Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Numerical and experimental investigations of connection for timber-steel hybrid system Azim, Md Riasat 2014

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

Item Metadata

Download

Media
24-ubc_2014_september_azim_md riasat.pdf [ 4.72MB ]
Metadata
JSON: 24-1.0165954.json
JSON-LD: 24-1.0165954-ld.json
RDF/XML (Pretty): 24-1.0165954-rdf.xml
RDF/JSON: 24-1.0165954-rdf.json
Turtle: 24-1.0165954-turtle.txt
N-Triples: 24-1.0165954-rdf-ntriples.txt
Original Record: 24-1.0165954-source.json
Full Text
24-1.0165954-fulltext.txt
Citation
24-1.0165954.ris

Full Text

  NUMERICAL AND EXPERIMENTAL INVESTIGATIONS OF CONNECTION   FOR TIMBER-STEEL HYBRID SYSTEM  by Md Riasat Azim  B.Sc., Bangladesh University of Engineering & Technology, 2011  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  MASTER OF APPLIED SCIENCE in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Civil Engineering)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  August 2014  © Md Riasat Azim, 2014 ii Abstract In recent years, hybrid systems have grown in popularity as potential solution for mid-rise construction. There is also an increased interest in using timber for such systems. The lack of established design guidance, however, has tabled the practical implementation of timber-based hybrid structures. The aim of this thesis is to address the existing knowledge gap regarding the detailed connection design of hybrid systems through combined experimental and numerical investigations on a novel timber-steel system called “FFTT”. The FFTT system relies on wall panels of mass timber such as Cross-Laminated-Timber (CLT) for gravity and lateral load resistance and embedded steel beam sections to provide ductility under seismic loading. A vital step towards practical implementation of the FFTT system is to obtain the proof that the connections facilitate the desired ‘strong column – weak beam’ failure mechanism.  The numerical work applied the software ANSYS; a parametric study based on the results of previous tests was conducted to obtain a suitable connection configuration for improved structural performance. The experimental work, carried out at FPInnovations, consisted of quasi-static monotonic and reversed cyclic tests on two different connection configurations: fully and partially embedded ASTM wide flange sections in combination with 7 ply CLT panels. The combination of partial embedment length and full embedment depth, even when using the smallest wide flange section, did not facilitate the desired behavior. The connection performance was significantly improved when reducing the embedment depth (to avoid creating stress peaks on a weak cross layer) and increasing the embedment length (larger center to center distance between bearing plates). The used small size steel beam, however, is not practical for a real structure; therefore, further studies with larger beams and a modified geometry are recommended before the FFTT system can be applied in practice.   iii Preface The numerical analysis section of chapter 3 has been accepted for publication at the proceedings of World Conference in Timber Engineering:  “Bhat, P., Azim, M.R., Tannert, T., Popovsky, M. “Experimental and numerical investigation of novel steel-timber-hybrid system”, Proceedings of World Conference in Timber Engineering, Quebec City, August 10-14, 2014.”  I conducted the numerical studies and wrote that portion of the manuscript. The main section on “Experimental Investigation” was drafted by Bhat and revised by Tannert and Popovski.   iv Table of Contents Abstract .................................................................................................................................... ii Preface ..................................................................................................................................... iii Table of Contents.................................................................................................................... iv List of Figures ......................................................................................................................... ix Acknowledgements ............................................................................................................... xiii Dedication.............................................................................................................................. xiv Chapter  1: INTRODUCTION .............................................................................................. 1 1.1  Tall Timber Structures: Timber-Steel Hybridization ................................................. 1 1.2  Research Need ............................................................................................................ 2 1.3  Research Objective ..................................................................................................... 3 Chapter  2: LITERATURE REVIEW .................................................................................. 4 2.1  Timber and Steel as Structural Materials .................................................................... 4 2.1.1  Cross-Laminated-Timber .................................................................................... 6 2.1.2  Material Modelling of CLT ................................................................................. 8 2.2  Hybrid Construction ................................................................................................. 10 2.2.1  Component Level Hybridization ....................................................................... 10 2.2.2  System Level Hybridization .............................................................................. 11 2.2.3  Hybrid Connections ........................................................................................... 11 2.3  Seismic Force Resisting System Design Principles.................................................. 12 2.3.1  Force-Based Design Approach .......................................................................... 12 2.3.2  Displacement-Based Design Approach ............................................................. 14 2.3.3  Selection of Design Strategy for Hybrid Systems ............................................. 15 v 2.3.4  Capacity Design Concept .................................................................................. 16 2.4  Lateral Load Resisting Systems for Timber Steel Hybrid Structures ....................... 16 2.4.1  Overview of Lateral Load Resisting Systems ................................................... 16 2.4.2  Infill Wall Systems ............................................................................................ 18 2.4.3  Mass Timber Construction ................................................................................ 19 2.5  Recent Experimental Research on CLT and Hybrid Systems .................................. 19 2.5.1  Ceccotti et al (2010) .......................................................................................... 19 2.5.2  Popovski & Karacabeyli (2011) ........................................................................ 20 2.5.3  Fragiacomo et al (2011) .................................................................................... 21 2.5.4  Numerical Investigations on Hybrid Systems ................................................... 22 2.6  FFTT System ............................................................................................................ 24 2.6.1  Structural System .............................................................................................. 24 2.6.2  Experimental Investigations on FFTT Connection ........................................... 26 Chapter  3: NUMERICAL INVESTIGATION ON FFTT SYSTEM .............................. 29 3.1  Finite Element Model Development ......................................................................... 29 3.1.1  Modelling of CLT Panels .................................................................................. 29 3.1.2  Modelling of Steel Beams ................................................................................. 30 3.1.3  Contact Simulation ............................................................................................ 30 3.1.4  Boundary Conditions ......................................................................................... 30 3.1.5  Load Application ............................................................................................... 31 3.1.6  Post Processing .................................................................................................. 31 3.2  Numerical Results ..................................................................................................... 32 3.2.1  Configuration 1: Partially Embedded Wide Flange Section ............................. 32 3.2.2  Configuration 2: Fully Embedded Wide Flange Section .................................. 35 3.2.3  Configuration 3: Fully Embedded Section with Reduced Cross Section .......... 38 vi 3.2.4  Configuration 4: Full Embedment Length of Hollow Steel Section ................. 40 3.2.5  Configuration 5: Reduced Embedment Length of Hollow Steel Section ......... 42 3.2.6  Summary on Model Results from Previous Tests ............................................. 45 3.3  Numerical Study to Improve Connection Configuration .......................................... 45 3.3.1  Geometry ........................................................................................................... 46 3.3.2  Parameter Variation ........................................................................................... 48 3.3.3  Parametric Study Results ................................................................................... 48 3.3.4  Parametric Study with Partial Embedment Depth ............................................. 54 3.4  Discussion on Numerical Analysis and Optimization Studies ................................. 56 Chapter  4: EXPERIMENTAL INVESTIGATION ON FFTT SYSTEM ...................... 63 4.1  Introduction............................................................................................................... 63 4.2  Materials ................................................................................................................... 63 4.3  Specimen Description ............................................................................................... 64 4.4  Test Procedure .......................................................................................................... 66 4.5  Experimental Results ................................................................................................ 68 4.5.1  Series 1: Monotonic Test on Fully Embedded Beam ........................................ 68 4.5.2  Series 1: Cyclic Test on Fully Embedded Beam ............................................... 72 4.5.3  Series 2: Monotonic Test on Partially Embedded Beam ................................... 76 4.5.4  Series 2: Cyclic Test on Fully Embedded Beam ............................................... 78 4.6  Discussion on Experimental Investigations .............................................................. 83 4.6.1  Comparison between Experimental and Numerical Results ............................. 83 4.6.2  Point of Rotation of Beam ................................................................................. 84 4.6.3  Ductility and Force Modification Factor ........................................................... 85 4.6.4  Hysteretic Behavior ........................................................................................... 87 4.6.5  Energy Dissipation ............................................................................................ 92 vii Chapter  5: CONCLUSIONS ............................................................................................... 93 5.1  Summary ................................................................................................................... 93 5.2  Recommendation for Further Studies ....................................................................... 95 References .............................................................................................................................. 96   viii List of Tables Table 1: Material Properties of Steel and Structural Timber (Yalda, 2009) .................................. 4 Table 2: Physical properties of CLT (Gagnon & Pirvu, 2011) ...................................................... 8 Table 3: Elastic properties of CLT (Gsell et al., 2007) .................................................................. 9 Table 4: FFTT System Options .................................................................................................... 26 Table 5: Properties of CLT ........................................................................................................... 29 Table 6: Properties of Steel beam ................................................................................................. 30 Table 7: Results from previous experimental tests and numerical simulation ............................. 45 Table 8: Parameter range for numerical study ............................................................................. 48 Table 9: Results of parametric study (Beam: W 150 x 29.8) ....................................................... 49 Table 10: Results of parametric study (Beam: W 130 x 23.8) ..................................................... 50 Table 11: Results of parametric study (Beam: W 100 x 19.3) ..................................................... 51 Table 12: Test specimen description ............................................................................................ 64 Table 13: Comparison between Experimental results and their numerical simulation ................ 83 Table 14: Ductility ratio and force modification factor ............................................................... 86 Table 15: Cyclic tests results ........................................................................................................ 87  ix List of Figures Figure 1: Stress-Strain Relationship- Structural Steel .................................................................... 5 Figure 2: The 3 directions for timber properties (Holtz, 2002) ...................................................... 5 Figure 3: Cross-Laminated-Timber ................................................................................................ 7 Figure 4: Nonlinear material model of timber in compression (Grosse and Rautenstrauch, 2004):   a) parallel to grain and b) perpendicular to grain ........................................................................... 9 Figure 5: Nonlinear material model of timber in shear and tension (Multiplas, 2013) ................ 10 Figure 6: Component level hybridization (left: filch Beam, right: Glulam with steel plate) ....... 11 Figure 7: Concept of Hybrid Connection (Yalda, 2009) .............................................................. 12 Figure 8: Force–deformation relationship of a typical plastic hinge (ASCE 41, 2006) ............... 15 Figure 9: Hysteretic model at near-collapse (Ceccotti & Karacabeyli, 2002) ............................. 17 Figure 10: Masonry infill walls Model (Yousuf & Bagchi, 2009) ............................................... 18 Figure 11: CLT Wall Response to Lateral Loading (Schneider, 2009) ....................................... 20 Figure 12: Semi-static CLT Wall Tests - Effect of Connection between Panels: (left) single panel CLT wall, (right) three panel CLT wall (Popovski and Karacabeyli 2011) ....................... 21 Figure 13: 7 Story CLT Shake Table Test (Fragiacomo et al. 2011) ........................................... 22 Figure 14: Solid Panel Core and Intersecting Ductile Steel Link Beams (Green and Karsh, 2012) ...................................................................................................................................................... 24 Figure 15: Type 3 Lateral Load Resisting System for FFTT (Green and Karsh, 2012) .............. 25 Figure 16: Type 3 Lateral Load Resisting System for FFTT (Green and Karsh, 2012) .............. 25 Figure 17: Typical Setup and Instrumentation (Bhat, 2013) ........................................................ 27 Figure 18: Load-deformation plot of the test configuration 4 (Bhat, 2013) ................................. 28 Figure 19: Finite Element Model of test configuration 1 ............................................................. 31 Figure 20: Shear stress plot test configuration 1 .......................................................................... 33 Figure 21: Compressive stress plot test configuration 1 .............................................................. 33 Figure 22: Comparative load deformation plot of test configuration 1: embedded portion ......... 34 x Figure 23: Comparative load deformation plot of test configuration 1: cantilever portion ......... 34 Figure 24: Finite element model of test configuration 2 .............................................................. 35 Figure 25: Compressive stress plot test configuration 2 .............................................................. 36 Figure 26: Shear stress plot test configuration 2 .......................................................................... 36 Figure 27: Comparative load deformation plot of test configuration 2: embedded portion ......... 37 Figure 28: Comparative load deformation plot of test configuration 2: cantilever portion ......... 37 Figure 29: Test configuration 3 .................................................................................................... 38 Figure 30: Comparative load deformation plot of test configuration 3: embedded portion ......... 39 Figure 31: Comparative load deformation plot of test configuration 3: cantilever portion ......... 39 Figure 32: Finite element model of test configuration 4 .............................................................. 40 Figure 33: Comparative load deformation plot of test configuration 4: embedded portion ......... 41 Figure 34: Comparative load deformation plot of test configuration 4: cantilever portion ......... 41 Figure 35: Finite element model of test configuration 5 .............................................................. 42 Figure 36: Shear stress plot test configuration 5 .......................................................................... 43 Figure 37: Compressive stress plot test configuration 5 .............................................................. 43 Figure 38: Comparative load deformation plot of test configuration 5 ........................................ 44 Figure 39: Finite element model for numerical optimization 1.................................................... 47 Figure 40: Details of the steel beam with bearing and side plates ............................................... 47 Figure 41: Load-deformation plot at different points of interest for the optimization study when the W100 x19.3 beam was fully embedded with 150 mm bearing length and 350 mm spacing . 53 Figure 42: Contour plot of compressive stress parallel to grain inside the CLT panel when the W100x19.3 beam was fully embedded with 150 mm bearing length and 350 mm spacing ........ 53 Figure 43: Finite element model for numerical optimization 2.................................................... 54 Figure 44: Contour plot of compressive stress parallel to grain numerical model when the W100 x19.3 beam was partially embedded with 150 mm bearing length and 665 mm spacing ............ 55 Figure 45: Load-deformation plot at different points of interest for the optimization study when the beam was W100 x19.3 with 150 mm bearing length and 665 mm spacing ........................... 56 xi Figure 46: Variation in compressive stress with embedment length of beam (Beam: W 100 x 19.3) .............................................................................................................................................. 60 Figure 47: Variation in compressive stress with length of bearing plate (Beam: W 100 x 19.3) 60 Figure 48: Variation in shear stress with embedment length of beam (Beam: W 100 x 19.3) .... 61 Figure 49: Variation in shear stress with length of bearing plate (Beam: W 100 x 19.3) ............ 61 Figure 50: Variation in displacement with embedment length of beam (Beam: W 100 x 19.3) . 62 Figure 51: Variation in displacement with length of bearing plate (Beam: W 100 x 19.3) ......... 62 Figure 52: Experimental setup for test series 1 ............................................................................ 65 Figure 53: Full embedment of the steel beam inside the CLT panel for test series 1 .................. 65 Figure 54: Full embedment of the steel beam inside the CLT panel for test series 2 .................. 66 Figure 55: CUREE loading protocol for series 1 ......................................................................... 67 Figure 56: CUREE loading protocol for series 2 ......................................................................... 68 Figure 57: Yielding of beam during experimental series 1 .......................................................... 69 Figure 58: Deformation inside the CLT panel during experimental series 1 ............................... 70 Figure 59: Load-displacement curve: Series-1, monotonic test-1 ................................................ 71 Figure 60: Load-displacement curve: Series-1, monotonic test-2 ................................................ 71 Figure 61: Rolling shear failure in CLT panel during cyclic test of series 1 ............................... 73 Figure 62: Cyclic test: Series-1, LVDT-1 .................................................................................... 74 Figure 63: Cyclic test: Series-1, LVDT-2 .................................................................................... 75 Figure 64: Cyclic test: Series-1, LVDT-4 .................................................................................... 75 Figure 65: Yielding of beam during experimental series 1 .......................................................... 76 Figure 66: Load-displacement curve: Series-2, monotonic test-1 ................................................ 77 Figure 67: Load-displacement curve: Series-2, monotonic test-2 ................................................ 78 Figure 68: Out of plane buckling of the steel beam during cyclic test of series 2 ....................... 79 Figure 69: Damage in the CLT panel during cyclic test of series 2 ............................................. 80 xii Figure 70: Cyclic test: Series-2, LVDT-1 .................................................................................... 81 Figure 71: Cyclic test: Series-2, LVDT-2 .................................................................................... 81 Figure 72: Cyclic test: Series-2, LVDT-3 .................................................................................... 82 Figure 73: Cyclic test: Series-2, LVDT-4 .................................................................................... 82 Figure 74: Points of rotation of beams for series 1 and 2 ............................................................. 85 Figure 75: Cyclic test: Series-1, LVDT-1 (with backbone curve) ............................................... 88 Figure 76: Cyclic test: Series-1, LVDT-2 (with backbone curve) ............................................... 88 Figure 77: Cyclic test: Series-1, LVDT-4 (with backbone curve) ............................................... 89 Figure 78: Cyclic test: Series-2, LVDT-1 (with backbone curve) ............................................... 90 Figure 79: Cyclic test: Series-2, LVDT-2 (with backbone curve) ............................................... 90 Figure 80: Cyclic test: Series-2, LVDT-3 (with backbone curve) ............................................... 91 Figure 81: Cyclic test: Series-2, LVDT-4 (with backbone curve) ............................................... 91 Figure 82: Energy dissipation during reverse cyclic tests ............................................................ 92    xiii Acknowledgements First of all, I would like to express my sincere gratitude to Dr. Thomas Tannert, my thesis supervisor, for his valuable guidance, support and encouragement throughout my graduate studies. It has been a pleasure and honor to work under his supervision in this project. I would like to thank Dr. Marjan Popovski from FP Innovations for his constant support and valuable suggestions. I also acknowledge Mr. Paul Simons, who through his time and effort made the experimental investigation a success. I extend my gratitude to Johannes Schneider, whose fabrication skills and support was very valuable for this research project. Also, the UBC technicians Mark Rigolo, George Lee and Harald Schrempp were most helpful at different stages of my work. I thank my fellow MASc students Michael Fairhurst and Alexandra Cheng from the Department of Civil Engineering for helping me during experimental investigation and proof reading my thesis, respectively.  I would like to thank Structurlam for providing the timber products for the experimental program.  Finally, I acknowledge NSERC for the financial support provided through the NewBuildS network.  xiv Dedication   To my loving parents and my brother, without their support, I could not have achieved anything.   1 Chapter  1: INTRODUCTION 1.1 Tall Timber Structures: Timber-Steel Hybridization Hybrid construction combines the structural and architectural features of components made from different materials. In hybrid construction, various materials may work independently or act together, in such a way that they combination is advantageous compared to either single material. During the last decade, much research has been conducted on applications of hybrid structures; the information on and details for steel and wood hybrid structures, however, are dispersed and not readily accessible to builders. As part of this thesis, a literature study on existing hybrid steel and wood structural systems was conducted to identify current techniques of hybridization along with the benefits and challenges associated with them. The literature review has highlighted the opportunity for wood-steel hybrid buildings and existing knowledge gaps. Tall wood buildings are not a new concept: 19 story wooden pagodas were built in Japan 1400 years ago and are still standing in one of the highly seismic regions in the world. The Stadthaus project, London (2008) is an example of an innovative system; it is a nine story building constructed entirely with timber. Its structural system is made of a Cross-Laminated Timber (CLT), which offers an effective solution for construction of large-scale and tall wood buildings. In North America, however, the use of structural wood in construction of new high-rise buildings is not common. History of losses due to fire has regulated the limitations on the building area and height for timber structures in various structural building codes. In recognition of improved fire-fighting measures, the BC Building Code (BCBC, 2009) allows the construction of light-frame wood structures to a maximum of six storeys since 2009; before that the limit was four storeys. 2 A highly ductile material such as steel, when combined with timber, can enhance the post-yield behavior of timber structures. A good engineering design of a hybrid system that combines the merits of the two materials can overcome the limitations of light-frame wood construction and revoke the building height restrictions currently placed on timber buildings. During the past few years, extensive research has targeted the construction of timber-based hybrid structures in order to increase their performance and also, owing to the demands of sustainable construction. One such system is the FFTT system (Green and Karsh, 2012), which is predominantly a mass-timber vertical system with embedded steel beam sections that provide ductility in the system. This system is discussed in detail in Chapter 2. 1.2 Research Need Mass timber and steel hybrid systems have the potential to impact the building industry, address issues of climate change and pose a challenge to concrete and steel structures. However, the current building codes provide no guidelines on seismic design and parameters for the construction of hybrid systems. Due to lack of design values and guidelines and understanding of the global behaviour of hybrid systems, the implementation of a large scale timber-steel hybrid system has not yet been possible in Canada. Analytical and experimental studies that verify the system performance, identify the challenges and optimize the connections for hybrid systems are necessary in order to establish design guidelines and enable implementation.  Recently, through collaboration between the University of British Columbia Vancouver (UBC) and FPInnovations, experimental investigations have been carried at the component level of the FFTT system. Different connection configurations were tested using quasi-static monotonic as well as reversed cyclic loading. Though these tests provided valuable information, they need to 3 be complemented by numerical analyses. Numerical modelling of timber can be very complex due to the fact that the material is anisotropic and that its behavior varies with the type of loading and also with the direction of loading. Design codes use only elastic properties of timber for structural design purposes; however, as will be discussed later, timber does exhibit good post yield inelastic behavior in compression and, consequently, nonlinear modelling can better capture the system behavior when the structure is subjected to overload (wind and earthquake), which an elastic model cannot accurately predict. Only considering the elastic properties of timber is conservative for design. Therefore, there is a need for conducting non-linear numerical investigations on the FFTT system to understand its behavior at the component level. 1.3 Research Objective The purpose of this study is to investigate numerically and experimentally the component level behaviour of the FFTT system and propose a connection layout that can facilitate its successful implementation in mid-rise and high-rise wood-hybrid structures.  The numerical investigations, described in Chapter 3, complement the results from previous experiments and improve the connection layout for further experiments. These experiments, as described in Chapter 4, include monotonic and cyclic loading tests. Based on the numerical and experimental results, conclusions are drawn and future research needs are outlined in Chapter 5.  4 Chapter  2: LITERATURE REVIEW  2.1 Timber and Steel as Structural Materials   The response of a structural system is influenced by the nature and behavior of the used construction materials. Hence, for designing timber-steel hybrid structures, it is important to understand the properties of the individual materials and their potential incompatibility. The properties of timber vary considerably with species. The properties of Spruce Pine SS (as representative of timber) and steel, used as construction materials, are summarized in Table 1. Table 1: Material Properties of Steel and Structural Timber (Yalda, 2009) Material Density (kg/m3) Elastic Modulus (MPa) Compressive Strength (MPa) Tensile Strength (MPa) Steel 7,800 200,000 400-1000 400-1000 Spruce Pine SS  400-500 10,500 Parallel 10 Perpendicular 3 Parallel 6 Perpendicular 1  Steel is a homogeneous and isotropic material. It has high tension and compression strengths along with high stiffness and ability to sustain large inelastic deformation without fracture. Steel exhibits linear stress-strain relationship up to yielding (Figure 1) and a very good post-yield behavior providing ductility to the system. This linear region is elastic and the slope of the curve is the elastic modulus of the material. Beyond yielding, stress increases with increasing deformation due to strain hardening till ultimate strength after which the material fractures.   The in-elastic force-deformation response of a structure depends on the hysteresis response under cyclic deformation of the structural materials and components due to inelastic behavior. The area under the hysteresis loop represents the dissipation of energy. Structural steel dissipates great 5 amounts of energy under cyclic loads. If designed efficiently, steel structures exhibit extreme ductile behavior during an earthquake event. Timber is an anisotropic material; that is, the mechanical properties vary in three mutually perpendicular directions: Longitudinal, Tangential and Radial (Figure 2). The strength properties are strong parallel to grain and weaker across the grain. Timber exhibits ductile failure in compression and brittle failure in tension and shear.   Figure 1: Stress-Strain Relationship- Structural Steel  Figure 2: The 3 directions for timber properties (Holtz, 2002) 6 There are inherent uncertainties in the structural properties of timber. Wood is a hygroscopic material, loss and gain of moisture affects its dimensional stability and strength. In addition, the properties are dependent on the species and characteristics of the tree from which the timber was harvested. The growing conditions and local imperfections (like knots) have an impact on the strength properties (Keenan, 1986). Therefore, engineers use conservative strength properties based on timber grades as specified in CSA 086 (CSA, 2010). The stress-strain relationship of wood under compression is non-linear with good post yield behavior under compressive loading. When subjected to tensile or shear forces, however, timber exhibits brittle failure. Unlike steel, no cyclic energy dissipation can be observed for structural wood when loaded in tension or shear. 2.1.1 Cross-Laminated-Timber CLT is a relatively new product which is gaining in popularity in Europe and recently also in North America. CLT panels are usually made of an odd number of wood layers glued together in a cross-layer pattern, where each layer is oriented in alternating 90 degree angles. CLT panels are generally made of three, five, seven etc. layers of softwood glued together (Figure 3). The gluing is done along the full surface of each panel. Panels are usually manufactured with their outer layers oriented in the direction that the CLT is going to span (Gagnon & Pirvu, 2011). Material properties of CLT, e.g. strength in bending and shear, vary according to manufacturer and raw materials.  7  Figure 3: Cross-Laminated-Timber Moisture content generally has a significant effect on wood performance due to shrinkage. Moisture content at delivery for CLT is typically 8–14%. Surface quality of CLT is important for architectural features and structural use. The estimation of design properties of the CLT not only depends on the species and quality of wood used, but also the number, orientation, and thickness of the layers. Classification of the surface quality of the panels is as in following: • Non-visible Grade: The surface is planned. Such panels are suitable for lining. •  Residential visible: The surface is planned and sanded. These panels are suitable for residential internal exposure. •  Industrial visible: The surface is planned and lightly sanded. Such panels are suitable for exposed industrial internal structure. CLT is generally manufactured with the properties as shown in Table 2.    8 Table 2: Physical properties of CLT (Gagnon & Pirvu, 2011) Width up to 4 m Length up to 16 m Thickness 19 mm, 27.5 mm, 35 mm and 42 mm Pre-cutting Any cuts for windows, doors and so on Wood types Spruce (Pine and Larch on request) Grading C24/C16 (in line with DIN 4074); higher grades on request Moisture content 12% +/- 2% Adhesive Formaldehyde free adhesive for edge and surface bonding, finger jointing  Optical qualities Standard and visible quality Surface finish Sanded  2.1.2 Material Modelling of CLT  The modeling of material properties for CLT is complex, owing to the fact that these properties vary with species and quality of wood, number of individual layers, their orientation and thicknesses. For the purpose of numerical modelling, CLT is often considered as a linear elastic orthotropic material. The various properties of the panel are obtained from experimentation or by using engineering theorems like Gamma Method, Shear Analogy or Composite Theory. According to Gsell et al. (2007), the assumption of linear elastic orthotropic material behavior of CLT is accurate enough to evaluate strength and stiffness properties of panels. The CLT properties as derived from their study are shown in Table 3 where the x, y, and z subscripts refer to three mutually orthogonal directions and E0 and E90 refer to elastic modulus of stiffness parallel and perpendicular to grain direction. These properties can be used for numerical analyses if linear elastic orthotropic behavior is considered.   9 Table 3: Elastic properties of CLT (Gsell et al., 2007) Properties Value (MPa) Properties Value  Ey (E0) 8210 γyx  0.090 Ex (E90) 4630 γzx 0.040 Ez 500 γyz 0.364 Gxz 949 γxy 0.051 Gxy 747 γxz 0.380 Gyz 54 γzy 0.022  Not many studies have been carried out regarding the nonlinear modelling of CLT (and timber in general) owing to the complex behavior of timber post yielding. Grosse and Rautenstrauch (2004) proposed a five stage nonlinear material model for timber incorporating degradation as shown in Figure 4 (a) and (b) for compression parallel and perpendicular to grain, respectively. The shear and tension behavior is usually modelled as linear elastic as shown in Figure 5 (Multiplas, 2013). Grosse’s procedure can be used to model the post-yield inelastic stress-strain behavior of CLT. However, as of now, no experimental data is available for CLT to numerically model such behavior.  Figure 4: Nonlinear material model of timber in compression (Grosse and Rautenstrauch, 2004):   a) parallel to grain and b) perpendicular to grain 10 T  Figure 5: Nonlinear material model of timber in shear and tension (Multiplas, 2013) 2.2 Hybrid Construction All timber structures, to some extent, are hybrid structures since connections are made using steel and foundations are usually concrete. However, true hybridization is the process of combining two or more materials to form a system by making use of the strength of each material and overcome their weaknesses. Hybridization can be classified as component level and system level hybridization (Yalda 2009). 2.2.1 Component Level Hybridization Component level hybridization exists when two different materials are combined together to act as a single structural unit (Figure 6). Common examples for this hybridization are hybrid bridge decks, hybrid slab/diaphragms, hybrid columns and hybrid beams (such as flitch beams). 11  Figure 6: Component level hybridization (left: filch Beam, right: Glulam with steel plate) 2.2.2 System Level Hybridization System hybridization combines different materials at the structural level to share the loads acting on them. Common examples for this type of hybridization are mixed vertical systems where the first few stories are built from a material different from that of the upper stories, hybrid roof trusses where timber is placed at the top of the truss and steel as bottom chord, and hybrid frames where wood and steel share both gravity and lateral loads. Limited research results are currently available on the response and behavior of steel-timber hybrid structures. 2.2.3 Hybrid Connections Due to material and structural differences between steel and wood, and efficient connection between the materials is of high priority (Figure 7). While combining steel with wood, dimensional changes like thermal expansion/contraction of steel and wood shrinkage/swelling may occur with time. Steel plates are commonly used for connections in wood/steel hybrid structures. Johansen’s yield model (Johansen, 1949) is adopted in CSA 086 (CSA, 2010) for the design of dowel-type connections. An ideal connection between steel and timber should lead to yielding of the steel connectors before the wood crushes. Splitting of wood is a brittle failure, and hence, should be avoided.  12  Figure 7: Concept of Hybrid Connection (Yalda, 2009) 2.3 Seismic Force Resisting System Design Principles The two main principles of designing Seismic Force Resisting Systems (SFRSs) are Force-Based Design and Displacement-Based Design.  2.3.1 Force-Based Design Approach In the force-based design approach, the maximum force experienced by the system is evaluated, which is the structure’s base shear. This force is reduced by seismic reduction factors accounting for ductility and over-strength and redistributed proportionally along the height of the building. Maximum Base Overturning Moments are calculated. The system is then designed to resist these forces and moments. NBCC 2010 (NRC, 2010) uses the Equivalent Static Force procedure to determine Base Shear and distribution of story shear. Base shear (Vbase) is calculated using: ࢂ࢈ࢇ࢙ࢋ ൌ ࡿሺࢀࢇሻࡹ࢜ࡵࢋࡾࢊࡾ࢕ ࢃ ,---------------------------------------------------------------(1) Where Sa(T) is the building acceleration, Mv are higher mode effects, Ie is the importance factor, Rd is the ductility factor, Ro is the over-strength factor, and W is the weight of the building. 13 The building fundamental period T is estimated using empirical formulae and limits the building period calculated from analytical model to certain value to account for non-structural components adding stiffness, model inaccuracies and to ensure minimum strength.  A hybrid of timber-steel is lighter than a regular steel frame structure. With this reduction in weight, seismic performance of the structure can be enhanced. The elastic forces evaluated are modified to “Design Forces” by reduction factors namely ductility and over-strength factors. This approach allows for inelastic deformation in the structure dissipating energy during a seismic event. NBCC classifies ductility levels into four categories- Ductile (D), Moderately Ductile (MD), Limited Ductility (LD), and Conventional Construction (CC). Systems with high ductility have specific design requirements and demand rigorous detailing.  The ductility factor is given by the ratio of ultimate roof drift to yield roof drift: ࡾࢊ ൌ ࢾ࢛ࢾ࢟    --------------------------------------------------------------------------------(2) Where δ୷	is defined as the point of first yield anywhere in SFRS and δ୳ is the ultimate drift (the deformation at the point of “near collapse”).  The over-strength factor accounts for the available over strength in the system. It is defined as the ratio of maximum base shear resistance (Vmax) to the design base shear (V) (FEMA, 2009). ࡾࡻ ൌ ࢂ࢓ࢇ࢞ࢂ 				--------------------------------------------------------------------------------------(3) There is no guideline provided so far in NBCC 2010 about the values of Rd and Ro for CLT shear wall buildings, however as per FPInnovations (Gagnon and Pirvu, 2011), these values can be conservatively assumed as 2.0 and 1.5, respectively. 14 2.3.2 Displacement-Based Design Approach The displacement-based design approach evaluates the maximum deformation experienced by the structure, and the system is the designed to resist this deformation either elastically or plastically. Plastic design ensures dissipation of energy during a seismic event; it results in larger deformation in which case the acceptance criteria are set to determine allowable damage in the structure without leading to collapse. This method is known as Performance Based Plastic Design (Wang et al., 2011). Performance is defined as the acceptable level of damage in the system. The estimation of the structural performance involves several uncertainties like variation in ground motion characteristics and the capacity of the components of the system to resist the imposed demands. Therefore, performance-based design follows a probabilistic design philosophy with the probability of exceedance of a certain desired performance.  The performance-based design approach is supported in ASCE 41 (2006) for seismic revaluation and rehabilitation of structures. Hinges are defined as the point of plastic yielding. Each point on the hinge behavior model (Figure 8) corresponds to different performance levels that define acceptance criteria of plastic deformation for each level. Immediate Occupancy (IO) Level occurs just after plastic yielding (Point B) while Life Safety (LS) level occurs significantly before point of total collapse (Point C). Prevention of Collapse (CP) Level corresponds to deformation just before the failure point. For the structure to be operational, the deformation is expected to be below Point B. Typically, for hinges under bending, the acceptance criterion is indicated in terms of rotations or curvatures. 15  Figure 8: Force–deformation relationship of a typical plastic hinge (ASCE 41, 2006) 2.3.3 Selection of Design Strategy for Hybrid Systems Current force-based design procedures use spectral acceleration to determine the lateral strength required by the system to remain elastic and then applies seismic reduction factors that account for inherent ductility and over-strength (ASCE 41, 2006). One shortcomings of this approach lies in the determination of fundamental period of the system. Empirical formulae for elastic fundamental period available in the design code are not particularly tailored for hybrid systems.  In order to minimize damage in wood frame buildings, inter-story drift can be used as key parameter for seismic design. Although the limitations of force-based procedure are alleviated, this approach is not extensively used in the design of timber buildings. This approach requires knowledge of global nonlinear monotonic load-displacement behavior of the building and viscous damping at a target displacement. In addition to sophisticated structural analysis models, system testing is necessary to obtain the required information for the design. With further research and test results on the global behavior of timber structures, this design procedure can be proven valuable in controlling damage in timber buildings resulting from seismic events. 16 2.3.4 Capacity Design Concept Capacity design is a principle that is based on the hypothetical behavior of the structure under seismic load. The system is designed so as to trigger a desired mechanism during a seismic event and suppress the undesired response. This behaviour is achieved by predetermining the weak link in the system and then designing to initiate dissipation of energy by yielding of those members of higher ductile nature and limit inelastic behavior of other components to avoid potential brittle failure (Mitchell, et al., 2003). The system is detailed to accommodate large deformations during an expected duration of strong ground motion without significant loss of lateral strength and ensuring the integrity of the system to sustain gravity loads.  The main difference between force based design and capacity design is that, in the former a particular force is calculated and the structure is proportioned to resist that load while for the latter, the required performance of the structure is known and the force is calculated. This design philosophy appears to be useful in the design of hybrid structures in order to avoid complex techniques of determining potential collapse mechanism. This design strategy helps to develop a hierarchy of capacity among the components of the structure.  2.4 Lateral Load Resisting Systems for Timber Steel Hybrid Structures 2.4.1 Overview of Lateral Load Resisting Systems In steel structures, moment frames mostly form the primary SFRS, often combined with bracing. In timber structures, lateral loads are transferred to the foundation by vertical bracing achieved mainly by shear walls with panel sheathing. Wood moment frames are not usually preferred since it is difficult to achieve a moment connection between wood members. Hybrid systems 17 could be used to resist the combination of lateral and gravity loads to enhance seismic performance of the timber structures. The use of structural panels is one of the most efficient ways of providing lateral support (Dickof et al., 2012). Plywood and OSB panels can be used for horizontal diaphragms and shear walls to brace the building for wind and seismic loads. Floor diaphragms are assumed to behave as deep I-Beams and the supporting shear wall transfers the loads to the foundation. The connections between the shear walls and diaphragm must be efficiently engineered and the wall should be anchored adequately to ensure systematic load transfer and avoid overturn under lateral loads.  The performance of timber structures during a seismic event is highly dependent on the behavior of its connections under cyclic loading. Wood in tension behaves linearly and elastically under cyclic loads and failure is brittle in nature with no dissipation of energy. Steel connections in timber structures are designed to be “semi-rigid” connections instead of perfectly rigid allowing for plastic deformation and energy dissipation. The pinching hysteric model of wood wall system developed at the University of Florence (Ceccotti & Karacabeyli, 2002), is shown in Figure 9. The force-deformation curve is initially steep till its elastic limit, and then the curve becomes non-linear and less steep reaching a peak, where the maximum connection capacity may be found Fmax. Ultimate displacement at “near collapse” criterion was taken as 0.8Fmax.   Figure 9: Hysteretic model at near-collapse (Ceccotti & Karacabeyli, 2002)  18 2.4.2 Infill Wall Systems Common infill wall systems include masonry infill walls in steel or concrete moment frames, see Figure 10. Previous studies have confirmed the increase in stiffness and strength of the frame; but on the other hand, they also decrease the system ductility (Kodor et al 1995).    Figure 10: Masonry infill walls Model (Yousuf & Bagchi, 2009) Typically, infill walls are not accounted for in the structural design of the system, but only the contributing addition mass is considered. However, addition of relatively stiff masonry infill wall in Ductile Steel Moment Frame has a significant impact on the seismic performance of the system due to high flexibility of steel frame and high stiffness of masonry walls. Yousuf and Bagchi (2009) confirmed that infill walls reduce the deflection and ductility in the system and hinging occurred in columns at locations other than the base. Therefore, it is necessary to isolate the infill walls from moment frame and be designed as structural components.  Masonry infill walls are typically designed as diagonal struts as shown in Figure 10. The CLT infill panels are found to provide higher strength and stiffness than OSB/Plywood shear walls. The reduction in ductility is least severe for low ductility moment frames and no evident benefit was found in choosing high ductility over low ductility moment frame. More detailed parametric studies are required to optimize the member sizes in order to get maximum ductility in the 19 system. Further research and experimental testing, mainly seismic reduction factors and connection behavior, need to be carried out for successful implementation of such a hybrid system (Dickof et al, 2012). 2.4.3 Mass Timber Construction Tall wood buildings are not a new concept. One example are Pagodas as high as 19 story buildings in Japan, built 1400 years ago and still standing in one of the highly seismic regions in the world (Green and Karsh, 2012). The Stadthaus project, London (2008) is an example of using mass timber in multi-story construction. It is a nine story building constructed entirely with timber (CLT), claimed as the world’s tallest pure timber residential building (at the time of completion). Mass timber construction is an approach of combining mass timber panels with structural technology to produce a system whose behavior is significantly different from light wood system. They behave more like concrete structures. Mass timber such as Laminated Strand Lumber (LSL), Laminated Veneer Lumber (LVL) and CLT are not only stronger and stiffer than conventional timber but also easier to design with due to their higher uniformity. 2.5 Recent Experimental Research on CLT and Hybrid Systems 2.5.1 Ceccotti et al (2010) Large scale dynamic tests have been performed to evaluate the ductility and overstrength factor for CLT panel structures; e.g. a three storey CLT building test was performed on a unidirectional shake table by NIED and CNR-IVALSA in Japan (Ceccotti et al, 2010). Tests were performed using Kobe, El Centro, and Nocera Umbra ground motions adjusted to peak ground accelerations for 0.15 g and 0.5 g. The test building was approximately 7 × 7 m in plan and 10 m tall. The walls were composed on 85 mm thick wall panels and 142 mm thick floor panels. No damage 20 was observed in any component at a peak ground acceleration of 0.5 g. When the ground acceleration was increased to 0.8 g, slight deformation was noticed in the screws at the vertical joints between the panels. Hold down failure was observed through pull out and bending of the nails when the peak ground acceleration was increased to 1.2 g deformation in the screws between the panels was also observed. 2.5.2 Popovski & Karacabeyli (2011) To determine the structural properties of CLT, Popovski and Karacabeyli (2011) performed a number of semi-static tests on CLT walls. The set up included varying connectors at the base. Single CLT wall and three panel CLT walls were tested. For both the walls, the height and length were 2.3 m and 3.45 m, respectively. The three panel wall was step jointed by screwing between panels. At the base of both walls, Type B brackets of 3.9 mm diameter and 89 mm length were used. Upon cyclic loading in accordance with CUREE protocol, three types of responses were observed, overturning, rocking or combination of the two, see Figure 11. Rocking and deflection of connection caused the maximum energy dissipation and subsequent failure. Hysteretic pinching behavior was observed as shown in Figure 12.   Figure 11: CLT Wall Response to Lateral Loading (Schneider, 2009) 21  Figure 12: Semi-static CLT Wall Tests - Effect of Connection between Panels: (left) single panel CLT wall, (right) three panel CLT wall (Popovski and Karacabeyli 2011) CLT walls showed enhanced seismic performance compared to light frame construction: CLT construction is far less susceptible to “Soft Story” mechanism than the platform frame systems since the panels are also vertical loading carrying components and remain in place without complete collapse (Popovski & Karacabeyli, 2011).  2.5.3 Fragiacomo et al (2011) A seven story building made of CLT slabs and walls was tested at the E-defense shake table as shown in Figure 13. The building was subjected to 100% of the Kobe earthquake with a peak ground acceleration of 0.82 g in one direction and 0.6 g in the perpendicular direction. The building responded with limited structural damage. Some damage to the connectors in the hold downs were noticed, although no failure occurred. Additionally, with appropriately ductile connections between the wall panels, an Rd of 3.0 is achieved (Fragiacomo et al., 2011). This finding was supported by other research with appropriate ductile connections (Yeoh et al, 2011). 22  Figure 13: 7 Story CLT Shake Table Test (Fragiacomo et al. 2011) 2.5.4 Numerical Investigations on Hybrid Systems Rinaldin et al. (2011) created a numerical model of a single CLT panel with connecting brackets. Two non-linear hysteretic behavior models were created for the brackets: one for shear only, to represent the angle bracket connections, and one for tension and compression in the hold-down connections. The timber panel was modeled as a shell element. The cross section was defined as five layers of linear elastic orthotropic wood material assuming that the all plastic deformations would occur in the connectors. Contact springs were also placed at the base of each shell along the bottom of the wall. The results from this model were compared with the results from wall tests and were found to match closely in hysteretic behavior as well as total energy dissipated. Ceccotti (2008) performed an analysis to predict the results of the 3D three story building shake table test. An analytical model, created in Drain3D, was modified to allow for the type of non-23 linear behavior of timber connections. The model consists of three major components, rigid panels modeled as stiff braced frames, and two types of non-linear springs: one type to represent the angle brackets, with symmetric nonlinear pinching behavior to match experimental data and the other type to represents the hold-downs, with non-symmetric behavior. Non-linear pinching behavior is modeled in compression and very stuff linear elastic behavior is modeled in tension. Dickof (2013) numerically studied CLT-steel hybrid systems at three, six, and nine story heights, examining the seismic response of this type of hybrid SFRS in regions with moderate to high seismic hazard indices. A non-linear model of a 2D in-filled frame system was developed and compared to the behavior of a similar plain steel frame at each height. Parametric analyses were performed to determine the effect of the panels and the connection configuration, steel frame design, and panel configuration in a multi-bay system. Static pushover loading was applied alongside semi-static cyclic loading to allow a basis of comparison to future experimental tests. Dynamic analyses were run using ten ground motions linearly scaled to the uniform hazard spectra for Vancouver, Canada with a return period of 2% in 50 years as, 10% in 50 years, and 50% in 50 years to examine the effect of infill panels on the interstory drifts. The ultimate and yield strength and drift capacity were used to determine the overstrength and ductility factors as described in the NBCC (NRC, 2010). It was observed that strength and stiffness of the system increased almost linearly with addition of each CLT panel, while at the same time interstory drift was reduced. The results showed CLT infill panels are better suited to low ductility systems. Ductility factor of 3.0 and overstrength factor of 1.3 have been recommended for such system.  24 2.6 FFTT System 2.6.1 Structural System  A new innovative system called FFTT- “Finding Forest through Trees”, predominantly a mass-timber vertical system bolted with partially embedded steel beam section, has been introduced by Green and Karsh (2012). No concrete is used beyond grade level and the system relies on steel sections for ductility. Steel beams have their sections reduced at the desired location to initiate plastic hinges under seismic loads. Due to the combination of high strength to weight ratio of mass timber and possible enhancement of lateral strength and ductile behavior due to steel sections, this system can serve as a viable option for the construction of high-rise timber structures in future.  The FFTT System consists of large timber panels acting as the vertical system. Beam elements made of steel sections are bolted to the wall panels and they act as the ductile weak link of the system (Figure 14). Beams are designed to have reduced cross-section near the end of the beam, such that plastic hinging occurs at these weak sections at or near design load levels. This provides the required ductile behavior and resistance to ground shaking.    Figure 14: Solid Panel Core and Intersecting Ductile Steel Link Beams (Green and Karsh, 2012) 25 The four combinations of SFRSs proposed for FFTT System based on the number of stories are listed in Table 4. The SFRS combination considered in this study is ‘Type 3’, which is a combination of Structural Core Wall and Perimeter Wall System. The schematic sketch of the system is shown in Figures 15 and 16. The FFTT system is laterally supported by core wall and perimeter structural wall. Steel beams run all across the perimeter wall supporting the panels over the opening and contributing to the overall ductility of the system. The wall is anchored down using ductile hold downs or dampers and rigid (elastic) shear connectors.   Figure 15: Type 3 Lateral Load Resisting System for FFTT (Green and Karsh, 2012)  Figure 16: Type 3 Lateral Load Resisting System for FFTT (Green and Karsh, 2012) 26 Table 4: FFTT System Options  Option Lateral Load Resisting Combination Storeys 1 Structural Core Wall – Glulam Perimeter Columns 12 2 Structural Core Wall – Interior Shear Walls – Glulam Perimeter Columns 20 3 Structural Core Wall – Perimeter Moment Frame 20 4 Structural Core Wall – Interior Walls and Exterior Moment Frame 30  A good engineering design of hybrid system like FFTT could overcome the challenges faced by the performance of light frame timber structures and set the standard for the development of construction technology for safe high-rise timber structures. Further structural analyses, testing and diligent peer review, however, are necessary to satisfy all code requirements before the successful implementation of the FFTT system. Advanced dynamic non-linear analyses, understanding of moment-frame behavior, detailed connections and cost analyses, fire performance testing, construction and erecting engineering are recommended as future studies. 2.6.2 Experimental Investigations on FFTT Connection Recently, through collaboration between UBC and FPInnovations, the effect of steel embedment length on the FFTT connection system was experimentally investigated (Bhat, 2013). The experimental program included 7 layer CLT panels as primary lateral force resisting system, connected by steel beams to provide ductility. To investigate the effect of embedment length on the load deformation response, a total of five different combinations of beam-wall connections were tested. Three of these lay-outs involved wide flanged section as steel beam while for the remaining two series, hollow steel sections were used. The CLT panels were 3 m long and 914 mm wide. Among the three tests conducted with wide flange section, one was partially embedded, next one was fully embedded and the last was also fully embedded but with reduced 27 cross section near the beam wall joint. The embedment length of the beam in each of these experiments was the total width of the CLT panel (914 mm). Another two series were conducted using hollow-steel sections with varying embedment lengths. In all five cases, the overhanging length of the beam was kept constant at 762 mm. The experimental setup is shown in Figure 17. The test specimens were subjected to quasi-static monotonic and reversed cyclic loading. At six different locations on the beam (three on the overhanging portion and three inside the CLT panel), load-deformation responses were obtained.  Figure 17: Typical Setup and Instrumentation (Bhat, 2013) The set-ups with wide flange beams showed damage to the CLT panel when the moment reached around 34.3 kN-m. This force produced excessive compressive stress on the CLT panels. Those set-ups with HSS sections as steel beam reached an ultimate moment of around 13.7 kN-m, and thereby did not produce stresses to cause noticeable damage to the CLT panels.  28 Cyclic tests showed good hysteretic behavior and a maximum moment of 33.9 kN-m. A typical load-deformation response from the tests for HSS section is shown in Figure 18.  Figure 18: Load-deformation plot of the test configuration 4 (Bhat, 2013) Bhat (2013) investigated a total of five different configurations. Even though HSS sections behaved well, these are not practical for mid-rise building construction. HSS sections are very small in size and building construction demand significantly larger beam sections. For construction purpose W sections are preferred. W sections are available at larger sizes to suit the demand of high-rise buildings. In her tests, she used only one W section size (W 150) and did not vary the embedment length which could be a very important design parameter. So, it is imperative to conduct further experiments with W sections incorporating variation in beam size as well as embedment length and depth to find out if these sections are suitable for the FFTT system. Also only experimental studies are not adequate to draw conclusion and formulate design guidelines for a new system. These results must be complimented by numerical studies. Therefore, further studies (both numerical and experimental) needed to be conducted. 29 Chapter  3: NUMERICAL INVESTIGATION ON FFTT SYSTEM 3.1 Finite Element Model Development To complement the experimental studies conducted by Bhat (2013), finite-element-analyses (FEA) were conducted on all test configurations. For this purpose, three-dimensional (3D) models were developed using the commercial software package ANSYS 14.5 (ANSYS Inc, 2013). The details of modelling assumptions are described in the following. 3.1.1 Modelling of CLT Panels For modelling of the CLT panels, SOLID186, a higher order 3D, 20-node solid element, was used that exhibits quadratic displacement behavior. The element is defined by 20 nodes having three degrees of freedom per node. The element supports plasticity, hyperelasticity, creep, stress stiffening, large deflection, and large strain capabilities. It also has mixed formulation capability for simulating deformations of nearly incompressible elastoplastic materials, and fully incompressible hyperelastic materials. The wood material has been modelled as being a linear elastic orthotropic material. The mechanical properties of CLT used in the model are shown in Table 5. The x and y direction properties were altered to represent the different layers of CLT. The layers of CLT panels are glued together so that force transfer occurs between layers. Table 5: Properties of CLT Elastic Moduli(MPa) Poisson Ratio Shear Moduli (MPa) Ex 11000 vxy 0.40 Gxy 700 Ey 5500 vyz 0.40 Gxy 500 Ez 600 vzx 0.04 Gxy 70 30 3.1.2 Modelling of Steel Beams Similar to CLT, SOLID 186 elements have been used to model the steel beam. Bilinear isotropic elasto-plastic material properties have been used to accommodate the post-yield inelastic response of the steel beam. The material properties are shown in Table 6. Table 6: Properties of Steel beam Modulus of Elasticity, E (MPa) 210,000 Yield Strength, fy (MPa) 310 Post-yield Stiffness, α (MPa) 5,000 Ultimate Strength, fu 420 Wide Flange Section W 150 x 26 Hollow Steel Section HSS 100 x 50  3.1.3 Contact Simulation  During the experiments, the steel beam came in contact with the CLT panel as it was pushed. To simulate this behavior, surface to surface contact technology has been used. This type of contact provides linear traction-separation, standard contact behavior after debonding and has capability of modeling unloading and reloading phase. The ANSYS Contact Manager was used to define areas of contact; the coefficient of friction (μ) between steel and wood has been set to 0.3. 3.1.4 Boundary Conditions All degrees of freedom were constrained at the base and at the top of the CLT panel. The steel beam was prevented against lateral buckling by restraining its translation along longitudinal and lateral direction inside the CLT panel. But the beam was allowed to rotate about its longitudinal axis. A typical finite element model is shown in Figure 19. 31  Figure 19: Finite Element Model of test configuration 1 3.1.5 Load Application A concentrated load was applied at the free end of the beam to simulate the actuator load during the experiments. The load was applied stepwise with small increments of time (0.05 seconds) to allow the solution to converge.  3.1.6 Post Processing Upon completion of analysis, results were extracted using the post-processing tool of ANSYS. The parameters of interest are compressive and shear stress inside the CLT panel, maximum deformation of the panel, maximum deformation of the steel beam at the free end and load-deformation behavior. The stress and deformation plots were obtained using “General Post-processing” feature of ANSYS. The load-deformation curves were constructed by obtaining the stepwise load and corresponding deformation values using the “Time History Post-processing” feature of ANSYS. 32 3.2 Numerical Results  3.2.1 Configuration 1: Partially Embedded Wide Flange Section For this configuration, the wide flange beam was partially embedded inside the CLT panel. During the experiment, beam yielding occurred at the panel beam interface at average load of 40 kN. The maximum load of 45.8 kN were observed (Bhat, 2013). These values correspond to 30.5 kN-m and 34.9 kN-m bending moment at the same interface, respectively. This configuration was numerically analyzed and a maximum force of 45.8 kN was applied. The deformation values were computed at the same six locations as during the experiment. The shear and compressive stress plots as obtained from ANSYS are shown in figures 20 and 21, respectively. The load-deformation plots for the cantilever and embedded portions are shown in Figures 22 and 23. It is observed that the load-deformation curves obtained from the numerical analysis are in good agreement with the experimental results for both cantilever and embedded portion, thereby validating the numerical model. However, the degrading portion of the curve was not captured because of using the bilinear steel material model.  The observed maximum compressive and shear stresses inside the wood were 68 MPa and 14 MPa, respectively. According to the CLT handbook (Gagnon and Pirvu, 2011), the maximum elastic compressive and shear strength values for CLT are 11.5 MPa and 5.5 MPa, respectively. Therefore, the observed values have gone well beyond the elastic range indicating that plastic deformations have occurred. These results also indicate that the use of elastic material model for CLT is not adequate to obtain actual stress and deformation results. A plastic CLT material model which considers the post yield behavior of CLT would provide better results.  33  Figure 20: Shear stress plot test configuration 1  Figure 21: Compressive stress plot test configuration 1 34  Figure 22: Comparative load deformation plot of test configuration 1: embedded portion  Figure 23: Comparative load deformation plot of test configuration 1: cantilever portion 05101520253035400 2 4 6Moment (kN‐m)Deformation (mm)Loc1_exploc1_numloc2_exploc2_numloc3_exploc3_num05101520253035400 10 20 30 40 50 60Moment (kN‐m)Deformation (mm)loc4_exploc4_numloc5_exploc5_numloc6_exploc6_num35 3.2.2 Configuration 2: Fully Embedded Wide Flange Section For this configuration, as shown in Figure 24, the wide flange beam was fully embedded inside the CLT panel. During the experiment, beam yielding occurred at the top flange of panel beam interface at 41 kN and the maximum load was 45.4 kN (Bhat, 2013). These values correspond to 31.25 kN-m and 34.6 kN-m bending moment at the wall beam interface, respectively. Similar to configuration 1, a force of 45 kN was applied. The deformation values were measured at the same six locations as during the experiment and compared with. The shear and compressive stress plots are shown in Figures 25 and 26, respectively. The load-deformation plots for the cantilever and embedded portions are shown in Figures 27 and 28 and are found to be reasonable. However, as previously stated, the degrading portion of the curve was not captured because of using bilinear steel material model. The maximum compressive and shear stress inside the wood are 83 MPa and 51 MPa, respectively. So again, the observed values have gone well beyond the elastic range indicating that plastic deformation have occurred.  Figure 24: Finite element model of test configuration 2 36  Figure 25: Compressive stress plot test configuration 2  Figure 26: Shear stress plot test configuration 2 37  Figure 27: Comparative load deformation plot of test configuration 2: embedded portion  Figure 28: Comparative load deformation plot of test configuration 2: cantilever portion 05101520253035400 2 4 6 8 10Moment (kN‐m)Deformation (mm)loc1_exploc1_numloc2_exploc2_numLoc3_exploc3_num05101520253035400 10 20 30 40 50 60Moment (kN‐m)Deformation (mm)loc4_exploc4_numloc5_exploc5_numloc6_exploc6_num38 3.2.3 Configuration 3: Fully Embedded Section with Reduced Cross Section The elements of test configuration 3 are shown in Figure 29. It was conducted on fully embedded wide flange I-sections with reduced cross-section near the beam-panel interface. During the experiment, beam yielding occurred at the panel beam interface at average load of 44.5 kN (Bhat, 2013). This value corresponds to 33.9 kN-m bending moment at the wall beam interface. After numerically analyzing this configuration, the deformation values were computed and compared with the experimental results. The load-deformation plots for the cantilever and embedded portions are shown in Figures 30 and 31, respectively. The load-deformation curve obtained from the numerical analysis is in good agreement with the experimental result for both cantilever and embedded portion. Reducing the section at the interface did not have significant effect on the overall behavior of the system. The compressive and shear stress plots show that the values were lower because of the reduction in steel flange at the panel-beam interface. The maximum compressive and shear stress inside the wood are 45 MPa and 24 MPa, respectively. Still, the observed values have gone well beyond the elastic range.   Figure 29: Test configuration 3 39  Figure 30: Comparative load deformation plot of test configuration 3: embedded portion  Figure 31: Comparative load deformation plot of test configuration 3: cantilever portion 05101520253035400 3 6 9 12Moment (kN‐m)Deformation (mm)Loc1_exploc1_numloc2_exploc2_numloc3_exploc3_num05101520253035400 5 10 15 20 25 30 35 40 45 50Moment (kN‐m)Deformation (mm)loc4_exploc4_numloc5_exploc5_numloc6_exploc6_num40 3.2.4 Configuration 4: Full Embedment Length of Hollow Steel Section For this configuration, hollow structural steel sections were fully embedded inside the CLT panel. The embedment length was the total width of the panel. During the experiment, beam yielding occurred at the panel beam interface at 17 kN and the maximum load was 18.5 kN (Bhat, 2013). These values correspond to 13.0 kN-m and 14.1 kN-m bending moment at the wall beam interface, respectively. This configuration was numerically analyzed; the finite element model of this configuration is shown in Figure 26. The deformation values were measured at the same six locations as during the experiment and compared with. The load-deformation plots for the cantilever and embedded portions are shown in Figures 33 and 34, respectively. It is observed that the load-deformation curve obtained from the numerical analysis is in good agreement with the experimental result for both cantilever and embedded portion. No damage in the CLT panel was observed during the experiment; from the numerical model it was found that both horizontal and vertical stress values are too small to cause any damage in the panel.  Figure 32: Finite element model of test configuration 4 41  Figure 33: Comparative load deformation plot of test configuration 4: embedded portion  Figure 34: Comparative load deformation plot of test configuration 4: cantilever portion 05101520253035400 2 4 6 8 10Moment (kN‐m)Deformation (mm)loc1_exploc1_numloc2_exploc2_numLoc3_exploc3_num02468101214160 10 20 30 40 50 60 70 80 90 100Moment (kN‐m)Deformation (mm)loc4_exploc4_numloc5_exploc5_numloc6_exploc6_num42 3.2.5 Configuration 5: Reduced Embedment Length of Hollow Steel Section For this configuration (Figure 35), hollow structural steel sections were fully embedded inside the CLT panel. However, the embedment length was reduced to two-third of the width of the panel. During the experiment, beam yielding occurred at the panel beam interface at 17.1 kN and the maximum was 18.5 kN (Bhat, 2013). These values correspond to 14.1 kN-m and 14.1 kN-m bending moment at the wall beam interface, respectively. This configuration was numerically analyzed and the deformation values were measured at the same six locations as during the experiment and compared with. The shear and compressive stress plots are shown in Figures 36 and 37, respectively. The load-deformation response for this configuration is shown in Figure 38. It is observed that the load-deformation curve obtained from the numerical analysis is in good agreement with the experimental result for both cantilever and embedded portion.   Figure 35: Finite element model of test configuration 5 43  Figure 36: Shear stress plot test configuration 5  Figure 37: Compressive stress plot test configuration 5 44  Figure 38: Comparative load deformation plot of test configuration 5 The compressive and shear stress plots for configurations 4 and 5 are similar. The values were lower than those corresponding to configurations 1, 2 and 3. The HSS sections had a very small section modulus compared to the W sections. Therefore, the peak loads for configurations 4 and 5 were much smaller (around 18.5 kN compared to over 40 kN for W sections). The maximum compressive and shear stress inside the wood for configuration 4 was 22 MPa and 9 MPa, respectively; while these values were 24 MPa and 10 MPa for configuration 5. Still, the observed values have gone well beyond the elastic range indicating that the use of an linear-elastic material model for CLT is not adequate to obtain actual stress results. A non-linear CLT material model which considers the post yield behavior of CLT would provide better results.    02468101214160 10 20 30 40 50 60 70 80 90 100Moment (kN‐m)Deformation (mm)Loc1_exploc1_numloc2_exploc2_numloc3_exploc4_exploc4_numloc5_exploc5_num45 3.2.6 Summary on Model Results from Previous Tests A summary of stress and deformation results for each configuration is presented in Table 7.  Table 7: Results from previous experimental tests and numerical simulation Series Maximum deformation inside CLT at peak load (mm) Stresses from numerical analysis (MPa) Experimental (Bhat, 2013) Numerical Compressive Shear Configuration 1 4.0 4.0 68.4 13.2 Configuration 2 7.5 7.0 83.7 21.1 Configuration 3 6.0 5.5 45.3 16.1 Configuration 4 0.55 0.25 24.1 11.6 Configuration 5 5.0 7.0 28.3 13.5  Overall, it is observed that the load-deformation curves obtained from the numerical analyses are in close agreement with those extracted from experimental investigations. However, the stress plots from the numerical analyses show that both compressive and shear stresses for each configuration were beyond the elastic strength limit of timber. Therefore, the connections have undergone plastic deformation and the linear elastic material model for CLT is no longer adequate to evaluate the stress magnitudes. Nonlinear material models for CLT have to be developed to obtain more realistic stress results.  3.3 Numerical Study to Improve Connection Configuration The experimental tests conducted by Bhat (2013) were significant, since they marked the beginning of research on the FFTT system at the component level. The results indicated that the HSS section allowed the desired failure mechanism to form, while the wide flange section 46 caused damage to the CLT panel. However, the hollow section was very small and, therefore, not suitable for high-rise construction. Moreover, in the tests conducted by Bhat (2013), the beams were not properly restrained against buckling, which is a critical consideration. Considering these shortcomings, an attempt has been made to improve the connection configuration. Only wide flange sections were considered since bigger sized sections can be used.  3.3.1 Geometry The CLT panels considered were similar to those used for testing by Bhat (2013). ASTM A992 WF beams with 350 MPa yield strength were chosen, additionally, bearing plates and side plates of the same steel property were included. A total of four bearing plates (each 150 mm in length and 6.25 mm thick) were placed at top and bottom of the beam to avoid stress concentration at the face of the panel beam interface. Moreover, to prevent buckling, four side plates of 6.25 mm thickness were also placed along the web of the beam. Also, the beam was supported against lateral movement. The material properties are the same as those shown in Tables 4 and 5. The model is shown in Figure 39. In Figure 40, the beam with the bearing and side plates is shown in detail. 47  Figure 39: Finite element model for numerical optimization 1  Figure 40: Details of the steel beam with bearing and side plates 48 3.3.2 Parameter Variation For the purpose of the numerical study; the dimensions of the CLT panel and side plates were kept constant. The parameters which were varied include spacing between bearing plates, embedment length and steel beam size and the bearing plate length. The ranges within which these parameters were varied are shown in Table 8. Table 8: Parameter range for numerical study Parameter Range Embedment length (mm) 500, 600, 700, 800, 900 Spacing between bearing plates ( mm) 250, 300, 350, 400, 450 Steel beam  W100 x 19.3, W130 x 23.8, W150 x 29.8 Bearing plate length (mm) 100, 125, 150  3.3.3 Parametric Study Results A number of analyses have been carried out to observe the behavior of the connection by varying the parameters as shown in Table 8. The maximum compressive and shear stresses parallel to grain and deformations inside the CLT panel for various combination of parameters are summarized in Tables 9, 10 and 11.   49 Table 9: Results of parametric study (Beam: W 150 x 29.8) Parameters Results Embedment length (mm) Plate   length (mm) Plate Spacing (mm) Compressive stress* (MPa) Shear   stress* (MPa) Deformation inside CLT (mm) 500 100 400 275 145 62 125 375 267 142 59 150 350 259 141 56 600 100 500 234 136 58 125 475 229 136 56 150 450 217 132 52 700 100 600 202 129 55 125 575 198 127 51 150 550 197 124 47 800 100 700 179 124 48 125 675 173 121 43 150 650 169 120 41 900 100 800 154 119 44 125 775 146 117 40 150 750 141 115 35 *Parallel to grain     50 Table 10: Results of parametric study (Beam: W 130 x 23.8) Parameters Results Embedment length (mm) Plate   length (mm) Plate Spacing (mm) Compressive stress* (MPa) Shear   stress* (MPa) Deformation inside CLT (mm) 500 100 400 243 134 46 125 375 237 131 45 150 350 231 128 43 600 100 500 217 130 47 125 475 212 126 43 150 450 207 125 42 700 100 600 197 128 45 125 575 188 125 41 150 550 179 121 41 800 100 700 165 115 40 125 675 157 112 37 150 650 151 110 33 900 100 800 131 106 34 125 775 121 101 30 150 750 119 97 26 *Parallel to grain     51 Table 11: Results of parametric study (Beam: W 100 x 19.3) Parameters Results Embedment length (mm) Plate   length (mm) Plate Spacing (mm) Compressive stress* (MPa) Shear   stress* (MPa) Deformation inside CLT (mm) 500 100 400 195 107 33 125 375 191 104 31 150 350 186 103 29 600 100 500 165 101 30 125 475 161 96 27 150 450 156 95 26 700 100 600 137 90 28 125 575 132 86 25 150 550 129 83 23 800 100 700 112 81 24 125 675 106 74 22 150 650 104 72 21 900 100 800 91 73 20 125 775 84 65 17 150 750 79 59 16 *Parallel to grain    52 It is observed that the parallel to grain compressive stress values remain very high and indicate crushing. A reduction in stress values has been achieved by increasing the embedment length and spacing between bearing plates. However, stresses still remain over the elastic limit. It means that for the given panel dimension and strength property, larger wide flange sections might be too strong and will cause crushing in the panel. However, this conclusion is drawn based on the elastic material model of CLT which is not sufficient to represent the actual behavior.  Upon reaching this conclusion, the smallest commercially available Wide Flange section (W 100 x 19) was chosen for the further analyses. To extract the load-deformation behavior, four points were chosen marked as LVDT-1 and LVDT-2 (inside the CLT panel) and LVDT-3 and LVDT-4 (in the beam at the cantilever portion). During the subsequent experiments (as reported in chapter 4), these locations were used to instrument the test specimens with Linear Variable Differential Transformers (LVDTs) and compare the experimental results to the numerical results.  In the model, the system yielded at an applied moment of 30.2 kN-m at the beam-wall interface. The maximum deformation computed at the end of the beam was 15 mm. The connection continued to pick up load up to 42.1 kN-m. The deformations inside the CLT panel were small as can be seen from the load-deformation plot (Figure 41). Observing the sign of displacement values, it can be concluded that the beam rotated about a point between LVDT- 1and LVDT-2. The compressive stresses are still very high (in the region of 80 MPa). The compressive stress plot is shown in Figure 42. However, this large stress occurred only within a very small region at the beam-panel interface. Therefore, unless the CLT panel dimension and strength properties are increased; only small Wide-Flange sections might provide the expected ductile behavior. Hence, to experimentally validate this outcome, W100 x19 section was chosen for the subsequent experimental investigations.  53  Figure 41: Load-deformation plot at different points of interest for the optimization study when the W100 x19.3 beam was fully embedded with 150 mm bearing length and 350 mm spacing   Figure 42: Contour plot of compressive stress parallel to grain inside the CLT panel when the W100x19.3 beam was fully embedded with 150 mm bearing length and 350 mm spacing  051015202530354045‐5 5 15 25 35 45 55 65Moment (kN‐m)Deformation (mm)LVDT‐1LVDT‐2LVDT‐3LVDT‐454 3.3.4 Parametric Study with Partial Embedment Depth The previous results have been obtained when the embedment depth is the full depth of the beam. During experimental testing of this configuration, rolling shear failure occurred as explained in Chapter 4. CLT is very weak against rolling shear; to avoid such failure, another numerical analysis has been carried out with a partial embedment depth of the steel beams. The embedment depth considered was 85 mm instead of 102 mm. Also the distance between bearing plates and consequently the embedment length were increased to reduce the bearing force. For this analysis the center to center distance between bearing plates was increased to 665 mm (from 350 mm). The model is shown in Figure 43.  Figure 43: Finite element model for numerical optimization 2  55 The compressive stress plot for this improved configuration is shown is Figure 44. The results show significantly reduced compressive and shear stress values. The compressive stress (parallel to grain) reduced from 186 MPa to 102 MPa while the shear stress (parallel to grain) dropped to 63 MPa from 76 MPa. This is due to the fact that the bearing force has been reduced. A ductile failure mode is predicted with large deformation in the steel beam. The free end of the beam deformed up to 125 mm whereas the deformation inside the CLT panel is around 2 mm. The load-deformation plot for this improved configuration is shown in Figure 45.   Figure 44: Contour plot of compressive stress parallel to grain numerical model when the W100 x19.3 beam was partially embedded with 150 mm bearing length and 665 mm spacing   56  Figure 45: Load-deformation plot at different points of interest for the optimization study when the beam was W100 x19.3 with 150 mm bearing length and 665 mm spacing  3.4 Discussion on Numerical Analysis and Optimization Studies The numerical studies complemented the experimental investigation by Bhat (2013). It demonstrated that the HSS section is well suited to FFTT connection system. The HSS section being very small, caused small compressive and shear stress to the CLT. It also failed in a ductile manner with large deformation which is desirable for the connection. The HSS section, however, is not practical for the actual construction of structures owing to its very small size. From a practical point of view, wide flange sections are better suited since larger sections are available. However, when wide flange sections were used for the study, the wood was subjected to excessive stress which may lead to failure before steel. The obtained compressive and shear stresses were beyond the elastic strength limit specified by the CLT Handbook (Gagnon and Pirvu, 2011).  0510152025303540‐5 20 45 70 95 120Moment (kN‐m)Deformation (mm)LVDT‐4LVDT‐3LVDT‐2LVDT‐157 To search for a better connection configuration, a numerical parameter study was carried out. The study revealed that by incorporating bearing and side plates the connection behavior can be significantly improved. As can be seen from Tables 9, 10 and 11 all the parameters of interest impact the amount of stress and deformation magnitudes inside the CLT panel.  The size of the beam is a major factor. The greater the beam size, the more stress and deformation it causes to the beam which is expected. During this parametric study, three beams were chosen with the largest being W 150x 29.8 and smallest being W100 x19.3, while the other section being W 130x 23.8. The difference in elastic section modulus between these beams is quite significant. W 150x29.8 has elastic section modulus of 218.4 cm3. The corresponding elastic section moduli of W 130x 23.8 and W100 x 19.3 are 139.5 cm3 and 89.9 cm3 respectively. The W 150x 29.8 section caused much higher compressive and shear stresses as well as larger deformations inside the CLT panel than the smaller sections given that other parameters remain same. For the W150 x29.8 section, the maximum parallel to grain compressive and shear stresses observed were 275 MPa and 145 MPa, respectively. This was observed when 100 mm bearing plates at 400 mm spacing was considered. It is also noticeable that by increasing the bearing plate length to 150 mm and spacing to 775 mm, the compressive stress could be reduced to 141 MPa (49% decrease). The shear stress also reduced from 145 MPa to 115 MPa (21% decrease).  For the W130 x23.8 section, the maximum compressive and shear stresses observed were 243 MPa and 134 MPa, respectively. These stresses are much lower than those caused by the W 150 x 29.8 section. This was observed when 100 mm bearing plates at 400 mm spacing was considered. By increasing the bearing plate length to 150 mm and spacing between them to 775 mm, the compressive stress could be reduced to 119 MPa (49% of maximum value). The reduction in shear stress was 20 % (from 134 MPa to 107 MPa). 58 For the W100 x19.3 section, which is the smallest commercially available wide flange section, the maximum compressive and shear stresses observed were 195 MPa and 107 MPa, respectively. Similar to the behavior observed during optimization with bigger sections, the highest stress values were observed when the lengths of the bearing plates were the shortest (100 mm) and the spacing between the plates was 400 mm. Again, by increasing the bearing plate length to 150 mm and spacing between them to 775 mm, the compressive stress could be reduced to almost 41% (from 195 MPa to 79 MPa). The shear stress also reduced significantly (from 107 MPa to 59 MPa). Form Tables 9, 10 and 11 and the preceding discussion, it is obvious that the length of the bearing plates, the embedment length of beam inside CLT and, consequently, the spacing between bearing plates are critical factors for connection design optimization. The effect of increasing the embedment length and therefore spacing between bearing plates are very significant as can be seen in Figure 46. This figure shows the variation in compressive stress with increasing embedment length of beam for a W 100x 19.3 section. Significant reduction in compressive stress was achieved by bearing plate length same but increasing embedment length. However, the variation in compressive stress with bearing plate length is not as pronounced which can be seen from Figure 47. This figure shows the variation in compressive stress with bearing plate length for the W100 x 19.3 beam. The reduction in stress when just increasing the bearing plate length is negligible.  Similar to the compressive stress, the shear stress can also be significantly reduced by increasing the spacing between bearing plates and the length of the plates. Again, the effect of increased embedment length and spacing between plates is more pronounced (Figure 48), while the influence of bearing length is negligible (Figure 49). As the stresses are reduced significantly, so 59 do the displacement values with increasing spacing between plates and the length of the plates. The variation of displacements with embedment length and with bearing plate length are shown in Figures 50 and 51 for the W100 x 19.3 beam. When the spacing between bearings plates are increased, the lever arm for resisting the applied moment increased, therefore, the forces transferred through the bearing plates are significantly reduced. The effect of increased bearing length means greater bearing area for force transfer between beam and CLT panel. However, when the length of plate is increased, the center to center spacing between them is decreased if the embedment length of the beam is kept constant. This might be the cause of not significant reduction in stresses even though the bearing area is increased. For the preceding discussion, the limitation of the linear-elastic material model for CLT has to be kept in mind: all the stress values are purely numerical results and can only be used for comparative purposes. In reality, the localized stress peaks would lead to wood crushing and plastic deformation with a resulting stress redistribution over a larger area but with a significantly reduces stress magnitude.  Overall, on the basis of the linear elastic CLT material model, it was found that even the smallest wide flange steel section was stronger than the CLT panel. While it implies that wide flange section is stronger than the CLT and would not provide ductile failure mode; it also indicates that the linear elastic model for CLT used for the numerical study is not entirely sufficient. Nonlinear material model for CLT will better simulate the actual behavior of the system.    60  Figure 46: Variation in compressive stress with embedment length of beam (Beam: W 100 x 19.3)  Figure 47: Variation in compressive stress with length of bearing plate (Beam: W 100 x 19.3) 6080100120140160180200500 600 700 800 900Compressive stress (Mpa)Embedment length (mm)100 mm bearing length125 mm bearing length150 mm bearing length6080100120140160180200100 125 150Compressive stress (Mpa)Bearing plate length (mm)500 mm embedment length600 mm embedment length700 mm embedment length800 mm embedment length900 mm embedment length61  Figure 48: Variation in shear stress with embedment length of beam (Beam: W 100 x 19.3)  Figure 49: Variation in shear stress with length of bearing plate (Beam: W 100 x 19.3) 5060708090100110500 600 700 800 900Shear stress (Mpa)Embedment length (mm)100 mm bearing length125 mm bearing length150 mm bearing length5060708090100110120100 125 150Shear stress (Mpa)Bearing plate length (mm)500 mm embedment length600 mm embedment length700 mm embedment length800 mm embedment length900 mm embedment length62  Figure 50: Variation in displacement with embedment length of beam (Beam: W 100 x 19.3)  Figure 51: Variation in displacement with length of bearing plate (Beam: W 100 x 19.3)  10152025303540500 600 700 800 900Deformation inside CLT (mm)Embedment length (mm)100 mm bearing length125 mm bearing length150 mm bearing length10152025303540100 125 150Deformation  inside CLT (mm)Bearing plate length (mm)500 mm embedment length600 mm embedment length700 mm embedment length800 mm embedment length900 mm embedment length63 Chapter  4: EXPERIMENTAL INVESTIGATION ON FFTT SYSTEM 4.1 Introduction This chapter describes the experimental tests conducted on the improved connection configuration obtained by the numerical study described in Chapter 3. The tests evaluated the behavior of embedded wide flange sections through quasi-static monotonic and reverse cyclic tests. Component tests with two different configurations were conducted in the Structural Laboratory of FPInnovations, Vancouver. The objective of the experimental study was to observe if the new connection layouts initiate the desired “Strong column-week beam” failure mode.  4.2 Materials Two 7-ply CLT panels of grade S-P-F No.1/No.2 of 0.9 m wide and 4 m long were used. The outer laminations were 32 mm while the inner laminations were 35 mm thick because the surfaces were planned. The overall thickness of the panel was 239 mm. The design material properties listed by the manufacturer of the CLT product (Structurlam) used in the project are summarized in Table 8. ASTM A992, W 100 x 19 sections were chosen. The yield strength and ultimate strength of the steel specimens were 350 MPa and 460 MPa, respectively. The modulus of elasticity was taken to be 210 GPa. As bearing and side plates, rectangular flat steel bars of 150 x 100 x 6.25 and 87 x 50 x 6.25 were used, respectively. 64 4.3 Specimen Description The slots, into which the beam sections were embedded, were pre-cut in the CLT panel. A total of 3 slots were cut in each CLT panel to facilitate two static and one cyclic test. The beams embedded into these slots were held in place using two 9.5 mm lag bolts in 12.7 mm drill holes, at 250 mm and 457 mm from wall beam interface for series 1 and 2, respectively. The experimental setup for series 1 along with the position of the four LVDTs (red arrows) is shown in Figure 52. In Figure 53, a side view of the embedment of the beam is shown.The force transfer through bearing of bolts was assumed to be negligible. Complete load transfer occurred through the bearing of steel beams alone. Two series of tests were conducted with two replicates of monotonic test and one cyclic test for each series as shown in Table 12. Table 12: Test specimen description Series Embedment Embedment Length Bolted Connection Distance between plates1 102 mm 500 mm 9.5 mm diameter bolt at 250 mm from interface 350 mm c/c 2 85 mm 914.4 9.5 mm diameter bolt at 457 mm from the  interface 664.4 mm c/c  65  Figure 52: Experimental setup for test series 1  Figure 53: Full embedment of the steel beam inside the CLT panel for test series 1 Series 2 was conducted by embedding the wide-flange I-section 85 mm into the outer 3 plies of the panel. The experimental setup for the partially embedded beams in series 2 is shown in 66 Figure 54. Series 2 was conducted by embedding the section 85 mm into the outer three plies in order to avoid rolling shear failure and to observe if avoiding the rolling shear failure can improve the behavior of the system.  Figure 54: Full embedment of the steel beam inside the CLT panel for test series 2 4.4 Test Procedure The panels were bolted down to the floor at both ends to restrain them from translation, rotation or uplift movement during the experiments. For series 1, four LVDTs were attached to the embedded beam with the first LVDT placed 152.4 mm from the edge of the panel (as shown by red arrows in Figure 52). For series 2, similarly four LVDT s were attached to the beam. The two LVDTs placed inside the CLT panel were located at the center of bearing plates. The LVDTs placed at the cantilever portion were located at 350 mm and 700 mm away from the beam-wall interface.  In the quasi static monotonic tests, the load was applied at the end of the projecting beam through a calibrated actuator (225 kN capacity). The loading was maintained constant at a rate of 12.7 mm/min. For series 1, the load kept on increasing without dropping. Hence no peak load 67 was reached and the loading for the monotonic tests was discontinued when the applied load reached 62 kN, which caused 43.4 kN-m at the beam-wall interface. However, the system showed well defined yield point at 52 kN force (36.4 kN-m). For series 2, similar behavior was observed with well-defined yield point at 45.0 kN force (32.6 kN-m) and the load continued to increase to 56 kN (40.6 kN-m) without degrading. The deformations at 90% yield load from the monotonic tests were chosen as target displacements (100%) for the subsequent reversed cyclic loading tests. The CUREE protocol (Krawinkler et al., 2001) was used for the cyclic loading for each test series (Figure 55 and 56). The loading was programmed to continue with an increment of 20 % beyond the target displacement until 200% of the target displacement. The cyclic tests were conducted at a loading rate of 5 mm/min (equivalent to a rotation of the beam of 0.007 rad/min).  Figure 55: CUREE loading protocol for series 1 ‐180‐160‐140‐120‐100‐80‐60‐40‐200204060801001201401601800 100 200 300 400 500 600 700 800Displacement (% of Max)Time (seconds)Cyclic Displacement ScheduleCUREE Test Protocol100% displacement = 1.125"Load rate= 0.2"/min68  Figure 56: CUREE loading protocol for series 2 4.5 Experimental Results 4.5.1 Series 1: Monotonic Test on Fully Embedded Beam During the tests, beam yielding occurred at the panel-beam interface (Figure 57) at an average 35.1 kN-m bending moment. Both tests showed well defined yield points. However the deformations were larger in test 1 compared to test 2. The system kept on taking load without dropping and no peak load was identifiable. Both tests were terminated when the moment reached 42.8 kN-m at the beam-wall interface.  -180-160-140-120-100-80-60-40-200204060801001201401601800 200 400 600 800 1000 1200 1400Displacement (% of Max)Time (seconds)Cyclic Displacement ScheduleCUREE Test Protocol100% displacement = 2.000"Load rate= 0.2"/min69  Figure 57: Yielding of beam during experimental series 1 Compared to the load-displacement curve obtained from numerical study (Figure 41), these monotonic test results are in close agreement. The curve shapes are similar. But the experimental yield point (36.4 kN-m moment) was 20% higher than numerical value (30.1 kN-m moment). Also the yield displacement at the location of LVDT 4 is similar to numerically obtained value (17 and 15 mm). The deformation that occurred inside the CLT panel during monotonic test of series is shown in Figure 58. 70  Figure 58: Deformation inside the CLT panel during experimental series 1 During both tests, the maximum deformation at the location of LVDT- 4 was around 60 mm which occurred at the 42.8 kN-m moment. At the location of LVDT-3, the deformation at yield moment was around 6 mm, whereas the maximum deformation at highest moment was 25 mm. The difference between yield and maximum displacement indicates that the system has good ductility. The load-displacement plots of the monotonic tests 1 and 2 of series 1 are shown in Figures 59 and 60 respectively. 71  Figure 59: Load-displacement curve: Series-1, monotonic test-1  Figure 60: Load-displacement curve: Series-1, monotonic test-2 05101520253035404550‐5 5 15 25 35 45 55 65Moment (kN‐m)Deformation (mm)LVDT‐1LVDT‐2LVDT‐3LVDT‐405101520253035404550‐5 5 15 25 35 45 55 65Moment (kN‐m)Deformation (mm)LVDT‐1LVDT‐2LVDT‐3LVDT‐472 The in-plane deformation of the embedded portion of the beam was measured at two locations. LVDT-1 corresponds to the LVDT attached at the far end of the embedded beam section. The deformation values obtained from both the LVDTs inside the CLT panel were very small (around 2 mm). The negative displacement value of LVDT-1 and positive displacement values of LVDT-2 indicate that the beam was rotated about a point between these two. The data acquired at LVDT-1 during test 1 was erroneous due to slot fabrication imperfections (e.g. slight spaces between beam and wood) that existed prior to the testing. This error was avoided during test 2 by improving the quality of fabrication. The stiffness of both curves were similar.  There were in-plane deformations inside the panel causing damage to wood before the beam reached yield load. Even though the damage was negligible, it still indicated that the beam is stronger than the CLT panel. The beam chosen, being the smallest commercially available section, reinforces the fact that this connection layout is not ideal for the FFTT system. Further studies need to be conducted with improved connection configuration and nonlinear plastic material properties for CLT, before this system can be considered. 4.5.2 Series 1: Cyclic Test on Fully Embedded Beam No peak load was identifiable from the monotonic tests, but yield point was well defined. Therefore, the deformation at 90% of the peak load was considered as target displacement. The value of this displacement was 28 mm. The setup was similar to monotonic tests. However, the failure mode was rolling shear followed by crushing. The beam rotated about the point where bolt was inserted. Rolling shear crack was seen at the weaker layer of the CLT panel as shown in Figure 61. Cracking and crushing began inside the panel before the beam began to yield. The load was continued to reach 180% of the target displacement. 73 The observed maximum bending moment at the beam-wall interface was 50.1 kN-m. The hysteresis behavior of the top face flange at the location of LVDTs is presented in Figures 62 through to 64. Based on the monotonic and cyclic tests, it can be deduced that the point of rotation of the beam is between LVDT-1 and LVDT-2. The maximum deformation at LVDT-1 which is located 425 mm inside from the beam-wall interface was found to be very close zero (3.8 mm), with negligible energy dissipation. The maximum deformation at LVDT-2 which is located 75 mm inside from the beam-wall interface was also found to be very small (4.4 mm), with negligible energy dissipation.   Figure 61: Rolling shear failure in CLT panel during cyclic test of series 1 74 The hysteretic curves obtained from LVDT-1 and 2 (Figures 62 and 63) show a stiffer slope on one side and flatter slope on the opposite. Very little energy dissipation occurred inside the CLT panel. The readings from LVDT-3, located 375 mm away at the cantilever portion of the beam, were erroneous and therefore not shown. Hysteresis plots at the locations of LVDT- 4 (Figure 64) suggest that almost all energy under cyclic loading was dissipated through the deformation of the cantilever portion of the beam for which the maximum deformation was 48 mm. The cyclic test, like the monotonic test showed that damage occurred in the CLT panel in the form of rolling shear before the steel beam yielded.   Figure 62: Cyclic test: Series-1, LVDT-1 ‐60‐50‐40‐30‐20‐1001020304050‐4 ‐3 ‐2 ‐1 0 1Moment (kN‐m)Deformation (mm)LVDT‐175  Figure 63: Cyclic test: Series-1, LVDT-2  Figure 64: Cyclic test: Series-1, LVDT-4 ‐60‐50‐40‐30‐20‐1001020304050‐5 ‐4 ‐3 ‐2 ‐1 0 1 2Moment (kN‐m)Deformation (mm)LVDT‐2‐60‐50‐40‐30‐20‐1001020304050‐60 ‐50 ‐40 ‐30 ‐20 ‐10 0 10 20 30 40Moment (kN‐m)Deformation (mm)LVDT‐476 4.5.3 Series 2: Monotonic Test on Partially Embedded Beam During the tests, beam yielding occurred at the panel-beam interface (Figure 65) at an average 33.4 kN-m bending moment. Both the tests showed well defined yield points. The system kept on taking load without dropping and no peak load was identifiable. Both the tests were terminated when the moment reached 38.5 kN-m at the beam-wall interface. Compared to series 1, the yield and maximum load obtained at series 2 are slightly lower. The longer distance between the bearing plates and consequently the greater lever arm for resisting moment contributed to the system yielding at a lower moment.  Figure 65: Yielding of beam during experimental series 1 77 Comparing the experimental curve to the load-displacement curve obtained from numerical optimization (Figure 45), these monotonic test results are in close agreement. The curve shapes are similar. But the experimental yield point was slightly higher than numerical value. This discrepancy indicates that the actual system might be stiffer than the numerical model. Also the yield displacement at the location of LVDT-4 is very similar to numerically obtained value (26 and 29 mm respectively). During test 1, the maximum deformation value of at the location of LVDT- 4 was around 180 mm which occurred at 38.5 kN-m moment. From test 2, the deformation at the same location was 140 mm. At the location of LVDT-3, during first monotonic test, the deformation at yield moment was around 12 mm, whereas the maximum deformation at highest moment was 75 mm. The large difference between yield and maximum displacement indicates that, the system has greater ductility than that of series 1. The load-displacement plots of the monotonic test 1 and 2 of series 1 are shown in Figures 66 and 67 respectively.  Figure 66: Load-displacement curve: Series-2, monotonic test-1 051015202530354045‐5 15 35 55 75 95 115 135 155 175 195Moment (kN‐m)Deformation (mm)LVDT‐1LVDT‐2LVDT‐3LVDT‐478  Figure 67: Load-displacement curve: Series-2, monotonic test-2 The in-plane deformation of the embedded portion of the beam was measured at two locations. The deformation values obtained from both the LVDTs inside the CLT panel were very small (around 2 mm). The negative displacement value of LVDT-1 and positive displacement values of LVDT-2 indicate that the beam was rotated about a point between these two. No out of plane buckling was observed during monotonic testing of series 2 configuration. There was in-plane deformation inside the panel causing damage to wood before the beam reached yield load. Even though the damage was negligible, it still indicated that the beam is stronger than the CLT panel. 4.5.4 Series 2: Cyclic Test on Fully Embedded Beam Similar to monotonic test of series 1, no peak load was identifiable from the monotonic tests for series 2. However, considering greater ductile behavior of this system as demonstrated by monotonic tests, a higher deformation (corresponding to 100 % yield load) was considered as target displacement. The value of this displacement was 50 mm. The setup was similar to monotonic tests. However, the failure mode was yielding of the beam followed by out of plane 051015202530354045‐5 15 35 55 75 95 115 135 155Moment (kN‐m)Deformation (mm)LVDT‐1LVDT‐2LVDT‐3LVDT‐479 buckling. No rolling shear failure was observed during cyclic test of series 2 which is a big improvement towards search for ideal connection configuration. The beam rotated about the point where bolt was inserted. The out of plane buckling failure mode of the beam during cyclic test are shown in Figure 68. Cracking and crushing began inside the panel when the load was beyond the yield load of the system. It was planned to continue the load up to 200% of target displacement. However at 160% of the target displacement, out of plane buckling occurred in the beam and it lifted up from its position. At this point, the application of load was discontinued.   Figure 68: Out of plane buckling of the steel beam during cyclic test of series 2 Minor damage in the form of splitting of CLT at the location of bearing plate was also observed as shown in Figure 69. The observed maximum bending moment at the beam-wall interface was 38.5 kN-m. The hysteresis behavior of the top face flange at the location of LVDTs is presented in Figures 70 through to 73. Based on the monotonic and cyclic tests, it can be deduced that the 80 point of rotation of the beam is between LVDT- 1 and LVDT-2. The maximum deformation at LVDT-1 which is located 675 mm inside from the beam-wall interface was found to be around 10 mm with small energy dissipation. The maximum deformation at LVDT-2 which is located 125 mm inside from the beam-wall interface was also found to be small (8 mm), with little energy dissipation. Upon completion of cyclic test a plastic deformation of 5 mm was observed at the CLT layer in contact with the bearing plate. Overall, series 2 showed ductile behavior with steel beam yielding before any significant damage to CLT. So, in terms of performance, the partially embedded system is better compared to the fully embedded system. However, the beam used during the experiment was still the smallest one commercially available. Therefore, further testing is required with larger beam size before it can be concluded that partially embedded system is an ideal configuration for the FFTT system.  Figure 69: Damage in the CLT panel during cyclic test of series 2 81  Figure 70: Cyclic test: Series-2, LVDT-1  Figure 71: Cyclic test: Series-2, LVDT-2 ‐50‐40‐30‐20‐1001020304050‐12 ‐10 ‐8 ‐6 ‐4 ‐2 0 2 4 6Moment (kN‐m)Deformation (mm)LVDT‐1‐50‐40‐30‐20‐1001020304050‐10 ‐8 ‐6 ‐4 ‐2 0 2 4Moment (kN‐m)Deformation (mm)LVDT‐282  Figure 72: Cyclic test: Series-2, LVDT-3  Figure 73: Cyclic test: Series-2, LVDT-4 ‐50‐40‐30‐20‐1001020304050‐40 ‐30 ‐20 ‐10 0 10 20 30 40Moment (kN‐m)Deformation (mm)LVDT‐3‐50‐40‐30‐20‐1001020304050‐80 ‐60 ‐40 ‐20 0 20 40 60Moment (kN‐m)Deformation (mm)LVDT‐483 4.6 Discussion on Experimental Investigations To substantiate the findings from the numerical parameter study, two experimental test series were conducted. Both series included two monotonic and one reversed cyclic test. The experimental results are discussed in the subsequent paragraphs. 4.6.1 Comparison between Experimental and Numerical Results A comparative summary of experimental and numerical results is shown in Table 13. It can be observed that for both the series, the numerical yield moments are around 12% lower than those obtained from experiments. However, the peak moments are almost identical. Therefore, the numerical model seems appropriate although a little less stiff than the actual connection. It is also noticeable from both experimental and numerical studies that a partially embedded beam with greater embedment length (series 2) yielded at a lower load than the system with full embedment of beam with reduced embedment length (series 1). The longer embedment length of beam inside CLT resulted in a longer lever arm for resisting the external force. Also, partial embedment caused the beam to lift up from its position due to out of plane buckling, therefore might be a limiting factor.  Table 13: Comparison between Experimental results and their numerical simulation Series Yield Moment (kN-m) Peak Moment (kN-m) Maximum Deformation at LVDTs (mm) Exp. Num. Exp. Num. 1 2 3 4 Exp. Num. Exp. Num. Exp. Num. Exp. Num. 1 36.4 30.1 42.8 42.4 -3.2 -2.2 1.4 1.8 34.2 30.2 60.4 68.0 2 33.4 29.5 38.5 38.4 -1.4 -1.2 0.8 0.9 74.4 55.0 175.0 130.0  84 The 1st and 2nd LVDTs were placed inside the CLT and the other two at the cantilever portion of the beam. The experimental deformation values inside the CLT for series 1 are greater than those obtained from series 2. This can be explained by the greater lever arm for the resisting moment in case of series 2 causing smaller compressive forces inside the CLT and consequently smaller deformations. This fact is supported by the numerical simulation which showed lower deformation inside CLT for series 2. The other two LVDTs represent the deformation of the steel beam only. For these two locations, series 2 showed lower deformations than series 1. This can be attributed to the fact that the fully embedded beam resulted in a stiffer system with no out of plane buckling, while partial embedment of beam in series 2 caused the beam to buckle out of plane resulting in significantly greater deformation of beam. The numerical analyses showed similar behavior but the values were lower than the corresponding experimental results. This is because the numerical model considered steel as a bilinear material without degradation while the actual steel exhibits degrading behavior. Overall though, the experimental and numerical results for the monotonic tests are in reasonable agreement. 4.6.2 Point of Rotation of Beam The point of yielding was at the beam-wall interface for both the test series. The point of beam rotation inside the CLT however varied between series. The point of rotation of beam was established based on the load-deformation behavior of the LVDTs attached inside CLT (LVDT-1 and LVDT-2). The point of rotation of beam is the location at which the deformation and consequently energy dissipation is zero. By observing the signs of displacement values inside CLT from Table 13, it can be concluded that the points of rotation of beam for both the systems lie between LVDT-1 and LVDT-2. For both series, deformation values from LVDT-1 are negative and the values from LVDT-2 are positive. A closer inspection of the values also 85 revealed that the point of rotation is closer to LVDT-2 for both systems. In Figure 74, the displacement values of LVDT-1 and 2 are plotted against the location of LVDTs. The value 1 and 2 in the horizontal axis depicts the position of LVDT-1 and 2 respectively. By assuming linear variation of displacement, it can be showed that, for series 1, axis of the beam rotation is at 1.7 times the distance from LVDT-1 towards LVDT-2. For series 2 this location is at 1.6 times the distance between two LVDTs from LVDT-1 towards LVDT-2.  Such assumption is reasonable considering very small displacement values at these locations. The points of rotation are illustrated in Figure 74. These points are of interest for the calculation of the stress transfer between steel and CLT through the bearing of the embedded beams on the wall panels. Therefore, these points are important for obtaining the location of bearing plates to achieve optimal performance of the connection.   Figure 74: Points of rotation of beams for series 1 and 2 4.6.3 Ductility and Force Modification Factor Ductility (µ) can be calculated from the load-displacement curves from the monotonic tests. It is an important parameter for seismic design. The ratio of ultimate and yield displacement can be ‐4‐3‐2‐10121 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2Deformation (mm)LVDT 1 and 2 locationSeries1Series286 considered as ductility for the connection (Munoz et al., 2009). The ductility related force modification factor (Rd) specified in NBCC 2010 (NRC, 2010) can be obtained from the connection ductility (Boudreault et al., 2007): ࡾࢊ ൌ ඥ૛ࣆ െ ૚…………………………………………………………………………… (4) Table 14 shows the values obtained from the experiments and applying the above equation. Table 14: Ductility ratio and force modification factor Series Ductility ratio (µ) Force modification factor (Rd) 1-1 3.00 2.24 1-2 3.16 2.31 2-1 5.86 3.27 2-2 5.93 3.30  The series 1 configuration exhibited an average ductility factor of 2.3; while for series 2, this value is 3.3. Even though both system exhibit ductility, it is noticeable that series 2 is almost 1.4 times as ductile as series 1. So, in terms of desirable ductile failure mode for FFTT system; the partially embedded connection with full embedment length is better than the fully embedded connection with reduced embedment. The greater embedment length and larger lever arm for series 2 caused less force transfer to the CLT and therefore, beam yielding occurred before any significant damage to CLT. And in that series, after yielding, the system continued to pick up load as steel beam undergo large post yield deformation.  The ductility factors obtained are not exact; rather, these are minimum ductility values for the connection because the monotonic tests were discontinued upon reaching the peak load without capturing the full degradation. Therefore, the deformation values used to calculate the connection 87 ductility are not the ultimate displacements. The ultimate displacements could be significantly higher than the displacements at peak load. So, the actual ductility values for the connections are expected to be higher than the values as listed in Table 14.  4.6.4 Hysteretic Behavior The results of the cyclic tests are summarized in Table 15. It is noticeable that during the reverse cyclic tests, series 1 was subjected to a greater peak moment (43.8 kN-m) than series 2 (38.5 kN-m). Series 1 test was stopped after 45 load cycles while series 2 was discontinued after 53 cycles due to beam uplifting from its longitudinal axis.  Table 15: Cyclic tests results Series  Cyclic Tests Peak Load (kN) Corresponding Peak Moment (kN-m) Number of Cycles Energy Dissipated (Joules) LVDT-1 LVDT-2 LVDT-3 LVDT-4 1 62.3 43.6 45 75 187 N/A 2204 2 56.1 38.9 53 120 170 3230 6120  Series 1: The hysteretic curve obtained from LVDT-1 of series 1 test showed an initial flat portion followed by sharp increase is load (Figure 62). This is due to a slight gap between beam and CLT that was initially there due to fabrication error resulting in deformation without increase in load. The slight arbitrary portion in otherwise a standard hysteretic curve obtained from LVDT-2 can be attributed to erroneous reading (Figure 63). LVDT-3 readings were totally erroneous and hence not considered. LVDT-4 produced a very well defined hysteretic curve (Figure 64). These curves with their backbone are reproduced in Figures 75, 76 and 77.  88  Figure 75: Cyclic test: Series-1, LVDT-1 (with backbone curve)  Figure 76: Cyclic test: Series-1, LVDT-2 (with backbone curve) ‐60‐50‐40‐30‐20‐1001020304050‐4 ‐3 ‐2 ‐1 0 1Moment (kN‐m)Deformation (mm)LVDT‐1‐60‐50‐40‐30‐20‐1001020304050‐5 ‐4 ‐3 ‐2 ‐1 0 1 2Moment (kN‐m)Deformation (mm)LVDT‐289  Figure 77: Cyclic test: Series-1, LVDT-4 (with backbone curve) Series 2: The hysteretic curves obtained from LVDT-1 and LVDT-2 of series 2 which are inside CLT showed well behaved hysteretic curves (Figures 70 and 71). However, there occurred some sudden spikes in horizontal direction in these two curves. These happened due to lifting up of beam from its longitudinal axis caused by out of plane buckling. LVDT- 3 and LVDT-4 produced well behaved hysteretic curves. These curves with their backbone are reproduced in Figures 78, 79, 80 and 81. ‐60‐50‐40‐30‐20‐1001020304050‐60 ‐50 ‐40 ‐30 ‐20 ‐10 0 10 20 30 40Moment (kN‐m)Deformation (mm)LVDT‐490  Figure 78: Cyclic test: Series-2, LVDT-1 (with backbone curve)  Figure 79: Cyclic test: Series-2, LVDT-2 (with backbone curve) ‐50‐40‐30‐20‐1001020304050‐12 ‐10 ‐8 ‐6 ‐4 ‐2 0 2 4 6Moment (kN‐m)Deformation (mm)LVDT‐1‐50‐40‐30‐20‐1001020304050‐10 ‐8 ‐6 ‐4 ‐2 0 2 4Moment (kN‐m)Deformation (mm)LVDT‐291  Figure 80: Cyclic test: Series-2, LVDT-3 (with backbone curve)  Figure 81: Cyclic test: Series-2, LVDT-4 (with backbone curve) The backbone curves obtained can be used to develop nonlinear hinge properties of the connection for dynamic analysis of structure at global level. These can also be used to study seismic performance criteria and checking suitability of such system. ‐50‐40‐30‐20‐1001020304050‐40 ‐30 ‐20 ‐10 0 10 20 30 40Moment (kN‐m)Deformation (mm)LVDT‐3‐50‐40‐30‐20‐1001020304050‐80 ‐60 ‐40 ‐20 0 20 40 60Moment (kN‐m)Deformation (mm)LVDT‐492 4.6.5 Energy Dissipation For both series, energy dissipation occurred mainly through yielding of the steel beams. Very little energy dissipation occurred inside the CLT. This is due to the fact that very little deformation occurred inside the CLT during cyclic test while the beam underwent large post yield deformation. The energy dissipated through different locations are reported in Table 15. During testing of series 1, LVDT-1 and LVDT-2 dissipated 74.7 and 187.3 Joules of energy while through beam yielding, 2204 Joules of energy were dissipated. The readings from LVDT-3 were erroneous during testing of series 1 and therefore not considered. Series 2 dissipated significantly higher energy than series 1 at all LVDT locations. Maximum energy dissipated during testing of this series was 6120 Joules which is almost 2.8 times the value obtained from series 1. This is due to the fact that, large deformation was observed with ductile behavior in case of series 2. So, series 2 performed better than series 1 and should be considered for further studies. The amount of energy dissipation for both series is shown in Figure 82.  Figure 82: Energy dissipation during reverse cyclic tests  01000200030004000500060001 2 3 4Energy Dissipated (Joules)LVDTsSeries1Series293 Chapter  5: CONCLUSIONS  5.1 Summary This research focused on the component level performance of the steel beam to CLT panel connection of the proposed hybrid FFTT system under quasi-static monotonic and reversed cyclic loads. The combined numerical and experimental work yielded following main results: 1) The numerical investigation included the modelling of five previously tested configurations and simulated the load-displacement behavior obtained by Bhat (2013). The numerical and the experimental curves were in good agreement with the numerical curves being slightly stiffer. This is due to the fact that in numerical modelling, fabrication imperfections were not considered while these existed in the tested specimens. Nevertheless, the numerical model was deemed appropriate to model the global deformation behaviour. On the material level, however, it was shown that the linear-elastic timber model was insufficient to model the local plastic deformation incurred in the timber, and, as a result, the obtained stress values, were unrealistic. 2) A numerical parameter study was conducted to recommend an improved connection geometry which included steel bearing and side plates. Parameters of interest were embedment length and depth, beam size and spacing of bearing plates. It was found that even the smallest wide flange beam could cause excessive stresses and crushing in the CLT panel before yielding the beam.  3) The stress values from the numerical study indicate that the CLT panels were stressed beyond yield and therefore, the linear elastic model was no longer sufficient. A nonlinear CLT material model would simulate the post yield behavior more realistically. Such models can be developed as mentioned in Chapter 2 (Grosse and Rautenstrauch, 2004). But post yield stress-strain data is required for CLT which is currently not available. 94 4) Experiments (two quasi-static monotonic and one reversed cyclic test per series) were conducted on two improved connection layouts. The first series consisted of fully embedded (102 mm deep) wide flange beam with 350 mm spacing between bearing plates. The second series consisted of partially embedded (85 mm deep) wide flange beam with 665 mm spacing between bearing plates. The monotonic tests resulted in little damage and cracking to the panel before steel yielding. The cyclic test on series 1 led to rolling shear failure in the weak CLT layer. This failure mode was avoided in series 2 by partially embedding the beam and increasing the spacing between bearing plates. 5) The experimental studies showed that both the system exhibit reasonable ductility with series 2 exhibiting higher ductility. The ductility factors were 2.28 and 3.29. These ductility factors are minimum values established based on deformations at peak load rather than ultimate deformations. The actual ductility for both systems are expected to be significantly higher.  6) The cyclic tests revealed that for both systems energy dissipation occurred mainly through yielding of beam with very little dissipation happening inside CLT. This is expected and desirable. The much higher energy dissipation of series 2 makes that series 2 better suited. 7) The load-deformation curves obtained from the study can be used to develop backbone curves of the connections to define plastic hinge properties for nonlinear modeling of the FFTT system.  8) Overall, the study concludes that by using a partially embedded connection configuration with large distance between bearing plates, the connection performance can be improved. However, choosing the smallest steel section is not practical from a constructional point of view, where significantly bigger sections would be required. If using bigger sections do not result in the desired performance; then CLT may not be the ideal material for the FFTT system and LVL may become a better option. 95 5.2 Recommendation for Further Studies Future studies that can advance the knowledge on timber-steel hybrid systems include:  Developing nonlinear stress-strain curve for CLT panel for numerical modelling as elastic strength properties do not account for post yield inelastic behavior.  A finite element numerical model with nonlinear timber properties which might simulate better behaviour of the current system.  Conducting experiments with larger wide flange beams with different trial configurations by varying embedment length and depth, CLT layer thickness, bearing area and distance etc.  Numerical and experimental investigation can be conducted with other mass timber products which are stronger than CLT like LVL.  A wall testing program that includes static pushover as well as time history analysis to observe system level behavior of the FFTT system.  96 References ANSYS 14.5. (2012) Copyright, SAS IP, Inc. Arbeitsgemeinschaft Holz. (2001). Konstruktive Holzwerkstoffe – Informationsdienst Holz holzbau handbuch, Reihe 4: Baustoffe, Teil 4: Holzwerkstoffe, Folge 1: Konstruktive Holzwerkstoffe, ISSN-Nr. 0466-2114  Ashtari, S. (2009). In-Plane Stiffness of Cross-Laminated Timber Floors, MASc Thesis, University of British Columbia, Vancouver, Canada. BCBC. (2009). Building Code of British Columbia. National Research Council of Canada.  Calder, K., & Senez, P. (2008). A Historical Perspective on Building Heights and Areas in the British Columbia Building Code, Senez Reed Calder Fire Engineering Inc, Richmond, BC, Canada.  Bhat, P. (2013). Experimental Investigation of Connection for The FFTT, A Timber-Steel Hybrid System, MASc Thesis, University of British Columbia, Vancouver, Canada. Boudreault, F.A., Blais, C., Rogers, C.A. (2007). Seismic Force Modification Factors for Light-gauge Steel-frame-wood Structural Shear Walls. Canadian Journal of Civil Engineering, Vol. 34, 56-65. Ceccotti, A., & Karacabeyli, E. (2002). Validation of Seismic Design Parameters for Wood Frame Shearwall Systems. Canadian Journal of Civil Engineering, Vol. 29, 484-498.  Ceccotti, A., Lauriola, M., Pinna, M., & Sandhaas, C. (2006). SOFIE Project -Cyclic Tests on Cross-Laminated Wooden Panels. In Proceedings of the 9th World Conference on Timber Engineering. Portland, USA. Ceccotti, A., Sandhaas, C., & Yasumuro, M. (2010). Seismic Behaviour of Multistory Cross-Laminated Timber Buildings. International Convention of Society of Wood Science and Technology, Geneva, Switzerland 97 CSA S-16. (2009). Design of Steel Structures. Ottawa: Canadian Standards Association.  CSA-O86. (2010). Engineering Design in Wood. Ottawa: Canadian Standards Association. Dickof, C., (2013). CLT Infill Panels In Steel Moment Resisting Frames As A Hybrid Seismic Force Resisting System, MASc Thesis, University of British Columbia, Vancouver, Canada. Dickof, C., Stiemer, S., & Tesfamariam, S. (2012). Wood-Steel Hybrid Seismic Force Resisting Systems: System Ductility. World Conference on Timber Engineering. Auckland.  Federal Emergency Management Agency. (2000). FEMA 356: Prestandard and Commentary for the Seismic Rehabilitation of Buildings. Washington, D.C.  Filiatrault, A., & Folz, B. (2002). Performance Based Seismic Design of Wood Framed Buildings. Journal of Structural Engineering, ASCE, Vol 128(1), 39-47. Filiatrault, A., Christovasilis, I., Wanitkorkul, A., & van de Lindt, J. (2010). Experimental Seismic Response of a Full-Scale Light-Frame Wood Building. Journal of Structural Engineering, ASCE, Vol 136 (3), 246-254.  Gagnon, S., Pirvu, C. (2011). Cross Laminated Timber (CLT) Handbook. FPInnovations, Vancouver, Canada. Green, M., & Karsh, E. (2012). TALL WOOD: The Case for Tall Wood Buildings. Vancouver: Wood Enterprise Coalition- (Used under a Creative Commons (CC) License-Attribution Non-Commercial ShareAlike). < http://wecbc.smallboxcms.com/database/rte/files/TallWood.pdf> Grosse, M., Rautenstrauch, K. (2004). Numerical Modeling of Timber and Connection Elements Used in Timber-Concrete Composite Constructions, University of Weimer, Germany. Gsell, D., Feltrin, G., Schubert, S., Steiger, R. & Motavalli, M. (2007). Cross laminated timber plates: Evaluation and verification of homogenized elastic properties. Journal of Structural Engineering, ASCE, Vol 133(1), 132-138. Johansen, K.W. (1949). Theory of Timber Connections. Publications of International Association of Bridge and Structural Engineering, Zurich, Switzerland, Vol 9, 249-262. 98 Keenan. (1986). Limit States Design of Wood Structures. Toronto, Ontario: Canada Morrison Hershfield Ltd.  Khorasani. (2010). Feasibility Study of Hybrid Wood Steel Structures, MASc Thesis, University of British Columbia, Vancouver, Canada. Kodor, V., Erki, N., & Quenneville, J. (1995). Seismic Design and Analysis of Masonry Infilled Frames. Canadian Journal of Civil Engineering, Vol 22, 576-587. Krawinkler, H., Parisi, F., Ibarra, L., Ayoub, A., & Medina, R. (2001). Development of a Testing Protocol for Woodframe Structures. Richmond, CA: CUREE Publication No. W-02.  Multiplas, (2013). Elastoplastic Material Models for ANSYS. Dynamic Software and Engineering GmbH, Germany. Mitchell, D., Tremblay, R., Karacabeyli, E., Paultre, P., Saatcioglu, M., & Anderson, D. (2003). Seismic force modification factors for the proposed 2005 edition of the National Building Code of Canada. Canadian Journal of Civil Engineering Vol. 30, 308-327.  Mohammad, M., Gagnon, S., Karacabeyli, E., & Popovski, M. (2011). Innovative Mid-Rise Timber Structures Offer New Opportunities for Designers. Structural Engineers Association of California (SEAOC). Munoz, W., Mohammad, M., Salenikovich, A., Quenneville, P. (2008). Determination of Yield Point and Ductility of Timber Assemblies: In Search for a Harmonized Approach. Journal of Engineered Wood Products Association (EWPA), USA. NBCC. (2010). National Building Code of Canada. Ottawa: National Research Council of Canada.  Palermo, A., Pampanin, S., & Buchanan, A. (2006). Experimental Investigations on LVL Seismic Resistant Wall and Frame Subassemblies. First European Conference on Earthquake Engineering and Seismology, Geneva, Switzerland. 99 Palermo, A., Pampanin, S., Buchanan, A., & Newcombe, M. (2005). Seismic Design of Multi-Storey Buildings using Laminated Veneer Lumber (LVL). New Zealand Society Earthquake Engineering (NZSEE). Wairakei, New Zealand. Pang, W., & Rosowsky, D. (2007). Direct Displacement Procedure for Performance- Based Seismic Design of Multistory Wood Frame Structures. NEESWOOD Report NW-02, National Science Foundation.  Paulay, T. (1995). The Philosophy and Application of Capacity Design. Scientia Iranica, Vol 2, 117-136.  Popovski, M., & Karacabeyli, E. (2008). Force Modification Factors and Capacity Design Procedures for Braced Timber Frames. In proceedings of the 14th World Conference on Earthquake Engineering. Beijing, China. Popovski, M., Schneider, J., & Schweinsteiger, M. (2010). Lateral Load Resistance of Cross-Laminated Wood Panels. World Conference on Timber Engineering (WCTE). Schneider, J. (2009). Connections in Cross-Laminated-Timber Shear Walls Considering the Behaviour under Monotonic and Cyclic Lateral Loading, MASc Thesis, University of Stuttgart, Germany Smith, T., Ludwig, F., Pampanin, S., Palermo, A., Fragiacomo, M., Buchanan, A., & Deam, B. (2007). Seismic Response of Hybrid-LVL Coupled Walls under Quasi-Static and Pseudo-Dynamic Testing. New Zealand Society of Earthquake Engineering (NZSEE).  Smith, T., Pampanin, S., Buchanan, A., & Fragiacomo, M. (2008). Feasibility and Detailing of Post-tensioned Timber Buildings for Seismic Areas. New Zealand Society for Earthquake Engineering. Structurlam. (2012). Cross Laminated Timber Design Guide-Version 7. Vancouver, BC, Canada, <http://structurlam.com/products/cross-laminated-timber/> 100 van de Lindt, J., & Pei, S. (2011). Seismic Numerical Modeling of a Six-Story Light-Frame Wood Building: Comparison with Experiments. Journal of Earthquake Engineering, Vol 15, 924-941.  Yawalata, D., & Lam, F. (2011). Development of Technology for Cross Laminated Timber Building Systems. Vancouver, BC: University of British Columbia. Yousuf, M., & Bagchi, A. (2009). Seismic Design and Performance Evaluation of Steel-Frame Buildings Designed Using the 2005 National Building Code of Canada. Canadian Journal of Civil Engineering, Vol 36, 280-294. 

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.24.1-0165954/manifest

Comment

Related Items