimens. Damage was observed in the anchors, and not in the fibres themselves. Several shear tests were performed with many variables in the reinforcing [Ehsani et al , 1997], examining effects of type of fabric, orientation and anchorage length. It displayed the typical increase in overall ductility and strength due to the FRP application, and found that although the angle of the fibres affected the stiffness in the specific direction, the ultimate strength was not. A series of small scale walls were subjected to four-point loading tests [Triantafiliou, 1998]. It found that in cases where the axial loads on the specimens were low, the shear capacity of the reinforced specimens was much higher than expected. A series of static tests were performed, at the Mid-America Earthquake centre [Moon et al., 2001] on a two story full-scale U R M building. The test project involved comparisons between two types of reinforcement, FRP and vertical prestressing. The testing involved crack testing, repairs and retesting, as well as tests to ultimate failure. The test was done as part of an extensive analytical modelling program, which exposed many difficulties in the modelling process. 20 Chapter 2 2.3 Pseudo-Dynamic Testing Literature Review A full-scale three-dimensional U R M building was tested under a simulated earthquake loading [Paquette, 2002] at the University of New York at Buffalo. The loading applied is called "pseudo-dynamic", applied by a series of hydraulic actuators at various locations on the building, in the two horizontal directions. Each actuator is capable of applying a different load, which allows for multi-directional shaking. Also this loading is applied at slower speeds than in a shake-table test, which allows for easier observation of the results. This test was to determine the behaviour of U R M in a complete building, and examine the connections to a wood diaphragm. The building was loaded until significant cracking was observed in various locations. Then the wall was repaired with fibreglass strips, intended to improve the rocking behaviour of the walls. During the U R M tests, cracking began at an acceleration of 0.15g, with many bricks falling out at 1.5g and a complete failure at 1.75g. The Tyfo reinforced specimen had some debonding at 2.0g, but the wall remained intact. 2.4 Shake-Table Testing An in-plane shake-table test was performed on a concrete block wall [Al-Chaar and Hasen, 1998] at the U.S. Army Construction Engineering Research Laboratories. The objective of this test was to test the effectiveness of using the Tyfo fibre system when it is used in one-sided applications. The test setup involved two U R M walls, mounted on the shake-table, connected at the top by a concrete slab. The outside face of one wall was wrapped using the Tyfo web material. Out-of-plane tests were run first; and these damaged the unreinforced wall. For the in-plane test, the damaged wall was repaired with the FRP sheet. The test was then on two reinforced walls, one previously damaged and one not. The specimens were built using 200mm CMU's. The dimensions of the specimens were 2800mm wide by 1800mm high. They were spaced 2500mm apart, centre-to-centre. The walls were completely unreinforced, only with vertical dowels at the top and bottom. This resembles classic construction details. The walls were 21 Chapter 2 Literature Review instrumented with an array of displacement transducers, accelerometers and strain gauges. The earthquake time histories that were used were synthetically generated. These records were designed to fit the Tennessee Valley-wide design response spectra. The tests were intended to simulate the behaviour of concrete-block switch-house structures. During the applied in-plane loads, the walls performed very well. The only significant behaviour that was observed was an uplift motion of the concrete slab, proving that the weakest part of the structure was the construction detail between the wall and the slab. One conclusion of the testing was that while the excellent performance of the system could be attributed to the FRP reinforcement, the heavy slab also contributed significantly. Another conclusion is that while the FRP reinforcement causes a significant improvement in the performance, proper anchoring details must be imposed to achieve optimal results. 2.5 Out-of-Plane Testing Static and cyclic out-of-plane tests are the most common. These can be done simply with a hydraulic jack applying point loads to the wall. A simple evaluation of the Tyfo-W Fibre system was performed on a clay-brick masonry wall specimen [Reinhorn and Madan, 1995] at the University of New York at Buffalo. Three specimens were tested, an unreinforced wall and two reinforced walls. One specimen was reinforced with two layers of fibre, the other was reinforced with four. One side of the wall used two vertical strips and the other used the web fabric to cover the entire face. The load was applied to the wall by two rollers at its one-third points. The results indicated that both types of fabric reinforcement improved the "cracking" performance. The addition of fibre helped control the cracking, and the results suggested that this could help with the reduction of uncertainty in crack strength prediction due to the variability of construction quality. The failure of the fibre occurred in two stages: debonding from the U R M and then tearing of the material. 22 Chapter 2 Literature Review Another testing program was applied to 25 wall specimens [Tumialan et al., 2002], 12 that were concrete block walls and 13 that were clay-brick. The walls were reinforced using both glass-fibre and aramid-fibre strips. The test examined different surface preparation methods, and the resulting effects on the compressive strengths. Small specimens were simply supported on their face, and loaded by a vertically mounted cylinder in the centre of the wall. The FRP strengthening was done by the application of a single strip, and the width of the strip was varied. These tests illustrated 3 failure modes of the FRP/URM system: debonding of the FRP strip, a flexural failure, either through tearing of the fibre of crushing of the block, and a shear failure in the masonry near one of the supports. The debonding failure that was observed in both of these tests is an example of when the system breaks down. The FRP requires the connection to the wall for stability, and the system also requires bonding to transfer the tensile forces across the weak mortar joints. The test also showed a substantial increase in strength and "pseudo"-ductility of these walls. Another observation was that the maximum usable strain of the FRP's could be increased by as much as 33% through proper surface preparation. Another cyclic test was performed at the University of Alberta [Kuzik et al , 1999] on a series of full-scale masonry block walls, reinforced with FRP strips on both sides. Eight walls were tested, each laterally supported at the top and bottom with two lateral loading points in between. It was found that the flexural stiffness of the wall had more than doubled with the addition of the FRP strips. A shake-table test was performed on a concrete block wall [Al-Chaar and Hasen, 1998] at the U.S. Army Construction Engineering Research Laboratories. Two walls were mounted on the shake-table at once; one of these was reinforced with Tyfo Fibrwrap (one side only) and the other was left unreinforced. The test procedure used synthetically generated time histories, and was applied by multiple tests, each with increasing levels of acceleration. The two walls were vertically loaded and connected together by a concrete slab that spanned between them. The only significant failure that occurred during the out-of-plane tests was in the slippage of the block faces between the first and second courses. 23 Chapter 2 Literature Review This displayed a high out-of-plane resistance for the system, which was not expected. The single FRP sheet was essentially the only lateral support for both walls. All of these tests show the primarily flexural behaviour of the FRP/URM system in the out-of-plane direction. One important mode of failure is in the crushing of the blocks. With a very high-strength composite material, the crushing of the blocks, hence diminishing the strength of the system, can be the governing factor. Other out-of-plane testing includes a set of tests on six brick specimens [Almusallam et al., 2001] in which GFRP laminates were added to improve on shear and flexure behaviour. A significant strength increase was observed in all specimens, with a compression failure of the bricks being most the most common mode. A different test was performed on a set of larger scale walls, using GFRP and CFRP laminates [Mosallam et al., 2001]. It also examined varying thickness of laminates and orientations of the fibres. It proved that significant strength was gained even with only a few laminations. A set of larger concrete masonry walls, up to 4.7 m high were tested to capacity [Hamilton and Dolan, 2001]. A simple analytical method was applied to the GFRP reinforced walls, and it was found that the equations overpredicted the capacity of the test specimens, but by no more than 20%. Fifteen large-scale wall panels were tested to examine the behaviour of web fabric as well as uni-directional strips [Hamoush et al., 2001]. A lateral load was uniformly applied to each specimen using the A S T M E-72 airbag methodology. It found that the flexural capacity could be increased if the shear failure was controlled. Another study was performed on several half-scale models, [Ehsani et al , 1999] proving that these specimens, when reinforced, could sustain loads of up to 32 times their self-weight, and experience strains of up to 2%. 24 Chapter 3 ; Chapter 3 - Shake Table Testing Shake Table Testing 3.1 General This chapter will discuss all aspects of the shake-table testing, including the specimen design, test setup and test program. The experimental aspect of this thesis involved a significant amount of design of the test setup, and also preparation of the test program. Shake-table testing has been in use for many years and has many well-established methods. Full-scale testing, however, is not as common and there is a specific set of difficulties associated with it. 3.2 Specimen Design This section will present the design of the specimen including the intention of the design, its details, construction and details of the FRP wrapping. Detailed drawings of the test setup are presented in Appendix A. 3.2.1 Description of Specimen The objective of this project was to test single masonry walls subjected to simulated earthquake loadings. The intention was to simulate full-scale shear walls subjected to shear and flexural forces. The specimens used in the testing program are single wall units 2400mm high by 3000mm long by 200mm thick. These dimensions were chosen to simulate a wall of a typical room height. The length of the wall was governed by the resulting overall weight of the wall, which was limited by the lifting capacity of the laboratory equipment. This gives them an aspect ratio (height to length) of 0.8. They are made of 400mm long by 200mm thick standard C M U blocks, using running bond. The walls were to be tested in-plane, with restraint from out-of-plane motion. 25 Chapter 3 3.2.2 Design Details Shake Table Testing Each wall was constructed on a reinforced concrete basebeam. This was done to provide a base for each wall, which would be used to lift the walls onto the shake-table. These basebeams are 3200mm long by 600mm wide by 300mm in depth. The basebeams have vertical steel dowels, which are tied into the reinforcement cage and extend upwards 200mm, which is the height of one course of blocks. They are located at three locations, the two end blocks and one in the centre, and are intended to prevent sliding failure along the base. The basebeams also have four steel hangers used for lifting the beam and the wall. The beams were designed in accordance with CSA Standard A23.1-94, and structural details are presented in Appendix A. Figure 3.1 shows the wall details. r 800 —c»j 25nn THREADED .ROD NORTH gPOnrv CMU BLOCKS R/P BASEBEAM 3.eop Figure 3.1 Wall Design Details 26 Chapter 3 Shake Table Testing. At the top course of each specimen is a bondbeam. The bondbeam allows for the embedment of four 25mm threaded rods, which protrude up between 300mm and 550mm above the wall. The bars allow the surcharge load (Section 3.5.1) to be fastened to the wall, transmitting the shear load. The rods are tied to two horizontal bars, which run the length of the bond beam, and grouted with 30 MPa concrete. All of the descriptions throughout this thesis will make reference to the east face of the wall, which is shown in the figure, unless otherwise noted. All of the dimensions given in this thesis are in mm. 3.2.3 Construction The basebeams were constructed first, on the strong floor of the Structures Lab at UBC. The reinforcement cages were tied and placed into prefabricated plywood forms according to CAN-CSA A23.1. The beams were then poured with a minimum strength 30 MPa concrete. They were cured for a minimum of 10 days before the forms were stripped and then moved for blocking. A photo of the construction of the basebeams is shown in Figure 3.2. Figure 3.2 Basebeam Construction 27 Chanter 3 Shake Table Testing Due to space restrictions, two of the walls were constructed on the shake table, and four were constructed on the floor of the lab. They were built using standard 30 MPa blocks, using standard running bond. The walls used 12 MPa Type-S mortar, which was mixed on-site. Mortar cubes were made for each day of construction and tested according to CSA Standard A l 79-94 using the UBC facilities. A photo of the construction of the walls is shown in Figure 3.3. Figure 3.3 Wall Construction The walls were built in two groups, six in the first and two in the second. This was done because at the time the second set was ready for testing, four walls in the first set had already been tested, and those basebeams were available for reuse. The walls were cured for at least 28 days before testing. 28 Shake Table Testing Chapter 3 The fibre wrapping was done in three groups, the first two wrapped walls, then the second two, then the final wall. The fibre wrapping was allowed to cure for a minimum of 3 days before testing. A photo of the first wrapping is shown in Figure 3.4. Figure 3.4 Fibre Wrapping 3.2.4 FRP Materials and Layout This section describes the five different layouts of the FRP reinforcing. All of the U R M walls were built exactly the same; it was only the fibre-wrap layouts that varied. For each of the different layouts, there is a description of the layout, including its intended purpose. There is also a diagram showing the wrap layout, and a photo of the specimen before the test. The material used for the tests was the Tyfo SHE-51A fibreglass sheet, and it was applied with Tyfo S Epoxy. The composite is a custom-weave, uni-directional fabric using glass fibres. The glass fibres are oriented in the sheet strong direction, with additional "yellow" 29 Chapter 3 Shake Table Testing glass fibres oriented in the weak direction. The sheets are cut into strips for the tests, and the strong direction is oriented in the long direction of the strips. The ultimate tensile strength of the composite is 575 MPa in the strong direction and 20.7 MPa in the weak direction. It has a tensile modulus of 26100 MPa. The typical thickness of the composite with the epoxy matrix is 1.3mm. All of the strips that were used in these tests were 300mm wide. Specimen RI The first specimen was reinforced using two vertical strips, shown in Figures 3.5 and 3.6. This was a simple layout, intended to increase the flexural stiffness of the wall. V /// •'// / '••"/ // / / y / m // / / V V. * '•I '// / Y -/A ft / / / Figure 3.5: Specimen RI FRP Layout 30 Chapter 3 Shake Table Testing Figure 3.6: Photo of Specimen RI It was also intended to examine the shear resistance of the strips applied vertically. The strips were anchored to the basebeam using Tyfo fibre anchors, shown in Figure 3.7. Tyfo fibre anchors are made of a fibre rope, one end of which is placed into a hole in the base and epoxied, the other is splayed out and bonded to the FRP. The reinforcing layout was repeated on both sides. 31 Chapter 3 Shake Table Testing SPLAYED ANCHOR -~JS (TOPVIEW) ' * F*RP STRIP Figure 3.7: Tyfo Fibre Anchor Installation Specimen R2 The second specimen was reinforced using 8 strips laid horizontally on the face of the wall, shown in Figures 3.8 and 3.9. This layout was intended to have a much higher shear resistance. At the bottom of the wall, the strip was laid so that half of its surface was bonded to the basebeam, and half of its surface was bonded to the wall. The strips were anchored to the basebeam using the Tyfo fibre anchors. The layout was repeated on both sides of the wall. 32 Chapter 3 Shake Table Testing Figure 3.8: Specimen R2 FRP Layout 33 Chapter 3 Specimen R3 Shake Table Testing The third specimen was reinforced with three vertical strips, shown in Figures 3.10 and 3.11. There were three strips placed on each side. This was intended to examine the effect of spacing of the strips on the behaviour of the wall. The layout was similar to that of Specimen RI, with a centre strip added. The strips were attached to the base using Tyfo fibre anchors. This layout was repeated on both sides. / / * * 'fr/ / . t fr m fr * • // // // Z / •fry * fr' '//. fr; * '//. '/ / fr * fr /// 'fry' * / / / / < fr/ > V/, fry / fr, * '//. fr * fr Figure 3.10: Specimen R3 FRP Layout 34 Figure 3.11: Photo of Specimen R3 Specimen R4 The fourth specimen was reinforced similar to Specimen RI, utilizing two vertical strips on either end. This layout is shown in Figures 3.12 and 3.13. This wall differed from RI in two ways. The first is that the wall had been cracked before the test, horizontally along the top of the 3 r d course. This allows for the examination of the behaviour of the FRP's to a damage-repair application. It is very common for walls in practical situations to have existing cracks. The second difference is in the type of anchors used. This wall used steel plate anchors, which are laid upon the strip at the base, and bolted through into the base. The bolts used are Hilti™ HSL Heavy Duty Sleeve Anchors. They have a diameter of 8mm with a maximum plate thickness of 20mm. They have an allowable tensile strength in 30 MPa concrete of approximately 8 kN. The anchor detail is shown in Figure 3.14. 35 Chapter 3 Shake Table Testing <4 / • fr '//. '// * /A • y/ // y ,- 'fr, '//. / / / / / / y/. • fr, y/y • • fr Yy • '//. \ fr/ 'A • \ • 1 -fr Pre-tes-t Crack Loca-tion Figure 3.12: Specimen R4 FRP Layout Figure 3.13: Photo of Specimen R4 36 Chapter 3 Shake Table Testing URM WALL STEEL PLATE N. FRP STRTP ftNCHDR BD.LT' •A' f •A Figure 3.14: Anchor #1 Details (Specimen R4) Specimen R5 The fifth specimen was reinforced using a series of single strips, laid out in an X-pattern, with vertical strips on either end and one horizontal strip across the top of the wall. The layout is shown in Figures 3.15 and 3.16. This wrapping was only applied to one side of the specimen. The intention of this layout is to eliminate the tension failure in the wall due to flexural forces, and to reinforce the wall in the shear crack locations. It is also intended to examine the effect of a one-sided application, which would be common in many cases where access to both sides of the wall is not possible. This specimen utilizes a second type of steel anchors for the FRP, and a detail is shown in Figure 3.17. These differ from the Tyfo anchors in their function and application. First the FRP strips are applied differently in this specimen. Instead of extending down past the bottom of the wall, and curving on to the base (as in Figure 3.14), these strips are cut flush with the bottom of the wall. Then a 150mm steel angle is bolted to the base, and bolted through the FRP strip and the toe blocks. The anchor bolts were the same as in steel Anchor #1. 37 Chapter 3 Shake Table Testing Figure 3.16: Photo of Specimen R5 38 Chapter 3 Shake Table Testing URM WALL STEEL ANGLE N FRP STRIP 4 ^NCHDR EDLT •A: Figure 3.17: Anchor #2 Layout (Specimen R5) 39 Chapter 3 3.3 U B C E E R F Shake Table Shake Table Testing 3.3.1 Description The University of British Columbia Earthquake Engineering Research Facility (EERF) has two shake tables. One is a smaller table capable of six degrees of freedom, and one is much larger and capable of uni-axial motions only. The larger table, although limited by a single direction of motion, is very useful for testing of full-scale structures. The large table was used in this study. A diagram of the table is shown in Figure 3.18. Additional details of the shake-table are provided in Appendix A. Figure 3.18: Shake-Table This table was custom designed for a full-scale two-story house test [Kharrazi, 2001]. The frame was designed to take the loads of the house mostly through the outside perimeter of the frame, with a load capacity of 310 kN. The table is pushed (and pulled) from one side by a hydraulic actuator, and rides on four low friction rollers that are 40 Shake Table Testing Chapter 3 situated at the four corners. The table has two vertical frames at either end that were used as stops to prevent the house from collapsing completely during the testing. The table is 7500mm long by 6000mm wide and is made of series of 200 mm HSS tubes. These tubes are arranged in the frame shown in Figure 3.18 and A.3. The members along the perimeter of the frame are reinforced with 25mm steel plates top and bottom. There are 4 diagonal members, which are 125mm HSS, that are used to laterally brace the frame. These are used to distribute the point load applied by the actuator to the rest of the steel frame. The table has a pump with a maximum capacity of 0.53m3/min (140 gal/min) at 20 MPa pressure. This pump is driven by a 200 HP electric motor that can produce a rotational velocity of 1800 rpm. The table itself can displace +/- 450mm, with a maximum velocity of 45cm/s. The actuator, shown in Figure 3.19, has a maximum pushing force of 260 kN. Figure 3.19: Shake-Table Actuator The shake-table is displacement controlled. The hydraulic pressure, which controls the displacement position of the table, is electronically controlled. The output voltage sent to the actuator is provided by an MTS servo-controller. A command signal is sent to the 41 Chapter 3 Shake Table Testing servo-controller, and this controller determines the level of the voltage that is output to the actuator. The command signal can be generated from a variety of sources. A Hewlett-Packard waveform generator will create various waves, such as sine, saw and square waves. A desktop or notebook PC can be used to generate command signals, which are then amplified and output to the servo-controller. The PC available in this system uses a DasyLab program to generate simulated earthquake output signals. The displacements from a given or synthesized earthquake are normalized into a voltage value. These displacements come from integrated acceleration records. The software takes this voltage data, which is provided in an ASCII datafile, and sends it to the servo-controller. This control setup is shown in Figure 3.20. The computer on the left of the photo is part of the data acquisition system, described in Section 3.5.4. Figure 3.20: Shake-Table Control and Data Acquisition System The shake-table frame is not suited for smaller specimen testing. It was required to make modifications to the base of the table, in order to mount the wall. The wall was to be set up along the length of the centre HSS beam, and tested in-plane. The basebeam is 600mm 42 Chapter 3 . _^  Shake Table Testing wide, while the centre HSS beam was only 200mm wide. To provide a mounting base, a set of steel saddles were designed and fabricated. Details of these saddles are shown in Figure 3.21. Each saddle is a set of steel plates welded together to slide over the HSS beam and provide a larger mounting surface for the beam. The top plate is 25mm thick, and is 800mm wide. Underneath each HSS beam on the table are a series of 10mm plates that were intended to provide a tie-down point for the wood floor of the house. These were used in these tests as tie-down points for the base saddles. The lower part of each saddle is a series of welded 10mm plates that attach to these lower plates. A total of four saddles were fabricated, providing four bearing points for the specimens. 3.3.2 Table Modifications During the testing program undesirable table behaviour made it necessary to make further modifications. Two separate modifications were made, and these are described in this section. Table 3.1 presents codes for the modifications to the table. 43 Chapter 3 Shake Table Testing Table 3.1: Shake-Table Modification Summary Table Modification Code Configuration M-A Original Configuration M-B Bottom Rollers Added M-C Top Roller Added After the first test, there was a significant rocking in the in-plane direction at the base of the wall. The basebeam length was less than half of the length of the table. This created a concentrated load at the centre, and lowered the natural frequencies of the shake-table centre-beam. Of particular interest were the first two modes, the vertical and the rocking mode. The new frequencies were now within the range of the testing frequencies (3 to 5 Hz for Kobe, 8 to 12 Hz for the wall), and this generated a significant response of the table. From Figure 4.34 the period of the wall/centre-beam system is 0.35s and the damping is approximately 16%. A simple finite element model was created of the table with the wall mounted. The model had a fundamental mode at 0.25 s. It also had a rocking mode of a slightly smaller period. During testing, the rocking mode was excited. This created problems, because the table was absorbing the energy of the excitation in this manner. At the completion of the test, the table would move from the rocking mode into a pure vertical mode in free vibration. To eliminate this problem, two rollers were added at the one-third points of the centre HSS beam. This change was added to the finite element model, and it increased the stiffness by several orders of magnitude. The rollers were located at each end of the basebeam. These rollers are shown in Figure 3.22. The table was tested and the frequency of the rocking mode was now out of the range of the testing frequencies. 44 Chapter 3 Shake Table Testing Figure 3.22: Bottom Roller When the first reinforced specimen was tested, the shake-table showed further undesirable behaviour. Because of the increase in stiffness in the reinforced specimen, the table tended to lift at the actuator end during the test. To alleviate this problem, a roller was added to the top of the centre HSS beam, between the actuator and the wall, shown in Figure 3.23. The table was tested and the lifting behaviour, primarily at the actuator end, had been reduced. This table configuration was used for all subsequent tests. Figure 3.24 shows a schematic of the centre HSS beam with all of the modifications, cut through the centre of the table. 45 Chapter 3 Shake Table Testing Chapter 3 3.4 Ground Motions Shake Table Testing The ground motions selected were a key aspect of the experimental program. Several records were used as a trial set of earthquakes. These records were used for the first two U R M specimens to decide which are most suitable, and then a final set was determined, for use in the remaining tests. Two records were part of the preliminary set that was used on the first two specimens, but not in the final set. These are the Sherman Oaks and the Landers records. These records did not have any observable impact and were not used for the remaining tests. Details of these records are presented in Appendix B. The intention of the selected records was to choose a near-field and a subduction record. Both of these are earthquakes that can affect the Southwestern British Columbia region. It was also of interest to test a code-level record, and a higher, more extreme-level earthquake to examine the failure modes of the wall configurations and life-safety issues that can arise with U R M . Out of the five records chosen to be used, the three that became the final records are presented in this section. 47 Chapter 3 3.4.1 Kobe, Japan 1995 (Kobe) Shake Table Testing A record from the Kobe, Japan earthquake of January 17, 1995 was chosen as the near-field record to be used in these tests. The specific record used is from the JMA recording station1. The Kobe record has a section of strong motion approximately 12 seconds long and a peak acceleration of 0.60g. For use in this testing program it is scaled down to a peak acceleration of 0.35g. This brings it closer to the NBCC design acceleration for this region. It is noted that in the N B C C 1995 the design value is 0.25g, but it is anticipated to be revised in the upcoming version to a value between 0.35 and 0.45g. Figure 3.25 shows the acceleration time history, Figure 3.26 shows the displacement time history and Figure 3.27 shows the response spectrum. 5 10 15 20 25 30 35 40 Time (seconds) Figure 3.25: Kobe Acceleration Time History 1 The record was downloaded from http://peer.berkeley.edu/smcat/ 48 Chapter 3 Shake Table Testing .20 0 5 10 15 20 25 30 35 Time (sec) Figure 3.26: Kobe Displacement Time History 2.5 Frequency (Hz) Figure 3.27: Response Spectrum of the Kobe Record 49 Chapter 3 3.4.2 LlayLlay, Chile 1985 (Chile) Shake Table Testing A record from the Valparaiso, Chile earthquake of March 3, 1985 was chosen as the subduction record to be used in these tests. The specific record used is from the LlayLlay recording station . It has two separate sections of strong motion, with a length of approximately 60 seconds. The peak acceleration of this record is 0.5g and is scaled up to 0.9g for these tests. Figure 3.28 shows the acceleration time history, Figure 3.29 shows the displacement time history and Figure 3.30 shows the response spectrum. 0.6 0:48 Time (seconds) Figure 3.28: Chile Acceleration Time History 2 The record was downloaded from http://db.cosmos-eq.org/ 50 Chapter 3 Shake Table Testing 0 5 10 15 20 25 30 35 40 45 50 55 60 65 Time (sec) Figure 3.29: Chile Displacement Time History (I A I 'V / 2 4 6 8 10 12 14 16 IS 20 22 24 Frequency (Hz) Figure 3.30: Response Spectrum of the Chile Record 51 Chapter 3 3.4.3 Telcordia VERTEQII Synthetic Record (VERTEQII) Shake Table Testing The Telcordia VERTEQII time history is a synthetically generated record used for the testing of telecommunications equipment [Telcordia, 1995]. It is a part of the GR-63-CORE criteria, Generic Requirements documents (GR's) that provide the Telcordia Technologies view of proposed generic criteria for telecommunications equipment, systems, or services. VERTEQII was chosen to be the extreme-level record for the testing due to its high acceleration and frequency content. This was desirable because of the high stiffness and natural frequency of the wall specimens. The record was applied using approximately full-scale accelerations. The strong motion section of the record is approximately 15 seconds long. The acceleration time history is shown in Figure 3.31, the displacement time history is shown in Figure 3.32 and the response spectrum of the accelerations is shown in Figure 3.33. Time (seconds) Figure 3.31: VERTEQII Acceleration Time History 52 Chapter 3 Shake Table Testing \\ i ! 1 0 5 10 15 20 25 30 35 Time (sec) Figure 3.32: VERTEQII Displacement Time History : 1 — 1 1 1 1 I I I I | , 1 | | | | | | | 1 1 1 1 l i l t * * m . § 1 1 1 i 1 i l l 1 * a • » i • * • 1 1 1 1 1 1 1 H I 1 1 I t I J L J L L Frequency (Hz) Figure 3.33: Response Spectrum of the VERTEQII Record [After Telcordia] 53 Chapter 3 3.5 Experimental Procedure and Instrumentation Shake Table Testing This section will outline the test setup, the test procedure and the instrumentation used in this study. 3.5.1 Test Setup The specimens described in Section 3.2 were tested in-plane using simulated earthquake ground motions. They are mounted on the UBC EERE shake-table, described in Section 3.3. The test setup will be described in terms of how each specimen was prepared for the test. The test setup is shown in Figure 3.34, with some details omitted for clarity. First the specimen was lifted onto the shake table by the four rebar loops cast into the basebeam. It was placed on the four base saddles (Section 3.3) and centred in both directions. Once the wall was placed onto to the base saddles, a set of 8 steel brackets were drilled and mounted to the sides of the basebeam. They were anchored using Hilti™ HSL M8/20 bolts. The brackets were then tied down to the base shoes using 25mm diameter threaded rods. This anchored the basebeam to the shake-table. Next, a C-channel with four slotted holes was placed on top of the wall, over the four threaded rods embedded in the bond beam. This channel was used for the attachment of four cables to the top of the wall. This restrains the wall out-of-plane. The C-channel had four steel loops welded to the flanges, to allow for the attachment of the cables. The four cables were assembled in two parts, tied together with a turnbuckle. The turnbuckles were closed as tight as possible. At this point, the wall was firmly mounted at the base, and secured at the top. Next, a surcharge load was added to the top of the wall. The surcharge load was designed to create the necessary inertial forces in the wall during the simulated earthquakes. The load was intended to represent the partial weight of a second story resting on this wall. The surcharge load was achieved by the use of a set of steel plates. Six plates were used, each weighing approximately 4450 N each. Three plates were mounted over each of the 54 Chapter 3 Shake Table Testing sets of two threaded bars grouted into the top of the wall. Steel nuts were threaded on to the rods, and two eyebolts were used in each set of plates to tie them together. The eyes were connected to the overhead crane in the lab by a set of straps. This was the safety mechanism to catch the surcharge load if the wall below collapses completely. The total weight of the wall assembly, including the surcharge, the wall and the basebeam was approximately 55 kN. Figure 3.34: Test Setup 3.5.2 Shake Table Tests This section will summarize the testing program. A summary of the tests is presented in Tables 3.2 and 3.3. Table 3.2 develops a set of test codes and relates them to the earthquake records described in Section 3.4, as well as the date performed. It also will give the wall configuration, as outlined in Section 3.2, the shake-table configuration, as outlined in Section 3.3 and the FRP type used. Table 3.3 presents the various shake-table levels for each test. The first column shows the hardware attenuation level. This is the percentage of the computer output voltage that is sent to the actuator. The second column 55 Shake Table Testing Chapter 3 shows the earthquake amplification level. This value relates the output of the shake-table to the original record in terms of a percentage. The last two columns show the peak acceleration and displacement of the shake-table for each run. There is a difference between the hardware attenuation level and the earthquake amplification level. This is because the shake-table cannot perfectly simulate the response of higher-acceleration records. A displacement record is converted into a voltage signal, and this is input to the table. A hardware attenuation value of 100% will theoretically produce the exact displacements of the record. Depending on the record, the actual displacements will be less, and these are related to the displacement of the original record in the Earthquake Amplification ratio. It can be observed in Table 3.2 that the records with the higher acceleration have a greater difference between the Hardware Attenuation and Earthquake Amplification. In order to achieve displacements closer to the original record, the attenuation is increased. 56 Chapter 3 Shake Table Testing Table 3.2: Test Program: Input and Test Codes Test Code Date Input Wall Configuration Table Config FRP Anchor System Ul.S Dec 11 Sine Unreinforced (UI) M - A N/A U l . l Dec 11 Landers Unreinforced (UI) M - A N/A U1.2 Dec 11 Kobe Unreinforced (UI) M - A N/A U2.1 Feb 21 Sherman Oaks Unreinforced (U2) M-B N/A U2.2 Feb 21 Kobe Unreinforced (U2) M-B N/A U2.3 Feb 21 Landers Unreinforced (U2) M-B N/A U2.4 Feb 21 Landers Unreinforced (U2) M-B N/A U2.5 Feb 21 Kobe Unreinforced (U2) M-B N/A U2.6 Feb 21 Landers Unreinforced (U2) M-B N/A U2.7 Feb 21 Kobe Unreinforced (U2) M-B N/A U2.8 Feb 22 Kobe Unreinforced (U2) M-B N/A U2.9 Feb 22 Chile Unreinforced (U2) M-B N/A U2.10 Feb 22 Chile Unreinforced (U2) M-B N/A U3.1 May 15 VERTEQII Unreinforced (U3) M-C N/A R l . l Apr 3 Kobe Reinforced 1 (RI) M-B Tyfo R1.2 Apr 3 Chile Reinforced 1 (RI) M-B Tyfo R1.3 Apr 4 Kobe Reinforced 1 (RI) M-C Tyfo R1.4 Apr 4 Chile Reinforced 1 (RI) M-C Tyfo R1.5 Apr 4 VERTEQII Reinforced 1 (RI) M-C Tyfo R1.6 Apr 4 VERTEQII Reinforced 1 (RI) M-C Tyfo R2.1 Apr 8 Kobe Reinforced 2 (R2) M-C Tyfo R2.2 Apr 8 Chile Reinforced 2 (R2) M-C Tyfo R2.3 Apr 8 VERTEQII Reinforced 2 (R2) M-C Tyfo R2.4 Apr 8 VERTEQII Reinforced 2 (R2) M-C Tyfo R3.1 Apr 25 Kobe Reinforced 3 (R3) M-C Tyfo R3.2 Apr 25 Chile Reinforced 3 (R3) M-C Tyfo R3.3 Apr 25 VERTEQII Reinforced 3 (R3) M-C Tyfo R3.4 Apr 25 VERTEQII Reinforced 3 (R3) M-C Tyfo R4.1 May 3 Kobe Reinforced 4 (R4) M-C Steel Anchor 1 R4.2 May 3 Chile Reinforced 4 (R4) M-C Steel Anchor 1 R4.3 May 3 VERTEQII Reinforced 4 (R4) M-C Steel Anchor 1 R5.1 May 21 Kobe Reinforced 5 (R5) M-C Steel Anchor 2 R5.2 May 21 Chile Reinforced 5 (R5) M-C Steel Anchor 2 R5.3 May 21 VERTEQII Reinforced 5 (R5) M-C Steel Anchor 2 R5.4 May 21 Chile Reinforced 5 (R5) M-C Steel Anchor 2 R5.5 May 21 VERTEQII Reinforced 5 (R5) M-C Steel Anchor 2 57 Chapter 3 Shake Table Testing Table 3.3: Test Program: Simulated Earthquake Levels Hardware Earthquake Peak Peak Test Code Attenuation Amplification Acceleration Displacement [%] [%] [cm] Ul.S N/A N/A 0.99 3.32 U l . l 100 96 0.20 15.07 U1.2 100 93 0.13 18.59 U2.1 120 113 0.08 14.85 U2.2 120 N/A N/A N/A U2.3 180 174 0.33 27.30 U2.4 180 174 0.33 27.30 U2.5 200 163 0.28 32.48 U2.6 180 N/A N/A N/A U2.7 200 162 0.27 32.26 U2.8 120 111 0.16 22.04 U2.9 120 111 0.44 16.25 U2.10 200 148 0.82 21.79 U3.1 180 115 1.39 9.73 R l . l 200 164 0.37 32.69 R1.2 200 149 1.19 21.88 R1.3 200 164 0.38 32.70 R1.4 200 145 1.01 21.37 R1.5 180 117 1.26 9.87 R1.6 180 117 1.26 9.87 R2.1 200 164 0.35 32.64 R2.2 200 149 1.15 21.83 R2.3 180 117 1.33 9.78 R2.4 180 117 1.40 9.78 R3.1 200 164 0.39 32.66 R3.2 200 149 0.99 21.95 R3.3 180 118 1.31 9.91 R3.4 180 118 1.30 9.89 R4.1 200 164 0.36 32.71 R4.2 200 149 0.98 21.92 R4.3 180 118 1.28 9.91 R5.1 200 164 0.35 32.75 R5.2 200 148 1.01 21.81 R5.3 180 117 1.31 9.80 R5.4 200 148 1.04 21.76 R5.5 180 113 1.26 9.49 58 Chapter 3 3.5.3 Instrumentation Shake Table Testing The following section describes the instrumentation scheme used in these tests. Figure 3.35 shows the instrumentation layout on the test setup. Each instrument location is numbered. Table 3.4 presents the different instrument configurations that were used in the testing. For various reasons, such as failures or noise, different instruments were used throughout the tests. This table relates the locations on Figure 3.35 to the specific instruments used. For each specimen, the same instrumentation was used for all of the runs. The instruments are presented in a code, and these are described following this table. Technical details of these instruments are presented briefly. Accelerometer 2 ' < Displacement Transducer 2 Accelerometer 1 Load Cell *S7 Displacemeiit Transducer ! (%>-Strain Gukge 2 - Strain Guage 1 (Backside) North Figure 3.35: Instrumentation Layout 59 Chapter 3 Shake Table Testing Table 3.4: Instrumentation Layouts Test Acc 1 Acc 2 Disp 1 Disp 2 SGI SG 2 L C U I F B A F B A L V D T 1 L V D T 2 N / A N / A L C U 2 Piezo 1 Piezo 1 L V D T 1 L V D T 2 N / A N / A L C U3 Piezo 2/4 Piezo 3 L V D T 1 L V D T 2 N / A N / A L C R I Piezo 3 Piezo 4 L V D T 1 L V D T 2 S G S G L C R2 Piezo 3 Piezo 4 L V D T 1 L V D T 2 S G S G L C R3 Piezo 2/3 Piezo 3/4 L V D T 1 L V D T 2 S G S G L C R4 Piezo 2/4 Piezo 3 L V D T 1 L V D T 2 S G S G L C R5 Piezo 2/4 Piezo 3 L V D T 1 L V D T 2 S G N / A L C Accelerometers Five different accelerometers were used in the testing program. The following section summarizes these instruments, based on the abbreviations presented in Table 3.4, and presents some basic technical information for each. The accelerometers were hardware filtered to a high-pass of 25Hz. FBA - Kinemetrics FBA-11 force-balance type accelerometer. These instruments have an output range of +/- 2.5V and a sensitivity of 0.2g/V. Piezo 1 - ICSensors Model 3110 piezo-resistive type accelerometer. These instruments have an output range of +/- 2g. Piezo 2 - Crossbow L P Series C X L 0 4 M 1 Z piezo-resistive type accelerometer. These instruments have an output range of +/- 2.0V and a sensitivity of 2g/V. Piezo 3 - ICSensors Model 3110 piezo-resistive accelerometer. These instruments have an output range of +/- 5g. Piezo 4 - ICSensors Model 3110 piezo-resistive accelerometer. These instruments have an output range of +/- 1 Og. Displacement Transducers Two different displacement transducers were used in the testing program. The following section summarizes these instruments, based on the abbreviations presented in Table 3.4, 60 Chapter 3 Shake Table Testing and presents some basic technical information for each. The displacement transducers were hardware filtered with a highpass of 50Hz. LVDT1 - MTS Temposonics LP Position Sensor, model LPRCVU03601. These sensors have a measuring range of 25 to 2000mm analog, and 25 to 3650mm digital. LVDT 2 - Celesco Cable-Extension Position Transducer, Model PTT01-0150-111-1110. These displacement transducers have a full-scale measurement range of +/- 3810mm. The accuracy of these sensors is +/- 3.81mm. Strain Gauges The same strain gauges were used in all of the tests. They were applied to a lightly sanded section of the FRP strip. SG - Tokyo Sokki Kenkyujo (TML) FCA-3.350-11 Strain Gauges (Rosette). These gauges are a 2-element cross, oriented at 90° to each other. The gauges are 3mm long by 2mm wide, with a typical stain limit of 3%. These gauges have a resistance of 350Q. Load Cell A load cell was connected to the actuator of the shake-table. This gave the load at table level throughout the test. LC - MTS Load Cell Model 661.23F/E-01. This load cell has a force capacity of 500 kN. It is 203mm high by 152mm in diameter. 61 Chapter 3 3.5.4 Data Acquisition System Shake Table Testing The components of the data acquisition hardware utilized in this project are summarized in this section. Signal Conditioner The signal conditioner unit is used to improve the quality of the signals by removing undesired frequency contents (filtering) and amplifying the signals. In this test, the filter cards used were Kinemetrics AM-3. Analog/Digital Converter The amplified and filtered analog signals are converted to digital data using an analog to digital converter (Keithly Model 575 with an AMM2 board) prior to storing on the data acquisition computer. The analog to digital converter is controlled by a data acquisition computer, which uses a LabView program. This converter is capable of sampling up to eight channels at sampling frequencies from 0.2 to 2000 Hz. Data Acquisition Computer Signals converted to digital form are stored on the hard disk of the data acquisition computer in binary form. The data can then be transferred to a data analysis computer where numerical analysis of measured data can be done independently of the data acquisition process. In this way, preliminary on-site data analysis can be carried out concurrently with data acquisition. 62 Chapter 4 Chapter 4 - Results and Observations Results and Observations 4.1 General This chapter describes the results of the shake table tests that were performed on the eight specimens. Section 4.2 describes the visual observations from the tests. It includes various photos and diagrams illustrating the behaviour of the walls. Sections 4.3 - 4.5 present the processed data collected from the various instrumentation described in section 3.5.3. It includes various plots and tables of relevant data. Section 4.6 presents drift results computed from the displacements of Section 4.4. 4.2 Observations A total of 36 separate earthquake runs were performed, divided up over the 8 specimens. For every run, a complete description of the significant behaviour of each wall is provided. The descriptions were obtained from detailed study of the test video. For the unreinforced tests, there were 3 specimens, and a total of 14 runs. These are presented in section 4.2.1. For the FRP reinforced tests, there were 5 specimens, and a total of 22 runs. These are presented in section 4.2.2. A l l of the runs are designated by a test code, which is presented in Table 3.2. A l l of the test descriptions are from the east side of the wall, according to the "front-view" videos. To complement the descriptions, there are many figures included in the section. Some of the figures illustrate graphically the crack patterns on the various specimens. In these figures, the cracks are shown as darker lines in the specimen. 4.2.1 U R M Specimens T e s t U l . S Test U l . S was intended to be a sine sweep on the wall and test setup before the earthquake records were run. When the sine sweep was initiated, there was an error with 63 Chapter 4 Results and Observations the servo-controller that caused the table to move suddenly. This "impulse" displacement cracked the wall, and the test was terminated. The error in the servo-controller was due to a malfunctioning attenuator dial that did not work properly at lower amplitudes. In this test, the intended attenuation was at 0.5% of 10 volts. In the test, the dial was set to zero, and slowly increased to 0.5%. Instead of a smooth linear increase, the voltage jumped quickly displacing the table. In this test, cracking occurred during the impulse. The crack pattern is shown in Figure 4.1. The table moved from south to north, with a significant clockwise rotation of the entire wall about the base. This rotation was due to the flexibility of the shake-table. This rotation, combined with the sideways force, cracked the wall beginning at the top of the 6 th course on the southside (1), down to the 1st course on the northside. There was also a second crack running along the top of the 2 n d course (2). At the end of the run, the wall was standing and stable. m Figure 4.1: Test Ul.S Crack Patterns 64 Chapter 4 TestsUl.l andU1.2 Results and Observations After the failure due to the impulse, the wall was subjected to the earthquake records. The first earthquake was the Landers record (Appendix C). It was it run at an amplification of 96%. During the strongest peaks of the ground motion, the wall rocked considerably due to the flexibility of the shake-table. The wall showed no further cracking throughout the run. At the end of the run the wall was standing and stable. The second earthquake was the Kobe record (Section 3.6.1). This ground motion has a higher peak acceleration than the Landers record. The Kobe record was run at an amplification of 93%. Again, during the strongest peaks of the ground motion, the wall rocked considerably. The wall did not show any further cracking during the shaking, and upon completion of the run the wall was standing and stable. At this point the testing on Specimen UI was terminated. Tests U2.1, U2.2, U2.3 and U2.4 Due to the large rotations observed in the first tests, the shake-table was modified to reduce those effects. The details of this modification are presented in Section 3.3.3. Specimen U2 was mounted on the modified table. Specimen U2 was first subjected to the Sherman Oaks record (Appendix C). It was applied at an amplification of 113%. This record did not visibly damage the wall. The next three records that were applied to Specimen U2 were: Kobe, applied at an amplification of 111%); Landers, applied at an amplification of 174%> and Landers, applied at a amplification of 174%. All of these runs produced no significant damage in the wall. Test U2.5 The next record that was applied was Kobe, with an amplification of 163%). The wall was cracked during this record, at two separate times. These cracks are shown in Figure 4.2. The first strong northward peak cracked the wall from the top of the fourth course on the 65 Chapter 4 Results and Observations southside across the wall down to the base between the second and third blocks from the northside (1). A later peak in the southward direction cracked the wall in the opposite direction. The crack began at the top of the fourth course on the northside and it propagated down to the base between the second and third blocks from the southside (2). There were also some additional cracks on the northside. The wall "opened" (lifted up along the cracks) during several of the remaining large peaks, although no further cracking was observed. The wall remained in-plane during and after the test. Figure 4.2: Test U2.5 Crack Patterns Tests U2.6 and U2.7 After Test U2.5, the wall was subjected to further records to examine the cracked behaviour. The records were: Landers, applied at an amplification of 174% and Kobe, applied at an amplification of 162%. The Landers record showed no effect on the cracked wall. The Kobe record did not create any new cracking, although it did open up the wall along the existing cracks similar to what was observed in Test U2.5. 66 Chapter 4 Tests U2,8, U2.9 and U2.10 Results and Observations The following day, further tests were performed to examine the behaviour of the cracked wall. The Kobe record was applied first, at an amplification of 111%. This was done to correlate the tests with those of the previous day. No significant behaviour was observed. The wall was then subjected to the Chile record (Section 3.4.2), first at an amplification of 111%, followed by a second run at an amplification of 148%). These runs had no observable effect on the wall. Test U3.1 Specimen U3 was subjected to the Telcordia VERTEQII synthetic record (Section 3.6.3). The record was applied with an amplification of 115%. The wall failed completely during this test, in the first 7 seconds of the record. Details of the crack patterns, shown in Figure 4.3, were as follows. The initial cracking occurred on the first strong half-cycle. This occurred when the southside of the wall lifted, and a crack propagated from the top of the 2 n d course, extending 2 blocks inwards (1). During the same half-cycle, a crack propagated from the top of the 4 t h course, diagonally down through to the middle of the wall (2). During the next half-cycle, the first cracks opened on the northside. The top of the 3 r d course cracked inwards to the end of the half-block, down through the middle of the wall (3). On one of the subsequent strong cycles, the wall displayed a sliding failure. It slid along the crack at the top of the 2 n d course southward. The wall slid 2 or 3 cm. This was where the first half block fell from the wall at the third course on the north side. On the next cycle in the south direction, a new diagonal crack opened up, causing a smaller x-pattern to appear on the south side, in the lower half of the wall. This x-pattern cracked through the end blocks, rather than through the mortar (4,5). 67 Chapter 4 Results and Observations All subsequent cycles caused significant sliding in the wall. At this point, the wall had developed a definite x-pattern from corner to corner, across the entire face. The north side failed first when the blocks to the right of the main x-pattern fell. Then the blocks on the south side of the cross fell, and then the blocks above the cross collapsed. This was the complete failure of the wall. Figures 4.4 to 4.6 show photos of the wall during three stages of the run. NORTH—> 1 CD Figure 4.3: Test U3.1 Initial Crack Patterns 68 Chapter 4 Results and Observations Chapter 4 Results and Observations Figure 4.6: Test U3.1 - Collapse Mechanism 4.2.2 FRP Reinforced Specimens This section presents the results of the FRP reinforced specimen tests. In all of the tests, the Kobe record was applied with an amplification of 164% and the Chile record was applied with an amplification of 149%). The VERTEQII record was applied with an amplification of 117%>. TestRl. l andR1.2 Specimen RI was first subjected to the Kobe and Chile records. During these runs it was observed that there was some lifting of the shake-table at the ends of the basebeam, causing rotations in the shake-table centre beam. These records created no observable cracking in the specimen. A second modification to eliminate the lifting was made to the table (Section 3.3.3.). 70 Chapter 4 TestR1.3 andR1.4 Results and Observations The test program was restarted after the modifications were completed. Specimen RI was subjected to the Kobe and Chile records. There was no observable cracking on the wall after these two runs. TestR1.5 Specimen RI was subjected to the VERTEQII record. At the end of the run, some hairline shear cracks had appeared in a few places on the wall. Test R1.6 Specimen RI was subjected to a second run of the VERTEQII record. Figure 4.7 shows the crack patterns developed in the wall. Early in the run, cracks opened up at the base of the wall, from the southside towards the middle of the wall (1). Also early in the run, a diagonal crack opened along the mortar joints on the bottom north side of the wall, from the FRP strip along the top of the 1st course (2). Next, a similar diagonal crack opened from the bottom south to the top north part of the wall (3). Finally, the main diagonal crack opened from the top south to the bottom north part of the wall (4). In the bottom south to top north direction, three diagonal cracks had opened. These cracks showed the largest gap width. All of the Tyfo anchors had failed, but had not pulled from the concrete base completely. Figure 4.8 shows the anchor failure. At the end of the run, the wall was still standing. However, most of the mortar joints had been cracked. In Figure 4.7, only the significant cracks are shown. 71 Chapter 4 Results and Observations NORTH—^ Figure 4.7: Test R1.6 Crack Patterns Figure 4.8: Test RI .6 Anchor Failure 72 Chapter 4 TestR2.1 and R2.2 Results and Observations Specimen R2 was subjected to the Kobe and Chile records. This specimen was completely covered with horizontal FRP strips. It displayed no visible cracking on the wall after the first two runs. Test R2.3 Specimen R2 was next subjected to the VERTEQII record. It did not appear to have sustained any damage during the run. It was possible that hairline cracks developed in some places underneath the FRP, but this could not be confirmed by visual inspection of the specimen. Test R2.4 Specimen R2 was then subjected to a second run of the VERTEQII record. The wall failed along the top of the first course. Once this had occurred, the intact upper portion of the specimen moved independently of the base for the duration of the run. At the end of the run, the upper portion of the wall had come to rest off-centre. It appeared as though the joint between the lower two FRP strips were almost at the same location as the mortar joints underneath. Figure 4.9 shows the failure mode of this specimen. The top 11 courses were intact, and had moved approximately 4cm southwards along the top of the first course. 73 Results and Observations Figure 4.9: Test R.2.4 Failure Mode TestR3.1 and R3.2 Specimen R3 was subjected to the Kobe and Chile records. It differed from Specimen RI by the addition of a third vertical FTP strip. It showed no observable cracking during these two runs. Test R3.3 Specimen R3 was then subjected to the VERTEQII record. It showed no observable cracking during this run, although it was possible that some hairline cracks may have developed in the wall. This could not be confirmed by visual inspection of the specimen. 74 Chapter 4 TestR3.4 Results and Observations Specimen R3 was then subjected to a second run of the VERTEQII record. Figure 4.10 shows the crack patterns in the specimen. First, the fibre anchors on the southside began to fail. Then the center anchors failed, and lastly the northside anchors began to fail. Once all of the fibre anchors had pulled loose, the toe blocks at the ends began to crush. Next the southside fibres failed completely by debonding from the base, followed by the center anchor, then by the northside anchor. At this point the wall was cracked entirely along the base (1), and free to move, due to the failed anchors and the fact that the vertical dowels no longer provided the intended lateral sliding support, since the blocks in which they were grouted to had crushed and broken apart. N O R T H - * IlBl r— ® illBli llllliiiii CD Figure 4.10: Test R3.4 Crack Patterns 75 Chapter 4 Results and Observations Next, the wall moved significantly southward, enough for the dowels to catch on the next solid block in the first course. This caused a momentary transfer of shear into the wall, opening some diagonal cracks on the north panel (2). These diagonals occured between the three strips. Figure 4.11 shows the anchor failures. The cracks shown in the south panel were less significant than those on the northside. Figure 4.11: Test R3.4 Anchor Failure Tests R4.1 andR4.2 Specimen R4 was subjected to the Kobe and Chile records. This was a similar specimen to RI; differing by the fact that the FRP strips were tied down with steel anchors, and that the wall had already been cracked horizontally along its length. It displayed no observable cracking during these two runs. 76 Chapter 4 Test R4.3 Results and Observations Specimen R4 was then subjected to the VERTEQII record. Figure 4.12 shows the significant crack patterns in the wall. Early in the run, the fibre on the southside began to fail. It separated from the anchor towards the wall. Then the southside lifted from the base, three blocks inwards. On the next half-cycle, the northside began to separate in a similar way. Next, the southside began to crack above the first course (1). Then the southside began to crush, with the outside face of the toe block falling. Then the southside lifted significantly, displaying some cracking throughout the wall. Then the northside lifted again, and this opened a large diagonal crack from the bottom south corner to the top north corner. It began at the bottom, moves up at a 45° angle, slid along the top of the third course one block to the north (2), and then continued upwards in a diagonal (3). This "sliding" portion is along the pre-existing crack. From this point in the run, the wall sustained further cracking throughout on every southward cycle. Next the main diagonal continued cracking upwards towards the top north corner, this time moving in a continuous diagonal and not sliding along the pre-existing crack (4). Then the fibre at the top north corner began to fail, tearing downwards in the weak direction of the FRP strip. Then several diagonal cracks opened up from the top south to bottom north directions. Finally, there is a strong pulse that caused the entire wall to experience a significant northward sliding. After the ground motion had stopped, the wall was "swaying" in-plane. Figure 4.13 shows a typical anchor failure. Figure 4.14 shows the failure of the FRP strip on the northside. 77 Chapter 4 Results and Observations N O R T H — > ® ® Existing Horizontal Crack Figure 4.12: Test R4.3 Crack Patterns Figure 4.13: Test R4.3 Anchor Failure 78 Chapter 4 Results and Observations Figure 4.14: Test R4.3 FRP Strip Failure Tests R5.1 andR5.2 Specimen R5 was subjected to the Kobe and Chile records. This specimen was only reinforced on one side; the reinforcement was placed in an "X" pattern. It also incorporated a second type of steel anchors, which were bolted to the base and to the wall. It displayed no observable cracking during these two runs. Test R5.3 Specimen R5 was subjected to the VERTEQII record. First the anchors on the southside pulled out of the base. This opened a crack along the base towards the middle of the wall. Then a crack opened above the first course in the middle of the wall. Then a crack opened along the top of the second course, near the middle of the wall, and this crack was shorter than the one below it. Next, the northside anchor fails, but did not pull out as dramatically as the southside anchor. Then the southside continued to lift out, and this caused a compression failure of the south toe block, splitting it in half. This is shown in Figure 4.15. 79 Results and Observations Chapter 4 When observing from the FRP reinforced side, the cracking in the lower courses extended between the FRP strips only. At places where a crack propagated towards the FRP, it would turn down through a block, rather than crack under the FRP strip. Figure 4.15: Test R5.3 Anchor Failure Test R5.4 Specimen R5 was subjected to a second run of the Chile record. The wall rocked slightly on top of the basebeam, but no further damage was observed. Test R5.5 Specimen R5 was subjected to a second run of the VERTEQII record. During the second run, there was an apparent "grinding" effect of the east side block faces. There was a progressive crushing of those faces from the southside inwards. Once the entire first course had been crushed, the wall buckled out eastward. The wall hung from the FRP strips, and remained intact except for a few blocks on the northside that fell out. Figures 4.16 and 4.17 show the failed wall. 8 0 Chapter 4 Results and Observations Figure 4.16: Failure of Specimen R5 - East View r • Figure 4.17: Failure of Specimen R5 - North View 81 Chapter 4 4.2.3 Test Summary Results and Observations A total of 36 tests were run on the 8 specimens. When the specimen exhibited cracking due to the applied loads, it was defined as a "failure". From herein in this thesis, specific tests will be defined as failure tests. The 8 tests that had this classification are listed in the first column of Table 4.1. Further tables that refer to the "failure tests" will also present these same runs. 4.3 Displacement Results The following section presents the displacement results collected during this study. Displacements were recorded at the table level, the top of the walls in the in-plane direction and the along the wall diagonals. These instruments are shown in Figure 3.35. The displacements were recorded using cable-extension type transducers. During a test such as this one, there can be considerable cross-axis motion in the specimen, which can create noise in the displacement signal. As a result, some of the recorded displacement signals are of marginal quality. The signals of the lowest quality are disregarded in this study. Table 4.1 presents the peak displacement results obtained from the eight "failure" tests performed (Section 4.2.3). Columns 2 and 3 present the peak displacements recorded at the table level and columns 4 and 5 present the peak values recorded at the top of the wall (Figure 3.35). The columns entitled "permanent displacement" indicate if there was a final value of displacement that was not zero, or if that value was negligible. All of the displacements measured are "absolute" values, with the instruments anchored to points off of the shake-table. 82 Chapter 4 Results and Observations Table 4.1: Peak Displacement Results from the Failure Tests Table Level Table Level Top Level Top Level Permanent Test South Motion North Motion South Motion North Motion Displacement [cm] [cm] [cm] [cm] [cm] Ul.S 0.05* 3.32* 1.22 5.15 Negl. U2.5 24.11 32.48 22.78 32.63 Negl. U3.1 9.73 8.50 11.65 13.16 Negl. R1.6 9.87 8.53 9.46 10.21 Negl. R2.4 9.78 8.53 10.32 11.00 -3.50 R3.4 9.89 8.51 12.33 12.78 5.54 R4.3 9.91 8.55 10.22 13.62 3.25 R5.3 9.80 8.48 9.44 10.55 -0.52 * These values are from the impulse test The following figures (Figure 4.18 to 4.29) show the displacement-time plots of the failure tests. Figure 4.18 shows the table displacement during the impulse test (Ul.S). Figures 4.19 - 4.21 show the table displacements of the Kobe, Chile and VERTEQII records. These are the records that were used in the main part of the testing program. All of the remaining figures show the recorded top values obtained during the tests. Two of the figures (4.24 and 4.25) show only partial accelerations. In these tests, data was lost due to failure of the specimen, or excessive noise caused by those failures. In all of the figures, the positive values represent the displacements in the south direction, negative in the north direction. These figures are not plotted with equal scales. 1 • i i — i. V - _ ^ . : i_ : j r — i 1 0 2 4 6 8 10 12 14 Time (sec) Figure 4.18: Test Ul.S Table Displacement 83 Chapter 4 Results and Observations 40 20 e c £ o TO f Q -20: -40 10 15. 20 25 30 35 40 45 Time (sec) Figure 4.19: Kobe Table Displacement 30 40 Time (sec) 50 Figure 4.20: Chile Table Displacement 1.5 20 Time (sec) 25 A ... i i 1 | \ \ J 50 55 60 60 70 30 35 Figure 4.21: VERTEQII Table Displacement 84 Chapter 4 Results and Observations M ! i \ t WBIKHIIIBIIF (STRTimHIWIH! m"f -fmnrT!xi -4 4 6 8 10. Time (sec) Figure 4.22: Test Ul.S Top Displacement 12 14 Figure 4.23: Test U2.5 Top Displacement £ £ as f Q 11 12 Time (sec) Figure 4.24: Test U3.1 Top Displacement (Partial Record) 85 Chapter 4 Results and Observations -to Figure 4.25: Test RI .6 Top Displacement (Partial Record) -10 20 25 Time (sec) 30. Figure 4.26: Test R2.4 Top Displacement 35 40 -10 15 20 25 Time (sec) 30 35 40 Figure 4.27: Test R3.4 Top Displacement 86 Chapter 4 Results and Observations 20 25 Time (sec) Figure 4.28: Test R4.3 Top Displacement 20 25 Time (sec) 30 35- 40 Figure 4.29: Test R5.3 Top Displacement 4.4 Acceleration Results The following section presents the acceleration results collected during this study. Accelerations were recorded at the table level and at the top of the wall in the in-plane direction. This is shown in Figure 3.35. Table 4.2 presents the peak accelerations at the table level and at the top of the wall. These accelerations are for the "failure" tests (Section 4.2.3). Columns 2 and 3 present the peak accelerations recorded at the table level and columns 4 and 5 present the peak values recorded at the top of the wall. 87 Chapter 4 Results and Observations Due to high noise levels in the lab, some of the recorded accelerations are of poor quality. The noise was caused by a variety of sources, including the overhead lights, electrical equipment running in the lab, and by the shake-table power supply. The signals of the lowest quality were disregarded in this study. Figure 4.30 shows the table acceleration during the impulse (Ul.S). Figures 4.31 - 4.33 show the table accelerations of the Kobe, Chile and VERTEQII records. All of the remaining figures show the recorded top values obtained during the tests. Two of the figures (4.36 and 4.37) show only partial accelerations. In these tests, data was lost due to failures or excessive noise caused by the failures. In all of the figures, the positive values represent the accelerations in the south direction, negative in the north direction. The figures are not plotted with equal scales. Table 4.2: Peak Acceleration Results from the Failure Tests Table Level Table Level Top Level Top Level Test South Motion North Motion South Motion North Motion [gl [gl [gl Ul.S 0.99 0.56 0.58* 0.58 U2.5 0.28 0.16 0.36 0.29 U3.1 1.39 1.07 0.92 0.86 R1.6 1.26 1.14 1.64 1.38 R2.4 1.40 1.09 1.63 1.56 R3.4 1.30 1.03 1.58 1.40 R4.3 1.28 1.11 1.63 1.59 R5.3 1.31 1.01 1.66 1.63 *this was an incomplete reading due to resolution loss The peak acceleration at the top of the wall in the south direction during Test Ul.S is significantly smaller than the peak table acceleration. This implies a de-amplification of motion, however, based on inspection of Figure 4.34 it can be seen that this is not the case. In Figure 4.30, the large peak in the south direction corresponds to the acceleration of 0.99g. In Figure 4.34, the peak that corresponds to the acceleration of 0.58g, has been clipped due to a lack of resolution. If the signal was continued at that point, it would reach an acceleration close to the table value. 88 Chapter 4 .__ . Results and Observations All of the remaining tests show an amplification of the table motion, except for Test U3.1. This is due to the fact that the specimen was significantly cracked early in the run, separating the surcharge mass from the base, and reducing the transmission of motion from the base to the top. In the reinforced tests, Specimen R2.4 had the least amplification of motion, while Specimens RI, R4 and R5 had amplifications of about 1.3. Time (seconds) Figure 4.30: Test Ul.S Table Acceleration 0.25 10 15 20 25 30 35 40 45 50 Time (seconds) Figure 4.31: Kobe Table Acceleration 89 Chapter 4 Results and Observations 1 Q.8 0.6 0.4 0.2 0 -0.2 -0.4 -0.6 -0.8 -1 -—— — K 1 1 --—— # -III- pi; 10 15 20 25 30 35 40 45 50 55 60 65 Time (seconds) 70 Figure 4.32: Chile Table Acceleration 20 Time (seconds) Figure 4.33: VERTEQII Table Acceleration 90 Chapter 4 Results and Observations 0.6 0:49 0.36 0;24: ? 0.12 o -0.12 -0.24 -o:-36; "0J48 -o:e. ft 'L...„ A >—. / L A J L J ! \ y I I I 4 2.5 3.5 4 4.5 Time (seconds) 5.5 Figure 4.34: Test Ul.S Top Acceleration Figure 4.35: Test U2.5 Top Acceleration 91 Chapter 4 Results and Observations 2 1.6 1.2 2 1 2 i i l i I i i I 7 8 9 10 11 12 13 14 15 16 Time (seconds) Figure 4.36: Test U3.1 Top Acceleration (Partial Record) Time (seconds) Figure 4.37: Test R1.6 Top Acceleration (Partial Record) 92 Chapter 4 Results and Observations 15 20 25 Time (seconds) 30 35 40 Figure 4.38: Test R2.4 Top Acceleration 20 25 Time (seconds) 30 35 40 Figure 4.39: Test R3.4 Top Acceleration 93 Chapter 4 . Results and Observations 10 15 20 25 30 35 40 Time (seconds) Figure 4.40: Test R4.3 Top Acceleration Time (seconds) Figure 4.41: Test R5.3 Top Acceleration 94 Chapter 4 4.5 Force Results Results and Observations In this section, the various load results obtained are presented. Table 4.3 presents the peak loads obtained. The first column shows the equivalent inertial force at the top of the wall, calculated from the peak acceleration (Table 4.2) multiplied by the surcharge mass. The second column presents the peak strains from the gauges applied to the southside FRP strips in each of the reinforced tests. The last column presents the maximum table load cell reading during these tests. The load cell reading includes the inertial force induced by the mass of the shake-table, the basebeam and the wall. These are not included in the top inertial force, and that accounts for the large difference in columns 2 and 4. The total weight of the table, plus the basebeam and the wall is approzimately 140 kN. Table 4.3: Force Results for the Failure Tests Peak Top Peak Strain Values Peak Load Test Inertial Force From Southside Cell Reading [kNl [kNl Ul.S 15.6 N/A N/A U2.5 9.5 N/A 196 U3.1 24.5 N/A 189 R1.6 43.8 255 183 R2.4 43.9 70 185 R3.4 49.6 185 185 R4.3 43.5 240 183 R5.3 48.2 225 157 4.6 Drift Results This section presents the drift results from the selected tests. Figures 4.44 - 4.51 show the calculated drift plots. The tests that were analysed in this section were chosen as to best represent the data, and to give additional information regarding estimates of the stiffness of each specimen. The drift was calculated as the table displacement subtracted from the top displacement. This data is used in Section 5.4 to generate force-deflection diagrams, and for the estimation of specimen in-plane stiffness. 95 Chapter 4 Results and Observations 20 30 40 Time (seconds) 50 60 70 Figure 4.44: Test U2.3 Drift Time History 1 o;6 0.2 -0.2 -0.6 -1 -1.4 -1.8 -2-2 -m -3 1 i i 10 20 30 40 Time (seconds) 50 60 70 Figure 4.45: Test U2.5 Drift Time History 96 ; l : I i ! j I i ! i I i | .0 5 10 15 20 25 30 35 40 45 50 55 60 Time (seconds) Figure 4.47: Test R2.4 Drift Time History 97 Chapter 4 Results and Observations | 1 ,™„„^.„, .w r f 1 \ „ i — — i Ii Li - — — — — — 1 1 10 15 20 25 30 35 40 45 50 Time (seconds) 55 60 Figure 4.48: Test R3.3 Drift Time History 98 Chapter 4 Results and Observations 8.6 Y— i™ Time (seconds) Figure 4.50: Test R4.3 Drift Time History 10 15 20 25 30 35 Time (seconds) 40 45 50 55 60 Figure 4.51: Test R5.3 Drift Time History 99 Analysis of Results and Interpretation Chapter 5 Chapter 5 - Analysis of Results and Interpretation 5.1 General This chapter presents a discussion of the observations presented in Chapter 4. It will discuss the results of the individual tests, and it will present relative comparisons between similar tests. The chapter begins by examining the behaviour of all of the specimens. This will include crack patterns, and failure modes. Section 5.2 discusses the behaviour of the U R M specimens, which are intended to be the benchmarks for the study. Section 5.3 discusses the behaviour of the FRP reinforced specimens and some comparisons will be drawn between the behaviour of the FRP reinforced specimens and that of the U R M specimens. This section will incorporate the results of Sections 4.3, 4.4 and 4.5. Section 5.4 presents an analysis of Section 4.6, and discusses the relative stiffness among the wall configurations. Section 5.5 summarizes the analyses of the behaviour and performance of the various FRP configurations and focuses on the contributions of the specific components. 5.2 Failure Modes and Behaviour of U R M Specimens For this study, three U R M specimens were tested. Although the first specimen was not intended for use as an impulse test, it will be treated as such. Therefore, for this study we had three different loadings that cracked the U R M specimens. The first was an impulse test, the second was a code-level earthquake, and the third was an extreme-load level earthquake. 5.2.1 Specimen UI The impulse test (Ul.S) can provide information about the behaviour of these walls during the earthquake loadings. The impulse, which had a peak table acceleration of 100 Chapter 5 Analysis of Results and Interpretation 0.99g, created a crack pattern as shown in Figure 4.1. The crack pattern observed implied two distinct behaviours. One was a shear/sliding behaviour, indicated by the diagonal stepping crack. This crack began at the top of the sixth course propagating down the wall, with longer horizontal cracks in between, indicative of the sliding behaviour. The second was a flexural behaviour, which was indicated by a horizontal crack along the top of the second course. The shear/sliding behaviour was initiated by the overall shear motion of the specimen, while the flexural crack was caused by the overturning motion of the wall during the impulse. 5.2.2 Specimen U2 Specimen U2 failed during the second run of the Kobe (U2.5) record. The wall was cracked twice, once by a strong northward cycle, and then again later in the run by a strong southward cycle, each with accelerations of approximately 0.3g. The first crack (northwards) propagated from the top of the fourth course on the southside, down towards the bottom northside. This was similar to the shear/sliding behaviour seen in the first specimen, however it did not have the horizontal flexural crack. The second crack (southwards) propagated from the top of the fourth course on the northside down towards the base at the second block from the southside. This is evidence that there were two similar failures within the same test, one in each direction. This can be compared to two separate impulses, each displaying similar cracking in the wall. In tests Ul.S and U2.5 it was observed that the cracking occurred in the lower half of the wall. It would be expected in a pure shear failure to observe a 45° diagonal failure from the top corner to the opposite bottom corner. This has been shown in cyclic tests, some of which are summarized in Section 2.3. One possible reason for this unexpected behaviour is that the nature of the loading in the wall is not like cyclic testing. In a typical cyclic test, the load is applied as a point load at the top of the wall. This neglects the lateral load caused by the self-weight of the wall. In an actual earthquake, the loads in the wall are inertial and are distributed across its height. Because these loads are evenly distributed, the resulting shears will vary linearly across the height. The shear resistance also varies 101 Chapter 5 Analysis of Results and Interpretation with height, because there is a component of the resistance that is dependent on the deadload (Section 1.2.2). At the point where the shear force on the wall overcomes the resistance, the wall cracks. Although the cracks in the two specimens were similar in shape and location, they were not exactly the same. This effect could be attributed to a difference in load levels. The acceleration levels in test Ul.S were higher than in U2.5, and this could be the reason why there was both shear/sliding and flexural cracking. Another reason could be that during the impulse (Ul.S) there was more vertical motion (lifting of the base) in the wall at the peak acceleration, and that caused an exaggerated overturning force. This placed added tension in the southside, which caused the flexural crack. 5.2.3 Specimen U3 Specimen U3 was subjected to considerably higher loads than the first two specimens. Although the peak acceleration in Ul.S was higher, it was only a single pulse of this value. In VERTEQII, there were many peaks of that magnitude, with a strong motion section of the record approximately 15 seconds long. During this run (U3.1), the wall collapsed completely. By analysis of the videos frame by frame, it was determined how the crack formations compared with the first two specimens. The initial cracking was similar to the first two in both shape and location, and had essentially one major crack in each direction. The first crack occurred in two parts, one beginning at the top of the fourth course on the southside, moving downwards towards the bottom northside, and the other opening along the top of the second course. The next strong cycle opened a crack in the opposite direction, from the top of the third course on the northside towards the bottom southside. This was similar to the cracking observed in the first two walls. The VERTEQII record differed from runs Ul.S and U2.5 because it continued to shake the wall heavily after the initial cracking, and this was main reason for the complete failure. On the next few large peaks of VERTEQII the diagonal crack, which was the final crack in the first two specimens, continued upwards and downwards towards the 102 Analysis of Results and Interpretation Chapter 5 south edge. Both of these cracks split the edge block in a roughly 45° manner, shown in Figure 4.3. 5.2.4 Comparison of the U R M Specimens When comparing the three U R M tests, there was evidence that these earthquakes were applied to the U R M specimens like a series of impulse loads. This is due to the fact that all three tests showed very similar initial crack formation. Specimens U2 and U3 showed very similar crack formation on a large peak in the opposite direction as well. It was also apparent that the walls show a predominant shear/sliding behaviour with influence of the overturning moment. The influence of the overturning moment was apparent in two ways. One was by the change in the location of the shear cracks, being lower than expected. This behaviour was observed again in test U3.1. The second was by the appearance of additional horizontal cracks that seemed to be independent of the main shear crack. This was observed in test Ul.S, to a lesser degree in test U2.5 and again in test U3.1. The shear behaviour was evident in all three specimens during the initial cracking. During test U3.1 the wall exhibited shear behaviours above the initial cracks, as shown in Figure 4.6. Because of the extended duration of strong loading in test U3.1, the full behaviour of the specimen could be observed. The X-pattern cracking began above the initial shear cracks, and it formed full 45° diagonals from corner to corner. This behaviour was different from the initial behaviour. One possible reason for this is that the boundary conditions had changed, with the wall now free at the base. The transfer of shear to the base came from the physical contact between blocks that remained and the contact friction between courses. The sliding behaviour that was exhibited by all three specimens was likely caused by the base boundary conditions. The first course was grouted onto steel dowels extending up from the base. There was an unknown amount of mortar-lock caused by the grout being squeezed into the second course existing in each specimen, which increased the crack 103 Chapter 5 Analysis of Results and Interpretation resistance along those bed-joints. When the diagonal shear crack began to propagate downwards it followed the path of least resistance, which was a horizontal crack along the top of the second course. 5.3 Failure Modes and Behaviour of FRP Reinforced Specimens Five different configurations of FRP reinforced specimens were tested, as described in Section 3.2.4. All of them experienced some degree of structural failure after being subjected to the VERTEQII record. Most of them survived the first run, and were subjected to a second run of the same record. All of these tests are compared to the results of test U3.1, which was a U R M wall subjected to the VERTEQII record. 5.3.1 Specimen RI Specimen RI was the most basic of the configurations. It had two vertical FRP strips on the outer edges of the wall, each strip anchored into the base by means of TYFO fibre anchors. This can be seen as a first-step upgrade from the U R M walls. Specimen RI was first subjected to the Kobe (R1.3) and Chile (R1.4) records. This was done in each of the tests to assess the performance of the reinforced wall when subjected to the code and near-code records. The wall experienced peak top accelerations of 1.37g during run R1.3 and 1.15g during run R1.4. During each run there was no visible cracking in the specimen. This is an indication that the presence of the FRP reinforcing in configuration RI when subjected to the code and near-code level records is beneficial to the performance of the wall. Specimen RI was then subjected to the VERTEQII (R1.5) record, with a peak top acceleration of 1.65g. After this run, the wall exhibits some hairline cracking. These were diagonal stepped cracks in the bare wall sections between the fibre strip reinforced ends. There was a significant improvement in the performance of the wall in comparison to Specimen U3, where the wall cracks at the edges and propagates inwards. The addition of 104 Chapter 5 Analysis of Results and Interpretation the FRP strips prevented this from occurring, thus preventing the overall failure of the wall. This is an indication that the presence of the FRP reinforcing in configuration RI when subjected to one run of the extreme-level record is beneficial to the performance of the wall. Specimen RI was then subjected to a second run of the VERTEQII (R1.6) record, with a peak top acceleration of 1.64g. Very early in the run, the wall cracked along the base, which had not happened in any of the U R M wall tests. This could be because the main part of the wall was so much stiffer, and the cracking occurred at the weakest point, which would be along the base. Because of the initial cracking in the bare wall, the shear capacity of the wall had been reduced. During this run, the diagonal cracks which were observed after the first test opened to a larger degree, further reducing the shear stiffness of the wall. Both of the TYFO anchors had partially debonded at the intersection between the base and the first course of the wall, as shown in Figure 4.8. They had not completely debonded from the base, but were allowed to twist and as a result act as "pin" connections. The result was similar to having two stiff columns at either end of the wall, pin connected to the base. This limited the rotation in the surcharge mass, forcing shear to be transferred into the wall throughout the test. This was different than what was observed in test U2.5, where the wall rotated after initial cracking, preventing further damage. The remaining shear capacity of the wall was dependent on the friction and remaining bond between the bare wall blocks. When comparing the behaviour of Specimen RI to U3, two observations can be made. The first is that there was an obvious improvement in the performance of the retrofitted wall when subjected to very severe shaking. While Specimen U3 exhibited a catastrophic failure during the first 7 seconds of the VERTEQII record, Specimen RI survived two runs, and remained standing. The second observation is that the cracking in Specimen RI resembled typical shear behaviour throughout the entire run. Specimen U3 showed the initial shear/sliding crack that was observed in all of the U R M walls, but it did not appear in Specimen RI. 105 Chapter 5 5.3.2 Specimen R2 Analysis of Results and Interpretation Specimen R2 was intended to resist the shear forces more than the flexural forces. It was completed covered with FRP strips placed horizontally. The bottom strip was placed partially on the wall and partially on the base to allow anchoring of the Tyfo anchors. Specimen R2 was subjected to the Kobe (R2.1) and Chile (R2.2) records. The wall was subjected to peak top accelerations of 1.28g during run R2.1 and 1.23g during run R2.2. This specimen showed no visible cracking from these two records. This is an indication that the presence of the FRP reinforcing in configuration R2 when subjected to the code and near-code level records is beneficial to the performance of the wall. Based on the strain gauge data, it was seen that the strains in the Specimen RI were higher than in Specimen R2. This could be due to the fact that there was a more limited load path in Specimen RI, with the same loads applied. This would place more strains on the specimen with less reinforcing area. Specimen R2 was then subjected to the VERTEQII (R2.3) record, with a peak top acceleration of 1.59g. During this record there was no visible cracking. This is an indication that the presence of the FRP reinforcing in configuration R2 when subjected to one run of the extreme-level record is beneficial to the performance of the wall. Unlike Specimen RI, where there was some shear cracking after the first run, Specimen R2 showed none. This indicates better shear resistance, which would be because the strips were taking most of the loads in their strong direction. Specimen R2 was then subjected to a second run of the VERTEQII (R2.4) record, with a peak top acceleration of 1.63g. The wall performed similar to run R2.3 for much of this one, until it suddenly failed at the top of the first course along the entire length. Upon inspection of the wall, it was apparent that the lap joint between the lowest two FRP strips was placed too close to the mortar joint (at the top of the first course) underneath. Though this was an obvious flaw in construction detail, it is important to note that the wall survived the Kobe, Chile records and the first run of the VERTEQII record. This 106 Chapter 5 Analysis of Results and Interpretation was a significant improvement in performance over the U R M wall, and also Specimen RI. It is likely that with proper construction detailing the wall would have survived the second run of the VERTEQII record. It is also of note that in the U R M tests it was observed that the walls had a tendency to crack along the top of the second course, which is likely because of the mortar-lock between the first and second courses. In this test, the addition of FRP's made this joint the weakest in the specimen, and this was where it cracked, overcoming the resistance due to the mortar-lock. 5.3.3 Specimen R3 Specimen R3 was similar to Specimen RI, except for the addition of a third vertical strip placed in the centre of the wall. This test was designed to examine the effect of strip spacing on the wall. Specimen R3 was first subjected to the Kobe (R3.1) and Chile (R3.2) records. The wall was subjected to peak top accelerations of 1.58g during run R3.1 and 1.34g during run R3.2. It did not show any visible cracking during these records. This is an indication that the presence of the FRP reinforcing in configuration R2 when subjected to the code and near-code level records is beneficial to the performance of the wall. The strain values from Specimen R3 were less than in Specimen RI, which was expected due to the addition of a third vertical load path to the wall. This will be true only if the moments are imbalanced, which would likely be the case. The strain values in Specimen R3 were slightly higher than in Specimen R2. Since R3 was similar to RI, having reduced load paths compared to R2, this makes sense. Specimen R3 was then subjected to the VERTEQII (R3.3) record, with a peak top acceleration of 1.64g. This specimen did not show any visible cracking during the run. It appeared to be considerably stiffer than Specimen RI, based on the fact that RI was cracked after run R1.5. This is an indication that the presence of the FRP reinforcing in configuration R2 when subjected to one run of the extreme-level record is beneficial to the performance of the wall. 107 Chapter 5 Analysis of Results and Interpretation Specimen R3 was then subjected to a second run of the VERTEQII (R3.3) record, with a peak top acceleration of 1.85 g. Similar to Specimen R2, the increase in stiffness changed the behaviour of the wall, cracking along the base. In this test, all of the anchors failed, completely debonding from the base. This allowed the wall to move freely for the rest of the run. This behaviour is different than Specimen RI, where the anchors only partially separated from the base, maintaining some anchorage. This would be due to the fact that in Specimen RI, the force to crack the bare-wall section was lower than (or close to) the force to completely debond the anchors. In Specimen R3, the force to crack the bare-wall sections was much higher than the force to debond the anchors. Once the wall had separated from the base, the amount of shear being transferred was reduced. As the wall moved a significant distance on the base, the dowels make contact with a non-damaged block and then shear was induced on the wall. This opened some diagonal cracks in the bare wall sections. Two 45° diagonal cracks appear, which is expected since the aspect ratio of the bare sections has been doubled. Again there was an obvious improvement in Specimen R3 over the performance of U3. There was a two-sided change in performance in comparison to RI. The shear performance of the wall had been improved with the reduction of cracking in the bare wall, however the failure of the anchors could make this a less desirable configuration. 5.3.4 Specimen R4 Specimen R4 features the same FRP strip layout as specimen RI, but differs in two aspects. The first was the addition of steel plate anchors connecting the FRP strips to the base. The second aspect was that the wall was cracked before the test, along the top of the third course. This allowed the observation of the behaviour of the FRP strips in a damage repair application. Specimen R4 was subjected to the Kobe (R4.1) and Chile (R4.2) records. The wall was subjected to peak top accelerations of 1.64g during run R4.1 and 1.50g during run R4.2. 108 Chapter 5 Analysis of Results and Interpretation It displayed no visible cracking during these runs. It was apparent from that the strain values from this test that they are very close to the values from test RI. This would indicate that the mortar strength did not contribute very much to the tensile strength of the combination URM/FRP system. The addition of the FRP strips appeared to improve the performance of the wall, based on comparison to the performance of the U R M walls that were tested and damaged. Although there was no pre-cracked U R M wall tested, it is likely that a pre-cracked U R M wall would perform worse than the non-cracked U R M wall, if not completely collapse. Specimen R4 was then subjected to the VERTEQII (R4.3) record, with a peak top acceleration of 1.63g. In this run, the anchors failed first. They pulled from the base, between the steel plate and up the wall one course. This happened first on the southside and then on the northside. This behaviour was similar to Specimen RI. Once the anchors had become "pinned", the shearing of the bare wall section began. One observation from this test that was different from Specimens RI and R3 was that the fibres at the top of the wall on all four strips showed some degree of stress. On the northside, the fibres tore down the centre of the strip, down one course of blocks (Figure 4.14). This was likely caused by the addition of the steel anchors at the base of the strips. To explain this, consider one cycle of loading, which has a significant drift that causes the strips to take tensile forces. In Specimen RI and R3, the strip will pull taut, and depending on the magnitude of the force, it will continue debonding the strip. In Specimen R4, the strip will pull taut against the anchor, and depending on the force, there will be an impact. These "impacts" at several cycles during a run can cause two things to occur. High tension is placed in the strip, and large shear forces are put into the diagonals (The strains measured in Specimen R4 during the VERTEQII failure run was higher than in the RI and R3 similar runs, and this indicates the higher forces). The contribution of these forces, which were not present in RI and R3, were likely the cause of the strip failures. 109 Chapter 5 Analysis of Results and Interpretation A second observation was that the wall failed on the first run of the VERTEQII record, where in the first 3 specimens failure occurred on the second run of the record. This can be due to the existing horizontal crack, which reduced the initial shear capacity of the wall. In the first 3 specimens, one run of the VERTEQII record was required to reduce the capacity, since initially the walls were at full strength. 5.3.5 Specimen R5 Specimen R5 explored the use of an "X" pattern of FRP strips, which was intended to resist the diagonal cracking observed in specimens RI and R4. It also examined the effect of applying the fibre to only one side of the wall. This would be common in practice where access to both sides is limited, or the other side is an exterior wall, etc. This made this the first specimen with an eccentric stiffness, which introduces possible out-of-plane effects. Specimen R5 was subjected to the Kobe (R5.1) and Chile (R5.2) records. The wall was subjected to peak top accelerations of 1.50g during run R5.1 and 1.43g during run R5.2. The specimen displayed no visible cracking during these runs. This is an indication that the presence of the FRP reinforcing in configuration R5 when subjected to the code and near-code level records is beneficial to the performance of the wall. Specimen R5 was next subjected to the VERTEQII (R5.3) record, with a peak top acceleration of 1.66g. During the run, the wall itself remains intact, and shows no substantial cracking, however damage occurs at the southside anchor. It pulled out of the base, and the end block was crushed below the bolts on the east side of the wall. The northside anchors also pulled out of the base, but to a lesser extent. The performance of the anchors was poor in comparison to test R4. It must be noted however that the anchors in Specimen R4 took half of the load, because there were twice as many. There were some cracks visible in the lower part of the wall, in the triangular section of bare wall that is partitioned off by the FRP strips. These cracks extend only to the FRP 110 Chapter 5 Analysis of Results and Interpretation strips. One important observation was that the cracks did not propagate below the FRP strips. In cases where a mortar crack ran in a direct path towards the fibre, it changed direction and cracked through a block before it would crack underneath the fibre. This is an indicator of the high bond strength of the FRP to the blocks. At this point it was determined that for practical purposes the wall has failed, and was subjected to a run of the Chile (R5.4) record, to simulate the effect of an aftershock of a lesser magnitude. The peak top acceleration was 1.29g. There were no significant effects observed, the wall was still in one piece moving on the basebeam, with the anchors still in place. Although they had no remaining pullout capacity, the bolts were still resting in the holes providing some shear resistance. Specimen R5 was then subjected to a second run of the VERTEQII (R5.5) record, with a peak top acceleration of 1.64g. It appeared that on the unreinforced side, there was enough movement in the upper part of the wall to impose a "grinding" effect on the lower course. The outside faces of the blocks fell, and once they were gone the wall buckled outwards. One important observation was that in this test the wall buckled, but did not collapse completely, as was the case with Specimen U3. The bond strength of the FRP strips was enough to hold the blocks, even when they were hanging freely. The wall did not have any additional load carrying capacity, but in terms of life-safety, the occupants of an adjacent room could be spared from injury or death. I l l Chapter 5 5.4 In-Plane Stiffness Analysis Analysis of Results and Interpretation This section will present the results of an approximate analysis that was performed using the force and displacement results from Chapter 4, calculating the equivalent linear stiffness of the wall/table system. Force-deformation diagrams have been created, shown in Figures 5.1 to 5.9. The force values used in the analysis are the surcharge-inertia force in the specimens, calculated as in Table 4.3. The drift values used are taken from Section 4.6. All of the figures have been plotted with equal scales in order to compare the slopes of the force-deformation loops. The thicker black lines on the figures represent the slopes that were used for the calculation of the stiffness. The U R M tests were scaled differently, because of the significantly different magnitudes of those tests. They were scaled down with the same force/drift ratio, so the slopes are visually comparable. The stiffness values calculated are presented in Table 5.1. The table presents two values, the initial stiffness and the residual stiffness. The residual stiffness is what remains after failure has been observed in the specimen. In some of the specimens, the residual stiffness was observed in a separate test, such as Specimen U2. In others, the residual stiffness was observed in the same test, such as Specimen R2. The tests that only exhibited one stiffness throughout, one of the columns has no value, indicated by a tilda (~). There are two extraneous effects in the testing that distort the interpretation of these diagrams. One effect is the excess rotation of the entire wall due to the rotation of the base, which was observed during some of the tests. This exaggerates the drift, and is shown up on the diagrams as a lower frequency motion, depicted by the larger loops on the diagram. The second effect is due to the fact that the top displacement transducer was attached to the surcharge beam. In some of the tests, this beam became loose and started sliding during the test, which again exaggerates the drift. It is most evident in test R2.3, where there is a net displacement of approximately 1 .Ocm, which is due to the sliding of the beam. 112 Chapter 5 Analysis of Results and Interpretation Table 5.1: Specimen In-Plane Stiffness Values Test Stiffness* (Initial) [kN/m] Stiffness (Residual) [kN/m] Decrease [%] U2.3 450 kN/m U2.5 ~ 50 kN/m 89 R1.5 5000 kN/m R2.3 9000 kN/m R2.4 8500 kN/m 400 kN/m 95 R3.3 6000 kN/m ~ R3.4 ~ 450 kN/m 93 R4.3 4500 kN/m 3000 kN/m 33 R5.3 3800 kN/m 2500 kN/m 34 * These values are approximations based on the best slope taken from the force-deformation plots in this section Two observations can be taken from Table 5.1. The first deals with a relative comparison of the stiffness values. With the exception of Specimen R5, the order of stiffness makes sense with what was expected. The order of stiffness, highest to lowest, excluding R5, was Specimen R2, R3, RI, R4 and U2 (It is assumed that U2 is representative of the U R M Specimens). According to the approximations, Specimen R5 is the least stiff of all of the reinforced specimens. This could be due to the fact that it was reinforced only on one side, which could have caused some twisting of the specimen, affecting the results. The second observation deals with the residual stiffness after failure. This phenomenon was observed in all of the specimens (Specimen RI is not included in this list due to lack of data for test R1.6). In Specimens U2, R2, R3 and R5, the failure involved a complete separation of the wall from the base, although not always directly at the base. This type of failure essentially creates a discontinuity in the load path from the top to the base. In Specimen U2, the crack was at midheight. In Specimen R2, the crack was along the top of the 2 n d course. In Specimens R3 and R5, the separation was along the base. In all of these cases, there is no longer a material bond that prevents overturning, so all of the residual stiffness comes from the self-weight effects, being resistance to overturning and sliding friction. 113 Chapter 5 Analysis of Results and Interpretation This residual stiffness is apparent by a second slope on the force-deformation diagram (always indicated by S2). Although Specimen R5 had the lowest stiffness of the reinforced specimens, it also had the lowest decrease in stiffness after the initial failure. This is likely due to the anchors, which gave the intact wall additional sliding resistance, which is one component of the residual stiffness. Specimen R4 also had a higher residual stiffness, and this is likely also due to the anchors, which provided both sliding resistance, and overturning resistance. Specimens R2 and R3 both had a large decrease in stiffness. This is likely due to their higher initial stiffness, and the fact that the separation was in a continuous line along one course. This would provide little sliding resistance, with no interlock between blocks, as is the case with Specimen U2. 1.2-104 9000 6000 g 3000 5 0 •Uu Id £ -3000 I I -  L " " s i 4 J . . . X — - 4-St ™ |" j I I I i I I f I i I i. I J -2:5 -2 -1.5 -1 -0.5 0 0.5 1 115: 2 . 2.5 0 rift (cm) Figure 5.1: Test U2.3 Force-Deformation Diagram 114 Chapter 5 Analysis of Results and Interpretation 1.5 i t r 9000 6000 •3000 0 -3000 -6000 -9000 - i a ; i o 4 -1.5-10 4 • \ ^ £ f . — SI <• -2.5 -1.5 -1 -0.5 0 0.5 Drift (cm) 1.5 2.5 Figure 5.2: Test U2.5 Force-Deformation Diagram 115 Chapter 5 Analysis of Results and Interpretation 6 4.8 3.6 2:4' 10' •10" 10" •10" 1.2 10" -1.2-10" -2.4 -3:6 10" 10 •10" 540" -10 SI I W mm — wti M m % V -4 -2 0 2 Drift (cm) 10 Figure 5.4: Test R2.3 Force-Deformation Diagram Drift (cm) Figure 5.5: Test R2.4 Force-Deformation Diagram 116 Chapter 5 Analysis of Results and Interpretation Drift (cm) Figure 5.7: Test R3.4 Force-Deformation Diagram 117 Chapter 5 Analysis of Results and Interpretation 118 Chapter 5 5.5 FRP Configuration Effects Analysis of Results and Interpretation The following section summarizes the performance enhancements created by the addition of the various FRP strip configurations and anchoring systems. It brings together the observations and results of Chapter 4, and the discussions of Section 5.3. 5.5.1 Effect of Vertical Strips Vertical FRP strips were applied to specimen RI, R3, R4 and R5. The addition of the strips shows a significant improvement in performance over the unreinforced specimen. In all of the tests, the specimens performed well to the code and near-code level records. The U R M wall was cracked significantly during a run of the Kobe record. Although it did not collapse completely, the in-plane capacity of the wall was reduced, and the wall would be vulnerable to out-of-plane motion. In three of the four tests, the specimens survived one run of the VERTEQII record, before experiencing failure during the second run. Specimens RI, R3 and R4 were all standing at the end of the two runs as well. This is in contrast to the U R M wall, which collapsed in approximately 7 seconds. One characteristic exhibited by these specimens is that the addition of vertical strips reduced flexural behaviour and changed the shear cracking so that the diagonals opened from corner to corner in the specimens. Another effect of the vertical FRP strips is to create relatively rigid "columns" in the wall, hence creating bare infill walls between them. Once the anchors have failed and cracking begins in the bare wall, the system can behave as a series of pin-ended columns, with inelastic infill walls between them. The infill walls are inelastic because the stiffness is reduced as the cracking progresses. 119 Chapter 5 Analysis of Results and Interpretation 5.5.2 Effect of Spacing on Vertical Strips One way to observe the effect of shear enhancement created by the use of vertical strips, is to reduce the strip spacing. This was done by adding a third strip between the first two. This in effect creates three columns, with two slender infill walls. In the cases using two vertical strips, the shear capacity seemed to be dependent on the capacity of the infill walls. This is evident in the performance of RI versus R4, since the "pre-cracked" wall that was less stiff failed much quicker. The observation in Specimen R3 is that the wall is significantly stiffer than walls RI and R4, and the infill walls are only slightly damaged during the test. Upon completion of the runs, the wall itself still has significant shear resistance, as well as its vertical load carrying capacity. This means that the reduction of spacing between strips improves the shear capacity of the wall. The concern with this configuration is that by stiffening the system the weakest component becomes the anchors, and in this test they failed. The design of the anchors is a detail that must be addressed. 5.5.3 Effect of Anchoring of Vertical Strips The following section will discuss the relative effects of the three anchoring systems that were explored in these tests. Since they were applied to different configurations of the FRP strips, they are compared in a qualitative sense only. TYFO Anchors The TYFO anchors are simply a fibre rope, which is drilled into the base, splayed at the end and bonded to the underside of the FRP strip, shown in Figure 3.7. It was used in specimens RI, R2 and R3. During all runs of the code level records (Kobe and Chile), the anchors performed satisfactorily. This was also the case during the first run of the extreme-level record. During the application of the second extreme-load level record (VERTEQII), the anchors showed different levels of failure, depending on the load level and duration. In the two-120 Chapter 5 Analysis of Results and Interpretation strip case, the anchors failed partially, in part due to the energy dissipation that occurred due to the continuous wall cracking. In the three-strip case, the anchors failed completely, due to the lack of that energy dissipation mechanism. The failure load of the anchor is essentially the bond strength of the epoxy. Flat Steel Plate Anchors (Steel Anchor #1) The flat steel plate anchors consisted of a flat steel plate and two anchor bolts. The detail is shown in Figure 3.14. This allowed the radius of the FRP at the base of the wall to be free. During the VERTEQII loading, the FRP strip pulled away from the wall and the base until it reached the plate, which did not move. This effect was similar to the behaviour of the TYFO anchors in Specimen RI. The main difference between them was that the once the TYFO anchors had failed, there could be no more tensile load transferred into the base. With the steel plates, there was still some load transfer, and this is evident by the tearing of the FRP strips at the top corners of the wall. It was observed that the wall with the steel plate anchors performed better than the wall without, even though that wall had a pre-existing crack. Angle Steel Anchors (Steel Anchor #2) The fifth specimen used two steel angles, bolted to the base and to the first course. This is a typical approach used in practice to anchor the fibre. The primary difference between these anchors and the steel plate anchors, is that the FRP strips had been cut flush to the bottom of the wall, and are not curved and bonded to the base as in the other specimens. This means the anchor can only rely on the tensile strength of the bolts into the base, and not of the epoxy bond strength. The detail is shown in Figure 3.17. During the test there was a quick failure of the base anchor bolts. The very stiff wall transferred the load to the anchors and quickly reached the tensile capacity. This is different from the effect of the steel plates, where the FRP pulled from the base, but did not pull out the steel plates. In that case, the loads on the Hilti anchors are in tension and 121 Chapter 5 Analysis of Results and Interpretation shear, rather than purely in tension in the case of Specimen R5. Also, the FRP does not contribute to the anchoring, since it has no bonding to the base. The brittle failure of the anchors in Specimen R5 implies that the anchor used in Specimen R4 has more ductility. It allowed for some yielding of the anchors without complete debonding. This would provide some additional out-of-plane resistance if the wall was subjected to further loading. The anchors in Specimen R5 had already failed and could not provide that out-of-plane resistance. 5.5.4 Effect of Existing Cracks on Performance of Vertical Strips The existing horizontal crack in Specimen R4 had reduced the initial shear capacity of the wall. This was apparent because the wall failed during the first run of VERTEQII as opposed to the second in Specimen RI. This reduced capacity allows the load to be transferred to the anchors more quickly, causing the wall to enter the pin-ended column system phase more quickly. It also added the effect of sliding to the wall, which did not allow the wall to form diagonal cracks in both directions. The U R M wall subjected to the VERTEQII record failed quickly, and it is likely that a pre-cracked wall, without FRP would behave similarly. The addition of the FRP strips prevented the wall from collapse, and displayed its effectiveness in a damage-repair application. They could not however prevent damage in the first run, as was seen in Specimen RI. 5.5.5 Effect of Horizontal Strips The horizontal strips provide a significant increase in the shear capacity of the wall. In the test that was performed in this study, the wall failed prematurely because of a poor construction detail. The part of the wall about the failure however, did not show any signs of distress. This was not the case in the walls without the horizontal strips. The stiffness was calculated to be approximately double of those with vertical strips. 122 Chapter 5 Analysis of Results and Interpretation It is also noted that even though the joint between the strips was overlapping the mortar joint, it did survive one run of the VERTEQII record. 5.5.6 Effect of X-Pattern Strips The X-pattern seems to be a more economical use of this type of retrofit. Although Specimen R2 was much stiffer, R5 used much less materials and still showed an increased shear capacity. Specimen R5 did not show much stress during the first four runs. There was no visible cracking anywhere except at the bottom of the wall near the base, where there was no immediate FRP reinforcement. The blocks in the other 3 exposed triangles showed no cracking. 5.5.7 Effect of One-Sided Application Only The test with the X-pattern was applied to only one side of the wall, to simulate a practical case, in which there is limited access for application. For the code level records, this did not have much impact. During the second run of the VERTEQII record, the eccentric reinforcing caused a failure of one side of the first course. This resulted in an out-of-plane failure, which was not observed in any other test. 123 Chapter 6 Conclusions and Recommendations Chapter 6 - Conclusions and Recommendations The objectives of this thesis were to examine the effect of FRP reinforcing on U R M walls when subjected to: 1) Code and near-code level earthquakes 2) Extreme-load level earthquakes Since the objectives of this thesis were presented in two parts, this chapter will be presented in a similar way. It presents a brief summary of the tests and results, and then the conclusions and the recommendations that have come from this project. A l l of the conclusions presented in Section 6.2 are based on observations from single tests of the various configurations. Single tests may not be true representations of the absolute behaviour of every configuration. Recommendations based on the project results are presented in Section 6.3. 6.1 Summary This research project was initiated to examine the use of FRP composites to retrofit existing concrete-masonry buildings. The project was sponsored by Public Works and Government Services Canada, who are responsible for many federal buildings in the Southwestern British Columbian region. The project was co-sponsored by Fyfe Co., a San Diego company that produces FRP products for use in structural-retrofit applications, The Masonry Institute of British Columbia and The Masonry Contractors Association of British Columbia. The performance of U R M to past earthquakes has been poor, and there is a need to upgrade structures with this type of construction to meet the demands of future earthquakes. The research project tested eight large-scale U R M wall specimens. The tests were performed in-plane, with simulated seismic loads applied by a shake-table. Of the eight specimens, three walls were tested without reinforcement, subjected to three distinct loads. The three loads were an impulse, a code-level earthquake record, and an extreme-124 Chapter 6 Conclusions and Recommendations level earthquake record. The remaining five specimens were reinforced using FRP strips that were bonded to their faces. Each of the five specimens used a different reinforcing configuration, and two of the five used different anchoring systems. These specimens were subjected to a code level, near-code level, and then the extreme-level records. 6.2 Conclusions In general, certain effects on these walls were observed. These effects are based on the test program that was implemented and the specimens tested, and do not necessarily represent all masonry walls. One observed effect was that a U R M wall responds to earthquake loadings as a series of impulsive loads, with damage occurring on specific peaks of the record. This was evident by comparing the two U R M tests that were subjected to the earthquake records (Specimens U2 and U3) to the impulse test. In both of these tests, the first two cracks that the walls experienced were from a single pulse of the record, one in each direction. A second observation is that the U R M walls responded and failed in shear as the primary mode, with a strong influence from a flexural component. This flexural component is evident from the behaviour of the initial cracking that was exhibited by all three specimens. Additional evidence comes from the behaviour of Specimen U2, where after the initial cracking, further loading could not damage the wall. During runs after the cracking, the upper portion of the walls lift from the lower part, indicating a flexural component. The shear component is mostly evident during the run on Specimen U3. 6.2.1 Code and Near-Code Level Performance The tests on the U R M walls showed that their performance was not satisfactory. Although the first two walls did not collapse completely, they suffered significant damage when subjected to code-level accelerations, and have a reduced in-plane capacity, and have a strong possibility of collapse when subjected to the simultaneous 125 Chapter 6 Conclusions and Recommendations application of out-of-plane loading. All of the 5 FRP configurations that were tested performed well to these loads. Each was subjected to a code-level near-field earthquake, followed by a near code-level subduction-type aftershock. There was no observable damage or cracking due to these records. The three anchoring systems that were tested performed well to the code and near-code level records. The Tyfo-fibre anchors are a simple and effective way of adding additional anchoring strength to the base of the wall. While the two steel anchors also performed well during these tests, it is not possible to make any qualitative comparisons between the three anchors when subjected to these load levels. 6.2.2 Extreme-Level Performance The application of a higher extreme-level loading allows for a comparison between the effectiveness of the different reinforcement configurations. The U R M wall that was subjected to the VERTEQII record did not perform well. It collapsed completely after being subjected to the first 7 seconds of the record. The catastrophic failure that was observed poses a potential life-safety hazard. The initial cracking in the wall was similar to the response of the first two U R M specimens. The behaviour of the wall from the rest of the record was different that the first two, because of the higher accelerations and longer duration of this record. The higher acceleration resulted in higher shear forces, which subjected the wall to more loads after its initial flexural cracking. This continuous shear resulted in the collapse. All 5 of the FRP configurations provided a significant improvement in the performance of the wall in comparison to the U R M test. They also showed that the addition of the FRP reinforcing helped to control the failure mode, with cracking that better resembled typical diagonal cracking. The improvements from using these various configurations are summarized by the following points. 126 Chapter 6 Conclusions and Recommendations 1. The addition of vertical FRP strips improves the in-plane performance Tyfo SHE-51A reinforcing fabric applied in 300mm wide vertical strips to the face of an unreinforced-masonry wall makes a significant improvement to the in-plane resistance of the wall. Two walls were reinforced under normal conditions with vertical strips. In both cases the U R M wall survived the application of the extreme-level record, and although both were damaged during the second run, neither collapsed. This is a dramatic improvement considering the magnitude of the loading.' The wall with two vertical strips suffered damage to the bare wall section. The FRP performed well with very little debonding of the material, only at the base in the first one or two courses. The wall had an improved flexural capacity, which was evident by the additional shear cracking in the reinforced specimen. 2. Reducing the spacing between vertical strips improves the shear capacity of the system When the center-to-center spacing of the vertical strips is decreased from 2700mm to 1200mm there is an apparent improvement in the shear capacity of the wall. The wall reinforced with two strips was extensively damaged in the exposed region between them; the addition of a third strip almost eliminated the damage in the wall completely. Only a few diagonal stepped cracks were observed. There was much less debonding of the material in the lower part of the wall in the three-strip test. This allows for better transfer of load from the masonry to the fibre sheets. 3. The addition of vertical strips is effective in repairing damaged walls Tyfo SEH-51A reinforcing fabric applied in 300mm vertical strips to the face of a previously cracked unreinforced-masonry wall makes a significant improvement to the in-plane resistance of the wall. The wall was cracked horizontally along its entire width. This reduces the sliding resistance, and therefore the shear resistance of the U R M wall. At the crack locations, the FRP will transfer most of the shear load across the crack. There will be some resistance due to the friction between blocks. 127 Chapter 6 Conclusions and Recommendations 4. Complete coverage of the wall with horizontal strips is the most effective in improving shear capacity One wall was reinforced using Tyfo SEH-51A strips, applied horizontally to both faces of the wall. During the test, the wall failed due to a poor construction detail. This makes it difficult to make conclusions on the true shear performance of the material in this application. However, some observations can be made. The wall cracked into two pieces, between the second and third courses. The upper portion of the wall did not show any observable cracking after the second run of the extreme-level record. This is different than all of the other FRP configurations. This is an indication of the high shear-resistance of the configuration. It also survived the first run of the record with no visible damage. 5. Use of strips in an X-Pattern is an effective and efficient way to improve in-plane capacity One wall was reinforced using Tyfo SEH-51A strips, applied in an X-pattern to the wall. Vertical strips were added to either end of the wall, and one horizontal strip was added along the top. During the testing, the wall survived all of the runs with very little cracking that was due directly to shear. The wall failed out-of-plane, this was caused by a crushing failure of the lower courses. It is difficult to make definite conclusions about this configuration, although throughout most of the testing the wall remained intact and showed little cracking. 6. FRP's are effective for controlling failure behaviour It was seen in all of the reinforced specimen tests that although they suffered damage, the FRP strips prevented a total and potentially life-threatening collapse. The unreinforced specimen exhibited a total collapse when subjected to the extreme-level record. 7. Anchor details are important for the stiff URM/FRP systems A comparison of the two-strip to three-strip reinforcing illustrates this point. During the two-strip test, which was less stiff, the anchors yielded but did not fail completely. In the three-strip test, which was stiffer, the anchors completely failed. The 128 Chapter 6 Conclusions and Recommendations FRP's transfer the load effectively across the face of the U R M , when the loads are parallel to the reinforcement. When the FRP's change direction, the load can be applied perpendicular to the reinforcement. This is the case at the base of the walls. The strength of the epoxy in tension, and the strength of the anchors become critical. 8. Ductility in the anchors is important to improve overall capacity of URM In comparing the three different anchors that were used in the tests, it is observed that ductility can be an issue in the overall performance of the wall. With the Tyfo fibre anchors, and the steel plate anchors, there was some yielding of the anchor system before failure. The angle steel anchors exhibited a brittle failure, pulling from the base quickly. 9. The steel-plate anchors were the most effective Of the three types of anchors tested, the steel-plate anchors (Steel Anchor #2) showed the best performance during the extreme-level record. 6.3 Recommendations In this section, several recommendations for design and application of this material, as well as recommendations for continued study are presented. The recommendations are based on the observations made during all of the tests. 6.3.1 Recommendations for Further Study To complement the tests performed in this study, further tests should be performed. These could include the following configurations. 1. A fourth U R M wall, subjected to the amplified Kobe record. This is done to confirm the response of the walls as two initial "impulse" cracks that occur in the lower half of the wall. 129 Chapter 6 Conclusions and Recommendations 2. A fifth U R M wall, subjected to the amplified VERTEQII record. The wall should be pre-cracked similar to Specimen R4, to provide a benchmark for comparison, similar to the other U R M specimens. 3. A sixth FRP reinforced specimen with the exact same configuration as Specimen RI. This is done to confirm the behaviour of the wall with two vertical strips. 4. A seventh FRP reinforced specimen with the same configuration as Specimen RI, except with the addition of steel plate anchors, similar to those used on Specimen R4. This is done to examine the behaviour of the wall with vertical strips attached with anchors. 5. An eighth FRP reinforced specimen with the same configuration as Specimen R3, except with the addition of steel plate anchors, similar to those used on Specimen R4. This is done to examine the effect of spacing on the vertical strips, and to better compare Specimens R7 and R8 because they have similar anchors. It would be expected that the anchors in the three strip case should not fail. 6. A ninth FRP reinforced specimen with the same configuration as Specimen R2, with attention to construction details made, so that a proper evaluation of the shear capacity can be made. Tyfo-fibre anchors should be used. 7. A tenth FRP reinforced specimen with the same configuration as Specimen R5, except with the addition of a horizontal strip placed at the bottom on each side. This is done to identify the extent of the out-of-plane behaviour observed in Specimen R5. 8. An eleventh FRP reinforced specimen with the same configuration as Specimen R5, except with both sides reinforced in the same way. This is done to allow for comparison to Specimen R2, and to better evaluate the shear capacity of that configuration. 130 Chapter 6 Conclusions and Recommendations To further study this retrofit technique, and to aid in the development of design and application guidelines in the use of this material, additional in-plane testing should be done. Certain aspects of the URM/FRP system and its behaviour should be addressed. Some of these are: • The aspect ratio of the wall that will determine whether or not the strips should be used horizontally, vertically or a combination of both • The boundary conditions at the ends of the wall that will determine the behaviour of the strips in different circumstances • The effect of additional stiffness created by upper stories as well as the additional mass • The boundary conditions at the top of the wall, to restrict rotations while being subjected to seismic loads applied by a shake-table • Wall assemblies, such as I-shape configurations and corners It would also be useful to examine some out-of-plane URM/FRP configurations, such as: • Behaviour of horizontal-strip layouts to out-of-plane forces • One-sided applications • Single-strip applications 6.3.2 Recommendations for Design and Application Based on the results of this study, several general design recommendations can be made. The first involves the orientation of the FRP strips. Although it is typical in current practice to orient the strips horizontally, such as in Specimen R2, it was seen that the performance of the wall having vertical strips only was adequate. It was observed that the specimens reinforced with horizontal strips, with two vertical strips and with three vertical strips, all performed well during the application of three records, the code-level, near-code and extreme-level records. 131 Chapter 6 Conclusions and Recommendations Secondly, it was observed that the smaller spacing between the vertical strips helped reduce the cracking in the wall after the second run of the extreme record. The wider spacing, however, was adequate to protect the wall after the first three records. A third recommendation is that the X-pattern configuration (Specimen R5) is comparable to the horizontal strips, even when applied to only one side. The configuration survived both of the code-level and near-code level records. Although that during the application of the extreme-level record, the anchors failed, the wall itself performed well. A fourth recommendation is that the use of vertical strips to repair and strengthen a cracked wall is a viable option. The cracked wall tested in this study survived both the code-level and near-code level records with no damage, and survived the extreme-level record without a complete collapse. 132 References ABK, - "Methodology for Mitigation of Seismic Hazards in Existing Unreinforced Masonry Buildings: Categorization of Buildings", A B K Topical Report 01, December 1981 ABK, - "Methodology for Mitigation of Seismic Hazards in Existing Unreinforced Masonry Buildings: The Methodology", A B K Topical Report 08, January 1984 Al-Chaar, G. and Hasen, H., "Masonry Bearing Walls and Shear Walls Retrofitted with Overlay Composite Material - Seismic Evaluation and Dynamic Testing", USACERL Technical Report 98/86, June 1998 Almusallam, T., Al-Salloum, Y., Alsayed, S., Mosallam, A., "Behaviour of Unreinforced Masonry Walls (URM) Strengthened with FRP Composite Materials ", International SAMPE Symposium and Exhibition (Proceedings), v46 I, p 473-484, 2001 Anastasiadis, A., Demosthenous, M . , Karakostas, C., Klimis, N., Lekidis, B., Margaris, B., Papaioannou, C., Papazachos, C., Theodulidis, N., "The Athens (GREECE) Earthquake of September 7,1999: Preliminary Report on Strong Motion Data and Structural Response ", Institute of Engineering Seismology and Earthquake Engineering (ITSAK) Report, September 1999 Arya, A., "Damage Performance of Masonry, Wood and Adobe Buildings During Earthquakes", Workshop on Seismic Hazard Mitigation of Non-Engineered Structures, Hyderabad, India, June 1996 Applied Technology Council (ATC), "FEMA 306 - Evaluation of Earthquake Damaged Concrete and Masonry Wall Buildings - Basic Procedures Manual", (ATC-43 Project), 1998 Applied Technology Council (ATC), "FEMA 307 - Evaluation of Earthquake Damaged Concrete and Masonry Wall Buildings - Technical Resources", 1998 Applied Technology Council (ATC), "FEMA 308 - Repair of Earthquake Damaged Concrete and Masonry Wall Buildings ", 1998 Cheung, M . , Foo, S., Granadino, J., "The Use of Innovative Technologies in the Seismic Retrofit of Canadian Federal Buildings ", 8th Canadian Conference on Earthquake Engineering (Proceedings), Vancouver, Canada, p537-542, 1999 133 Ehsani, M . , Saadatmanesh, H., Al-Saidy, A., "Shear Behaviour of URM Retrofitted with FRP Overlays", Journal of Composites for Construction, vl nl , pi 7-25, February 1997 Ehsani, M . , Saadatmanesh, H., Velazques-Dimas, J., "Behaviour of Retrofitted URM Walls under Simulated Earthquake Loading", Journal of Composites for Construction, v3 n3, pl34-142, August 1999 EQE International, "The October 17, 1989 Loma Prieta Earthquake: Commercial Structures", EQE Quick Look Report, October 1989 EQE International, "The January 17, 1994 Northridge, CA Earthquake: Commercial Structures", An EQE Summary Report, March 1994 EQE International, "The October 9, 1995 Manzanillo, Mexico Earthquake: Summary of Structural Damage ", An EQE Summary Report, 1995 EQE International, "The Adana-Ceyhan, Turkey Earthquake of 26 June 1998", An EQE International Summary Report, 1998 EQE International, "Western Washington Earthquake of July 2, 1999", EQE Report, 1999 EQE International, "Izmit, Turkey Earthquake of August 17, 1999 (M7.4)", An EQE Briefing, 1999 EQE International, "ChiChi, Taiwan Earthquake of September 21, 1999 (M7.4) ", An EQE Briefing, 1999 Federal Emergency Management Agency (FEMA), "NEHRP Guidelines for the Seismic Evaluation of Existing Buildings", Report FEMA-178, 1992 Federal Emergency Management Agency (FEMA), "NEHRP Guidelines for the Seismic Rehabilitation of Existing Buildings", Report FEMA-273, 1996 Fyfe Co., "Tyfo CIS: Corrosion Inhibitor Type S", Fyfe Product Literature, 1998 Fyfe Co., "Tyfo Systems for Unreinforced Masonry (URM) and Reinforced Concrete/Masonry Wall Strengthening", Fyfe Product Literature, 1999 Fyfe Co., "Tyfo Fire-Resistant Systems'', Fyfe Product Literature, 2000 Glanville, J., Hatzinikolas, M . , Ben-Orman, H., "Engineering Masonry Design - Limit States Design", Winston House, Winnipeg, Canada, 1996 134 Goel, R., "Performance of Buildings during the Januaray 26, 2001 Buhj, India Earthquake ", EERI Report, Dept of Civil Engineering, California Polytechnic State University, 2001 Hamilton, H., Dolan, C , "Flexural Capacity of Glass FRP Strengthened Concrete Masonry Walls", Journal of Composites for Construction, v5 n3, pl70-178, August 2001 Hamoush, S., McGinley, M . , Mlakar, P., Scott, D., Murray, K., "Out-of-Plane Strengthening of Masonry Walls with Reinforced Composites ", Journal of Composites for Construction, v5 n3, pl39-145, August 2001 International Conference of Building Officials (ICBO), "Guidelines for the Seismic Retrofit of Existing Buildings ", Draft, 2000 Kharrazi, M . , "Vibration Characteristics of Single-Family Woodframe Buildings", Master's Thesis, The University of British Columbia, August 2001 Kuzik, M . , Cheng, J., Elwi, A., "Full-Scale Out-of-Plane Tests of Masonry Walls with External GFRP Sheets ", Annual Conference of the Canadian Society for Civil Engineering (Proceedings), Regina, Canada, p345-354, June 1999 Langenbach, R., "Bricks, Mortar and Earthquakes: Historic Preservation vs. Earthquake Safety", Journal of the Association for Preservation Technology, vXXI n3&4, p30-43, 1989 MacRae, G., and Lehman, D., "The Nisqually Earthquake of 28 February 2001: Preliminary Reconnaissance Report", Nisqually Earthquake Clearinghouse Group, The University of Washington, March 2001 Moon, F., Tianyi, Y., Leon, R., Kahn, L. , "Retrofit of Unreinforced Masonry Structures with FRP Overlays and Post-Tensioning", Rehabilitating and Repairing the Buildings and Bridges of the Americas: Hemispheric Workshop on Future Directions (Proceedings), Mayaguez, Puerto Rico, p20-36, April 2001 Mosallam, A., Haroun, M . , Almusallam, T., Faraig, S., "Experimental Investigation on the Out-of-Plane Response of Unreinforced Brick Walls Retrofitted with FRP Composites ", International SAMPE Symposium and Exhibition (Proceedings), v46 II,pl364-1371,2001 135 Paquette, J., "Seismic Testing of URM Building with Flexible Diaphragm", Doctoral Thesis, Ottawa-Carleton Institute for Civil Engineering, February 2002 Public Works and Government Services Canada (PWGSC), "Procedure for Seismic Assessment of Existing Buildings", 1998 Public Works and Government Services Canada (PWGSC), Pacific Region, "Guidelines for the Seismic Assessment of Stone-Masonry Structures ", July 2000 Public Works and Government Services Canada (PWGSC), Pacific Region, "Guidelines for the Seismic Upgrading of Stone-Masonry Structures ", July 2000 (Draft) Reinhorn, A., Madan, A., "Evaluation of Tyfo-W Fiber Wrap System for Out of Plane Strengthening of Masonry Walls ", Test Report No. A M R 95-0001, Department of Civil Engineering, State University of New York at Buffalo, March 1995 Reinhorn, A., Madan, A., "Evaluation of Tyfo-W Fiber Wrap System for In Plane Strengthening of Masonry Walls ", Test Report No. A M R 95-0002, Department of Civil Engineering, State University of New York at Buffalo, August 1995 Saadatmanesh, H., "Extending Service Life of Concrete and Masonry Structures with Fibre Composites", Construction and Building Materials, v l 1 n5-6, p327-335, July-Sept 1997 Telcordia Technologies, "Network Equipment-Building System (NEBS) Requirements: Physical Protection", Telcordia Technologies Generic Requirements, GR-63-CORE, Issue 1, October 1995 Triantafiliou, T., "Strengthening of Masonry Structures using Epoxy-Bonded FRP Laminates", Journal of Composites for Construction, v2 n2, p96-103, May 1998 Triantafiliou, T., "Composites: A New Possibility for the Shear Strengthening of Concrete, Masonry and Wood", Composites Science and Technology, v58 n8, pi285-1295, August 1998 Tumialan, J., Morbin, A., Micelli, F., Nanni, A., "Flexural Strengthening of URM Walls with FRP Laminates ", Third International Conference on Composites in Infrastructure (ICCI 2002 - Proceedings), San Francisco, CA, June 2002 136 Weng, D., Lu, X., Ren, X., Lu, Z., "Experimental Study on Seismic Resistant Capacity of Masonry Walls ", 4 t h Multi-Lateral Workshop on Development of Earthquake and Tsunami Disaster Mitigation Technologies and their Intergration for the Asia-Pacific Region (EQTAP), Kakamura, Japan, December 2001 137 Appendix A - Test Setup Details This appendix presents detailed drawings of various components of the test setup and the shake-table. These are found in Figures A . l to A.8. All of the dimensions shown on the drawings are in mm, and they are not shown here to scale. Table A . l presents a list of the drawings. Table A. 1: Drawing List Figure A. 1 Drawing 1 Test Setup Elevation Views Figure A.2 Drawing 2 Test Setup Expanded View Figure A.3 Drawing 3 Shake Table Details (1) Figure A.4 Drawing 4 Shake Table Details (2) Figure A. 5 Drawing 5 Shake Table Details (3) Figure A.6 Drawing 6 Basebeam Details Figure A.7 Drawing 7 Base Saddle Details Figure A.8 Drawing 8 Top Beam and Cable Details 138 S U R C H A R G E S D U T H VIEW in 3,000 •laJL™ " • • • " " • i - ^ ••JLJ^  =0= o EAST VIEW Figure A . 1: Test Setup Details 139 S U R C H A R G E M A S S S U R C H A R G E B E A N T H R E A D E D R O D S tr tr B O N D B E A M U R M W A L L G R O U T E D C O R E S (3) S T E E L D O W E L S B A S E B E A M M O U N T I N G B R A C K E T S B A S E S A D D L E S L O W E R H S S B E A M P L A T E S S H A K E T A B L E C E N T R E B E A M Figure A.2: Test Setup Expanded View 140 moav KWAS Figure A.3: Shake-Table Details (1) 141 Figure A.4: Shake-Table Details (2) 142 a o i o 5 <*»g> I— !±f ft5 . 3 il-ea a i, e S o "«:0S s o O LLJ CO Figure A.5: Shake-Table Details (3) 143 T Q E ^ P L M E 1 \ _ S H E A R _ ± L A T E S -GUSSET P L A T E i i Q f L J i D L E GUSSET P L A T E Tap PLATE T -S H L A R P L A T E BQrfflM^PLjWE \ GUSSET PLATE ^ 9 P J T Q H PLATE Figvire A.7: Base Saddle Details 145 Figure A. 8: Top Beam and Cable Details Appendix B: Additional Earthquake Record Details This appendix presents the two additional earthquake records that were not included in the main program of testing. The Landers1 record is from the Landers, California earthquake of June 28, 1992. The record is taken from the Joshua Tree station. The Sherman Oaks2 record is from the Northridge, California earthquake of January 17, 1994. The record is taken from the Sherman Oaks station. 0 10 20 30 40 50 60 70 Time (sec) Figure B . l : Landers Displacement Time History 1 The Landers record was downloaded from http://peer.berkeley.edu/smcat/ 2 The Sherman Oaks record was downloaded from http://db.cosmos-eq.org/ 147 Time (seconds). Figure B.2: Landers Acceleration Time History 1.2 2 4 6 8 10 12 14 16 18 20 22 24 Frequency (Hz) Figure B.3: Response Spectrum of the Landers Record 148 15 10 -5 -10 • - — — • i 10 1 5 20 Time (sec) 25 30 35 40 Figure B.4: Sherman Oaks Displacement Time History 15 20 25 Time (seconds) 30 35 40 Figure B.5: Sherman Oaks Acceleration Time History 149 15 5 10 15 20 25 Frequency (Hz) Figure B.6: Response Spectrum of the Sherman Oaks Record 150 Appendix C - Complete Displacement Results This appendix presents the complete displacement results, following the same descriptions as what was presented in Section 4.3. Tables C.l and C.2 present the complete results, similar to Table 4.1. Figures C.l to C.32 plot the displacement results, similar to Figures 4.22 to 4.29. Table C.l : Complete Displacement Results of URM Specimens Table Table Top Top Net Test Max South Max North Max South Max North Disp fcml fcml fcml fcml fcml Ul.S 3.32 0.04 5.15 1.22 U l . l 9.40 15.07 9.28 15.04 U1.2 11.90 18.59 12.44 18.81 U2.1 7.21 14.85 6.93 14.83 U2.2 ~ ~ U2.3 27.30 17.04 26.51 16.62 U2.4 27.30 17.05 26.56 16.41 U2.5 32.48 24.11 32.63 22.778 U2.6 U2.7 32.26 24.02 32.06 23.35 U2.8 22.04 14.40 21.60 14.23 U2.9 15.78 16.25 15.44 16.01 U2.10 16.32 21.79 16.25 21.76 U3.1 8.50 9.73 13.16 11.65 151 Table C.2: Complete Displacement Results of Reinforced Specimens Table Table Top Top Net Test Max South Max North Max South Max North Disp [cm] [cm] [cm] [cm] [cm] R l . l 32.69 23.98 33.08 22.58 ~ R1.2 21.88 15.81 21.47 15.99 R1.3 23.91 32.20 23.13 32.51 R1.4 21.37 16.18 21.11 16.24 R1.5 9.87 8.53 10.59 8.79 R1.6 8.53 9.87 9.46 10.21 R2.1 23.89 32.64 22.79 32.98 R2.2 21.83 16.01 21.36 16.61 R2.3 9.78 8.55 10.23 8.63 R2.4 8.53 9.78 10.32 11.00 -3.50 R3.1 23.93 32.66 23.38 33.01 R3.2 21.95 15.85 21.86 16.11 R3.3 9.91 8.55 11.23 15.46 R3.4 8.51 9.89 12.33 12.78 5.54 R4.1 23.93 32.71 23.34 32.97 R4.2 21.92 15.83 21.87 16.56 R4.3 8.55 9.91 10.22 13.62 3.25 R5.1 23.91 32.75 22.93 33.21 R5.2 21.81 15.92 21.93 16.33 R5.3 8.48 9.80 9.44 10.55 -0.52 R5.4 21.76 16.23 21.96 17.53 R5.5 9.41 8.50 ~ ~ Ml 1 « 1* ' I I III i lllj | WilHiHMiHH isnmnTf ii! i» mi i : g I I i i I ! < ! _ 0 2 4 6 3 10 12 14 Time (sec) Figure C . l : Test Ul.S Top Displacement Time (sec) Figure C.2: Test U l . l Top Displacement —AAML.X^.. _ _/ w / 1 0 5 10 15 20 25 30 35 40 45 50 55 60 Time (sec) Figure C.3: Test U1.2 Top Displacement 153 t --10 1S 20 26 30 35 Time (sec) Figure C.4: Test U2.1 Top Displacement 0 5 10 15 20 25 30 35 40 45 50 55 Time (sec) Figure C.5: Test U2.3 Top Displacement Time (sec) Figure C.6: Test U2.4 Top Displacement 154 20 a. w D -20 -40 .....A... y 1 * j \ 10 15 20 25 30 35 40 45 50 55 60 Time (sec) Figure C.7: Test U2.5 Top Displacement 40 20 n IA _ A J s -20' -40. 10 15 20 40 25 30 35 Time (sec) Figure C.8: Test U2.7 Top Displacement .46 50 55 e o (0 % • 5 25 30 35 40 Time (sec) Figure C.9: Test U2.8 Top Displacement 45 50 55 60, 155 so: 10 20 30 40 Time (sec) 50 60 70 Figure CIO: Test U2.10 Top Displacement 10 13 14 11 12 Time (sec) Figure C. l 1: Test U3.1 Top Displacement (Partial Record) 15 16 24 26 28 30 32 34 36 38 Time (sec) Figure C.12: Test R l . l Top Displacement 40 156 Q 20 22 24 28 28 30 32 34 38 38 40 Time (set) Figure C.14: Test R1.3 Top Displacement 157 9 10 11 12 13 14 15 16 17 18 19 Time (sec) Figure C.l6: Test R1.5 Top Displacement (Partial Record) 9 10 11 12 13 14 15 16 17 18 19 20 Time (sec) Figure C.l7: Test R1.6 Top Displacement (Partial Record) , a. ._ ! I l i I — \ I ^  10 15 20 30 25 Time (sec) Figure C.l8: Test R2.1 Top Displacement 35 40 158 30 20 10 15 20 25 30 '35 40 45 50 55 60 65 TO Time (sec) . Figure C.l9: Test R2.2 Top Displacement I L i i k j i h M I 1 If 10 15 20 25 30 35 40 Time (sec) Figure C.20: Test R2.3 Top Displacement 10 15 20 25 30 35 40 Time (sec) Figure C.21: Test R2.4 Top Displacement 159 •40. 6 8 10 12 14 16 18 20 22 24 26 28 30 Time (sec) Figure C.22: Test R3.1 Top Displacement j | 'ji , , i 'i 1 : 20 I : i » I I I \ i i I i ] 10 15 20 25 30 35 40 45 50 55 60 65 70 Time (sec) Figure C.23: Test R3.2 Top Displacement 20 10 -10 -20 5 10 15 20 25 30 35 40 Time (sec) Figure C.24: Test R3.3 Top Displacement 160 A 1 10 15 30 20 25 Time (sec) Figure C.25: Test R3.4 Top Displacement 35 40 _ _ A _ _ k A J\i 10 15 30 20 25 Time (sec) Figure C.26: Test R4.1 Top Displacement 35 40 10 15 20 25 30 35 40 45 50 55 60 65 70 Time (sec) Figure C.27: Test R4.2 Top Displacement 161 20 25 Time (sec) Figure C.28: Test R4.3 Top Displacement 40 irir 10 15 30 20 25 Time (sec) Figure C.29: Test R5.1 Top Displacement 35 40 10 15 20 25 30 35 40 45 50 55 60 65 70 Time (sec) Figure C.30: Test R5.2 Top Displacement 162 20 25 Time (sec) Figure C.31: Test R5.3 Top Displacement 40 30 20 « 10 -10 -20 10 15 20 25 30 35 40 45 50 55 60 Time (sec) Figure C.32: Test R5.4 Top Displacement (Partial Record) 65 70 163 Appendix D - Complete Acceleration Results This appendix presents the complete acceleration results, following the same descriptions as what was presented in Section 4.4. Tables D. l and D.2 present the complete results, similar to Table 4.2. Figures D. l to D.33 plot the displacement results, similar to Figures 4.30 to 4.41. Table D . l : Complete Acceleration Results Table Table Top Top Test Max South Max North Max South Max North [gl [gl [gl [gl Ul.S 0.56 0.99 0.58 0.58 U l . l 0.16 0.20 0.28 0.38 U1.2 0.11 0.13 0.15 0.17 U2.1 0.07 0.08 0.17 0.18 U2.2 ~ ~ U2.3 0.23 0.33 0.33 0.37 U2.4 0.23 0.33 0.35 0.39 U2.5 0.13 0.28 0.29 0.36 U2.6 ~ ~ ~ ~ U2.7 0.18 0.27 0.40 0.25 U2.8 0.09 0.16 0.19 0.19 U2.9 0.44 0.37 0.52 0.51 U2.10 0.73 0.82 0.67 0.66 U3.1 1.07 1.39 0.86 0.92 164 Table D.2: Complete Acceleration Results Table Table Top Top Test Max South Max North Max South Max North [gl [gl [gl [gl R l . l 0.25 0.37 1.35 1.20 R1.2 1.19 0.96 0.96 1.25 R1.3 0.37 0.37 1.37 1.20 R1.4 1.01 0.96 1.15 0.99 R1.5 1.26 1.14 3.18 2.53 R1.6 1.13 1.26 1.38 1.64 R2.1 0.35 0.24 1.28 0.89 R2.2 1.15 0.95 1.23 0.93 R2.3 1.33 1.16 1.57 1.59 R2.4 1.09 1.40 1.63 1.64 R3.1 0.39 0.26 1.58 0.95 R3.2 0.99 0.96 1.34 1.01 R3.3 1.30 1.10 3.08 2.98 R3.4 1.03 1.30 1.55 1.85 R4.1 0.36 0.22 1.65 0.99 R4.2 0.98 0.94 1.50 1.03 R4.3 1.10 1.28 1.63 1.59 R5.1 0.35 0.21 1.49 0.92 R5.2 1.01 0.91 1.43 0.97 R5.3 1.01 1.31 1.66 1.80 R5.4 1.04 0.90 1.12 1.29 R5.5 1.26 1.12 ~ 0.75 0.6 0.45 0.3 0.15 0 -0.15 -0.3 -0.45 -0.6 -0.75 (I 1 — i 1 —-• -A 7 — V 2.5 3.5 4.5 5 Time (seconds) 5.5 6.5 Figure D. l : Test Ul.S Top Acceleration Figure D.2: Test U l . l Top Acceleration 166 0 5 10 15 20 25 30 35 40 45 50 55 60 Time (seconds) Figure D.3: Test U1.2 Top Acceleration 0 10 20 30 40 50 60 70 80 Time (seconds) Figure D.4: Test U2.1 Top Acceleration 167 0 5 10 15 20 25 30 35 40 45 50 55 60 Time (seconds) Figure D.5: Test U2.3 Top Acceleration Time (seconds) Figure D.6: Test U2.4 Top Acceleration 168 0.5 0.4 0.2 -0.4 1.5 ' 1 ! ' 1 ! ! 1 1 10 15 20 25 30 35 40 45 50 Time (seconds) Figure D7: Test U2.5 Top Acceleration i21 \ i i i i j i \ i i i i• ; 0 5 10 15 20 25 30 35 40 45 50 55 60 Time (seconds) Figure D8: Test U2.7 Top Acceleration 169 0 5 10 15 20 25 30 35 40 45 50 55 60 Time (seconds) Figure D.9: Test U2.8 Top Acceleration Time (seconds) Figure D.10: Test U2.9 Top Acceleration 170 0 1:0 20 30 40 50 60 70 Time (seconds) Figure D.l 1: Test U2.10 Top Acceleration 7 8 9 10 11 12 13 14 16 16 Time (seconds) Figure D.12: Test U3.1 Top Acceleration (Partial Record) 171 16 18 20 22 24 26 28 30 32 34 36 38 40 Time (seconds) Figure D.13: Test R l . l Top Acceleration 7 ' ' i •> 1 ! 1 i ! l I 15 20 25 30 35 40 45 50 55 60 65 70 : Time (seconds) Figure D.14: Test R1.2 Top Acceleration 172 Figure D.16: Test RT.4 Top Acceleration I I i ! ' > I ! ! ! ! I I 8 3 10 11 12 13 14 15 16 17 18 19 20 Time (seconds) Figure D.l7: Test R1.5 Top Acceleration (Partial Record) 1.6 -1.6 2 I i l i i i ! I ! I I > I I 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Time (seconds) Figure D.l8: Test R1.6 Top Acceleration (Partial Record) 174 2 1.6 10 15 20 25 30 35 40 Time (seconds) Figure D.19: Test R2.1 Top Acceleration 1.6 1.2 10 15 20 25 30 35 40 45 50 55 60 65 70 Time (seconds) Figure D.20: Test R2.2 Top Acceleration 10 15 20 25 30 35 40 Time (seconds) Figure D.21: Test R2.3 Top Acceleration 2 1.6 1.2 Time (seconds) Figure D.22: Test R2.4 Top Acceleration 176 l l nil If 7 I I 1 i I 1 I I 5 10 15 20 25 30 35 40 Time (seconds) Figure D.23: Test R3.1 Top Acceleration 2 1.6 2 I ! >• I ! ' I I ! I ! I 10 15 20 25 30 35 40 45 50 55 60 65 70 Time (seconds) Figure D.24: Test R3.2 Top Acceleration 177 Time (seconds) Figure D.25: Test R3.3 Top Acceleration - J I > ! ! ! ! I I 10 15 20 26 30 35 40 Time (seconds) Figure D.26: Test R3.4 Top Acceleration 178 w Ip W H O M t *mt$ P i l 1 •- - — ————— — ... . .—_ I i ! i I I ! I 5 10 15 20 25 30 35 40 Time (seconds) Figure D.27: Test R4.1 Top Acceleration -1.2 -1.6 10 15 20 25 30 35 40 45 50 55 60 65 70 Time (seconds) Figure D.28: Test R4.2 Top Acceleration 179 2 Time (seconds) Figure D.29: Test R4.3 Top Acceleration 2 1.8 ,n I I I i I I \ I 6 10 15 20 25 30 35 40 Time (seconds) Figure D.30: Test R5.1 Top Acceleration 180 35 40 45 Time (seconds) eo 65 TO Figure D.31: Test R5.2 Top Acceleration 15 20 25 Time (seconds) 30 35 40 Figure D.32: Test R5.3 Top Acceleration 181 182 Appendix E: Complete Force Results This appendix presents the complete force results, following the same descriptions as what was presented in Section 4.5. Tables E . l and E.2 present the complete results, similar to Table 4.3. Table E. 1: Complete Force Results for U R M Specimens Peak Top Peak Load Test Inertial Force Cell Reading [kNl [kNl Ul.S 15.9 N/A U l . l 10.4 185 U1.2 4.6 192 U2.1 4.9 187 U2.2 N/A N/A U2.3 10.1 188 U2.4 10.6 187 U2.5 9.7 196 U2.6 N/A N/A U2.7 10.8 195 U2.8 5.3 189 U2.9 14.1 92 U2.10 18.2 95 U3.1 25.0 189 183 Table E . l : Complete Force Results for Reinforced Specimens Peak Top Peak Strain Peak Strain Peak Load Test Inertial Force Values (South) Values (North) Cell Reading [kN] [kN] R l . l 36.9 186 ' 124 191 R1.2 33.9 193 141 96 R1.3 37.4 189 130 190 R1.4 31.4 215 135 93 R1.5 37.4 235 170 182 R1.6 38.1 255 188 183 R2.1 35.0 58 44 191 R2.2 33.5 66 29 92 R2.3 43.4 74 60 184 R2.4 44.7 70 51 185 R3.1 42.9 44 N / A 192 R3.2 36.5 ; 75 N / A 94 R3.3 51.2 177 . N / A 185 R3.4 50.4 185 N / A 185 R4.1 44.8 58 71 192 R4.2 40.9 78 N / A 99 R4.3 44.3 240. 210 183 R5.1 40.8 55 N / A 189 R5.2 38.9 j 78 N / A 94 R5.3 49.0. 225 N / A 157 R5.4 35T; 33 N / A 92 R5.5 N / A N / A N / A 162 .1 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0063871/manifest

Comment

Related Items