UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Innovative energy dissipating system for earthquake design and retrofit of timber structures Yung, Willy Chi Wai 1991

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

Item Metadata

Download

Media
831-UBC_1991_A7 Y86.pdf [ 11.09MB ]
Metadata
JSON: 831-1.0062628.json
JSON-LD: 831-1.0062628-ld.json
RDF/XML (Pretty): 831-1.0062628-rdf.xml
RDF/JSON: 831-1.0062628-rdf.json
Turtle: 831-1.0062628-turtle.txt
N-Triples: 831-1.0062628-rdf-ntriples.txt
Original Record: 831-1.0062628-source.json
Full Text
831-1.0062628-fulltext.txt
Citation
831-1.0062628.ris

Full Text

INNOVATIVE ENERGY DISSIPATING SYSTEM FOR EARTHQUAKE DESIGN AND RETROFIT OF TIMBER STRUCTURES by Willy Chi Wai Yung B.A.Sc, University of British Columbia, 1988 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE in THE FACULTY OF GRADUATE STUDIES Department of C i v i l Engineering We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA January, 1991 ° Willy Chi Wai Yung, 1991 I n p r e s e n t i n g t h i s t h e s i s i n p a r t i a l f u l f i l m e n t of the requirements f o r an advanced degree a t the U n i v e r s i t y of B r i t i s h Columbia, I agree t h a t the L i b r a r y s h a l l make i t f r e e l y a v a i l a b l e f o r r e f e r e n c e and study. I f u r t h e r agree t h a t p e r m i s sion f o r e x t e n s i v e copying of t h i s t h e s i s f o r s c h o l a r l y purposes may be granted by the head of my department or by h i s or her r e p r e s e n t a t i v e s . I t i s understood t h a t copying or p u b l i c a t i o n of t h i s t h e s i s f o r f i n a n c i a l g a i n s h a l l not be allowed without my w r i t t e n permission. Department of C i v i l E n gineering The U n i v e r s i t y of B r i t i s h Columbia 2324 Main M a l l Vancouver, Canada V6T 1Y3 Date: 3/3- 3rY0 <M ABSTRACT This thesis presents the results obtained from a preliminary investigation into the potential application of the f r i c t i o n damping concept to wood structures to improve their seismic response. Sliding f r i c t i o n devices which contain heavy duty brake lining pads have been proposed in order to enhance a wood structure's seismic performance. The devices are mounted onto a structure's shearwalls to dissipate seismic energy input during the wall's deformation in an earthquake. Prototypes of the four f r i c t i o n damping devices were tested to examine their hysteretic behaviour. Conventional f u l l scale, 2.44 x 2.44 m (8 x 8 ft) timber shearwalls, typical of ones used in residential and light-commercial building applications, and ones retrofitted with the f r i c t i o n damping devices were tested on a shake table. Three set of tests were conducted. They involved loading the walls under unidirectional racking, s t a t i c -c y c l i c and simulated earthquake loads. Test results from the two types of shearwalls were compared against each other and against the findings from the computer programs SADT and FRICWALL. SADT i s a f i n i t e elements program which computes the load-deformation behaviour of shearwalls. FRICWALL i s an inelastic time-history dynamic model which computes the response time-history of a shearwall under a simulated seismic event. The cycl i c tests of the f r i c t i o n damping devices showed that they exhibited very stable and non-deteriorating hysteretic behaviour. The shake table tests of the f u l l scale timber - i i -s h e a r w a l l s showed t h a t the f r i c t i o n damped w a l l s were s t i f f e r , can s u s t a i n an average of 23.7 % h i g h e r r a c k i n g l o a d and d i s s i p a t e an average of 42.9 % more energy than the con v e n t i o n a l ones before a d u c t i l e f a i l u r e . F a i l u r e i n the con v e n t i o n a l w a l l s was b r i t t l e . These r e s u l t s were i n agreement w i t h the SADT f i n d i n g s . Under slow c y c l i c l o a ds, they d i s s i p a t e d more energy, but because t h e i r o v e r a l l h y s t e r e t i c behaviour was s t i l l pinched, they were j u s t as i n e f f i c i e n t as the co n v e n t i o n a l w a l l s a t d i s s i p a t i n g energy. On the average, t h e i r s e i s m i c performance was o n l y m a r g i n a l l y b e t t e r than t h a t of the con v e n t i o n a l w a l l , w i t h an average drop of 9.6 % i n peak w a l l d e f l e c t i o n . T h i s i s f a r s h o r t of the average o f 29.5 % computed by FRICWALL. D e t a i l e d a n a l y s i s of the r e s u l t s show t h a t due t o bending i n the framing members of the s h e a r w a l l , the load necessary t o cause s l i p p a g e of the f r i c t i o n devices was not achieved u n t i l w a l l d e f l e c t i o n s i n the order of 25.4 mm (1.0 in) was reached. Since o n l y a t the peak or near-peak e x c i t a t i o n l e v e l s of an earthquake d i d shearwall d e f l e c t i o n s surpass t h i s magnitude, the devices were not able t o c o n t r i b u t e t o the energy d i s s i p a t i o n of the sh e a r w a l l s d u r i n g the m a j o r i t y p o r t i o n of a s e i s m i c event. - i i i -TABLE OF CONTENTS ABSTRACT /• i i TABLE OF CONTENT iv LIST OF FIGURES . . .... . . . . . v i i LIST OF TABLES x i i ACKNOWLEDGEMENT x i i i 1. INTRODUCTION 1.1 Background 1 1.2 Objective and Scope 4 2. BEHAVIOUR OF CONVENTIONAL AND FRICTION DAMPED  TIMBER SHEAR WALLS 2.1 Introduction 6 2.2 Conventional Timber Shearwalls 6 2.3 Friction Damped Timber Shearwalls 9 3. TEST SPECIMENS 3.1 Introduction 13 3.2 Friction Damping Devices 13 3.3 Full Size Timber Shearwalls 16 3.4 Friction Device to Shearwall Connection Design 20 4. STATIC FINITE ELEMENTS MODEL 4.1 Introduction 23 - iv -4.2 Racking Behaviour of Conventional Timber Shearwall . . . . 25 4.2.1 Finite Element.Mesh Generation . . . . . . 25 4.2.2 Shearwall Component Modelling 29 4.2.2.1 Framing Members 29 4.2.2.2 Sheathing Panels 30 4.2.2.3 Connections Between Framing Members 31 4.2.2.4 Sheathing-To-Framing Connections . 3 3 4.2.3 Numerical Results 35 4.3 Racking Behaviour of Friction Damped Timber Shearwall 38 4.3.1 Finite Element Mesh Generation 4 0 4.3.2 Shearwall Component Modelling 43 4.3.3 Numerical Results 4 6 4.4 Summary 51 5. CYCLIC TEST OF FRICTION DAMPING DEVICES 5.1 Introduction 52 5.2 Test Equipment and Instrumentation 52 5.3 Test Procedure 55 5.4 Results and Discussion 59 5.5 Summary 69 6. INELASTIC TIME-HISTORY DYNAMIC ANALYSIS 6.1 Introduction 70 - v -6.2 Numerical Analysis 77 6.3 Results and Discussion 84 6.3.1 Optimal S l i p Load Study . . . . . . . . . 84 6.3.2 Detailed Analysis 90 6.4 Summary 100 7. FULL SIZE SHEARWALL TESTS 7.1 Introduction 101 7.2 Test Equipment and Instrumentation 101 7.3 Specimen Setup 106 7.4 Test Procedures 108 7.4.1 Racking Tests 108 7.4.2 S t a t i c C y c l i c Test I l l 7.4.3 Dynamic Earthquake Simulation 112 7.5 Results and Discussion 114 7.5.1 Racking Tests 114 7.5.2 S t a t i c C y c l i c Tests 126 7.5.3 Dynamic Earthquake Simulation 140 7.6 Summary 164 8. CONCLUSION 8.1 Summary and Conclusions 166 8.2 Future Research . . 168 BIBLIOGRAPHY 169 - v i -LIST OF FIGURES 2.1 Deformation Pattern of a Conventional JTimber Shearwall 7 2.2 Hysteretic Behaviour of Conventional Timber Shearwall 9 2.3 Timber Shearwall Incorporating F r i c t i o n Damping Devices 10 2.4 Forces Acting on F r i c t i o n Devices Due t o Shearwall Deformation 12 3.1 F r i c t i o n Damping Device 15 3.2 Layout of Shearwall Construction ."Y 17 3.3 Double End Stud Reinforcement 19 3.4 Bracket and Pin Connectors f o r F r i c t i o n Damping Device 20 3.5 Lag Screw Layout Pattern for Hinged Legs of F r i c t i o n Damping Device 22 4.1 Framing F i n i t e Element Mesh for Conventional Timber Shearwall 27 4.2 Sheathing Panel F i n i t e Element Mesh f o r Conventional Timber Shearwall 28 4.3 Curve Modelling the Non-Linear Behaviour of the Connectors Between the Framing Members 31 4.4 Load Deformation Behaviour for Sheathing-to-Framing Connection 34 4.5 Racking Behaviour of a Conventional Timber Shearwall as Predicted by SADT . . . 37 4.6 Idealized Load-Deformation Curve of a F r i c t i o n Damping Device 38 4.7 Framing F i n i t e Element Mesh for F r i c t i o n Damped Timber Shearwall 41 4.8 Sheathing Panel F i n i t e Element Mesh f o r F r i c t i o n Damped Timber Shearwall 42 - v i i -4.9 Non-Linear Connector ,Between Friction Damping Device and Wall Framing 45 4.10 Racking Behaviour of a Friction Damped and a Conventional Timber Shearwall as Predicted by SADT 48 4.11 Maximum Strut Force Developed in the Non-Slipping Friction Devices During Each Lateral Load Increment 50 5.1 General Setup of MTS Testing Equipment for Cyclic Test of the Friction Joints 54 5.2 Test Specimen as Mounted in the Loading Frame of the MTS Testing System 56 5.3 Time-History Plots Illustrating the Pattern of the Load Application on the Test Specimens 57 5.4 Load-Deformation Behaviour of a Friction Joint Under a 50 Cycle Sinusoidal Load 59 5.5 Hysteresis Loops for Friction Joint Tests Conducted at the Frequencies of 0.5, 2.0, 3.5 and 5.0 Hz 62 5.6 Hysteresis Loops for Friction Joint Tests Conducted at the Frequency of 1.0 Hz 63 5.7 Slip Load Calibration Curves for Friction Damping Devices 66 5.8 Rotation of Friction Joint from Eccentric Strut Loads 68 6.1 Numerical Modelling of a Friction Damped Timber Shearwall 71 6.2 Hysteretic Behaviour of Shearwall as Modelled by FRICWALL 73 6.3 Hysteretic Behaviour of Friction Damping Device as Modelled by FRICWALL 74 6.4 Absolute Acceleration Response Spectra for Seismic Events Used in Simulations 78 - v i i i -6.5 Ground Acceleration Time-Histories 79 6.6 Free Body Diagram of Shearwall Testing System 81 6.7 FRICWALL Results for Peak Top-of-Wall Displacement - El Centro Earthquake @ 0.35 g P.G.A. - 85 6.8 FRICWALL Results for Peak Top-of-Wall Displacement - El Centro Earthquake @ 0.6 g P.G.A. - 86 6.9 FRICWALL Results for Peak Top-of-Wall Displacement - Romania Earthquake § 0.2 g P.G.A. - 87 6.10 Relative Top-of-Wall Displacement Time-Histories Computed by FRICWALL 91 6.11 Hysteresis Loops from FRICWALL Simulation - E l Centro Earthquake @ 0.35 g P.G.A. - 93 6.12 Hysteresis Loops from FRICWALL Simulation - El Centro Earthquake @ 0.6 g P.G.A. - 94 6.13 Hysteresis Loops from FRICWALL Simulation - Romania Earthquake @ 0.2 g P.G.A. - 95 6.14 Energy Time-Histories from FRICWALL Simulation - El Centro Earthquake @ 0.35 g P.G.A. - 97 6.15 Energy Time-Histories from FRICWALL Simulation - El Centro Earthquake @ 0.6 g P.G.A. - 98 6.16 Energy Time-Histories from FRICWALL Simulation - Romania Earthquake @ 0.2 g P.G.A. - 99 7.1 General Setup of Shake Table for Full Scale Testing 104 7.2 Setup of Data Acquisition & Control System and AST 80286 Micro-Computer for Full Scale Testing .... 105 7.3 Double End Stud Anchor Plate Detail 107 7.4 Ramp Load for Unidirectional Racking Tests 108 7.5 Measuring Instruments for the Friction Damping Device 110 - ix -7.6 Sinusoidal Loading Waves for the Static Cyclic Tests 112 7.7 Racking Load-Deformation Behaviour of Timber Shearwalls • 116 7.8 Load-Deformation Behaviour of Friction Damping Devices 120 7.9 Time-Histories of Axial Strut Loads on the Friction Damping Devices 122 7.10 Hysteresis Loops from Static Cyclic Tests of Timber Shearwalls 127 7.11 Energy Dissipation Efficiency of Timber Shearwalls 132 7.12 Hysteresis Loops of Friction Damping Devices - Static Cyclic Test, =/- 12.7 mm (0.5 in) - 134 7.13 Hysteresis Loops of Friction Damping Devices - Static Cyclic Test, =/- 25.4 mm (1.0 in) - 135 7.14 Hysteresis Loops of Friction Damping Devices - Static Cyclic Test, =/- 50.8 mm (2.0 in) - 136 7.15 Relative Top-of-Wall Displacement Time Histories - El Centro Earthquake § 0.35 g P.G.A. - 141 7.16 Relative Top-of-Wall Displacement Time Histories - El Centro Earthquake.@ 0.6 g P.G.A. - 142 7.17 Relative Top-of-Wall Displacement Time Histories - Romania Earthquake @ 0.2 g P.G.A. - 143 7.18 Hysteresis Loops from Earthquake Simulations - El Centro Earthquake @ 0.35 g P.G.A. - 150 7.19 Hysteresis Loops from Earthquake Simulations - El Centro Earthquake @ 0.6 g P.G.A. - 151 7.20 Hysteresis Loops from Earthquake Simulations - Romania Earthquake @ 0.2 g P.G.A. - 152 7.21 Acceleration Time-Histories from Earthquake Simulations 156 - x -7.22 Energy Time-Histories from Earthquake Simulation - E l Centro Earthquake @ 0.35 g P.G.A. - 157 7.23 Energy Time-Histories from Earthquake Simulation - E l Centro Earthquake @ 0.6 g P.G.A. — 158 7.24 Energy Time-Histories from Earthquake Simulation - Romania Earthquake § 0.2 g P.G.A. - 159 - xi -L I S T OF T A B L E S 4.1 .Material Properties f o r SPF Framing 30 4.2 Material Properties f o r Douglas F i r Plywood Panel 30 4.3 Parameters Defining the Non-Linear Behaviour of the Connectors Between the Framing Members 32 4.4 Material Properties f o r F r i c t i o n Damping Devices 43 5.1 S l i p Loads of the F r i c t i o n Damping Devices as Determined by the C y c l i c Tests 65 6.1; Summary of FRICWALL Inputs 83 6.2 Summary of Marginal Drop i n Peak Relative Top-of-Wall Displacement 89 7.1 Racking Behaviour of Timber Shearwalls 116 7.2 Summary of S l i p Loads from the Racking Tests of the F r i c t i o n Damped Timber Shearwalls 120 7.3 Parameters Defining the V i r g i n Load-Deformation Path of a Conventional Timber Shearwall 125 7.4 Summary of Peak Racking Resistance from S t a t i c C y c l i c Tests 128 7.5 Summary of Hysteretic Behaviour of Timber Shearwalls 130 7.6 Summary of S l i p Loads f o r A l l F r i c t i o n Devices from S t a t i c C y c l i c Tests 138 7.7 Parameters Defining the Pinched Hysteresis ; Loops of a Conventional Timber Shearwall 140 7.8 Summary of Peak Top-of-Wall Relative Deflection from Earthquake Simulation 146 7.9 Summary of Energy Time-Histories from Earthquake Simulations 160 - x i i -ACKNOWLEDGEMENT The author would like to thank the Forintek Canada Corporation and the Natural Sciences and Engineering Research Council of Canada for providing the - research scholarship and operating grant in support of this project. The expert guidance and numerous advices contribute by my supervisor Dr. A. F i l i a t r a u l t throughout this project are gratefully acknowledged. The assistance from Dr. R.O. Foschi i s much appreciated. The inputs from Dr.. J.D. Dolan at the i n i t i a l stage of this project i s acknowledged. The professional work by the technical staff in the C i v i l Engineering workshop, particularly Dick Postgate, Bernie Merkli, Guy Kirsch and Max Nazar i s much appreciated. A special thanks is extended to the Earthquake Engineering and Structural Dynamics Research Laboratory technicians Howard Nichol and Michael Penn. Howard's countless hours of assistance during the experimental phase of this project kept this project on track. Michael's data acquisition software immensely aided the management of the substantial amount of data collected during the experiment. Last, but not least, I wish to thank my family and friends for their support throughout my graduate career. Willy C.W. Yung January, 1991 Vancouver, Bri t i s h Columbia - x i i i -1 . INTRODUCTION 1 l . l Background In North America, wood frame c o n s t r u c t i o n i s commonly used i n r e s i d e n t i a l and li g h t - c o m m e r c i a l b u i l d i n g a p p l i c a t i o n s . These wood s t r u c t u r e s u t i l i z e wood-based panels, such as plywood or waferboard, fastened t o wood framing w i t h n a i l s , screws or adhesives, t o serve as s t r u c t u r a l elements. H o r i z o n t a l elements such as f l o o r s and r o o f s are r e f e r r e d t o as "diaphragms". V e r t i c a l elements, such as w a l l s and p a r t i t i o n s are r e f e r r e d t o as " s h e a r w a l l s " . .. A s h e a r w a l l i s r e q u i r e d t o t r a n s m i t a l l the loads imposed on a wood s t r u c t u r e t o i t s foundation. G r a v i t y loads such as snow, occupancy and the s e l f - w e i g h t of the s t r u c t u r e are t r a n s f e r r e d by i t s v e r t i c a l framing members. L a t e r a l f o r c e s from wind and earthquake are c a r r i e d by i t s sheathing panels. In the past, wood s t r u c t u r e s have performed w e l l when subjected t o s e i s m i c a c t i v i t i e s . T h i s i s a t t r i b u t e d t o the high strength-to-weight r a t i o of wood, the d u c t i l i t y of the s t e e l f a s t e n e r s and the redundancy of the s t r u c t u r a l system. However, today's t r e n d s toward longer spans between w a l l s , concrete t i l e s on r o o f s and concrete o v e r l a y on f l o o r s have l e a d t o l a r g e r and h e a v i e r wood s t r u c t u r e s . In a d d i t i o n , new and he a v i e r m a t e r i a l are being added t o the upper f l o o r s f o r sound c o n t r o l , f i r e p r o t e c t i o n , a e s t h e t i c and economic reasons. Due t o these r e l a t i v e l y r ecent developments i n wood s t r u c t u r e s , l i t t l e o r no i n f o r m a t i o n i s a v a i l a b l e regarding t h e i r s e i s m i c performance. 2 Past experiences based on older wood structures are no longer reliable. This lack of information, coupled with the use of traditional construction techniques on these larger and heavier structures raise concerns regarding their a b i l i t y to withstand seismic loads. To address this concern, countries such as New Zealand, Japan, Canada and the United States have been engaged in a surge of research to study and improve the aseismic design of wood structures since the late 1970*s. Current methods of aseismic design place heavy reliance on a- structure's a b i l i t y to supply sufficient d u c t i l i t y (i.e. the a b i l i t y to withstand large inelastic deformations without collapse). In a wood structure, this d u c t i l i t y can be developed mainly by the sheathing-to-framing connections of i t s shearwalls. Under cyclic loading from an earthquake, the wood around the sheathing-to-framing connections i s severely crushed. This permanent damage results in a progressive loss of lateral stiffness (pinching) in the shearwalls. Consequently a wood frame structure must undergo larger lateral deformations in order to provide the d u c t i l i t y necessary for dissipating a given amount of seismic energy. For larger and heavier buildings, these deformations can be quite large, leading to greater damage and possibly, structural collapse. By decreasing the amount of seismic energy the shearwalls must dissipate, the damage level they sustain can be lessened. Consequently, the structure's rate of degradation in lateral stiffness i s reduced, leading overall to lower structural damage. This reduced damage level is possible by providing a wood structure with an alternate 3 means of seismic energy dissipation, in addition to i t s existing shearwalls. In recent years, the concept- of f r i c t i o n damping has been proposed to improve the seismic performance of steel braced frame buildings ( F i l i a t r a u l t , 1985)". Sliding f r i c t i o n devices added to 'the bracing system of a steel braced frame building were able to enhance i t s earthquake resistance and damage control potential ( F i l i a t r a u l t , 1985). Under normal service conditions and moderate earthquakes, the devices do not s l i p . However, under severe seismic excitations, the f r i c t i o n devices s l i p at a predetermined load, allowing a large portion of the seismic energy input to be dissipated mechanically in f r i c t i o n rather than by the inelastic deformation of the primary structural components. The addition of f r i c t i o n damping devices reduced the level of structural damage to a building, minimizing a repair cost which could be as significant as the replacement of a collapsed structure. In this investigation i t i s proposed to consider the concept of f r i c t i o n damping for wood structures. Sliding f r i c t i o n devices similar to ones used for steel braced frame buildings are added to the four corners of a timber shearwall. They are intended to assist the shearwall in dissipating the seismic energy input. The potential application of this concept to improve the seismic performance of wood structures i s explored in this research project. 4 1.2 O b j e c t i v e and Scope The primary objective of this research i s to explore the potential application of the f r i c t i o n damping concept to wood structures for improving their seismic performance. This objective was achieved through experimental means by testing f u l l scale 2.44 m (8 ft) square conventional light frame timber shearwalls against ones retrofitted with f r i c t i o n damping devices. The walls were tested under racking, s t a t i c - c y c l i c and dynamic earthquake simulations on the shake table. Results from these tests were analyzed to compare and evaluate the seismic performance between conventional light frame timber shearwalls and ones retrofitted with f r i c t i o n damping devices. This project was intended to be a preliminary study, to serve as a basis for further research and development into the application of the concept of f r i c t i o n damping in wood structures. The series of tasks carried out to meet the objective were as follows: i) Predict the racking behaviour of conventional and f r i c t i o n damped timber shearwalls, prior to the slippage of the devices. An existing f i n i t e elements program, SADT (Foschi, 1977), i s used for this purpose, i i ) Perform cycl i c tests on four f r i c t i o n damping devices to investigate their hysteretic behaviour and t o calibrate each unit for the f u l l scale tests. i i i ) Design the connections for attaching the four f r i c t i o n damping devices to the framing members of a shearwall. iv) Compare the racking and static c y c l i c performance of f u l l scale 2.44 m (8 ft) square timber shearwall specimens of both types, v) Determine the optimal s l i p load to •'tune" each f r i c t i o n damping device to. Predict the response time history of both types of shearwalls under the influence of different ground excitations. A numerical inelastic time-history dynamic model, FRICWALL (F i l i a t r a u l t and Dolan, 1989), is used for these tasks. Hysteretic parameters for this model are derived from the racking and static cyclic tests on the f u l l scale shearwalls. vi) Perform dynamic earthquake simulation tests on f u l l scale 2.44 m (8 ft) square timber shearwalls of both types to compare their performance under the influence of different ground motions. Since this i s a preliminary investigation, only shearwalls constructed from 38x140mm ( 2 x 6 in nominal) Spruce-Pine-Fir (S-P-F) framing, sheathed with 9.5 mm (3/8 in) Douglas F i r plywood panels using hot dipped galvanized nails are considered. The use of different framing members, wall panels, nails and nai l spacings are beyond the scope of this study. In-depth investigation of different material combinations and r e l i a b i l i t y based studies are needed before modifications may be made to the existing design and construction procedures of wood structures. 6 2. BEHAVIOUR OF CONVENTIONAL AND FRICTION DAMPED TIMBER SHEAR WALLS 2.1 Introduction A shearwall carries the .vertical and lateral loads imposed on a wood structure to i t s foundation. This chapter describes the role of the individual components of a conventional and a f r i c t i o n damped timber shearwall. The manner in which the two load carrying elements dissipate the seismic energy input from an earthquake ground motion i s discussed. 2.2 Conventional Timber Shearwalls A conventional timber shearwall i s constructed of wood-based sheathing panels, such as plywood or waferboard, fastened to the wood framing with steel sheathing-to-framing connectors. The sheathing panels, with their high in-plane r i g i d i t y provide lateral strength and stiffness to the structure. The framing, basically a configuration of pin ended members, carries vertical loads and provides out-of-plane stiffness to the sheathing. It offers practically no stiffness in the plane of the wall, and therefore does not contribute significantly to the lateral stiffness and strength of a wood structure. When subjected to a lateral load, i t has been observed that a conventional timber shearwall deforms mainly in shear. The framing distorts into a parallelogram with the header and sole plates remaining essentially horizontal, while the sheathing panels rotate essentially as r i g i d bodies. This deformation pattern i s shown in Figure 2.1. During this deformation, heavy-reliance i s placed on the ductile sheathing-to-framing connectors to deform i n e l a s t i c a l l y and absorb the seismic energy inputted.by an earthquake. By limiting i n e l a s t i c deformation to the ductile connectors, the chances of a b r i t t l e failure in the framing and sheathing panels are minimized. HEADER PLATE STUD L A T E R A L L O A D SHEATHING PANEL L A T E R A L L O A D SOLE PLATE F i g u r e 2.1: Deformation Pattern of a Conventional Timber Shearwall 8 This method of imposing a mode of failure in a structural element i s a requirement of the capacity design approach used in shearwall design. This approach reduces the likelihood of an unexpected failure mode which may be detrimental to the structure. From an energy perspective, this design approach for shearwalls i s i n e f f i c i e n t because i t leads to a pinched hysteretic behaviour as illustrated in Figure 2 . 2 . The pinching i s typical of conventional timber shearwalls, resulting from their loss in energy dissipating a b i l i t y . After the wood surrounding the sheathing-to-f raming connectors has been crushed during i t s plastic deformation in the i n i t i a l load cycle, i t s a b i l i t y to dissipate energy in subsequent load cycles i s drastically reduced. In successive load cycles, new wood must be crushed to compensate for this reduction such that a certain level of energy dissipation i s maintained. As more and more wood is damaged, the lateral stiffness of a shearwall progressively decreases. Consequently, a structure must undergo a large lateral deformation in order to dissipate the required quantity of seismic energy. For traditional wood structures, the rate of loss in lateral stiffness i s not c r i t i c a l enough to cause any structural problems. However, for large and heavy buildings, where individual shearwalls are supporting higher upper storey loads, the progressive loss in lateral stiffness can be much more rapid, leading to higher p o s s i b i l i t i e s of structural failure during an earthquake. 9 40 -60 -40 -20 0 20 40 60 DISPLACEMENT (mm) F i g u r e 2.2: Hysteretic Behaviour of Conventional Timber Shearwall 2 .3 F r i c t i o n Damped Timber Shearwalls The proposed f r i c t i o n damped timber shearwall i s composed of a conventional shearwall construction coupled with f r i c t i o n damping devices introduced in the four corners of the framing. The devices are compact and readily accommodate any construction 10 or architectural requirements. They can be incorporated in the design of new shearwall structures or i n the r e t r o f i t of existing ones. This simple modification, as i l l u s t r a t e d in Figure 2.3, provides an alternate mean of seismic energy dissipation in a. shearwall, in addition to i t s connections. FRICTION DAMPING DEVICE Figure 2.3: Timber Shearwall Incorporating Friction Damping Devices 11 When subjected to lateral loads, a f r i c t i o n damped timber shearwall deforms in the same manner as a conventional timber shearwall. As the framing distorts in a parallelogram shape, ther: geometry of the deformation allows forces to be transmitted into the overlapping diagonal struts of the f r i c t i o n damping devices from the framing. Two diagonally opposing devices are subjected to a tensile force while the other two are subjected to a compressive force as illustrated in Figure 2.4. Under normal service conditions and moderate earthquakes, the devices do not -slip. However, under severe seismic excitations, the f r i c t i o n devices s l i p at a predetermined optimal load, prior to any severe crushing of wood around the framing-to-sheathing connectors. This slipping of the device enables a portion of the seismic energy input to a shearwall to be mechanically dissipated in f r i c t i o n . Consequently, this reduces the plastic deformation in the connectors and the damage to the surrounding wood. The degradation in a shearwall's lateral stiffness i s reduced, thus a structure need not undergo a large lateral deformation to dissipate a given amount of seismic energy. The quantity of seismic energy dissipated by the f r i c t i o n devices depends on the magnitude of forces transmitted to the diagonal struts and the amount of slippage generated. Therefore the response of a structure can be "tuned" by adjusting the s l i p load on each device. 12 F i g u r e 2 . 4 : F o r c e s A c t i n g on F r i c t i o n D e v i c e s Due t o S h e a r w a l l D e f o r m a t i o n 13 3. TEST SPECIMENS 3.1 Introduction Four f r i c t i o n damping device prototypes and twenty-six f u l l scale timber shearwall - • specimens were tested in - this experimental investigation. This chapter provides a description of the f r i c t i o n devices, the shearwalls, and the design of the connection between these-two components. 3.2 Friction Damping Devices The four f r i c t i o n ' damping devices considered for this investigation were modifications of similar devices used in an earlier experiment on a braced frame steel structure ( F i l i a t r a u l t , 1985). Each device consisted of four mild steel plates hinged together in the form of a triangle. Figure 3.1 shows a detailed drawing of the device. Two 6 x 76 x 457 mm (1/4 x 3 x 18 in) plates formed the legs for mounting the device to the wood framing of a shearwall. The hinge at the intersection of these legs helped reduce the amount of framing separation in the corners during horizontal distortion of the wall. The framing separation arises from the build-up of high tensile forces in the end studs resulting from overturning moments. It was found from an earlier investigation (Dolan, 1989), that the addition of steel corner connectors could reduce the amount of framing separation. Hinged steel connectors were adopted for the devices since they did not contribute to the moment capacity of 14 the framing at the corners of the shearwall. The remaining two 9.5 x 51 x 318 mm (3/8 x 2 x 12.5 in) plates formed the diagonal struts for the device. These -struts were checked to ensure that they would not buckle under axial compression-before axial yielding of the members occurred. One end of each strut was connected to a leg, 356 mm (14 in) away from the hinge. The remaining end from each strut contained a slotted hole. These holes were overlapped onto each other and bolted in between an assembly of plates and washers using a 19 mm (3/4 in) diameter ASTM A490 high strength bolt. A pair of heavy -duty brake li n i n g pads (No. 55B by "ASBESTONOS"), as illu s t r a t e d in Figure 3.1, was glued around one of the holes using plasti-lock glue. The energy input from a severe seismic excitation i s dissipated by the brake lini n g pads, in f r i c t i o n , during the relative displacement between the diagonal struts of the device. By tightening the bolt to a specified torque, the s l i p load to cause this relative displacement can be controlled. The procedure to determine the s l i p load - torque relationship for each device i s discussed in Chapter 5. HEAVY DUTY BRAKE LINING PADS SLOTTED HOLE Figure 3 . 1 : Friction Damping Device 3 . 3 F u l l S i z e Timber Shearwalls The timber shearwall specimens considered i n the experimental i n v e s t i g a t i o n were constructed of material found i n t y p i c a l North American r e s i d e n t i a l and light-commercial bu i l d i n g s . Figure 3.2 shows a detailed-construction layout of a shearwall specimen. The 2.44 x 2.44 m (8 x 8 ft) specimens were framed with 38 x 140 mi (2 x 6 i n nominal) Spruce-Pine-Fir (S-P-F) members, #2 or better. (In some of the specimens, f i n g e r - j o i n t e d S-P-F members were used f o r the i n t e r i o r v e r t i c a l studs.) These members were oriented with t h e i r strong bending axis orthogonal to the plane of the wal l . The framing configuration consisted of two header plates and one sole plate, with studs spaced i n between at 610 mm (24 in) . Double studs were used at the two ends of the wall. The frame was assembled with lOd or 76 mm (3 in) common n a i l s penetrating from the header and sole plates into the ends of the studs. The second header plate was attached a f t e r the f i r s t header plate was na i l e d to the studs. The framing was sheathed on one side with two sheets of 1.22 x 2.44 m (4 x 8 f t ) , 3 ply, 9.5 mm (3/8 in) th i c k Douglas F i r plywood panels. They were fastened with the face grain i n the v e r t i c a l orientation, using 8d or 63.5 mm (2.5 in) hot dipped galvanized common n a i l s . Spacing of these n a i l s were at 102 mm (4 in) around the perimeter of the sheathing panels and 152 mm (6 in) along the i n t e r i o r studs. The shearwall construction was based on s p e c i f i c a t i o n s from the National Building Code of Canada (National Research Council of Canada, 1985). 17 HEADER PLATES PERIMETER 8d NAILS @ 102 mm SPACING DOUBLE END STUD INTERIOR 8d NAILS @ 152 mm SPACING INTERIOR STUD SOLE PLATE 9.5 mm 3 PLY DOUGLAS FIR PLYWOOD 0.61 m 0.61 m 0.61 m 0.61 m 2.44 m -L- I NOTE: All framing members are of 38 x 140 mm SPF, #2 or better Figure 3.2: Layout of Shearwall C o n s t r u c t i o n 18 In i n i t i a l tests of the shearwalls retrofitted with f r i c t i o n damping devices, bending of the double end studs at points where the diagonal struts from the devices intersected the framing was noticed. Since the end studs were oriented with their weak bending axis in the plane of the wall, they offered limited bending resistance when subjected to the concentrated forces introduced by the struts during the wall's shear deformation. To increase their bending stiffness, reinforcements were added. A 38 x 140 nm ( 2 x 6 in nominal) S-P-F stud, #2 or better, with i t s strong bending axis in the plane of the wall, was added to each pair of end studs for subsequent tests. They were fastened with two staggered columns of lOd or 76 mm (3 in) common nails spaced at 204 mm (8 in) along the double end studs. This modification was j u s t i f i a b l e because the reinforcements could represent the end studs from adjoining partition walls, or adjoining shearwalls at the corners of a timber framed building. These studs were of the same length as the ones being reinforced and therefore added no additional stiffness or strength to the corners of the shearwalls. Figure 3.3 shows an elevation and a cross-sectional view of a f r i c t i o n damped timber shearwall with the end stud reinforcements. 19 E L E V A T I O N V I E W ® • k \ ' ^ . j Q O < X X x ? 0 S E C T I O N A - A VERTICAL STUD DOUGLAS FIR PLYWOOD 38 x 140 mm END STUD REINFORCEMENTS Figure 3 . 3 : Double End Stud Reinforcement 3.4 F r i c t i o n Device t o Shearwall Connection Design Bracket and p i n connectors, as i l l u s t r a t e d i n Figure 3.4, were designed t o t r a n s f e r a x i a l f o r c e s i n t o the diagonal s t r u t s of the f r i c t i o n damping dev i c e s . Each connection c o n s i s t e d of two b r a c k e t s and one p i n . One bracket was welded t o the end of a diagonal s t r u t , w h i l e the second was secured t o the hinged l e g u s i n g f o u r 6 x 32 mm (1/4 x 1-1/4 in) cap screws. A p i n was snugly i n s e r t e d between the brackets f o r the l o a d t r a n s f e r . T h i s connection was designed t o comply w i t h CSA Standards, CAN3-S16.1-M84 (CISC, 1986). F i g u r e 3 .4 : Bracket and P i n Connectors f o r F r i c t i o n Damping Device 21 Brackets were cut from a bar of mild structural steel and pins were cut from a high strength carbon d r i l l rod material. These parts were machined to produce a tight f i t in order that the slightest distortion in the framing would enable forces to develop in the diagonal struts of the devices. Cyclic tests on the diagonal strut assembly of one device showed that i t s hysteretic behaviour drastically degraded when load transfer to the diagonal struts was via a loose connection. Such tolerances resulted in play between the connection parts and diminished the amount of slippage in the devices. Consequently, the level of seismic energy-dissipated by the devices i s reduced. The f r i c t i o n damping devices were mounted to the wood frame of the shearwall with lag screws designed in accordance with CSA Standard CAN3-086.1-M84 (Keenan, 1986). The design load was based on results from the static f i n i t e elements model discussed in Chapter 4. The lag screw mounting patterns are l a i d out in Figure 3.5. For the double end studs and header plates, six 9.5 x 76 mm (3/8 x 3 in) lag screws were required for the load transfer. For the sole plate, ten 9.5 x 38 mm (3/8 x 1-1/2 in) lag screws were used. Holes were d r i l l e d in the hinged legs at the locations where the screws were to be attached. Additional screws were added to secure the legs along their lengths. 22 23 4 . STATIC FINITE ELEMENTS MODEL 4.1 Introduction A finite-elements program, Static Analysis of Diaphragms and Trusses (SADT)> was used to numerically compute the racking load-deformation behaviours for the conventional and the fr i c t i o n damped timber shearwalls prior to any testing. SADT analyses wood shearwalls or diaphragms, taking into account the non-linear deformation of i t s connectors (Foschi, 1977). In the analysis, the . framing members, the sheathing panels, the connections.between adjoining framing members and the sheathing-to-framing connections are considered as four separate structural components. A beam element i s used in SADT to model each framing member, or segment of a framing member defined by the nodes of a f i n i t e element mesh. This element assumes a cubic function to approximate i t s lateral deformation and a linear function to approximate i t s axial displacement. Adjacent framing segments are assumed to be continuous, with compatible displacements, rotations and stresses. Each sheathing panel i s modelled by a plane stress, 12 node, cubic isoparametric f i n i t e element with in-plane displacements. The material for this plane element are assumed to be elasti c with orthotropic principal axes of elasti c symmetry. Each connection between adjoining framing members i s . modelled as a single discrete connector element. Non-linear load-deformation and moment-rotation properties are assigned to 24 this connection. The sheathing-to-framing connectors are modelled as uniform lines of smeared connectors, with non-linear*load-deformation properties. The number of connectors within each framing member or segment of a framing member defined the density! (ie. nails per unit length) of each smeared line of connectors. In the analysis, SADT uses an iterative process to balance the external energy from the applied la t e r a l load with the total internal strain energy from the deformation of the non-linear wall connectors. For the connectors between adjoining framing members, the strain energy arises from the relative displacement and rotation between the intersecting nodes of these adjoining members. For the sheathing-to-framing connectors, the strain energy arises from the relative displacement between the framing members and the sheathing panels. Since these connectors are modelled as uniform lines of smeared connectors, SADT can perform a numerical integration along the entire length of the framing member segment (ie. along the smeared connector line) to compute the work done by these connectors. This method of determining the strain energy in the sheathing-to-framing connectors allowed f l e x i b i l i t y in the f i n i t e element mesh generation procedure by not requiring the nodes on the framing f i n i t e element mesh to coincide with those of the sheathing panel f i n i t e element mesh. The energy balance process i s repeated during each load increment u n t i l the error in energy balance between the internal and external work i s below an acceptable level of tolerance. The analysis terminates when a framing element reaches a prescribed failure stress level, when a connector exceeds a previously defined level of d u c t i l i t y or when the programs reaches the applied lateral: load level without any failure in the wall. At this point, SADT provides with an output tracing the node displacements, and the framing element forces and stresses computed for each load increment. This chapter discusses the generation of the f i n i t e element mesh and the modelling of each component of the shearwall for the SADT analysis. The . numerical results for both the conventional and the f r i c t i o n damped timber shearwalls are presented. The Comparison between these numerical results and the experimental findings are discussed in Chapter 7. 4.2 Racking Behaviour of Conventional Timber Shearwall The generation of the f i n i t e element mesh, modelling of the shearwall components and numerical results from the SADT analysis of a conventional timber shearwall subjected to a uniaxial racking load are discussed in this section. 4.2.1 Finite Element Mesh Generation A two-dimensional X-Y coordinate system was used to reference a l l points on the shearwall to an origin at the lower l e f t corner of the wall. A f i n i t e element mesh with lines spaced 610 mm (24 in) apart in both the X and Y directions divided each framing member into four equal length segments. The nodes on this framing f i n i t e element mesh were identified by a numbering system while the framing member segments in between these nodes were numbered using a separate system. The X-Y coordinate system and the numbering systems for the nodes and the framing member segments on the framing f i n i t e element mesh of a conventional timber shearwall are displayed in Figure: 4.1. A * separate f i n i t e element mesh was generated for the sheathing panels. To accommodate the 12 node, cubic isoparametric f i n i t e element representing each sheathing panel, f i n i t e element mesh lines divided each panel into three equal sized sections in both the X and Y directions. Based on.the vertical" orientation of the sheathing panels, this division produced mesh lines spaced 407 mm (16 in) apart in the X direction, and 813 mm (32 in) apart in the Y direction. Each node on this mesh coinciding with the perimeter of a sheathing panel was assigned a number using the numbering scheme previously adopted for the nodes on the framing f i n i t e element mesh. The X-Y coordinate system and the numbering system for nodes on the sheathing panel f i n i t e element mesh are displayed in Figure 4.2. 27 FRAMING MESH NODE FRAMING MESH NODE NUMBER E o C O E E o C O o to E E o C O 26 10 12 FRAMING FINITE ELEMENT MESH FRAMING MEMBER SEGMENT \ 18 FRAMING MEMBER SEGMENT NUMBER 10 •9-~8 8 16 4 4 , 15 43-14 22 49 21 20 ....7..... (..1.2) U-7-) 13 19 5 /—\ 11 / — x 17 \ 23 6 ) (11) (16 25 28 0 27 23 26 25 21 ® 610 mm 610 mm 610 mm 610 mm Figure 4.1: Framing Finite Element Mesh for Conventional Timber Shearwall 28 PERIMETER OF SHEATHING PANEL E E co CO E E CO CO E CO CO 30 SHEATHING.PANEL MESH NODE NUMBER 34 36 SHEATHING PANEL FINITE ELEMENT MESH 41 45 40 44 39 43 38 42 47 49 46 48 53 52 51 50 407 mm 407 mm 407 mm 407 mm 407 mm 407 mm Figure 4 .2: Sheathing Panel Finite Element Mesh for Conventional Timber Shearwall 29 4 . 2 .2 Shearwall Component Modelling The modelling of the framing members, sheathing panels, connections between framing members and sheathing-to-framing connections are discussed in this section. The boundary conditions and the points of lateral load application on the shearwall model are also discussed. 4 .2.2.1 Framing Members Two different cross-sectional sizes were assigned to the beam elements used for modelling the framing members. The interior studs and the sole, plate were sized as 38 x 140 mm (1.5 x 5.5 in actual) elements. The double header plates and end studs were sized as 76 x 140 mm (3 x 5.5 in actual) elements. These double members were assumed to behave li k e a single member with the same net cross-sectional area. Both sizes of elements were 610 mm (24 in) in length. They were modelled with their strong bending axis oriented in the plane of the wall. The material properties for both sizes of beam elements pertained to that of a S-P-F member, graded #2 or better. These properties are l i s t e d in Table 4.1. Table 4.1 shows two different values- for the maximum bending strength of the S-P-F framing member. The values of 18 MPa (2660 psi) and 40 MPa (5750 psi) correspond, respectively, to the non-parametric 5th percentile and mean value bending strengths for the member. These values are based on analysis of bending strength data obtained from the Canadian Wood Council 1s (CWC) In-Grade Testing Program (Foschi, Folz and Yao, 1989). 30 Due to the se n s i t i v i t y of the shearwall's failure load to the bending strength of i t s framing members, the analysis was performed for both bending strength values. Table 4.1: Material Properties for SPF Framing MAX BENDING STRESS (|) MAX SHEAR " STRESS <FV) MODULUS OF ELASTICITY (E) 5th PERCENTILE VALUE MEAN VALUE 18 MPa (2660 psi) 40 MPa (5750 psi) 2 MPa (300 psi) 9660 MPa (1400000 psi) 4.2.2.2 Sheathing Panels The two sheathing panels were modelled with two plate elements, each 1.22 x 2.44 m (4 x 8 ft) in size. The material properties assigned to these plate elements were from a 9.5 mm (3/8 in) thick, 3-ply Douglas F i r plywood panel, with i t s principle axis oriented in the v e r t i c a l (Y-axis) direction. These properties are l i s t e d in Table 4.2. Table 4.2: Material Properties for Douglas F i r Plywood Panel MODULUS OF ELASTICITY SHEAR MODULUS (G) POISSON'S RATIO (Ey) ^yx 6400 MPa (933000 psi) 4600 MPa (664000 psi) 690 MPa (100000 psi) 0.18 0.25 4.2.2.3 Connections Between Framing Members Four d i f f e r e n t connector elements were used t o model the connections between t h e framing members a t th e f o u r c o r n e r s of the s h e a r w a l l . These elements a l l o w e d f o r the r e l a t i v e deformation and r o t a t i o n between t h e framing members which occur d u r i n g t h e deformation o f the s h e a r w a l l . F i g u r e 4 .3 shows a curve i l l u s t r a t i n g how the connector element's n o n - l i n e a r l o a d -deformation and moment-rotation behaviours a r e modelled i n SADT. Figure 4.3: Curve M o d e l l i n g the Non-Linear Behaviour o f the Connectors Between the Framing Members The parameters P0, and Kg shown in Figure 4.3 represent, respectively, the intercept of the asymptote with slope Kj, the stiffness for large relative displacement or rotation and the i n i t i a l connector stiffness. These parameters were determined based on the number of fasteners used at a corner between the framing members, and the type of load introduced to the corner as a result of the shearwall's racking deformation. The manner in which a wall specimen was mounted to the testing frame was also considered in the choice of parameters for each connector element. Table 4.3 l i s t s the parameters used in the non-linear curves for modelled the behaviour of the framing connectors at the four corners of a conventional timber shearwall. Table 4.3: Parameters Defining the Non-Linear Behaviour of the Connectors Between the Framing Members B W A L L C O R N E R D A B C D p * 61.19(13.75) 40.05 (9.0) 5.56(1.25) 40.05 (9.0) 0(0) 0(0) 0(0) 0(0) K0 54.75 (312.5) 87.6 (500.0) 5.47(31.25) 87.6 (500.0) Po 66.75 (15.0) 11.13(2.5) 11.13(2.5) 66.75 (15.0) K2 0(0) 0(0) 0(0) 0(0) K0 52.56 (300.0) 10.95 (62.5) 10.95 (62.5) 52.56 (300.0) Po 6.68 (1.5) 6.68 (1.5) 6.68 (1.5) 6.68(1.5) K2 0(0) 0(0) 0(0) 0(0) K0 52.56 (300.0) 19.71 (112.5) 19.71 (112.5) 19.71 (112.5) o i CO <! z o I cc co co LU Z u. LL CO CO at z It .g CC Q cr a * NOTE: P0 in units of [kN(kips)]; K a& K |n units of [kN/m(kips/in)] Non-linear connector elements were not considered for the connections between the interior vertical studs and horizontal framing plates. With the absence of these connector elements, the nailed joints between the framing members at these locations were assumed to behave in a r i g i d manner during the wall deformation, with no relative rotation between the framing members. This assumption contradicted the "pin ended member" analogy previous used to describe the frame distortion, where relative rotation between the intersecting framing members i s allowed. However this simplified analogy really described the overall frame distortion. The actual behaviour at the joint i s quite complex because of the interaction with the sheathing panels, which adds inherent r i g i d i t y to the joint. Additional moment capacity i s introduced by the fasteners between the framing members. With the combination of non-linear connectors between the framing member at the four corner joints, r i g i d connectors at the interior framing joints and the use of four beam elements to represent each framing member, SADT was able to adequately predict the load-deformation behaviour of a conventional timber shearwall under a racking load. 4.2 .2 .4 Sheathing-To-Framing Connections The sheathing-to-framing connections were modelled with non-linear connectors exhibiting the load-deformation behaviour shown in Figure 4 . 4 . The parameters from this curve were determined from single shear connection tests of 8d or 63.5 mm (2.5 in) hot dipped galvanized common nails (Dolan, 1989). These tests were performed using a different grade of framing material. However, since the lateral and withdrawal capacities of nails are only distinguished for different species of wood but not different grades (Canadian Wood Council, 1982), no adjustments were made to these parameters for the analysis. The values shown for the parameters on thi s curve were used to simulate the connector behaviour for loading orientations p a r a l l e l and perpendicular to the grain direction of the framing members. P(N) Figure 4.4: Load Deformation Behaviour for Sheathing-to-Framing Connection 35 4.2.2.5 Boundary Conditions A l l nodes on the framing f i n i t e element mesh, coinciding with the base of the shearwall model, were restrained against translational motion in both the v e r t i c a l >or horizontal directions. No restrictions were applied to their rotational degree of freedom. A l l of the remaining framing nodes were allowed f u l l translational and rotational degrees of freedom. 4.2.2.6 Load Application Loading for the shearwall was applied at the five nodal points located at the top of the wall. Each of these nodes were assigned an i n i t i a l l ateral load of 445 N (100 lbs) in the positive X direction. During each load increment, 445 N (100 lbs) were added to each node until the analysis terminated. No gravity loads were used in the analysis. 4.2.3 Numerical Results The two curves in Figure 4.5 il l u s t r a t e s the racking load-deformation behaviour of a conventional timber shearwall as predicted by SADT. The points from the curves correspond to the maximum lateral deformation in the wall during each load increment. Two analyses were carried out, using separate allowable bending stresses for the framing members, to produce the two load-deformation curves. Both curves exhibit the same load-deformation characteristics, with the wall analyzed with the higher strength framing members f a i l i n g at a higher lateral displacement. The curve shown by the solid line terminated at a peak racking load of 28.9 kN (6.5 kip) with a late r a l displacement of 34.5 mm (1.4 in). This failure-load corresponded to the non-parametric 5th percentile allowable bending stress of 18 MPa (2660 psi) in the framing members. The wall failure was init i a t e d by framing member segment #16 (see Figure 4.1), where the allowable bending stress was exceeded. The racking strength of 95% of shearwalls tested i s expected to exceed this load capacity. The curve shown by the dashed line terminated at a peak racking load of 42.3 kN (9.5 kip) with a late r a l displacement of 87.9 mm (3.5 in). This higher failure load corresponded to the mean allowable bending stress of 40 MPa (5750 psi) in the framing members. This wall failure was also initiated by framing member segment #16 where the allowable bending stress was exceeded. This peak racking load represents the average load capacity expected from a large sample of shearwall tests. Since SADT did not have the capability to predict the post-failure behaviour of a shearwall, both load-deformation curves only illustrated the wall behaviour up to the failure load. The comparisons between these numerical results and the experimental findings are discussed in Chapter 7. 37 50 I i PEAK LOAD = 28.9 kN ~y • (ALLOWABLE BENDING / STRESS = 18 MPa) / LATERAL LOAD ON WALL 40 30 20 10 -/ / \ / \ \ PEAK LOAD = 42.3 kN ^ Jt (ALLOWABLE BENDING ^ STRESS = 40 MPa) 0 • i 0 20 40 60 80 RELATIVE TOP OF WALL DISPLACEMENT (mm) Figure 4 . 5 : Racking Behaviour of a Conventional Timber Shearwall as Predicted by SADT 38 4.3 Racking Behaviour of Friction Damped Timber Shearwall Under axial loading, a f r i c t i o n damping device exhibits an Ela s t i c load-deformation response u n t i l i t s s l i p load i s reached. At this point, i t enters the p l a s t i c region of the load-deformation curve as the diagonal struts s l i p relative to each other under a constant load plateau. This elasto-plastic property of the f r i c t i o n damping device i s i l l u s t r a t e d by the idealized load-deformation curve in Figure 4 .6. PLASTIC REGION Q < O ELASTIC REGION SLIP LOAD PLATEAU K =• 200000MPa-(30000000 psi)-AXIAL DEFORMATION (mm) Figure 4 . 6 : Idealized Load-Deformation Curve of a Friction Damping Device The elasto-plastic behaviour of the f r i c t i o n damping devices could not be modelled by SADT because i t s non-linear capability was limited to the connectors of a shearwall. For this reason, SADT could not be used to predict the load-deformation behaviour of a f r i c t i o n damped timber shearwall in racking. Instead, SADT was used to determine the design load required by the connections between a f r i c t i o n device and the wall framing. An SADT analysis was performed by modelling each pair of overlapping diagonal struts on the four f r i c t i o n devices as a single steel member. This member could be interpreted at as a strut assembly which has been set to a very high s l i p load such that forces transmitted by the shearwall could not cause i t to s l i p . During the analysis, this member's high strength and stiffness would allow i t to remain elastic and not f a i l prior to any other component of the shearwall. By remaining elastic prior to the failure of the shearwall, loads transmitted to the devices would increase with the shearwall deformation rather than level off at a plateau corresponding to some s l i p load. When the wall failed, the axial load developed in the strut assembly would be at i t s peak value, surpassing any level of s l i p load the devices may be set to during i t s working condition. This strut force was used to estimate the design load for the connectors between the f r i c t i o n damping devices and the shearwall framing. This section discusses the generation of the f i n i t e element mesh and the modelling of the shearwall components for the SADT analysis of a f r i c t i o n damped timber shearwall. The design load chosen for the connectors between the f r i c t i o n devices and the wall framing i s presented. 4.3.1 Finite Element Mesh Generation The f i n i t e element meshes previously generated for the conventional timber shearwall formed the bases for the f i n i t e element meshes of the f r i c t i o n damped timber shearwall. To accommodate the f r i c t i o n damping devices, additional mesh lines, nodes and elements were added to the framing f i n i t e element mesh. For the sheathing panel f i n i t e element mesh, only a reassignment of node numbers was necessary. The X-Y coordinate system and the numbering systems for the nodes and the framing member segments on the framing f i n i t e element mesh of a f r i c t i o n damped timber shearwall are illustrated in Figure 4 . 7 . The X-Y coordinate system and the numbering system for the nodes on the sheathing panel f i n i t e element mesh are displayed in Figure 4 . 8 . 41 Figure 4.7: Framing Finite Element Mesh for Friction Damped Timber Shearwall 42 PERIMETER OF SHEATHING PANEL E E co 5 E E co CO CO CO 46 SHEATHING PANEL MESH NODE NUMBER SHEATHING PANEL FINITE ELEMENT MESH 57 61 56il60 55 59 63 65 69 68 50 52 54 58 62 64 66 407 mm 407 mm 407 mm 407 mm 407 mm 407 mm Figure 4.8: Sheathing Panel Finite Element Mesh for Fri c t i o n Damped Timber Shearwall 43 4.3.2 Shearwall Component Modelling The modelling of framing members, sheathing panels, connections between framing members, sheathing-to-framing connections, boundary conditions and the locations of l a t e r a l load application were identical to those used in the SADT analysis of a conventional timber shearwall. Section 4.2.2 contains the details on how these components were modelled. 4.3.2.1 F r i c t i o n Damping Devices Three beam elements were used to represent the steel plates of the f r i c t i o n damping devices. The two hinged legs of a device were treated as 6 x 76 x 356 mm (1/4 x 3 x 14 in) elements. Its two overlapping diagonal struts were treated as a single 9.5 x 51 x 503 mm (3/8 x 2 x 19.8 in) element. The strength properties for CSA G40.21-M, structural quality steel were assigned to these elements. These values are l i s t e d in Table 4.4. Table 4 .4: Material Properties for Friction Damping Devices MAX BENDING STRESS a y MAX SHEAR STRESS (Fv) MODULUS OF ELASTICITY (E) 300 MPa (43500 psi) 300 MPa (43500 psi) 200000 MPa (30000000 psi) The hinges connecting the legs and the diagonal struts of the f r i c t i o n damping devices were modelled by the connector elements. Parameters assigned to these elements enabled them to exhibit a high stiffness for their translational degrees of freedom and a zero stiffness for their rotational degree of freedom. These stiffness values eliminated any relative displacement between the adjoining members of the f r i c t i o n devices but allowed them to rotate freely relative to each other. The parameters used for modelling these spring stiffness are shown on the non-linear curves in Figure 4 . 9 . Figure 4 . 9 also i l l u s t r a t e s how these hinge connections were modelled in the SADT analysis. The four wall corners which contained the hinge between the two legs of each f r i c t i o n device were modelled in the same manner as in the conventional timber shearwall. Refer to Table 4 . 3 for details. i TRANSLATIONAL SPRING STIFFNESS P 0= 120.2 kN (27 kips) K 2 = 0 KQ = 118.3 kN/mm (675 kips/in) ROTATIONAL SPRING STIFFNESS p0 = 4.45 N (1 lb) K 2 = 0 K 0 = 0 ROTATIONAL SPRING TRANSLATIONAL SPRINGS MODELLED AS IN THE CONVENTIONAL SHEARWALL Figure 4.9: Non-Linear Connector Between Friction Damping Device and Wall Framing 46 4.3.3 Numerical Results The SADT predictions of the racking load-deformation behaviour of a f r i c t i o n damped timber shearwall, with non-slipping f r i c t i o n devices, are illustrated by the pair of thick curves i-rt. Figure 4.10. For comparison, the predictions for the conventional timber shearwall are shown by the two thinner curves. They are the same curves as the ones illustrated in Figure 4.5. The displacements shown on the curves correspond to the maximum lateral deformation of the wall during each load increment. The -curve shown by the thick solid line represented an analysis which was carried out using the non-parametric 5th percentile allowable bending stress value of 18 MPa (2660 psi) in the framing members. This wall failed at a racking load of 40.1 kN (9.0 kip) with a lateral displacement of 35.5 mm (1.4 in) . The failure was initi a t e d in framing member segment #16 (see Figure 4.7) which exceeded i t s allowable bending stress. The curve shown by the thick dashed line represented an analysis which was carried out using the mean allowable bending stress of 40 MPa (5750 psi) in the framing members. This wall fai l e d at a racking load of 53.4 kN (12.0 kip) at the lateral displacement of 83.8 mm (3.3 in). This failure was also in i t i a t e d in framing member segment #16 which exceeded i t s allowable bending stress. Comparison of the two sets of curves shows that the f r i c t i o n damped shearwall, retrofitted with non-slipping f r i c t i o n devices, offered greater lateral stiffness and sustain higher racking loads at failure than a conventional shearwall. The higher lateral stiffness and strength of the f r i c t i o n damped shearwall are attributed to the redistribution of shear forces from the shearwall to the devices. The share of lateral load harvested up by the struts alleviate the stresses in the shearwall, allowing i t to resist greater loads before failure. In the actual shake table tests, the f r i c t i o n devices were calibrated to s l i p at some load between zero (no device) and the maximum strut force computed from the analysis with non-slipping f r i c t i o n devices. Thus the load-deformation response revealed by the f u l l scale tests of the f r i c t i o n damped timber shearwalls would l i e somewhere between the two thick and the two thin curves shown in Figure 4.10. The two thick load-deformation curves could be interpreted as the upper bound solution, while the two thin curves could be regarded as the lower bound solution. The comparisons between these numerical load-deformation results and the experimental findings are discussed in Chapter 7. 48 Z XL i z o Q < o < UJ h -60 50 -40 -30 -20 -10 PEAK LOAD = 53.4 kN (ALLOWABLE BENDING STRESS = 40 MPa) PEAK LOAD = 40.1 kN (ALLOWABLE BENDING STRESS = 18 MPa) » ******* 0& PEAK LOAD = 42.3 kN (ALLOWABLE BENDING STRESS = 40 MPa) PEAK LOAD = 28.9 kN (ALLOWABLE BENDING STRESS = 18 MPa) 20 40 60 80 RELATIVE TOP OF WALL DISPLACEMENT (mm) Figure 4.10: Racking Behaviour of a Fri c t i o n Damped and a Conventional Timber Shearwall as Predicted by SADT Figure 4.11 shows the maximum axial strut force developed by the non-slipping f r i c t i o n damping devices during each increment of lateral load. For the wall analyzed with framing members at the 5th percentile bending strength of 18 MPa (2660 psi), a maximum strut force of 15.6 kN (3.5 kip) was computed by SADT. The solid line i l l u s t r a t e s the strut force to lateral load relationship for this strength of framing member. For the wall analyzed with framing jnembers at the mean value bending strength of 40 MPa (5750 psi), a maximum strut force of 24.5 kN (5.5 kip) was computed. The dotted line i l l u s t r a t e s the strut force to lateral load relationship for this strength of framing member. Since the f r i c t i o n devices were designed to s l i p during i t s working condition, i t was decided that the 15.6 kN (3.5 kip) strut force would be an adequate design load for the lag screw connections which attached the f r i c t i o n devices to the framing. This design load proved to be adequate as the f r i c t i o n devices were not calibrated to a slip-load in excess of 13.4 kN (3.0 kip) for any of the shearwall tests. The connection details are discussed in Chapter 3. LATERAL LOAD ON WALL (kN) F igure 4.11: Maximum Strut Force Developed in the Non-Slipping Fri c t i o n Devices During Each Lateral Load Increment 51 4.4 Summary The SADT analyses have been performed to predict the racking behaviour of timber shearwalls. Two sets of solutions, the lower and upper bounds, were produced. The lower bound solution corresponded to a conventional shearwall (ie. s l i p Toad = 0) . The upper bound solution corresponded to a f r i c t i o n damped shearwall with non-slipping diagonal struts (ie. s l i p load = axial yield load). The f r i c t i o n damped timber shearwalls to be tested, with a s l i p load somewhere between zero and the axial yield load of the struts, are expected to exhibit a racking behaviour which would l i e somewhere between these bounds. However, i t must be noted that due to the high v a r i a b i l i t y in the strength of wood, the racking behaviour shown by the f u l l scale test walls c a n f a l l outside of these bounds. The numerical results showed that the racking performance of a f r i c t i o n damped timber shearwall i s superior over that of a conventional shearwall, exhibiting higher in-plane stiffness and sustaining higher racking load before failure. As a note of caution, wall failure modes such as the withdrawal of the sheathing-to-framing connectors from the framing or the breaking of these connectors through the sheathing were not considered by SADT. Therefore the actual failure modes observed from the tests may not be reflected by the bending mode failure shown by the analysis. 52 5. CYCLIC TEST OF FRICTION DAMPING DEVICES 5.1 Introduction For the f r i c t i o n damping devices to function effectively, they must exhibit good energy .dissipating characteristics (hysteretic behaviour) under the dynamic loading which result from a seismic event. Due to factors such as ground condition, seismic intensity and the proximity of a seismic source, the duration, frequency content and magnitude of the dynamic loading a structure experiences would be unique for each earthquake ground excitation. This chapter discusses the cycli c tests which exposed the proposed f r i c t i o n damping devices to the possible range of dynamic loading conditions they may encounter during a seismic event. These tests examined the hysteretic behaviour of the fr i c t i o n devices and the s t a b i l i t y of this behaviour for different loading rates and s l i p load levels. From these tests, a correlation between the applied torques on each f r i c t i o n joint and the s l i p loads generated by these torques was also derived. This correlation was needed for calibrating each f r i c t i o n device to the desired s l i p load for the f u l l scale shake table tests. 5.2 Test Equipment and Instrumentation Cyclic tests of the f r i c t i o n devices were carried out on a MTS 810 Material Test System, located in the Timber Engineering Laboratory, at the University of British Columbia. The test system consisted of a load frame, a hydraulic pump and a MTS 53 458.20 MicroConsole as shown in Figure 5.1. The load frame contained a height adjustable cross-arm and a hydraulic actuator. The cross-arm could be locked at different heights to accommodate specimens of various sizes. A 250 kN (55 kip) capacity load c e l l mounted to the underside of the cross-arm measured loads on the specimen during the tests. The hydraulic actuator, with a load capacity of 250 kN (55 kip) , generated the loading force to the specimen being tested. Both static and dynamic loads could be produced. The available frequency range for the dynamic load i s between 0.1 to 300 hz. The maximum stroke of the actuator i s 152 mm (6 in) in the vert i c a l direction, decreasing with loading rates at the higher end frequencies. Travel of the actuator was measured by a LVDT (Linear Variable Displacement Transducer) located inside the hydraulic actuator. A steel platform connected to the top of the actuator f a c i l i t a t e d the mounting of test specimens. Hydraulic pressure to the actuator was supplied by a hydraulic pump. A pair of servo-valves regulated the f l u i d flow in and out of the actuator. The operation of the testing system was controlled, in a closed loop, by the MTS 458.20 MicroConsole. Via the MicroConsole, an IBM-At Personal Computer regulated the movement of the actuator to reproduce the desired loading pattern. The IBM computer also stored data recorded during the experiments. A Hewlett-Packard 7090A Measuring Plotting System connected to the MicroConsole produced instantaneous echoing of data as the tests were being conducted. 55 5 . 3 Test Procedure The diagonal strut assembly, which contained the f r i c t i o n joint of the f r i c t i o n damping device, was mounted in the load frame between the adjustable cross-arm and the platform of the hydraulic actuator as shown in Figure 5.2. The mounting brackets were machined to provide a tight f i t around the mounting bolts in order to reduce the tolerance in the connection between the specimen and the load frame. This connection detail was provided to minimize any loss in the hysteretic behaviour of the f r i c t i o n devices as a consequence of play between the mounting brackets and bolts. This low tolerance c r i t e r i a was adopted in the fabrication of the f r i c t i o n device and of the connections between the f r i c t i o n device and the shearwall discussed in Section 3.4. The hysteretic behaviour of the f r i c t i o n damping device and the s t a b i l i t y of this behaviour were examined by testing the strut assembly under repeated axial tensile and compressive loads for a duration of 50 cycles. A 1.0 hz, sinusoidal, displacement controlled wave, with an displacement amplitude of 9.5 mm (3/8 in) in both loading directions regulated the travel of the hydraulic actuator to produce the desired loading pattern. This load duration was chosen because i t encompassed the predominant period of strong motion of typical seismic ground excitations. It i s during this period of strong motion in which a f r i c t i o n device i s expected to dissipate the seismic energy input. For this i n i t i a l test, the f r i c t i o n joint on the strut assembly had been torque to 67.8 N-m (50 f t - l b ) . 56 Note: Each f r i c t i o n joint was covered with plastic to contain any asbestos dust generated during the tests. F i g u r e 5.2: Test Specimen as Mounted in the Loading Frame of the MTS Testing System The s e n s i t i v i t y of the device's hysteretic behaviour to different rates of loading was examined by repeating the 50 cycle test under sinusoidal frequencies of 0.5, 2.0, 3.5 and 5.0 hz. The frequency content exhibited by a typical earthquake ground excitation i s expected to f a l l somewhere within the 0.5 to 5.0 hz.frequency range. The same displacement controlled sinusoidal wave was used, but for o s c i l l a t i o n frequencies beyond 1.0 hz, the capability of the MTS testing system limited the amplitude of displacement in the actuator to a shorter stroke. Time history plots of the sinusoidal loading patterns employed in the tests are ill u s t r a t e d in Figure 5.3. Figure 5 . 3 : Time-History Plots Il l u s t r a t i n g the Pattern of Load Application on the Test Specimens 58 The hysteretic behaviour of the f r i c t i o n device under different s l i p loads was investigated by repeating the tests with the f r i c t i o n joint tightened to different levels of torques between the range of 27.1 to 203.4 N-m (20 to 150 f t - l b ) . For the f r i c t i o n joint being tested, this variation of torques generated s l i p loads ranging from 1.2 to 21.1 kN (0.3 to 4.7 kip) . The same 50 cycle duration of loading and set of sinusoidal loading frequencies were used at each torque level. Aside from investigating the hysteretic behaviour of the f r i c t i o n device sunder the influence of different s l i p loads, this last set of tests was also used to obtain the relation between the s l i p load of the device and the applied torque on i t s f r i c t i o n joint. This correlation was needed for tuning the device to the desired s l i p load when i t was mounted onto the shearwall for the f u l l scale tests. For the remaining three f r i c t i o n damping devices constructed for this study, the same testing procedure was repeated, but only at the sinusoidal frequency of 1.0 hz. This was done because the findings from the cyclic tests on the f i r s t f r i c t i o n device showed that i t s hysteretic behaviour was frequency independent. Measurements of the axial strut load and relative displacement (slippage) between the struts were taken during each test. These measurements were acquired by the MTS testing system's load c e l l and LVDT, and stored by the IBM computer. A Hewlett-Packard plotter graphically displayed the load and displacement measurements during the course of the tests. 59 5.4 Results and Discussion The load and displacement measurements from a l l of the tests were analyzed to examine the hysteretic behaviour of a l l four devices under different loading frequencies and s l i p loads. Figure 5 .4 i l l u s t r a t e s a load-displacement plot from one of these tests. This set of hysteresis loops was taken from a c y c l i c test which was conducted at the sinusoidal frequency of 1.0 hz with the f r i c t i o n joint torqued at 67.8 N-m (50 f t - l b ) . -8 --10 - 1 1 1 1 i 1 1 r -12 -8 -4 0 4 8 12 DISPLACEMENT (mm) Figure 5 .4: Load-Deformation Behaviour of a Fr i c t i o n Joint Under a 50 Cycle Sinusoidal Load 60 The overall rectangular shape of the hysteresis loops ill u s t r a t e d the excellent energy dissipating a b i l i t y of the f r i c t i o n joint. The slight imperfections in the 2nd and 4th quadrants of the hysteresis loops occurred during the load reversal parts of the test. They result from play between the mounting brackets and bolts which attached the test specimen to the loading frame. (These losses in the hysteresis loops were found to be considerably more pronounced in earlier tests, where mounting brackets which f i t t e d loosely around the mounting bolts were utilized.) By manufacturing brackets which f i t t e d snugly around the mounting bolts, and tightly securing these bolts to further minimize any movement in the mounting connections, i t was possible to obtain these considerably better hysteresis loops which showed only slight areas of imperfections. This finding demonstrated the significance of minimizing connection tolerances in the manufacturing of the f r i c t i o n devices. As shown by Figure 5.4, the hysteresis loops remained consistently stable and did not deteriorate over the entire 50 cycle test duration. Even after the f r i c t i o n pads began t o heat up after a series of tests with the f r i c t i o n joint set at different torques, they s t i l l preserved their stable hysteretic behaviour with no noticeable fade throughout the entire sequence of tests. Figure 5.5 i l l u s t r a t e s hysteresis loops for the same f r i c t i o n device, at the same torque of 67.8 N-m (50 f t - l b ) , but loaded under different frequencies, namely 0.5, 2.0, 3.5 and 5.0 hz. A progressive drop in the displacement i s evident for the tests conducted at the higher frequencies. As mentioned before, this was due to constraints in the capacity of the testing equipment. The hysteretic behaviour exhibited by the f r i c t i o n joint under the different loading frequencies remained stable, with no deterioration throughout the entire duration of each test. ~ The s l i p loads generated from these four tests ranged from 6.2 to 6.5 kN (1.4 to 1.5 kip) which i s consistent with the s l i p load of 6.5 kN (1.5 kip) derived from Figure 5.4 for the test which was conducted at the 1.0 hz frequency. This consistency in s l i p loads showed that the energy dissipating a b i l i t y of the f r i c t i o n device was not influenced by the loading rate. Figure 5.6 ill u s t r a t e s three sets of hysteresis loops taken from the cyclic tests of the same f r i c t i o n joint. These test were conducted at the sinusoidal frequency of 1.0 hz, with the f r i c t i o n joint calibrated to 67.8, 135.6 and 203.4 N-m (50, 100 and 150 ft-lb) of torques. As illustrated by Figure 5.6, the hysteretic behaviour exhibited by the f r i c t i o n joint under the different torques remained stable, with no deterioration throughout the entire test duration. The only difference between the three sets of hysteresis loops i s the increase in s l i p load for the tests conducted for higher torques. As the torque on the f r i c t i o n joint i s increased, more pressure i s exerted on the f r i c t i o n joint, thus increasing the axial load required to cause the struts to s l i p . The higher s l i p loads resulted in hysteresis loops with greater area, representing a larger amount of energy dissipation. 62 10-8-6-4-f 2\ 3.2--4--6 -8 -10-10 8 6 4 f 2 o " ,-2--4 --6 --8 -FREQUENCY j= 0.5 Hz SLIP LOAD = 6.50 kN -12 1 r -8 -4 0 4 8 DISPLACEMENT (mm) FREQUENCY = 3.5 Hz SLIP LOAD = 6.32 kN •10 10 8 6 4^  z 2 12 J-2 -4 -6 -8 -101 FREQUENCY \= 2.0 Hz SLIP LOAD 1 r 6.32 kN •12 -8 -4 0 4 8 DISPLACEMENT (mm) - i r 10 8 6 4 I2 § 0 o -4 -6 -8 -10 FREQUENCY '= 5.0 Hz SLIP LOAD = 6.19 kN 12 - i r -12 -8 -4 0 4 8 12 DISPLACEMENT (mm) -12 -8 -4 0 4 8 12 DISPLACEMENT (mm) Figure 5.5: Hysteresis Loops for Friction Joint Tests Conducted at the Frequencies of 0.5, 2.0, 3.5 and 5.0 Hz < O 30 20 10 0 -30 TORQUE = 67.8 N-m SLIP LOAD = 6.45 kN -12 -8 -4 0 .4 DISPLACEMENT (mm) 12 < O 30 20 1 10 0 -10 -20 -30 30 20 1 10 0 -10 -20 -30 TORQUE = 135.6 N-m " SLIP LOAD = 13.57 kN e & a f a -•12 -8 -4 0 4 DISPLACEMENT (mm) 8 12 < o TORQUE = 203.4 N-m o u r sr a _,, -12 -8 -4 0 4 DISPLACEMENT (mm) 8 12 Figure 5.6: Hysteresis Loops for F r i c t i o n Joint Tests Conducted at the Frequency of 1 . 0 Hz Since the hysteretic behaviour of the f i r s t f r i c t i o n joint was found to be insensitive to the sinusoidal frequency at which i t was being tested, the remaining f r i c t i o n joints were only tested under a sinusoidal frequency of 1.0 hz. The same range of torques from 27.1 to 203.4 N-m (20 to 150 ft-lb) was applied on the remaining f r i c t i o n joints to investigate each one's s l i p load to torque relationship. Table 5.1 summarizes the cyclic test results for a l l four f r i c t i o n devices. It l i s t s the s l i p loads attained by the f r i c t i o n joint of each device, corresponding to the loading frequency and applied torque at which the test was conducted. These s l i p loads are obtained by averaging the s l i p loads from the tension and compression portions of each test. The stable and non-deteriorating hysteretic behaviour exhibited in the c y c l i c tests on the f i r s t f r i c t i o n joint was found to be typical for a l l four f r i c t i o n joints under a l l the torque levels which were investigated. Based on Table 5.1, the torque to s l i p load relationship for the four f r i c t i o n devices was established. Figure 5.7 displays this relationship along with the linear regression f i t of the data set from each device. Information from this figure was used for torquing the f r i c t i o n damping devices to the desired s l i p loads during the shake table tests. Table 5.1: Sl i p Loads of the Friction Damping Devices as Determined by the Cyclic Tests TORQUE DEVICE A W-m (ft-lb) 0.5 Hz 1.0 Hz 2.0 Hz 3.5 Hz 5.0 Hz 27:1 (20) * 1.16(0.26) 40.7 (30) 54.2 (40) 4.58 (1.03) 67.8 (50) 6.50 (1.46) 6.45(1.45) 6.32 (1.42) 6.32 (1.24) 6.19(1.39) 81.4 (60) 7.90 (1.78) 94.9 (70) 108.5 (80) 11.35 (2.55) 122.0 (90) 13516 (100) 13.71 (3.08) 13.57 (3.05) 13.54(3.04) 13.78 (3.10) 13.27(2.98) 149.2 (110) 162.7 (120) 15.91 (3.58) 176.3 (130) 189.8 (140) 203.4 (150) 21.01 (4.72) 20.69 (4.65) 20.76 (4.67) 21.05 (4.73) TORQUE DEVICE B DEVICE C DEVICE D N-m (ft-lb) 1.0 Hz 1.0 Hz 1.0 Hz 27.1 (20) 1.38 (0.31) 1.47 (0.33) 0.53 (0.12) 40.7 (30) 1.54 (0.35) 1.82 (0.41) 54.2 (40) 2.27 (0.51) 2.18 (0.49) 67.8 (50) 3.34 (0.75) 2.91 (0.66) 4.00 (0.90) 81.4 (60) 4.00 (0.90) 4.27 (0.96) 94.9 (70) 5.38 (1.21) 6.01 (1.35) 108.5 (80) 5.45 (1.23) 6.68 (1.50) 122.0 (90) 6.34 (1.43) 7.79 (1.75) 8.46 (1.90) 135.6 (100) 7.12(1.60) 8.79 (1.98) 149.2 (110) 10.01 (2.25) 162.7 (120) 7.68 (1.73) 176.3 (130) 12.02 (2.70) 189.8 (140) 203.4 (150) 10.24 (2.30) 13.80 (3.10) 13.80 (3.10) * NOTE: Slip Loads in units of [kN(kips)] The differences i n s l i p load generated by the four devices under the same torque may be attributed to deviations i n tolerance between the torquing bolts. This deviation required each bolt to be tightened to a different torque in order to exert the same pressure on the f r i c t i o n j o i n t . Another factor influencing the s l i p load response may be differences i n the fraying surfaces of the steel at the contact interface between the two struts. E z b o CO z o UJ ZD o O l— 200 180 160 140 120 100 80 60 40 20 0 DEVICE B DEVICE C DEVICE A - 1 1 1 r i i i 4 8 12 16 20 SLIP LOAD OF FRICTION PADS (kN) Figure 5 .7 : S l i p Load Calibration Curves for Frict i o n Damping Devices 6 7 During a l l of the cyclic tests, rotation of the f r i c t i o n joint of each device was noticed at every reversal of load from tension to compression and vice versa.<-> With the struts at the f r i c t i o n joint overlapping instead of meeting on a single plane, the lines of action of the strut forces, became eccentric to each other. This eccentricity created a moment about the shear plane where slippage between the two struts took place, thereby causing the rotation in the f r i c t i o n joint. The amount of rotation was insignificant for the lower s l i p loads, but increased with i t s magnitude. This gain in joint rotation was expected as the greater axial forces created a proportionally higher moment about the joint. To eliminate the joint rotation, the device could be redesigned to abolish any load eccentricity. However, the joint rotation did not seem to have any negative effects on the results of the cyclic tests, therefore the devices were used for this preliminary investigation without any modifications. Figure 5 .8 illustrates how the load eccentricity caused joint rotation to occur. 68 E C C E N T R I C I T Y C O M P R E S S I O N T E N S I O N Figure 5.8 : Rotation of Friction Joint from Eccentric Strut Loads 9 6 9 5.5 Summary The following findings were conceive from the c y c l i c tests on the f r i c t i o n joint of each of the four f r i c t i o n damping devices. - The hysteresis 7 loops produced by the f r i c t i o n damping devices were generally rectangular with l i t t l e degradation. - The hysteresis loops remained stable and did not deteriorate over the entire 50 cycle duration of each test. The sinusoidal frequencies at which the devices were being tested did not affect the s t a b i l i t y of their hysteretic behaviour. The s l i p loads generated by each device varied linearly with the applied torque on i t s f r i c t i o n joint. Each device exhibited a somewhat different torque to s l i p load relationship. These findings demonstrated that the f r i c t i o n damping devices are capable of dissipating seismic energy input under the varying dynamic loading conditions which may be expected from an earthquake ground excitation. To maximize this effective energy dissipating a b i l i t y , the f r i c t i o n devices must be manufactured under s t r i c t quality control to ensure that they are fabricated to a low level of tolerance. By the same token, they must be securely installed onto the framing of the shearwall to allow for the proper transfer of loads. 70 6. INELASTIC TIME-HISTORY DYNAMIC ANALYSIS 6.1 Introduction Prior to conducting any f u l l scale tests with the f r i c t i o n damped^timber shearwalls, the optimal s l i p load for"calibrating the f r i c t i o n damping devices needed to be determined. This was done using FRICWALL (F i l i a t r a u l t and Dolan, 1989), a simulation model which predicted the seismic performance of timber shearwalls. Based on an inelastic time-history dynamic analysis, FRICWALL computes the response time-history of a shearwall under a simulated seismic event. The v a l i d i t y of this model for conventional timber shearwalls has been verified through comparison with earlier experimental shake table results 4 ( F i l i a t r a u l t , 1989). FRICWALL considers a shearwall as an equivalent non-linear single-degree-of-freedom system as shown in Figure 6.1. The general equation of motion for the system i s given as: mft(t) + cx(t) + F c(t) + F d(t) - -mxg(t) (6.1) The mass m represents the i n e r t i a l mass supported by the shearwall. The mass of the shearwall i t s e l f i s assumed to be negligible. The non-linear spring force F c(t) is"the lateral force carried by the shearwall. This force i s essentially bared by the shearwall connectors. The non-linear spring force F d(t) i s the lateral force carried by the f r i c t i o n damping devices. The magnitude of F d(t) can be varied by inputting different s l i p loads for the f r i c t i o n devices. At zero s l i p load, F d(t) = 0 , and the analysis would be for a conventional timber shearwall. The viscous damping coefficient c accounts for a l l the supplemental energy dissipation, mechanisms in the shearwall with the exception of the hysteretic energy dissipated by the connectors and the energy dissipated by the f r i c t i o n damping devices. The relative top of wall displacement, x(t), generalizes the system's degree of freedom; x(t) and x(t) are the velocity and acceleration of the top of the wall relative to the moving base; and x g(t) i s the ground acceleration. F(t) x(t) Hinges^ P 33 Hinge V • Rigid Hinge I Connectors, F.I -7vNAAA-L___. ^ Devices, Fd(t) ^ft ^ ) ^ 1 Figure 6.1: Numerical Modelling of a Friction Damped Timber Shearwall 72 The load-deformation or hysteretic behaviour of the non-linear spring representing the shearwall i s modelled by a combination of degrading exponential functions as illustrated in Figure 6.2. The virgin load-deformation path of the shearwall follows an exponential envelope function, as shown by the dashed line, to a maximum displacement A. This function i s identified by the parameters Kg, K, and P0, representing respectively, the i n i t i a l stiffness, asymptotic stiffness and the asymptotic load intercept of the wall. This function i s similar to the one used to model the sheathing-to-framing connectors in the SADT f i n i t e elements program discussed in Chapter 4. Because the energy dissipation in a conventional timber shearwall comes mainly from i t s connectors, i t s overall hysteretic behaviour resembles that of the connectors. For deflections greater than A, the shearwall i s assumed to degrade linearly at a rate S2 u n t i l failure. In a l l subsequent unloading and reloading, the load-deformation of the shearwall i s modelled using the pinched hysteresis loop. Linear parameters are assigned to this loop's load intercepts and associated stiffness S^ while the connectivity between these portions of the loop and the virgi n load-deformation path are defined by exponential equations (Dolan, 1989). In FRICWALL, the hysteretic behaviour of the f r i c t i o n damped timber shearwall i s derived from the superposition of the hysteretic behaviours of the conventional shearwall and the f r i c t i o n devices. Therefore the parameters defining the wall's hysteretic behaviour could be derived from the racking and 73 s t a t i c c y c l i c tests of the conventional timber shearwalls. Figure 6.2: Hysteretic Behaviour of Shearwall as Modelled by FRICWALL The hysteretic behaviour of the f r i c t i o n damping devices i s modelled as an equivalent elasto-plastic system, as depicted by Figure 6 . 3 . The variables Fg, Kd and x s represent respectively, the l a t e r a l load i n i t i a t i n g slippage of the devices, the equivalent e l a s t i c stiffness of the devices and the l a t e r a l wall displacement causing slippage. The equations defining the 74 Figure 6.3: Hysteretic Behaviour of Fri c t i o n Damping Device as Modelled by FRICWALL variables are given as ( F i l i a t r a u l t , 1989): F 4 L ^ P 8 H ( L V 2 + L h 2 ) 1 / 2 8 (6.2) 4EA(L vL h) 2 H 2(L V 2+L h 2) 3 / 2 (6.3) I * P8H(L V2+Lh2) 8 Kd EAL vL h (6.4) where Ps i s the s l i p load of a f r i c t i o n device, and 1^  are lengths of the respective vertical and horizontal legs of the devices, E i s the Young's modulus of steel (material of the fr i c t i o n devices) and A i s the cross-sectional area of a device's diagonal struts. These variables are derived by f i r s t principles, assuming that the framing members are pin ended and w i l l deform as a parallelogram under lateral load. A l l four f r i c t i o n devices are presumed to be of the same size and exhibit the same elasto-plastic behaviour. During the analysis, an energy balance calculation i s made at each time-step to monitor the accuracy of the calculations. The energy balance equates the external energy input to the shearwall to i t s internal work. The equation i s derived by pre-multiplying the terms in the general equation of motion, Equation 6.1, by the velocity x(t) and then integrating each term over time. Thus, Equation 6.1 takes on the form of: m (6.5) Now: i ( t ) _ d * j t ) _ a n d i ( t ) . d x ( t ) 76 Substituting these two terms into Equation 6.5 and re-evaluating the limits of integration leads to the following: Rewriting Equation 6.6, the f i n a l form of the energy balance equation i s obtained as: where T(t) i s the kinetic energy of the wall, D(t) i s the energy dissipated by viscous damping, W(t) i s the energy absorbed by the sheathing-to-framing connectors (the wall), F(t) i s the energy absorbed by the f r i c t i o n devices and I(t) i s the seismic energy input. At the completion of the analysis, output f i l e s containing the acceleration, displacement, force and energy time-histories of the shearwall are generated. This chapter discusses the modelling of the timber shearwall and presents some of the results from the FRICWALL analyses. The choice of an optimal s l i p load for the f r i c t i o n damping devices i s made and a detailed comparison of the seismic response of the two types of shearwalls i s presented. The comparisons between these numerical results and the experimental findings are examined in Chapter 7. (6.6) T(t) + D(t) + W(t) + F(t) - K t ) (6.7) 77 6.2 Numerical Analysis The analyses using the FRICWALL model were carried out for f r i c t i o n device s l i p loads ranging from 0 to 15.6 kN (0 to 3.5 kip), in increments of 2.2 kN (0.5 kip). The 0 kN (0 kip) s l i p load corresponded to a lower bound solution, representing shearwalls without the f r i c t i o n devices (ie. conventional timber shearwalls). The 15.6 kN (3.5 kip) s l i p load corresponded to an upper bound solution which i s the design s l i p load determined from the SADT analysis discussed in Chapter 4. The s l i p load parameter was varied in order to investigate i t s influence on the in-plane displacements of the shearwall. This information was necessary for finding an optimal and practical s l i p load for calibrating these devices for the f u l l scale shake table tests. Since the amount of viscous damping in a timber shearwall could not be determined theoretically, the numerical analyses were carried out using viscous damping coefficients in the range of 0.002 to 0.03 kN-s/mm (0.01 to 0.15 kip-s/in). This range corresponds to viscous damping ratios between 0 to 12% of c r i t i c a l damping for conventional timber shearwalls which have a natural frequency of vibration at approximately 3 hz. Past experiments have shown that the viscous damping ratio for these shearwalls f e l l within this range. Ground acceleration records from the 1940 El Centro (N-S) and the 1977 Romania (Bucharest, N-S) earthquakes were used in the simulations. The absolute acceleration response spectra and the f i r s t 16 seconds of the acceleration time-histories of these earthquake ground excitations are shown in Figures 6.4 and 6.5. 78 The E l Centro was simulated at the f u l l scale peak ground acceleration (P.G.A.) of 0.35 g, and at 0.6 g, 171 % of f u l l scale. The Romania earthquake was simulated at the f u l l scale peak ground acceleration of 0.2 g. These levels of ground acceleration are the same, ones used i n the f u l l scale dynamic earthquake simulations. The peak ground accelerations of 0.6 g and 0.2 g for the E l Centro and the Romania earthquakes are the maximum peak ground accelerations achievable with the shake table at UBC. PERIOD (s) Figure 6 . 4 : Absolute Acceleration Response Spectra for Seismic Events Used in Simulations 4000 -4000 1940 EL CENTRO (N-S) PEAK GROUND ACCEL - 0.35 g (3417 mm/s~2) 6 8 10 12 14 16 TIME (s) CM «5 4000 3000 J , 2000 -2 £ A1000 -< UJ UJ 8 -1000 -< 5 -2000 § -3000 O -4000 1977 ROMANIA (Bucharest, N-S) PEAK GROUND ACCEL = 0.20 g (1980 mm/s~2) —i 1 i 1 1 1 4 6 8 10 12 14 16 TIME (s) Figure 6 . 5 : Ground Acceleration Time-Histories 80 I n the f u l l s c a l e dynamic s i m u l a t i o n t e s t s , the c e n t r e of g r a v i t y of the i n e r t i a l mass supported by the hinged s t e e l frame (see F i g u r e 7.1) i s l o c a t e d 457 mm (18 in) above the top of the s h e a r w a l l . I n the FRICWALL s i m u l a t i o n , t h i s - mass i s modelled as being l o c a t e d d i r e c t l y above the s h e a r w a l l . To accounted f o r t h i s h e i g h t d i f f e r e n c e , an e q u i v a l e n t mass which produces the same i n e r t i a l f o r c e as the mass i n the experimental s e t up was used i n the model. Figu r e 6.6 i l l u s t r a t e s a f r e e body diagram of the sh e a r w a l l t e s t i n g frame i n the d e f l e c t e d shape. The masses f o r the frame and the sh e a r w a l l are.assumed n e g l i g i b l e compared t o the i n e r t i a l mass, thus are l e f t out of the f r e e body diagram. From summation of the h o r i z o n t a l and v e r t i c a l f o r c e s on the f r e e body of the t e s t i n g frame, the f o l l o w i n g e q u i l i b r i u m equations are obtained: I F x - 0; A x + D x - mx + Rx + R2 (6.8) £ F y - 0; Ay + Dy - Wm (6.9) By summation of the moments about p o i n t s B and C of frame members A-B and C-D the f o l l o w i n g equations are produced: E M B - 0; AyX + R^h-hJcosB - A xhcos6 (6.10) E Mc - 0; D yx + R 2(h-h w)cos8 - D xhcos6 (6.11) Figure 6.6: Free Body Diagram of Sh e a r w a l l T e s t i n g System 82 Adding Equations 6.10 and 6.11, s u b s t i t u t i n g i n Equations 6.8 and 6.9, and c o l l e c t i n g terms, the following i s obtained: W^ x - (Ri+R-^h^cose - mxhcosS (6.12) The resistances R l and R2 are provided by the shearwall, thus: R x + R 2 - k w a l lA, where A - -^x (6.13) S u b s t i t u t i n g Equation 6.13 into Equation 6.12 and rearranging the terms, the following equilibrium equation i s obtained: (-r|i-) mxcos6 + k^nxcose - W m(-—)x - 0 (6.14) For t h i s experimental setup, h = 2896 mm (114 i n ) , h H = 2438 mm (96 i n ) , k w a l l = 0.84 kN/mm (4.8 kip/in) , Wm = 44.6 kN (10.0 k i p ) . At small l a t e r a l d e f l e c t i o n s , cos9 - 1, thus, Equation 6.14 takes on the form of: 1.41mx + 0.82x - 0 (6.15) Based on Equation 6.15, the FRICWALL analysis was ran using a seismic mass of 0.0077 kN-s2/mm (0.044 kip-s 2/i.n) which was 1.41 times the mass used i n the experiment. This scaled up seismic mass enabled the model to simulate the same i n e r t i a l force which would be produced i n the experiment. 83 Other main inputs to the program included the height of the shearwall, the parameters d e f i n i n g the h y s t e r e t i c behaviour of the wall and the device, and the acceptable tolerance i n the energy balance. Table 6;1 l i s t s these inputs. Derivation of •the 'WALL HYSTERETIC PARAMETERS' i n Table 6.1 are discussed i n Sections 7.5.1 and 7.5.2 of Chapter 7. Table 6.1: Summary of FRICWALL Inputs INPUT PARAMETERS VALUE SEISMIC MASS 0.0077 kN-s~ 2/mm (0.044 kip-s~2/in) VISCOUS DAMPING COEFF 0.002 - 0.03 kN-s/mm (0.01 - 0.15 kip-s/in) HEIGHT OF WALL 2438.4 mm (96 in) WALL HYSTERETIC PARAMETERS -Po 30.35 kN (6.82 kip) Pi 4.94 kN (1.11 kip) A 72.68 mm (2.86 in) K0 2.13 kN/mm (12.13 kip/in) Ki 0.13 kN/mm (0.72 kip/in) Si 0.53 kN/mm (3.0 kip/in) s 2 -1.00 kN/mm (-5.69 kip/in) DEVICE HYSTERETIC PARAMETERS -Ps 0-15.6 kN (0 - 3.5 kip) L v 356 mm (14.0 in) L h 356 mm (14.0 in) E 200000 MPa (30000 ksi) A 484 mm ~ 2 (0.75 in ~ 2) ENERGY BALANCE TOLERANCE 10% * Mass has been corrected to account for height difference between experimental set up and numerical model + Corresponds to viscous damping between 1 - 1 2 % of critical damping 84 6 .3 Results and Discussion This section discusses the choosing of an optimal s l i p load for c a l i b r a t i n g the f r i c t i o n devices and presents a d e t a i l e d comparison of the seismic performance between a conventional and a f r i c t i o n damped timber shearwall as computed by FRICWALL. 6.3.1 Optimal Slip Load Study To i l l u s t r a t e the s e n s i t i v i t y of the seismic response of a f r i c t i o n damped timber shearwall to v a r i a t i o n s i n the s l i p load, the peak r e l a t i v e top-of-wall displacements, along with the marginal displacement decreases at each increment of s l i p load, are p l o t t e d i n Figures 6.7, 6.8 and 6.9. Figures 6.7 and 6.8 show the r e s u l t s from the E l Centro earthquake simulation at the 0.35 g and 0.6 g peak ground acceleration l e v e l s . Figure 6.9 shows the r e s u l t s from the Romania earthquake simulation at the 0.2 g peak ground acceleration l e v e l . The peak r e l a t i v e top-of-wall displacement to s l i p load r e l a t i o n s h i p s shown i n the top graph of each figure are f o r the viscous damping r a t i o s shown i n t h e i r legends. A l l the graphs show a trend of progressively decreasing peak displacements with incrementing s l i p loads. This trend s i g n i f i e d that a higher s l i p load allowed the f r i c t i o n devices to d i s s i p a t e a greater proportion of seismic energy input a t t r a c t e d by a shearwall. Consequently, the share of seismic energy a shearwall must dissipate.-would be reduced, r e s u l t i n g i n a lower peak in-plane wall d e f l e c t i o n . The net r e s u l t would be an a l l e v i a t i o n of the damage sustained by a wall during a seismic event. 85 1940 EL CENTRO (N-S) PEAK GROUND ACCELERATION = 0.35 g (3417 mm/s ~ 2) VISCOUS DAMPING COEFF, c [kN-8/mm(klp-s/ln)] • ••1 0.002 (0.01) + 0.007 (0.04) o 0.012 (0.07) I S M 0.019 (0.11) X 1 1 0.026(0.15) SLIP LOAD = 6.7 kN (1.5 kip) 8 10 SUP LOAD (kN) 12 14 16 Figure 6.7: FRICWALL Results for Peak Top-of-Wall Displacement - E l Centro Earthquake @ 0.35 g P.G.A. -8 6 1940 EL CENTRO (N-S) PEAK GROUND ACCELERATION = 0.6 g (5886 mm/s~2) VISCOUS DAMPING COEFF, C (kN-8/mm(kip-a/1n)] • H i 0.002 (0.01) + FZ23 0.005 (0.03) o 0.009 (0.05) A 1 / 1 0.012 (0.07) X 0.019(0.11) V UZ3 0.026 (0.15) 6 8 10 12 SUP LOAD (kN) Q . 22 Q 55 Q . a. O CC a a CC •< 2 14 - i 12 -10 -8 -6 -4 2 H 0 < < < < < < < < < < < < < < •< >< 2.2 4.5 6.7 8.9 11.1 SUP LOAD (kN) 13.4 15.6 Figure 6 . 8 : FRICWALL Results for Peak Top-of-Wall Displacement - E l Centro Earthquake @ 0.6 g P.G.A. -87 1977 ROMANIA (BUCHAREST, N-S) PEAK GROUND ACCELERATION » 0.2 g (1980 mm/sA2) VISCOUS DAMPING COEFF, c [kN-a/mm(klp-8/ln)] • 0.002 (0.01) + 0.007 (0.04) © 0.012 (0.07) rssi 0.019(0.11) X i i 0.026(0.15) 8 10 SUP LOAD (kN) 20 -j 18 -0_ W Q if 0. z c_ o rr Q - i < z o < 2 16 -14 -12 10 8 6 4 2 x xx-x:? x » x>> xx-j I 6.7 8.9 11.1 SUP LOAD (kN) 13.4 15.6 Figure 6.9: FRICWALL Results for Peak Top-of-Wall Displacement - Romania Earthquake § 0.2 g P.G.A. -88 The inverse relationship between the magnitude of deformation in the wall and the s l i p load on i t s f r i c t i o n devices shown in the graphs suggests that the higher the s l i p load,, the better the seismic performance by the wall. Thus, the most suitable s l i p load would be the upper bound load of 15.6 kN (3.5 kip) which produced the greatest reduction in relative wall displacement, hence the least damage. However, the trend of the peak displacement to s l i p load relationship flattens out as higher s l i p loads are approached. This decreasing slope indicated that the margin of benefit from displacement reduction degrades as s l i p loads become too large. Thus, in order to obtain a more suitable s l i p load, the relative drop in peak displacement between each increment of s l i p load i s examined. The results of this investigation for s l i p load increments of 2.25 kN (0.5 kip) i s illustrated by the marginal drop in peak displacement versus s l i p load charts at the bottom of Figures 6.7, 6.8 and 6.9. The information shown in the charts are summarized in Table 6.2. Table 6.2 l i s t s the average marginal displacement drop in percentage, for each increment of s l i p load and each set of acceleration record simulated. The average values were calculated using a l l the viscous damping ratios being investigated. Using the three average values at each s l i p load, an overall average was determined. No weighting factor was used in the averaging process. Table 6.2 shows that an optimal s l i p load of 6.7 kN (1.5 kip) provided a 11% drop in the overall average marginal drop in lateral wall deflection. This was the highest marginal drop in wall deflection for the range 89 of s l i p loads considered in the investigation. Table 6.2: Summary of Marginal Drop in Peak Relative Top-of-Wall Displacement SLIP LOAD [kN(kip)] ACCELERATION 2.2 4.5 6.7 8.9 11.1 13.4 15.6 RECORD (0.5) (1.0) (1.5) (2.0) (2.5) (3.0) (3.5) EL CENTRO (0.35 g P.G A) * 8.3 7.7 7.0 7.9 5.1 4.7 4.4 EL CENTRO (0.6 g P.GA.) 12.3 11.6 8.8 6.4 7.8 6.8 4.0 ROMANIA (0.2 g P.GA.) 4.8 6.1 17.7 11.1 9.6 7.1 4.0 OVERALL AVERAGE 8.5 8.5 11.1 8.5 7.5 6.2 4.1 * NOTE: Displacement drops listed as a percentage In addition to a better marginal reduction in peak wall displacement, a lower s l i p load would also result in lower stress concentrations in the sheathing-to-framing connectors located in the v i c i n i t y of where the struts of the f r i c t i o n devices are mounted to the wall framing. These connectors were the main mechanisms for lateral load transfer from the sheathing panels to the f r i c t i o n devices. With a higher s l i p load, a greater shear force would have to be carried by these connectors to i n i t i a t e slippage of the devices. The higher forces would cause more severe crushing of the wood around the connectors. With this irreversible damage, the slack developed around the connectors after just a few load cycles would render the devices ineffective as they would not get the proper transfer of forces. This build-up of slack was inevitable, regardless of the 90 magnitude of s l i p load, because the wood damage i s an inherent part of the behaviour of a shearwall under lateral load. However, the integrity of the wall and the effectiveness of the f r i c t i o n devices could be maintained much longer by reducing the rate of slack development. Based on the marginal reduction in peak wall deflection and the physical interaction between the connectors and the wood elements of the wall, a s l i p load of 6.7 kN (1.5 kip) was chosen as the optimal s l i p load for the f r i c t i o n devices. 6.3.2 Detailed Analysis This section provides a detailed comparison of the FRICWALL results for a conventional and a f r i c t i o n damped timber shearwall with devices set at the optimal s l i p load of 6.7 kN (1.5 kip). The results shown are for the two E l Centro and the one Romania earthquake simulations using a viscous damping coefficient of 0.012 kN-s/mm (0.07 kip-s/in), which corresponds to 5% of c r i t i c a l damping. The relative displacement time-histories, hysteretic behaviour and energy time-histories of both types of shearwalls are presented. Figure 6.10 shows the relative top-of-wall displacement time-histories from the three simulations. The dotted lines represent the deflection time-histories for a conventional wall while the solid lines represent those for a f r i c t i o n damped wall. Drops of 33 %, 19 % and 29 % in peak wall deflection are respectively achieved for the f r i c t i o n damped wall at peak ground accelerations of 0.35 g and 0.6 g under the El Centro 91 «• 50 E g -i CL CO O d I o QL P LU 3  CC J, _i o. co Q d I o e LU 2: S LU CC 0. CO Q d o CL. LU 2: LU CC 40 -30 20 10 0 -10 -20 --30 --40 --50 EL CENTRO @ 0.35 g P.G.A. 120 80 40 -0 -40 -80 H -120 £ 2 0 10 5 0 -5 -10 --15 --20 ; 28.1 mm (FRICTION DAMPED WALL) 42.1 mm (CONVENTIONAL WALL) 2 3 4 TIME (s) EL CENTRO @ 0.6 g P.GA 89.4 mm (FRICTION DAMPED WALL) 110.2 mm (CONVENTIONAL WALL) 3 4 TIME (s) i\ 16. h ROMANIA @ 0.2 g P.GA. 0 mm (CONVENTIONAL WALL) 11.4 mm (FRICTION DAMPED WALL) 6 8 10 TIME (s) 12 14 16 Figure 6.10: Relative Top-of-Wall Displacement Time-Histories Computed by FRICWALL simulations, and 0.2 g under the Romania simulation. Aside from lower peak wall deflections, a f r i c t i o n damped wall i s shown to develop lower deflections than a conventional wall throughout the time history of the simulations. For the Romania earthquake simulation, negligible wall deflections are shown at the second half of the time-history. The hysteretic behaviour exhibited by the two types of walls are shown in Figures 6.11, 6.12 and 6.13. A f r i c t i o n damped wall i s shown to generate hysteresis loops which are less pinched than those of a conventional wall. It i s able to dissipate the required amount of.seismic energy input at lower restoring forces and deflections. Its better hysteretic behaviour stems from the presence of the f r i c t i o n damping device which not only dissipate a large portion of the seismic energy input, but at the same time, also limit the quantity of this energy input to the shearwall. Referring back to Equation 6.7, each term of the energy balance equation i s calculated by integrating through the relative displacement of the wall. With the addition of a f r i c t i o n damping term to the l e f t hand side of the equation to contribute to the energy dissipation of the wall system, an energy balance can be achieved at lower relative wall displacements. A lower displacement results in a reduction in the magnitude of a l l the terms in the energy balance equation, including the seismic energy input term. The devices can be looked at as structural dampers or safety valves which reduce the deflection of a structure and limit the seismic energy input i t attracts. 93 Figure 6.11: Hysteresis Loops from FRICWALL Simulation - E l Centro Earthquake @ 0.35 g P.G.A. -EL CENTRO @ 0.6 g P.G.A. m o cr 2 (5 Z CC g w 111 CC i O i 60 30 -30 H -40 -60 CONVENTIONAL WALL -120 -80 -40 0 40 80 TOP-OF-WALL LATERAL DISPLACEMENT (mm) 120 z UJ o cr £ z CC w UJ CC d i uL O QL p 10 0 -40 --50 --60 FRICTION DAMPED WALL -120 -80 -40 0 40 80 TOP-OF-WALL LATERAL DISPLACEMENT (mm) 120 Figure 6.12: H y s t e r e s i s Loops from FRICWALL S i m u l a t i o n - E l Centro Earthquake § 0.6 g P.G.A. -9 5 Figure 6.13: Hysteresis Loops from FRICWALL Simulation - Romania Earthquake § 0.2 g P.G.A. -To show how the f r i c t i o n damping devices affect the energy balance of a shearwall during the duration of seismic loading, the time-histories from each term of the energy balance equation are plotted in Figures 6.14, 6.15 and 6.16 for both shearwall systems. The time-history plots of the f r i c t i o n damped wall reflect a substantial reduction in each of the terms in the energy balance equation as compared to the conventional wall. For example, seismic energy input reductions of 23 %, 21 % and 39 % were respectively calculated for this wall at peak ground accelerations of,0.35 g and 0.6 g at the end of the El Centro simulations and 0.2 g at the end of the Romania simulation. The amount of energy dissipated by the f r i c t i o n devices under these three simulations was quite significant, at 37 %, 25 % and 54 % of the energy input for the respective acceleration records. This detailed comparison of the FRICWALL results between a conventional and a f r i c t i o n damped timber shearwall shows that the f r i c t i o n damping devices could significantly improve the seismic performance of a timber shearwall. The devices not only contribute to the energy dissipating a b i l i t y of a shearwall, but also limit the quantity of seismic energy input i t attracts. Ultimately, the decline in energy requirements of a timber shearwall would enable a timber framed building to sustain less structural or non-structural damages during an earthquake. 97 EL CENTRO @ 0.35 g P.G.A. 3.2 2.8 2.4 t * i z i « a 5 1.2 z LU 0.8 0.4 0 (x10~3) 3.2 2.8 2.4 t * Z o 5 1.2 z LU 0.8 0.4 0 KINETIC ENERGY WALL ENERGY VISCOUS DAMPING DEVICE ENERGY INPUT ENERGY 3 TIME (s) FRICTION DAMPED WALL CONVENTIONAL WALL / A . - - — t •A A / r-J / / v — • • Figure 6.14: Energy Time-Histories from FRICWALL Simulation - E l Centro Earthquake @ 0.35 g P.G.A. -9 8 &(10A3) 13 EL CENTRO (c 5 0.6 g P.G.A. KINETIC ENERGY VISCOUS DAMPING WALL ENERGY DEVICE ENERGY INPUT ENERGY E E i Z CC UJ z 12 H 11 10 9 -8 -7 -6 -5 4 -3 -2 -1 -0 (x10~3) 13 -12 -11 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1 -o -• E E I Z oc III z UJ CONVENTIONAL WALL .r<y j ^ 1 : / \ / •\ : * v. 2 3 TIME (s) FRICTION DAMPED WALL \ / » 2 3 TIME (s) T " 4 Figure 6.15: Energy Time-Histories from FRICWALL Simulation - E l Centro Earthquake @ 0.6 g P.G.A. -99 ROMANIA @ 0.2 g P.G.A. (x10~3) 0.7 0.6 -0.5 H I 0.4 H & 0.3 -UJ * 0.2 H 0.1 H 0 (x10~3) 0.7 0.6 -0.5 | 0.4 I 0.3 UJ * 0.2 0.1 -KINETIC ENERGY WALL ENERGY VISCOUS DAMPING DEVICE ENERGY INPUT ENERGY CONVENTIONAL WALL i V . . \ i ' \ <\-v r, 8 10 TIME (s) 12 14 FRICTION DAMPED WALL T r-2 8 10 TIME (s) 16 Figure 6.16: Energy Time-Histories from FRICWALL Simulation - Romania Earthquake @ 0.2 g P.G.A. -100 6.4 Summary The FRICWALL analyses have been performed to determine an optimal s l i p load for calibrating the f r i c t i o n damping devices r for the f u l l scale shake table tests and to predict the seismic performance of timber shearwalls. The inverse relationship shown between the magnitude of deformation in a shearwall and the s l i p load on i t s f r i c t i o n devices suggested that improvements in the seismic performance of a timber shearwall could be maximized by setting i t s f r i c t i o n devices to the highest allowable s l i p load. But an examination of the marginal changes in wall deflection at each increment of s l i p load showed that the marginal benefit of reduced peak wall deflection diminishes as high s l i p loads are approached. Based on this finding and a consideration of the physical interaction between the connectors and the wood members of a shearwall, an optimal s l i p load of 6.7 kN (1.5 kip) was chosen for calibrating the f r i c t i o n devices for the f u l l scale shake table tests. Detailed analyses of the numerical results showed that the f r i c t i o n devices relieved the energy demand on a shearwall by limiting the seismic energy i t attracts and contributing to the dissipation of this energy. The inclusion of the devices allows a f r i c t i o n damped timber shearwall to exhibit superior seismic performance over i t s conventional counterpart. These results . suggest that a f r i c t i o n damped timber shearwall could serve as a very practical and effective lateral load resisting system. 101 7. FULL SIZE SHEARWALL TESTS 7.1 Introduction Numerical results from the SADT and the FRICWALL analyses have shown that a shearwall retrofitted with f r i c t i o n damping devices performed significantly better under lateral loads than a conventional timber shearwall. To verify these results, tests of f u l l scale conventional and f r i c t i o n damped timber shearwalls were conducted. These tests involved loading timber shearwalls under racking, static cyclic and dynamic earthquake simulation loads. The racking tests examined the load-deformation behaviour of shearwalls under unidirectional static loading. The s t a t i c cyclic tests investigated their hysteretic behaviour under slow cyclic loading. The dynamic earthquake simulations explored their seismic response under severe earthquake conditions. This chapter discusses the f u l l scale shearwall tests and the analysis of test results. -The comparisons between the experimental and numerical results are also presented in this chapter. 7.2 Test Equipment and Instrumentation The testing of the f u l l scale-timber shearwalls were carried out in the Earthquake Engineering and Structural Dynamics Research Laboratory at the University of British Columbia. The testing equipment consisted of a shake table, a hydraulic pump, a hinged steel frame, a data acquisition and control system, and 102 an AST 80286 micro-computer. The shake table consisted of a 3 x 3 m (10 x 10 ft) welded aluminum platform mounted on four support posts with swivel end bearings. A 155 kN (34.8 kip) capacity hydraulic actuator mounted at one end of the platform generated the uniaxial horizontal loading force to the specimen being tested. Static or dynamic loads with a limiting frequency of 30 hz may be reproduced. Hydraulic pressure to the actuator was supplied by a hydraulic pump. The horizontal motion of the actuator could be displacement, velocity or acceleration controlled. This motion was monitored by a LVDT mounted inside the actuator and an accelerometer mounted directly below the table. The displacement limit of the actuator (stroke) i s 150 mm (6.0 in) (75 mm (3.0 in) peak-to-peak), i t s peak velocity i s 130 cm/sec (5.1 in/sec) and i t s peak acceleration i s 2.5 g. Vertical motion of the table was restrained by the support posts while yawing or sideway movement not along the axis of loading was eliminated by three hydrostatic bearings at the two,.sides of the table. Bolted onto the shake table platform was a hinged steel frame, inside which the shearwalls were mounted. The frame was originally constructed for a previous shearwall experiment (Dolan, 1989). The frame was designed with hinges_such that i t would not interfere with the lateral deformation of the shearwall or contribute to i t s lateral strength or stiffness. Attached to the top of the frame were three concrete blocks with a total mass of 4,545 kg (311.3 slugs). The blocks served as 103 i n e r t i a l mass to the shearwalls during the dynamic earthquake simulation tests. The mass represented the i n e r t i a l load contribution from the upper two stories of a three-storey, North American-style apartment building based on the tributary area of the shearwall. Since these masses were supported by the steel frame, the shearwalls being tested were considered as partition walls, bearing no overhead dead loads. Mounted to one side of the steel frame, at the same height as the top of the shearwall, was a cross-beam which extended to a steel reaction wall, erected outside the shake table. The reaction wall restrained the top of the shearwall from any late r a l motion during the racking and static c y c l i c tests. A 90 kN (20 kip) capacity, Baldwin Lima Hamilton SR4 load c e l l attached to this cross-beam monitored the lateral resistance of the shearwall during these tests. Figure 7.1 shows the general shake table setup with the hinged steel frame, i n e r t i a l mass, timber shearwall and steel reaction wall. Via a 32 channel data acquisition and control system, an AST . 80286 micro-computer regulated the movement of the actuator to reproduce the desired loading pattern, and stored data recorded during the experiments. Figure 7.2 shows the general setup of the controller and micro-computer. 104 F i g u r e 7.1: General Setup of Shake Table f o r F u l l Scale T e s t i n g 105 F i g u r e 7 . 2 : Setup of Data A c q u i s i t i o n & C o n t r o l System and AST 3 0 2 8 6 Micro-Computer f o r F u l l S c a l e T e s t i n g 106 7.3 Specimen Setup Two hollow structural steel beams f a c i l i t a t e d the mounting of each shearwall to the hinged steel frame. The beams were each bolted to the shearwall's header and sole plates with four bolts for the transfer of shear forces into the specimen. The top beam was connected to the steel frame with brackets at i t s two ends. The bottom member was bolted along i t s length to the base of the steel frame. At the two ends of the shearwall, an additional steel plate anchor reinforced the connection between i t and the bottom steel beam. This reinforcement i s illustrated in Figure 7.3. It consisted of a 6 x 75 x 200 mm (0.25 x 3 x 7.9 in) mild steel plate welded to a 12 mm (0.5 in) mild steel rod. The steel plate was fastened to the double end studs using 25 4 x 76 mm (No. 8x3.0 in) drywall screws while the steel rod was bolted to the inside of the bottom beam. The anchors were designed in an earlier shearwall experiment (Dolan, 1989) to prevent vertical separation between the end studs and sole plate of the shearwall. The separation was ini t i a t e d by overturning moments in the shearwall which caused a build-up of high tensile forces in the double end studs. Both the conventional and the f r i c t i o n damped timber shearwalls were mounted in the same manner for a l l of the shake table tests. 107 F i g u r e 7.3: Double End Stud Anchor P l a t e D e t a i l 108 7.4 Test Procedures In total, twenty-seven f u l l scale shearwall specimens were tested. Seven were tested in racking, fiv e under static c y c l i c and fifteen under dynamic earthquake loads. This section describes the procedures for these tests. 7.4.1 Racking Tests The racking tests were performed by restraining the top of the shearwall with the cross-beam to the reaction wall, while unidirectionally displacing i t s base (ie. shake table) to force the shearwall to deform in a racking configuration. The displacement rate of 5.1 mm/min (0.2 in/min), as recommended by ASTM E564-76 Standard (1986) for testing building panels in racking, was used. In order to examine the load-deformation behaviour of the walls before and after fai l u r e , the base of each shearwall was displaced a total of 127.0 mm (5.0 in). For this displacement, each test took 25 minutes. The loading ramp used in the racking tests i s shown in Figure 7.4. DISPLACEMENT (mm) L u 0 TIME (min) 25 -Figure 7.4: Ramp Load for Unidirectional Racking Tests 109 During a l l of the racking tests, the l a t e r a l resistance of the shearwall was measured by the load c e l l in the cross-beam while the displacement at the base of the wall was monitored by the LVDT mounted inside the hydraulic actuator. For the f r i c t i o n damped timber shearwalls, the axial strain and the slippage or relative axial displacement between the struts for a l l four f r i c t i o n devices were also monitored. The axial strains were measured by a pair of Micro Measurements CEA-06-250UW-350 strain gauges, mounted back to back on the 51 mm (2 in) faces of one diagonal strut from each f r i c t i o n device. The slippage between each pair of diagonal struts was monitored by a Trans-tek LVDT, mounted directly on the struts. The sampling rate for a l l the instruments used in the racking tests was at 1 sample per second. The measuring instruments which monitored each f r i c t i o n damping device i s illustrated in Figure 7.5. 110 F i g u r e 7 . 5 : Measuring Instruments f o r the F r i c t i o n Damping Device I l l 7.4.2 Static Cyclic Test The static c y c l i c tests were performed by restraining the top of the shearwall with the cross-beam to the reaction wall,' while displacing i t s base with a sinusoidal displacement controlled wave to force the shearwall to deform in a racking configuration. The amplitude of the displacement wave initiated at +/" 12.7 mm (0.5 in), then increased to +/- 25.4 mm (1.0 in), +/- 50.8 mm (2.0 in), and concluded at +/- 76.2 mm (3.0 in). For displacements up to +/- 50.8 mm (2.0"in), the conventional shearwalls were displaced for two cycles at each amplitude while the f r i c t i o n damped shearwalls were displaced for three cycles at each amplitude. For the +/- 76.2 mm (3.0 in) displacement, both types of shearwalls were only displaced for two sinusoidal cycles. The f r i c t i o n damped timber shearwalls were subjected to the extra cycle at each of the displacement amplitudes in order to examine i f the addition of the f r i c t i o n devices affected the degradation of the wall's hysteretic behaviour. Similar to the racking tests, a table displacement rate of 5.1 mm/min (0.2 in/min) was used. At this rate, the testing of each conventional shearwall lasted 260 minutes while the testing of each f r i c t i o n damped shearwall lasted 330 minutes. Figure 7.6 shows the sinusoidal loading waves used in the static c y c l i c tests of both types of shearwalls. The measurements recorded during these tests were the same as the ones surveyed in the racking tests. A sampling rate of , 1 sample per second was also used for the measuring instruments in these tests. 112 -101.6 — i 1 1 1 1 1 1 1 1 1 1 1 1 1 1 r -0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 TIME (min) i i i i i i i i i i i i i i 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 TIME (min) Figure 7.6: Sinusoidal Loading Waves for the Static Cyclic Tests 7.4.3 Dynamic Earthquake Simulation The dynamic earthquake simulations were performed with the cross-beam restraining the top= of the shearwall detached while the base was displaced using a ground acceleration time-history which has been recorded from a seismic event. Two different 113 ground acceleration records were employed in the simulations. They were from the 1940 El Centro (N-S) and the 1977 Romania (Bucharest, N-S) earthquakes. The f u l l duration of 54 seconds for the El Centro earthquake and 16 seconds for the Romania earthquake was simulated in these tests. These ground acceleration records were chosen because of the differences in the frequency distribution of their seismic energy contents. The El Centro earthquake has i t s energy content distributed mainly over the low to mid period range of the acceleration response spectrum while the Romania earthquake has i t s energy content centred about the high period end of the spectrum. The absolute acceleration response spectra of these two seismic events and the f i r s t 16 seconds of their acceleration time-histories are illu s t r a t e d in Figures 6.4 and 6.5 of Chapter 6. Measurements recorded during these tests were the absolute accelerations and displacements at the base and at the top of the shearwall, and the axial strains and slippages in the diagonal struts of each f r i c t i o n device. The acceleration and displacement at the base of the shearwall were respectively measured by a Kistler 50 g accelerometer and the. LVDT mounted inside the hydraulic actuator. The acceleration and displacement at the top of the wall were respectively measured by a Statham 2.5 g accelerometer and a Celcsco,.PT101 string displacement transducer. Axial strains and slippages in the ^struts of the f r i c t i o n devices were monitored using the same instruments as in the racking tests. The sampling interval for a l l of the measuring instruments used in these tests was at 0.01 114 second (a rate of 100 samples per second). 7.5 Results and Discussion This section presents the experimental results from the shake table tests of the timber shearwalls. These results are analyzed to evaluate the difference in performance between the conventional and the f r i c t i o n damped timber shearwalls. The comparisons between the experimental and numerical results are also presented in this section. 7.5.1 Racking Tests In total, seven f u l l scale racking tests were performed. Three tests were conducted on conventional timber shearwalls and the other four on f r i c t i o n damped timber shearwalls. Of the four f r i c t i o n damped walls, three were tested with their f r i c t i o n devices calibrated at the optimal s l i p load of 6.7 kN (1.5 kip), while the forth was tested at a s l i p load of 13.4 kN (3.0 kip). The last test was conducted at the higher s l i p load to examine i t s impact on the behaviour of the wall. These tests were conducted to compare the racking behaviour of the conventional and the f r i c t i o n damped timber shearwalls. The results from the conventional timber shearwall tests were also needed to derive the parameters which defined a shearwall's virgi n load-deformation path for i t s hysteretic modelling in the FRICWALL program discussed in Chapter 6. Results from a l l the tests are also used to verify the SADT findings. The load-deformation behaviours of both types of shearwalls 115 are shown in Figure 7.7. The solid lines, Cl, C2 and C3, show the behaviour of the conventional walls, with peak loads ranging from 34.3 kN (7.7 kip) to 41.4 kN (9.3 kip) and averaging at 38.9 kN (8.7 kip) for the three specimens. The average of the peak loads i s in agreement with the numerical result of 42.3 kN (9.5 kip) from SADT which represents the average load capacity of a timber shearwall. The dashed lines, FI, F2, F3 and F4, show the behaviour of the f r i c t i o n damped walls, with peak loads ranging from 43.0 kN (9.6 kip) to 57.1 kN (12.8 kip) and averaging at 48.1 kN (10.8 kip) for the four specimens. The average of the peak loads i s slightly below the numerical result of 53.4 kN (12.0 kip) from SADT for a wall with non-slipping devices. The peak racking load sustained by each shearwall along with their corresponding displacement are summarized in Table 7.1. The energy dissipated by each shearwall up to failure, where failure i s defined as the point of maximum racking resistance, i s also l i s t e d in Table 7.1. 116 0 20 40 60 80 100 120 RELATIVE TOP OF WALL DISPLACEMENT (mm) Figure 7.7: Racking Load-Deformation Behaviour of Timber Shearwalls Table 7.1: Racking Behaviour of Timber Shearwalls _ „ . . . . , . , CONVENTIONAL TIMBER SHEARWALL SPECIMEN WALL# PEAK LOAD [kN(kip)] PEAK LOAD DISPL [mm(in)] ENERGY [kN-mm(kip-in)] C1 C2 C3 40.9 (9.2) 34.3 7.7) 41.4 (9.3) 84.3 (3.3) 65.1 (2.6) 68.7 (2.7) 2530.8 (22.4) 1651.5 (14.6) 2096.7 (18.6) AVERAGE 41.4 (9.3) 68.7 (2.7) 2096.7 (18.6) SADT 42.3 (9.5) 87.9 (3.5) 2598.1 (23.0) FRICTION DAMPED TIMBER SHEARWALL SPECIMEN WALL# SUP LOAD [KN(kip)] PEAK LOAD [kN(kip)] PEAK LOAD DISPL [mm(in)] ENERGY [kN-mrh (kip-in)] " F1 , F2 : F3 F4 6.7 (1.5) 6.7 (1.5) 6.7 (1.5) 13.4 (3.0) 57.1 (12.8) 47.6 (10.7) 43.0 (9.6 45.0 (10.1) 79.5 (3.1) 97.2 (3.8) 84.0 3.3 83.9 (3.3) 3318.1 (29.4) 3327.8 (29.4) 2596.3 (23.0 2719.3 (24.1) AVERAGE 48.2 (10.8) 86.2 (3.4) 2990.4 (26.5) SADT NO SUPPAGE 53.4 (12.0) 83.8 (3.3) 3276.4 (29.0) 117 Overall, the results show that the racking behaviour of the conventional and the fr i c t i o n damped timber shearwalls are i n i t i a l l y similar. But as wall deflections increase, the latter system becomes s t i f f e r and stronger with the contribution from the f r i c t i o n devices. The f r i c t i o n damped walls sustained an average peak racking resistance of 48.1 kN (10.8 kip), which i s 23.7% higher than the average of 38.9 kN (8.7 kip) for the conventional walls. The greater stiffness and strength allowed them to command a better energy dissipating capability. On the average, they dissipated 2990.4 kN-mm (26.5 kip-in) of energy at failure. This i s 42.9% better than the 2093.0 kN-mm (18.5 kip-in) of the conventional walls. Aside from the better racking performance and energy dissipating a b i l i t y , the f r i c t i o n damped timber shearwalls also failed in a ductile manner, maintaining a substantial amount of their load carrying a b i l i t y at post peak load displacements. In the conventional walls, the failure was b r i t t l e , with l i t t l e or no load carrying capacity after the maximum racking load was reached. The load-deformation behaviour of a l l the timber shearwall specimens, with the exception of one f r i c t i o n damped wall, FI, f e l l within the solution bound of the SADT numerical analyses. As discussed in Chapter 4, these solution bounds represented shearwalls with no f r i c t i o n devices at the lower bound, and with non-slipping devices at the upper bound. Therefore, a l l the experimental results should f a l l with in these bounds. However, the analyses were based on the 5th percentile and mean value 118 bending strengths of the framing members, thus a wall b u i l t with stronger members would a resist higher racking load than that which has been predicted by SADT. In addition, the v a r i a b i l i t y in the wood's Modulus of E l a s t i c i t y was not considered in the analysis, thus i t would be possible for a wall to exhibit a load-deformation behaviour which may be s t i f f e r or more flexible than the calculated results. Thus, the higher racking resistance and greater stiffness exhibited by shearwall FI may be explained as being attributed to stronger and s t i f f e r framing and sheathing members. Comparisons between the four f r i c t i o n damped timber shearwalls indicated no noticeable correlation between their racking strength and the s l i p loads calibrated on their f r i c t i o n devices. In fact, the average peak racking resistance of the three walls tested at the 6.7 kN (1.5 kip) s l i p load i s 49.2 kN (11.1 kip), which i s 10% higher than the wall tested at the higher s l i p load of 13.4 kN (3.0 kip). From an energy perspective, this finding should not-be'"possible because a higher s l i p load should allow for more energy dissipation, resulting in a wall system with greater racking strength. However, the inherent v a r i a b i l i t y of wood coupled with the testing of only one shearwall at the higher s l i p load makes i t d i f f i c u l t to see the true relationship between s l i p load magnitude and racking resistance. Thus, i t could not be concluded from these tests how the racking resistance of a f r i c t i o n damped timber shearwall i s affected by the s l i p load on i t s f r i c t i o n devices. 119 The load-deformation behaviour of the f r i c t i o n damping devices from shearwall specimen F2, for wall deflections up to 127.0 mm (5.0 in), are plotted in Figure 7.8. Wall F2 was chosen because i t s peak racking resistance matched closest to the average peak resistance of the four f r i c t i o n damped timber shearwalls tested. The elasto-plastic behaviour exhibited by these devices are typical of the results recorded in the other walls. Although these devices were set to s l i p at 6.7 kN (1.5 kip), Figure 7.8 shows the actually slippage occurring between 4.2 kN (0.9 kip) to 5.0 kN (1.1 kip). Since the expected s l i p load was determined from the linear regression f i t of the s l i p load to torque data from the c y c l i c tests of each device (see Section 5.4), differences between the expected and measured values are possible. Another source causing discrepancy between these values came from the torque wrench used for tightening the torquing bolt on each f r i c t i o n joint. The wrench only had an accuracy of +/- 10 % for the range of torques considered. Table 7.2 summarizes the s l i p load measured on each friction, damping device corresponding to the wall on which they were installed. Figure 7.8 shows that the total slippage in each device at the end of the test i s different. These differences arise because the behaviour in each corner of the wall can vary depending on the amount of separation between i t s framing and the amount of damage i t imparts during i t s deformation. SHEARWALL F2 1 2 0 g 4 H rr cv) 2 -z o I o LU Z o Q £ -4H -20 B t/-.-.-.-.-.-.v.-.\J DEVICE A DEVICE C DEVICE B —r -15 -10 -5 0 5 10 15 RELATIVE DISPL BETWEEN STRUTS (mm) 20 Figure 7.8: Load-Deformation Behaviour of Friction Damping Devices Table 7.2: Summary of Slip Loads from the Racking Tests of the Friction Damped Timber Shearwalls SLIP LOAD OF FRICTION DAMPING DEVICES [kN(kip)] WALL# DEVICE A DEVICE B DEVICE C DEVICE D F1 5.6 (1.3) 5.6 (1.3) 6.7 (1.5) 5.0 (1.1) F2 4.2 (0.9) 5.0 (1.1) 4.3 (1.0) 4.5 (1.0) F3 4.0 (0.9) 3.7 (0.8) 5.2 (1.2) 5.4 (1.2) F4 7.9 (1.8) 5.5 (1.2) 7.9 (1.8) 4.3 (1.0) 121 A time-history plot of the axial load in each f r i c t i o n device from shearwall F2 i s shown in Figure 7.9. The figure shows that the s l i p load in each devices was not reached u n t i l approximately 300 seconds (5.0 min) into the test. At a racking rate of 5.1 mm/min (0.2 in/min), this time delay resulted in a lateral wall displacement of 25.4 mm (1.0 in) before the devices began to dissipate energy input. Similar time delays were experienced for the f r i c t i o n devices on the other f r i c t i o n damped walls. Forces in the f r i c t i o n devices are developed from the parallelogram shape distortion of the framing as the wall deforms under a racking load. Even with reinforcements in the end stud, the axial stiffness of the devices i s much greater than the bending stiffness of the framing members, thus large bending deformations in the framing are required before the load being transferred to the devices can be great enough to cause them to s l i p . The weaker bending stiffness in the framing members contributed to the time delay before the f r i c t i o n devices could became effective energy dissipators. 122 ? 6 TIME (s) Figure 7.9: Time-Histories of Axial Strut Loads on the Friction Damping Devices During a l l of the tests, ela s t i c buckling of both plywood panels between the framing studs was observed. The buckling increased with progressing in-plane wall deformation. It was caused by in-plane forces developed in the sheathing panels as they resisted the l a t e r a l deformation of the wall. .Because the sheathing's interaction with the interior framing studs provided them with out-of-plane stiffness, the degree of buckling was limited, thus a catastrophic buckling failure of the sheathing which would lead to a sudden collapse of the wall was prevented. 123 Warping of the interior vertical studs was noticed during the deformation of the shearwalls. The warping was caused by the sheathing panels pulling on these studs as they rotated in r i g i d body mode during the deformation of the wall. Bending of the double end studs was also noticed in a l l of the f r i c t i o n damped timber shearwalls. The bending was caused by the f r i c t i o n devices which introduced point loads part way up each double end stud to resist the distortion of the framing. With the presence of these f r i c t i o n devices, the four wall corners behaved as moment resisting joints instead of pinned joints. This additional moment capacity enabled the framing to also contribute s l i g h t l y to the stiffness of the f r i c t i o n damped shearwalls. At the end of the tests, damage to the conventional shearwalls was mainly -found in one sheathing where most of the sheathing-to-framing connectors around the perimeter broke through the plywood panel. For the f r i c t i o n damped shearwalls, damage was spread to both plywood panels, but the level was less severe. The sheathing-to-framing connectors only broke through one or two veneers of the plywood panels as opposed to breaking completely through the entire thickness. No failure of any framing members was found. By comparison, the SADT analyses in Chapter 4 showed wall failures being ini t i a t e d by the exceedence of allowable bending stress in one segment of the framing members rather than by the failure of the sheathing-to-framing connectors. The different failure mode in the numerical model was attributed to the method 124 in which the shearwall's elements were modelled. A wall failure i n i t i a t e d by the sheathing-to-framing connectors had not been considered in the analyses due d i f f i c u l t i e s in modelling such a failure. The only way such a failure mode may be modelled was by assigning' a value to the global d u c t i l i t y parameter which detected yielding of these connectors. Due to the high redundancy of a shearwall, failure at a single sheathing-to-framing connector, whether by yielding or puncturing through the sheathing, merely dictates the redistribution of load to adjacent connectors. Only when enough connectors have failed w i l l the wall f a i l . Thus the value for such a d u c t i l i t y parameter would depend on the wall material, the density of the connectors and i t s load-deformation behaviour. Because of i t s complexity, i t was not considered in the analyses and the walls were modelled assuming that they contained adequate d u c t i l i t y . Thus the modelling technique used in SADT only allowed the walls to f a i l from the exceedence of bending stress in their framing members. Despite the different modelling technique, good agreement between the numerical and experimental results shows that an analysis carried out using the allowable framing member stress as the failure criterion produces an adequate prediction of the racking load-deformation behaviour of a timber shearwall. The racking load and deflection data from each of the conventional shearwall tests were also f i t t e d to an exponential curve to obtain parameters for defining the virgin load-deformation path of a shearwall used in the FRICWALL program discussed in Chapter 6 . Table 7 . 3 summaries the parameters from 125 the three wall along with their average values which were used in the program. The load-deformation curve defined by these parameters i s shown in Figure 6.2. Table 7.3: Parameters Defining the Virgin Load-Deformation Path of a Conventional Timber Shearwall PARAMETER WALL C1 WALL C2 WALL C3 AVERAGE P0[kN(kip)] 31.91 (7.17) 26.08 (5.83) 33.24 (7.47) 30.35 (6.82) P2 [mm(in)] 84.30 (3.32) 65.08 (2.56) 68.67 (2.70) 72.68 (2.86) Kg [kN/mm (kip/in)] 1.76(10.06) 2.23(12.73) 2.38(13.59) 2.13(12.13) K [kN/mm (kip/in)] 0.11 (0.65) 0.14 (0.78) 0.13 (0.74) 0.13 (0.72) S 2 [kN/mm (kip/in)] -1.04 (-5.93) -1.49 (-8.53) -0.46 (-2.63) -1.00 (-5.69) 126 7.5.2 Static Cyclic Tests A total of five f u l l scale static c y c l i c tests were performed. Two tests were conducted on conventional timber shearwalls while the other three were on f r i c t i o n damped timber shearwalls. A sj i p load of 6.7 kN (1.5 kip) was set on the f r i c t i o n devices of two of the f r i c t i o n damped walls while a s l i p load of 13.4 kN (3.0 kip) was used on the devices of the third wall. The results from these tests were analyzed to compare the energy dissipating a b i l i t y of the two types of timber shearwalls. The conventional timber shearwall test results were also used for deriving the parameters which modelled the pinched hysteretic behaviour of a shearwall in the FRICWALL analysis of Chapter 6. Figure 7.10 shows the hysteresis loops exhibited by conventional and f r i c t i o n damped timber shearwalls SCC1 and SCF1 during their static c y c l i c tests. The f r i c t i o n devices on wall SCF1 were set to a s l i p load of 6.7 kN (1.5 kip). The hysteretic behaviour shown i s typical of a l l the walls tested. Both walls generated pinched hysteresis loops, but the ones from the f r i c t i o n damped wall provided higher racking resistance at each displacement level than the conventional wall. The greater resistance arises because the f r i c t i o n devices act as struts to resist the deformation of the wall, thereby, increasing i t s i n -plane stiffness and strength. Table 7.4 l i s t s the peak racking resistances measured in each shearwall during each displacement cycle of the static c y c l i c tests. 127 CONVENTIONAL TIMBER SHEARWALL SCC1 50 "I ; ; ; ; ; ; 1 i ; • ; 40 -40 -50 1 ; i 1 1 r j ' i i 1 j i 1 -80 -60 -40 -20 0 20 40 60 80 RELATIVE WALL DISPL (mm) FRICTION DAMPED TIMBER SHEARWALL SCF1 (SLIP LOAD = 6.7 kN) 50 I i : • ; i ; i -50 i i i 1 1 i i 1 ! i i 1 i 1 i i -80 -60 -40 -20 0 20 40 60 80 RELATIVE WALL DISPL (mm) Figure 7.10: Hysteresis Loops from Static Cyclic Tests of Timber Shearwalls 128 Table 7.4: Summary of Peak Racking Resistances from Static Cyclic Tests PEAK RACKING RESISTANCE (kN(kip)] DISPLACEMENT [mm(ln)] + 12.7 (0.5) -12.7 (0.5) 4 25.4(1.0) -25.4(1.0) + 50.8 (2.0) • 50.8 (2.0) WALL CONVENTIONAL TIMBER SHEARWALL SPECIMEN SCC1 CYCLE 1 CYCLE 2 20.0 (4.5) 19.8 (4.4) -16.7 (-3.8) -16.3 (-3.7) 27.5 (6.2) 25.9 (5.8) -24.3 (-5.5) -23.4 (-5.3) 35.5 (8.0) 30.5 (6.9) -30.2 (-6.8) -27.4 (-6.2) SCC2 CYCLE 1 CYCLE 2 16.4 (3.7) 15.8 (3.6) -17.2 (-3.9) -16.8 (-3.8) 23.6 (5.3) 22.1 (5.0) -24.9 (-5.6) -23.8 (-5.3) 31.6 (7.1) 27.0 (6.1) -31.2 (-7.0) -28.2 (-6.3) WALL FRICTION DAMPED TIMBER SHEARWALL SPECIMEN SCF1 CYCLE 1 CYCLE 2 CYCLE 3 »- SUP LOAI 22.0 (4.9) 21.7 (4.9) 21.5 (4.8) - 8.7kN(1.5klp) -21.1 (-4.7) -20.7 (-4.7) -20.6 (-4.6) 32.4 (7.3) 31.2 (7.0) 30.3 (6.8) -31.8 (-7.1) -30.4 (-6.8) -29.6 (-6.6) 39.6 (8.9) 33.0 (7.4) 28.9 (6.5) -39.2 (-8.8) -34.7 (-7.8) -31.0 (-7.0) SCF2 CYCLE 1 CYCLE 2 CYCLE 3 »- SUP LOAI 18.9 (4.2) 18.8 (4.2) 18.4 (4.1) -8.7kN(1.5klp) -17.7 (-4.0) -17.6 (-4.0) -17.3 (-3.9) 25.9 (5.8) 24.8 (5.6) 24.0 (5.4) -29.5 (-6.6) -28.3 (-6.4) -27.8 (-6.2) 33.4 (7.50 28.8 (6.5) 26.2 (5.9) -38.8 (-8.7) -35.4 (-8.0) -32.5 (-7.3) SCF3 CYCLE 1 CYCLE 2 CYCLE 3 »- SUP LOAI 17.2 (3.9) 17.1 (3.8) 16.9 (3.8) - 13.4 kN(3.0 kip) -10.6 (-2.4) -10.5 (-2.4) -10.8 (-2.4) 31.6(7.1) 30.3 (6.8) 30.2 (6.8) -26.1 (-5.9) -24.7 (-5.6) -24.2 (-5.4) 38.4 (8.6) 34.9 (7.8) 32.9 (7.4) -41.5 (-9.3) -37.9 (-8.5) -36.4 (-8.2) 129 Aside from higher peak racking resistances, the f r i c t i o n damped timber shearwalls also dissipated more energy than the conventional walls as shown by their fatter hysteresis loops. For example, -the area inside the hysteresis loop of the f i r s t +/- 50.8 mm (2.0 in) displacement cycle of f r i c t i o n damped wall SCF1 in Figure 7.10 shows an energy dissipation of 2605.4 kN-mm (23.1 kip-in). This i s 41% greater than the 1851.7 kN-mm (16.4 kip-in) of conventional wall SCC1 at the identical displacement cycle. The amount of energy dissipated by the shearwalls during each displacement cycle i s summarized in Table 7.5. The amount of energy dissipation from each type of shearwall, at each displacement cycle, were averaged and are also shown in Table 7.5. These average values indicated that the amount of energy dissipated by both types of shearwalls were comparable during the +/- 12.7 mm (0.5 in) cycles. The f r i c t i o n damped walls showed only marginally improved hysteretic behaviour during the +/- 25.4 mm (1.0 in) cycles. Only during the +/- 50.8 mm (2.0 in) displacement cycles did noticeably better energy dissipating a b i l i t y was exhibited by the f r i c t i o n damped walls. Since a l l the walls failed during the +/- 76.2 mm (3.0 in) cycle, no energy calculations were made for this displacement level. The hysteretic behaviour of the f r i c t i o n damped timber shearwalls shows that the f r i c t i o n devices adds to the strength and stiffness of a shearwall, but in agreement with the racking test findings, did not contribute to i t s energy dissipation u n t i l wall deformations were beyond 25.4 mm (1.0 in). 130 Table 7.5: Summary of Hysteretic Behaviour of Timber Shearwalls ENERGY [kN-mm(klp-ln)] EFFIC-IENCY [%] ENERGY [kN-mm(klp-)n)] EFFIC-IENCY. ENERGY [kN-mm(klp-ln)] EFFIC-IENCY DISPLACEMENT [mm(In)] ±12.7(0.5) + 25.4(1.0) ±50.8(2.0) WALL CONVENTIONAL TIMBER SHEARWALL SPECIMEN SCC1 CYCLE 1 150.21 (1.33) 17.59 561.68 (4.97) 22.84 1851.68 (16.38) 29.04 CYCLE 2 118.00 (1.04) 14.26 421.03 (3.72) 18.06 1244.50 (11.01) 22.20 SCC2 CYCLE 1 180.33 (1.60) 19.36 640.32 (5.67) 24.30 1988.14 (17.59) 29.78 CYCLE 2 143.30 (1.27) 15.60 480.77 (4.25) 19.20 1343.40 (11.89) 22.85 AVERAGE VALUES CYCLE 1 165.27 (1.46) 18.48 601.00 (5.32) 23.57 1919.91 (16.99) 29.41 CYCLE 2 130.65 (1.16) 14.93 455.90 (4.03) 18.63 1293.95 (11.45) 22.53 FRICTION DAMPED TIMBER SHEARWALL SPECIMEN SCF1 — SUP LOAD - 8.7 kN(1.5 kip) CYCLE 1 194.92 (1.72) 17.80 752.31 (6.66) 23.06 2605.36 (23.05) 32.56 CYCLE 2 154.08 (1.36) 14.30 555.45 (4-91) 17.74 1771.80 (15.68) 25.76 CYCLE 3 143.83 (1.27) 13.46 510.67 (4-52) 16.81 1505.40 (13.32) 24.74 SCF2 — SUP LOAD - 8.3 kN(1.5klp) CYCLE 1 164.86 (1.46) 17.70 674.10 (5.96) 23.97 2231.55 (19.74) 30.45 CYCLE 2 152.13 (1.35) 16.49 498.22 (4.41) 18.47 1554.00(13.75) 23.85 CYCLE 3 142.09 (1.26) 15.66 461.69 (4.08) 17.55 1403.50 (12.42) 23.55 SCF3 SUP LOAD - 13, « kNp.o kip) CYCLE 1 92.49 (0.82) 13.09 682.24 (6.04) 23.31 2315.25 (20.48) 28.53 CYCLE 2 79.66 (0.70) 11.35 501.96 (4.44) 17.96 1485.70 (13.14) 20.10 CYCLE 3 76.53 (0.68) 10.87 464.03 (4.11) 16.79 1360.10 (12.03) 19.33 AVERAGE VALUES CYCLE 1 CYCLE 2 CYCLE 3 150.70 (1.33) 128.62 (1.14) 120.82 (1.07) 16.20 14.05 13.33 702.90 (6.22) 518.54 (4.59) 478.80 (4.24) 23.45 18.06 17.05 2384.05 (21.09) 1603.83 (14.19) 1423.00 (12.59) 30.51 23.24 22.54 131 To examine how e f f i c i e n t the two energy dissipating systems were, the ratios of the areas of their pinched hysteresis loops to those of ideal, unpinched loops were calculated for each displacement cycle of the" walls. The area for the ideal, unpinched -loops were calculated by taking a rectangular area bounded by the peak positive and negative racking loads and displacement spans at each cycle. The peak racking loads are l i s t e d in Table 7.4. For the same two hysteresis loops discussed above, efficiencies of 32.5% and 29.0% were respectively calculate for the f r i c t i o n damped and conventional walls. The efficiency of the hysteresis loop at each displacement cycle i s tabulated in Table 7.5 beside the amount of energy being dissipated. To better i l l u s t r a t e the changes in efficiency with respect to incrementing displacement spans, the efficiencies calculated from the f i r s t two cycles of each displacement span are averaged and plotted in Figure 7.11. 132 0 10 20 30 40 DISPLACEMENT SPAN (mm) Figure 7.11: Energy Dissipation Efficiency of Timber Shearwalls From Figure 7.11, i t can be concluded that both types of shearwalls are equally inefficient in the manner in which they dissipate energy input. Although the f r i c t i o n damped timber shearwalls could dissipate a higher amount of energy input, they did not perform this task with any more efficiency than the conventional timber shearwalls. Comparisons between the three f r i c t i o n damped shearwalls revealed no noticeable correlation between the amount of energy 133 dissipated during each displacement cycle and the s l i p loads on their f r i c t i o n devices. More tests would have to be made to obtain a better representation of any differences produced by the two s l i p loads. The hysteresis loops generated by the f r i c t i o n devices on wall SCF1 are illustrated in Figures 7.12, 7.13 and 7.14 for displacement cycles of +/- 12.7 mm (0.5 in), +/- 25.4 mm (1.0 in) and +/" 50.8 mm (2.0 in), respectively. The hysteresis loops in Figure 7.12 shows negligible areas, with peak axial loads which were not large enough to cause slippage of the :devices. This illustrates that the devices were not dissipating any energy at wall deflections up to +/- 12.7 mm (0.5 in). Figure 7.13 shows hysteresis loops with higher peak loads and some hysteretic area, indicating that the devices were just beginning to dissipate some energy although the s l i p load was not f u l l y reached by a l l the devices. Figure 7.14 shows rectangular hysteresis loops generated by a l l four f r i c t i o n devices, with s l i p loads ranging from 3.38 to 5.49 kN (0.76 to 1.23 kip). The nearly perfect loops, with only slight imperfections in the 2nd and 4 th quadrants, i l l u s t r a t e the effectiveness of the devices in dissipating energy input. Similar to the findings from the racking tests, the amount of slippage _in each device varies because of the different joint behaviour at each corner of the wall. The hysteresis loops generated by the f r i c t i o n devices on the other shearwalls are similar to the ones illustrated in Figures 7.12, 7.13 and 7.14. 134 8 7 6-5-4-z 3 Q 2-< 3 1 -0-£ -1 • to _l -2--3 -4--5--6--7--8 8-1 7 6 5 4 z 3 a 2 < 1 -o _J fe o-£ -1 • CO —1 -2 -3 -4--5--6 -7 -8 DEVICE B -3 -1 1 RELATIVE DISPL BETWN STRUTS (mm) DEVICE A 8-j 7-6-5-4-z 3-2-a < 1 -o _i fe o-£ -1 -CO -1 -2-i -3--4--5--6--7--8 4 DEVICE C - 5 - 3 - 1 1 3 5 RELATIVE DISPL BETWN STRUTS (mm) z <£. Q < 3 fe g 5 -3 -4 -5 -6 -7 -8 - DEVICE D -.....r. . ,—.— -1 1 RELATIVE DISPL BETWN STRUTS (mm) - 5 - 3 - 1 1 3 5 RELATIVE DISPL BETWN STRUTS (mm) Figure 7.12: Hysteresis Loops of Friction Damping Devices - Static Cyclic Test, +/~ 12.7 mm (0.5 in) -135 z Q < 3 I 3 8 7 6 5 4 3 2 1 0 -1 -2 -3 -4 -5 -6 -7 -8 DEVICE B • I z Q I 1 0 -1 -2 -3 DEVICE C i l -3 -1 1 -3 -1 1 RELATIVE DISPL BETWN STRUTS (mm) RELATIVE DISPL BETWN STRUTS (mm) a I S 3 8 7 6 5 4i 3 2 1 • 0 -1 -I -2 -3 -4--5 -6 -7i -8 DEVICE A —i 1 i 1 •5 "3 - 1 1 3 5 RELATIVE DISPL BETWN STRUTS (mm) Q I i 2 8 7 6 5 4 3 2 1 0 -1 -2 -3 -4 -5 -6 -7 -8 DEVICE D —i 1 1 1 1—] 1 1 r-5 - 3 - 1 1 3 5 RELATIVE DISPL BETWN STRUTS (mm) Figure 7.13: Hysteresis Loops of Friction Damping Devices - Static Cyclic Test, +/- 25.4 mm (1.0 in) -136 Figure 7.14: Hysteresis Loops of Friction Damping Devices - Static Cyclic Test, +/- 50.8 mm (2.0 in) -137 Table 7.6 provides a summary of the s l i p loads generated by the f r i c t i o n devices on the other shearwalls for the +/- 50.8 mm (2.0 in) displacement cycle. Peak axial loads in the f r i c t i o n devices for the two smaller displacement spans are not shown in Table 7.6 since no slippage, of the devices occurred at these wall displacements. This examination of the hysteretic behaviour of the individual f r i c t i o n devices explains why the hysteretic behaviour of the f r i c t i o n damped walls did not show significant improvement u n t i l the +/~ 50.8 mm (2.0 in) displacement cycles of the test. Recall from Table 7.5 that the energy dissipated by the f r i c t i o n damped walls were the same as the conventional wall during the +/- 12.7 mm (0.5 in) displacement cycles, marginally improved in the +/- 25.4 mm (1.0 in) displacement cycles and noticeable improved during the +/- 50.4 mm (2.0 in) displacement cycles. Findings from the static cyclic tests are in agreement with those from the racking tests. Both set of tests revealed that the f r i c t i o n damping devices were ineffective u n t i l wall deflections of approximately 25.4 mm (1.0 in) were reached. 138 Table 7.6: Summary of Slip Loads for A l l Friction Devices from Static Cyclic Tests SLIP LOAD OF FRICTION DAMPING DEVICES [kN(kip)] WALL# DEVICE A DEVICE B DEVICE C DEVICE D SCF1 4.71 (1.06) 5.43 (1.22) 5.49 (1.23) 3.38 (0.76) SCF2 4.55 (1.02) 3.67 (0.82) 4.63 (1.04) 3.70 (0.83) SCF3 7.05 (1.58) 6.52 (1.47) 7.58 (1.70) 7.68 (1.73) Visual inspection of each specimen during the tests revealed similar behaviours exhibited by both types of shearwalls. Ela s t i c buckling in the plywood panels between the framing studs was observed. The amount of buckling fluctuated proportionally with the magnitude of deformation in the wall. Warping of the interior v e r t i c a l studs, as they resisted the rotation of the sheathing panels was noticed. Bending of the double end studs of the f r i c t i o n damped walls, as seen in the racking tests, was also noticed during these tests. No surface damage was noticed in the sheathing panels during the +/- 12.7 mm (0.5 in) displacement cycles. However, as wall displacements incremented to higher levels, progressively more crushing of the panels by the sheathing-to-framing connectors was evident. The damage was especially noticeable for the connectors -located around the perimeter of each panel. The crushing occurred as a result of the rocking motion of the connector heads as they rotated in and out of the sheathing, in 139 sync with reversals in the wall displacement. Increasing separation between the framing and the perimeter of each sheathing was also noticed as the connectors pulled out of the framing with each successive load cycle. Despite the escalating damage level, the walls continued to maintain their load carrying a b i l i t y up to the end of the +/- 50.8 mm (2.0 in) displacement cycles. However, during the i n i t i a l +/- 76.2 mm (3.0 in) displacement cycle, shear failure of some sheathing-to-framing connectors combined with the complete pulling out of others from the framing, lead to the wall's failure. Both failure modes were not observed during ;the racking tests where the connectors were subjected to higher loads and deformations than in these sta t i c cyclic tests. Thus i t could be concluded that the connector shear failures were from low cycle fatigue, not from the lack of strength, while their pulling out from the framing was due to repeated prying action from the cycli c nature of loading. The hysteresis loops generated by the conventional timber shearwalls were also used to determine the load intercept and linear stiffness parameters for defining the pinched hysteresis loop used in the FRICWALL program of Chapter 6. Table 7.7 summarizes these parameters along with their average values which were used in the program. The pinched hysteresis loop defined by these parameters i s illustrated in Figure 6.2. 140 Table 7 . 7 : Parameters D e f i n i n g the Pinched H y s t e r e s i s Loop of a Conventional Timber Shearwall PARAMETER WALL 1 WALL 2 AVERAGE R, [kN(kip)] S1 [kN/mm(kip/in)] 4.72 (1.06) 0.52 (2.95) 5.12 (1.15) 0.53 (3.05) 4.92 (1.11) 0.53 (3.00) 7 . 5 . 3 Dynamic Earthquake simulation A t o t a l of f i f t e e n dynamic earthquake s i m u l a t i o n t e s t s were c a r r i e d out. Seven t e s t s were conducted w i t h the c o n v e n t i o n a l timber s h e a r w a l l s and e i g h t w i t h the f r i c t i o n damped timber s h e a r w a l l s . On the c o n v e n t i o n a l w a l l s , two t e s t s were conducted u s i n g the Romania earthquake and f i v e u s i n g the E l Centro earthquake. On the f r i c t i o n damped w a l l s , two t e s t s were conducted u s i n g the Romania earthquake and s i x u s i n g the E l Centro earthquake. The Romania earthquake was simu l a t e d a t a peak ground a c c e l e r a t i o n of 0.2 g w h i l e the E l Centro earthquake was s i m u l a t e d a t peak ground a c c e l e r a t i o n s o f 0.35 g and 0.6 g. The 0.2 g and 0.6 g peak ground a c c e l e r a t i o n s are the maximum a t t a i n a b l e l e v e l s w i t h the UBC shake t a b l e f o r these r e c o r d s . Devices on the f r i c t i o n damped w a l l s were s e t t o s l i p a t 6.7 kN (1.5 k i p ) f o r a l l o f the t e s t s except two E l Centro t e s t s which used a s l i p l o a d o f 13.4 kN (3.0 k i p ) . These t e s t s were conducted t o examine the s e i s m i c performance o f the two types of s h e a r w a l l s . F i n d i n g s from these t e s t s were a l s o compared a g a i n s t the numerical r e s u l t s from FRICWALL. 141 1940 EL CENTRO (N-S) @ 0.35 g P.G.A. E E, _ i Q _ 00 Q _1 _J Lu O i Cu H i > 1 LU CC E E, _j Cu o _l -J LU O 1 0-UJ > LU CC 50 40 -30 -20 -10 -0 -10 --20 --30 --40 --50 50 30 10 -20 -50 EXPERIMENTAL RESULTS 26.3 mm (CONVENTIONAL WALL ENW1) 22.3 mm (FRICTION DAMPED WALL EDW1) 6.7 kN) 3 TIME (s) T 4 2 3 TIME (s) FRICWALL . NUMERICAL RESULTS w/28.1 mm (FRICTION DAMPED WAL (SUP LOAD - 6.7 kN) : / 42.1 mm (CONVENTIONAL WALL) L) Figure 7.15: R e l a t i v e Top-of-Wall Displacement Time H i s t o r i e s - E l Centro Earthquake § 0.35 g P.G.A. -142 1940 EL CENTRO (N-S) @ 0.6 g P.G.A. 120 E E, - i D_ CO Q _1 - J I ul O i CL e UJ > 1 UJ CC 0 --40 --80 --120 E E, _ j Cu CO Q - J — J i O I Q. P UJ > 1 UJ CC 120 80 40 -0 --40 --80 --120 'r 35.0 mm (FRICTION DAMPED WALL EDWSy1 (SUP LOAD» 13.4 kN) 36.6 mm (FRICTION DAMPED WALL EDW3) (SLIP LOAD = 6.7 kN) 42.4 mm (CONVENTIONAL WALL ENW4) 3 4 TIME (S) FRICWALL NL IMERICAL RESULTS yj 8 y i n 9.4 mm (FRICTION DAMPED WALL) (SUP LOAD - 6.7 kN) 12 mm (CONVENTIONAL WALL) 3 4 TIME (s) Figure 7.16: Rela t i v e Top-of-Wall Displacement Time H i s t o r i e s - E l Centro Earthquake @ 0.6 g P.G.A. -143 1977 ROMANIA (BUCHAREST, N-S) @ 0.2 g P.G.A. Q_ C O Q LL O I Q_ e UJ > LU cr 20 15 -10 -5 -0 -5 -10 --15 --20 EXPERIMENTAL RESULTS 13.6 mm (CONVENTIONAL WALL RNW1) 3 12.2 mm (FRICTION DAMPED WALL RDW1) (SUP LOAD - 6.7 kN) TIME (S) FRICWALL NUMERICAL RESULTS ]\16.0 mm (CONVENTIONAL WALL) 11.4 mm (FRICTION DAMPED WALL) (SLIP LOAD = 6.7 kN) -i 1 1 1 1 1 1 r 2 4 6 8 10 12 14 16 TIME (s) Figure 7 . 1 7 : R e l a t i v e Top-of-Wall Displacement Time H i s t o r i e s - 1977 Romania Earthquake § 0.2 g P.G.A. -144 The f i r s t 6 seconds of the relative displacement time-histories from the conventional and the friction damped timber shearwalls, tested under the El Centro earthquake at peak ground accelerations of 0.35 g and 0.6 g, are respectively displayed at the top-of-Figures 7.15 and 7.16.- Only the first.6,seconds are shown since this is the strong motion portion of the El Centro earthquake. The entire 16 seconds of the relative displacement time-histories of a conventional and a friction damped timber shearwall, tested under the Romania earthquake at a peak ground acceleration of 0.2 g, are plotted at the top of Figure 7.17. The dotted lines represent the response of the conventional walls while the solid lines represent that of the friction damped walls. The time-histories shown in these figures are representative of the responses exhibited by the other shearwalls tested under these simulated earthquakes. The figures show that the friction damped walls exhibit slightly lower peak relative top-of-wall deflections than the conventional walls during a seismic event. For the time-histories shown in the above figures, peak deflections in the friction damped walls were 8.5 %, 13.6 % and 10.3 % lower than the conventional walls for, respectively, the El Centro earthquake at 0.35 g and 0.6 g, and the Romania earthquake at 0.2 g peak ground acceleration, at a slip load of 6.7 kN (1.5 kip). A displacement reduction of 17.6 % is exhibited under the El Centro earthquake at 0.6 g peak ground acceleration, with a 13.4 kN (3.0 kip) slip load. Other than lower peak deflections, the response time-histories of the friction damped walls 145 basically followed those of the conventional walls. Dynamic earthquake simulation tests on the other shearwalls revealed similar findings. The only noticeable difference between the seismic response of the conventional and the f r i c t i o n damped shearwalls i s sli g h t l y lower peak deflections in the latter type of walls. The peak top-of-wall displacement measured in each shearwall i s summarized in Table 7.8, along with the percentage reduction in peak displacements exhibited by the f r i c t i o n damped walls. Due to the v a r i a b i l i t y of the properties of wood and the small deflections involved, some of the f r i c t i o n damped walls show a slight higher peak deflections than the conventional walls when subjected to the same seismic excitation. But overall, the f r i c t i o n damped walls exhibited lower peak displacements than the conventional ones, with reductions in the order of 6.4 to 14.9 % and averaging at 9.6 % for the walls tested. In comparison to the FRICWALL results, an average drop of 29.5 % in peak wall deflection was computed based on the range of viscous damping ratios considered in the analyses. The results from the El Centro simulation at 0.6 g peak ground acceleration shows that a greater reduction in peak wall deflection was achieved by the 13.4 kN (3.0 kip) s l i p load than by the 6.7 kN (1.5 kip) s l i p load. But with an average difference in peak deflection of only 3.9 mm (0.1 mm) between the two s l i p loads, the advantage achieved by a higher s l i p load was insignificant. In the racking and static c y c l i c tests, i t was shown that the f r i c t i o n devices did not contribute to the strength or 146 stiffness of a shearwall u n t i l relative wall deflections of approximately 25.4 mm (1.0 in) were reached. These dynamic tests under simulated severe earthquake loads have shown that the relative, deflections of the shearwalls are quite small, exceeding 25.4 mm (1.0 in) only at the peak values. Due the limitation of the effectiveness of the f r i c t i o n devices .in timber shearwalls, the seismic performance of the f r i c t i o n damped timber shearwalls did not dramatically improve over the conventional ones. Table 7.8: Summary of Peak Top-of-Wall Relative Deflection from Earthquake Simulations PEAK TOP-OF-WALL RELATIVE DISPLACEMENT (mm[in]) ACCELERATION RECORD CONVENTIONAL TIMBER SHEARWALL SPECIMEN FRICTION DAMPED TIMBER SHEARWALL SPECIMEN SUP LOAD - 6.7 kN(1.5 kip) % REDUCTION WALL RECORDED AVERAGE WALL RECORDED AVERAGE ROMANIA (P.G A = 0.2 g) RNW1 RNW2 13.6 (0.5) 13.0 (0.5) 13.3 (0.5) RDW1 RDW2 12.2 (0.5) 12.5 (0.5) 12.3 (0.5) 7.5 EL CENTRO (P.GA = 0.35 g) ENW1 ENW2 ENW3 26.3 (1.0) 24.4 (1.0) 22.0 (0.9) 24.2 (1.0) EDW1 EDW2 22.3 (0.9) 21.5 (0.9) 21.9 (0.9) 9.5 EL CENTRO (P.GA = 0.6 g) ENW4 ENW5 42.4 (1.7) 49.2 (1.9) 45.8 (1.8) EDW3 EDW4 EDW5 EDW6 36.6 (1.4) 49.1 (1.9) 35.0 (1.4) 43.0 (1.7) 42.9 (1.7) 39.0 (1.5) 6.4 14.9 * SUP LOAD - 13.4 kN(3.0 kip) 147 For comparisons between the experimental and numerical results, the response time-histories computed by FRICWALL in Chapter 6, for a viscous damping coefficient ..of 0.012 kN-s/mm (0.07 kip-s/in), are displayed at the bottom of Figures 7.15, 7.16 and 7.17. Under the E l Centro simulations, the model readily reproduced the phases of the conventional timber shearwall's response time-histories, but sl i g h t l y overestimated their amplitudes, especially at the peak deflection values. For higher viscous damping ratios (within the range expected from a typical timber shearwall), the model was able to reproduce response amplitudes which were in agreement; with experimental results, except for the peak wall deflection which was slightly on the high side. Under the Romania simulation, both the phase and amplitudes of the numerical results for the conventional timber shearwall were in agreement with the experimental findings, except for the amplitude at peak wall deflection. For the f r i c t i o n damped timber shearwalls, the time-histories computed by the model followed in-phase with the experimental findings, but the reductions in deflection amplitudes observed during the tests were much lower than the levels computed by FRICWALL. The discrepancy between the numerical and experimental results arise mainly from the way in which the hysteretic behaviour of a f r i c t i o n damping device i s modelled. In the derivation of the equations describing the hysteretic behaviour of the f r i c t i o n devices (see Section 6.1), a wall's framing members are assumed to be r i g i d and do not bend from the 148 point loads of the f r i c t i o n devices as they resist the wall's distortion. Thus, only a slight deformation in the wall i s needed to create enough force* in the f r i c t i o n devices to cause slippage. For example, with the type and size of shearwalls and f r i c t i o n devices considered Jin the investigation, computation using Equation 6.4 shows that slippage could be in i t i a t e d with only a 0.325 mm (0.013 in) relative wall deflection. In an actual wall however, the framing members are not r i g i d . During a wall's deformation, the axial resistance from the devices causes i t s framing members to bend. Depending on the stiffness and size of the framing members and sheathing panels, size of the sheathing-to-framing connectors and their density, a certain amount of bending deformation must be reached before adequate forces can be developed to cause slippage of the devices. By the time slippage i s initiated, wall deflections would have reached a magnitude much higher than 0.325 mm (0.013 in). In fact, i t was shown in the racking and sta t i c c y c l i c tests that the devices did not actually s l i p u n t i l wall deformations in the order of 25.4 mm (1.0 in) were reached. In order for FRICWALL to more validly model the hysteretic behaviour of the f r i c t i o n devices, the equations which modelled their hysteretic behaviour should be modified to generate no hysteretic area u n t i l a certain wall deflection i s reached. The magnitude of this deflection would have to be determined from racking tests since i t depended on the stiffness and size of the framing members and sheathing panels, size of the sheathing-to-framing connectors and their density. Racking tests similar to 149 the ones carried out in this investigation could be used. With this modification to the hysteretic modelling of the f r i c t i o n devices, FRICWALL should be able to better reproduce the response time-history of a f r i c t i o n damped timber shearwall. Another less c r i t i c a l cause of discrepancy between the numerical and experimental results l i e s in the way the hysteretic behaviour of a f r i c t i o n damped timber shearwall i s modelled. FRICWALL derives i t s hysteretic behaviour from the superposition of those of a conventional wall and the f r i c t i o n devices. With the introduction of the devices, the framing no longer behaves li k e a configuration of pinned ended members. The four wall corners now have a moment capacity, thus the deformation pattern of this retrofitted wall would d i f f e r s l i g h t l y from i t s conventional counterpart. However, this difference i s not c r i t i c a l because the static c y c l i c tests have shown that the hysteretic behaviour of both types of wall are very similar. Therefore the use of the hysteretic behaviour of a conventional timber shearwall to arrive at the hysteretic behaviour of the f r i c t i o n damped timber shearwall i s quite v a l i d . 150 1940 EL CENTRO (N-S) @ 0.35 g P.G.A. 40 30 ^ 20 H z a S ° + LU S "10 CQ -20 -j -30 -40 -50 40 30 ^ 20 H z 10 H 0 rx a i co LU <g -10 CQ -20 -30 -40 -50 CONVENTIONAL WALL ENW1 -30 -10 10 30 RELATIVE LATERAL DEFORMATION (mm) 50 FRICTION DAMPED WALL (SLIP LOAD = 6.7 kN) EDW1 -30 -10 10 30 RELATIVE LATERAL DEFORMATION (mm) 50 Figure 7.18: Hysteresis Loops from Earthquake Simulation - E l Centro Earthquake § 0.35 g P.G.A. -120 S5 CD LU CO < CO 80 -40 -0 -40 --80 --120 -50 120 CC CO LU CO < CO 80 40 -\ 0 -40 --80 --120 -50 120 Z S X CO LU CO < CD 80 40 0 -40 -80 H -120 1940 EL CENTRO (N-S) @ 0.6 g P.GA CONVENTIONAL WALL ENW4 -30 -10 10 30 RELATIVE LATERAL DEFORMATION (mm) FRICTION DAMPED WALL (SLIP LOAD = 6.7 kN) EDW3 -30 -10 10 30 RELATIVE LATERAL DEFORMATION (mm) FRICTION DAMPED WALL (SUP LOAD = 13.4 kN) EDW5 -50 -30 -10 10 30 RELATIVE LATERAL DEFORMATION (mm) 50 50 50 Figure 7.19: Hysteresis Loops from Earthquake Simulation - E l Centro Earthquake @ 0.6 g P.G.A. -152 1977 ROMANIA (BUCHAREST, N-S) @ 0.2 g P.G.A. 20 _ 10 z 3 n X 0 CO UJ CO < CQ -10 H CONVENTIONAL WALL RNW1 -20 i i i i i i -16 -12 -8 -4 i i i i I I I I 0 4 8 12 16 RELATIVE LATERAL DEFORMATION (mm) 20 10 z I o CO UJ CO < CQ -10 H -20 FRICTION DAMPED WALL (SLIP LOAD = 6.7 kN) RDW1 ~ ~ i 1 1 I 1 i 1 I i 1 1 1 i 1 1— -16 -12 -8 -4 0 4 8 12 16 RELATIVE LATERAL DEFORMATION (mm) Figure 7.20: Hysteresis Loops from Earthquake Simulation - Romania Earthquake § 0.2 g P.G.A. -153 The load-deformation behaviour exhibited by the shearwalls of Figures 7.15, 7.16 and 7.17 are shown in Figures 7.18, 7.19 and 7.20, respectively. The f i r s t 6 seconds of the El Centro tests and the f u l l 16 seconds of the"Romania tests are shown in the figures. The load in the figures;.:represents the base shear of the wall while the deformation -corresponds to i t s relative lateral deformation. The base shear i s derived from the quantity (R1+R2) of Equation 6.12. - For small deflections, cos0 - 1, thus Equation 6.12 can be rearranged and rewritten as: Substituting these terms into Equation 7.1, the base shear at each time interval can be derived from the displacement and acceleration measurements using the relationship: - itlX K (7.1) Recall from geometry: therefore R1+R2 - J L [ wmA - hmA ] 2 (7.2) The hysteresis loops illustrated are typical of the other shearwalls tested during this investigation. They show that both types of shearwalls exhibit similarly inelastic hysteretic 154 behaviours, with the f r i c t i o n damped walls dissipating the required amount of seismic energy input at s l i g h t l y lower peak deflections than the conventional walls. Loads at the lower displacements are generally smaller relative to those at the higher d e f l e c t i o n s . T h i s i s caused by the pinching effect which i s typical of timber shearwalls. In comparison to the FRICWALL results, the differences in the amount of pinching between the two types of walls are not evident from these hysteresis loops. Since the contribution of the f r i c t i o n devices was quite limited, differences in the hysteretic behaviour of the two types of walls would not really be noticeable. The reductions in peak wall deflection computed by FRICWALL for the f r i c t i o n damped walls are also not reflected by these results. In the earlier comparisons of the wall's response time-histories, i t was shown that this discrepancy arises because the way in which the hysteretic behaviour of a f r i c t i o n damping device is modelled did not adequately reflect i t s behaviour in a shearwall. On some of the hysteresis loops, a few of the points seems to fluctuate e r r a t i c a l l y instead of changing smoothly with the displacement. To find the cause of these erratic points, the displacement and acceleration records from a l l of the tests were examined over their entire time-history. Each record showed a smooth transition throughout i t s time-history except for 6 top-of-wall acceleration records. The acceleration signals from these f i l e s were extremely noisy with many spikes, despite the high sampling rate of 100 points per second. Since the ve r t i c a l 155 or base shear coordinates of the hysteresis loops are a function of the accelerations, the erratic points are suspected of being caused by the spikes in the signal. An example of a smooth and a noisy record i s shown in Figure 7.21. The noisy records were fOTCind to produce hysteresis loops with drastic changes in loads between some of the displacements, while the smooth records were found to produce loops with a more gradual load fluctuation between displacements. Source of the disturbance in the acceleration signal may be from background noise in the data acquisition system or high frequency vibrations in the test frame being picked up by the sensitive accelerometers. Due to the inaccuracies in the acceleration readings, no comparisons of the peak top-of-wall accelerations between the two types of walls were made. 156 Figure 7.21: Acceleration Time-Histories from Earthquake Simulations 157 (10~3) 1.6 £ E 2 >• O C C L U 2 L U 0.4 (10 ~ 3) 1.6 E E 1.4 1.2 -1 -O 0.8 -J C C L U 2 0.6 -\ L U 0.4 0.2 i 0 1940 EL CENTRO (N-S) @ 0.35 g. P.GA -I N P U T E N E R G Y D E V I C E E N E R G Y CONVENTIONAL WALL ENW1 I 4 TIME (s) FRICTION DAMPED WALL EDW1 / (SUP LOAD » 6.7 kN) J TIME (s) F i g u r e 7 . 2 2 : E n e r g y T i m e - H i s t o r i e s f rom E a r t h q u a k e S i m u l a t i o n - E l C e n t r o E a r t h q u a k e @ 0 . 3 5 g P . G . A . -158 (10~3) e 1940 EL CENTRO (N-S) @ 0.6 g. P.G.A. O I «" 1 INPUT ENERGY ~ ' — DEVICE ENERGY I 3 " O 2 -rr LU CONVENTIONAL WALL ENW4 * w> . / 0 i c (10~3) s, .. i i I 2 i i i i 4 6 TIME (s) 3| 4-Z T -O 2 -cn LU o -I FRICTION DAMPED WALL EDW3 (SLIP LOAD = 6.7 kN) '" C (10~3) 5 -I i l l 2 4 e TIME (s) ! 4 -1 3 -5 2 -CC 2 1 -LU FRICTION DAMPED WALL EDW5 (SLIP LOAD = 13.4 kN) 0 i 0 i i 2 i i i 4 € TIME (s) i Figure 7.23: Energy Time-Histories from Earthquake Simulation - E l Centro Earthquake @ 0.6 g P.G.A. -159 (10~3) 0.2 E E i z & 0.1 H >• o CC LU z LU (10~3) 0.2 E E z >-o CC UJ z UJ 0.1 -1977 ROMANIA (BUCHAREST, N-S) @ 0.2 g. P.G.A. INPUT ENERGY DEVICE ENERGY CONVNETIONAL WALL RNW1 FRICTION DAMPED WALL RDW1 (SLIP LOAD = 6.7 kN) 1—n r 0 - — i — i — r - * i — i — i — i — i — i — i — i — i — i — i — r 0 2 4 6 8 10 12 14 16 TIME (s) i r \ r 6 8 10 TIME (s) — r — r 12 14 16 Figure 7.24: Energy Time-Histories from Earthquake Simulation - Romania Earthquake § 0.2 g P.G.A. -The time-histories of the seismic energy input and the energy dissipated by the f r i c t i o n damping devices for the shearwalls of Figures 7.15, .7.16 and 7.17 are shown in Figures 7.22, 7.23 and 7.24. The time-histories shown in these figures are representative of those;;;from the other shearwalls tested under the simulated seismic loads. Table 7.9 summarizes these energy quantities for a l l the shearwalls tested. The percentage of seismic input energy dissipated by the f r i c t i o n devices in each f r i c t i o n damped shearwall i s also shown in Table 7.9. Table 7 . 9 : Summary of Energy Time-Histories from Earthquake Simulations ENERGY (kN-mm[kip-in]) ACCELERATION RECORD CONVENTIONAL TIMBER SHEARWALL SPECIMEN FRICTION DAMPED TIMBER SHEARWALL SPECIMEN SUP LOAD - 6.7 kN(1.5 kip) INPUT ENERGY DISSIPATED BY FRICTION DEVICES [%] ROMANIA (P.GA = 0.2 g) WALL INPUT I WALL ENERGY j WALL DEVICE ENERGY INPUT ENERGY 0.0 0.0 RNW1 RNW2 37.4 (0.3) 39.4 (0.4) RDW2 RDW1 0.0 (0) 0.1 (0) 30.1 (0.3) 40.6 (0.4) AVERAGE 38.4 (0.3) 0.1 (0) 35.4 (0.3) 0.0 EL CENTRO (P.GA - 0.35 g) ENW1 ENW2 ENW3 702.0 (6.2) 1250.2 (11.1) 949.8 (8.4) EDW1 EDW2 107.4 (1.0) 144.9 (1.3) 1582.9 (14.0) 1067.7 (9.5) 6.8 13.6 AVERAGE 1100.0 (9.7) 126.2 (1.1) 1325.3 (11.7) 10.2 EL CENTRO (P.GA - 0.6 g) AVERAGE ENW4 ENW5 3278.8 (29.0) 3293.9 (29.1) EDW3 EDW4 682.4 (6.0) 485.6 (4.3) 3859.3 (34.1) 2882.6 (25.5) 17.7 16.9 3286.4 (29.1) 584.0 (5.2) 3371.0 (29.8) 17.3 EDW5 EDW6 249.2 (2.2) 365.8 (3.2) 2843.0 (25.2) 3261.6(28.9) 8.8 11.2 AVERAGE 307.5 (2.7) 3052.3 (27.0) 10.0 * SUP LOAD - 13.4 kN(3.0 kip) The seismic energy input was computed by multiplying the i n e r t i a l mass with the average ground acceleration and relative displacement of the wall in each time interval. The energy dissipated by the devices was derived from the summation of the product of each device's average axial load and slippage during each time interval. Each energy term calculated from the time intervals was cumulatively summed to produce their time-histories. From Table 7.9, the amount of seismic energy dissipated by the f r i c t i o n damping devices are shown to range from a high of 17.3 % to.a low of 0 % for the El Centro simulation at 0.6 gand the Romania simulation at 0.2 g peak ground acceleration, respectively. The s l i p load for both cases i s 6.7 kN (1.5 kip) . For a l l the f r i c t i o n damped shearwalls tested, an average energy dissipation of 9.4 % was provided by the f r i c t i o n devices. With the devices contributing very l i t t l e to the energy dissipation of a shearwall, the seismic response of both types of walls were quite similar. This alikeness resulted in both types of walls attracting comparable levels of seismic energy inputs. These energy inputs are much lower that the quantities calculated from FRICWALL because the shearwall deflections observed in the tests were much lower than those predicted. Visual inspection of the shearwalls during the tests show that the behaviour of both system was about the same. In a l l of the walls, e l a s t i c out-of-plane buckling of both plywood panels, warping of the interior vertical studs and damage to the wood around the sheathing-to-framing connectors were observed. Wood 162 damage around the connectors were limited to those at the perimeter of each plywood panel, with the wood at the corners of each panel suffering the more severe destruction. For the f u l l scale E l Centro and Romania earthquake simulations, no damages were found in either wall system at the end: of each test. For the 0.6 g peak ground acceleration E l Centro earthquake simulation, similar damage was notices in both wall systems. In the f i r s t conventional wall, a few perimeter nails in each sheathing were sheared off along one ve r t i c a l stud; more damage was evident in the second wall, where a l l the nails in one sheathing sheared off along one stud. In the f r i c t i o n damped walls, the damage ranged from none to a l l the nails in one sheathing shearing off completely along one stud. It was shown earlier that bending in the framing members prevented the f r i c t i o n devices from contributing to the energy dissipation of a shearwall u n t i l wall deflections in the order of 25.4 mm (1.0 in) were reached. During the tests of the f r i c t i o n damped walls, quite a b i t of relative motion between the sheathing-to-framing connectors and the plywood panel was observed at points of load transfer between the struts of the devices and the framing. Because force transfer between the struts and the framing were mainly done via these connectors, they were subjected to very high stresses. The higher stresses increased the amount of wood crushing in these regions, causing play between the framing members and the sheathing panels. Consequently, for subsequent displacement cycles, the wall was forced to undergo further deflections before the devices could 163 develop the necessary s l i p load to become effective. Due to a shearwall's modest response during the majority portion of an earthquake's time-history., deflections causing slippage of the devices were only reached at periods of peak or near-peak excitation. This slack: in the connections along with the bending in the end studs offset the benefits of the devices, thus reductions in peak wall deflection observed in the f r i c t i o n damped walls were not as high as expected. 164 7.6 Summary The following observations were made from the shake table tests conducted on the f u l l scale conventional and f r i c t i o n damped timber shearwalls. The f r i c t i o n damped timber, shearwalls exhibited s t i f f e r in-plane behaviour and sustained higher racking loads than the conventional timber .. shearwalls. They failed in a ductile manner, unlike the b r i t t l e failure in the conventional timber shearwalls. -'.-Both the conventional and the f r i c t i o n damped timber shearwalls exhibited pinched hysteretic behaviour. But the pinching in the f r i c t i o n damped walls was sli g h t l y less, thus the energy dissipation in these walls was better than in the conventional walls. The pinched hysteretic behaviour makes both wall systems equally inefficient when i t comes to energy dissipation. - When retrofitted with f r i c t i o n damping devices, the seismic performance of a timber shearwall was only marginally improved. Other than a marginally lower peak wall deflection, the seismic response of the f r i c t i o n damped walls basically followed those of the conventional wall. The level of damage in both wall systems after the same simulated seismic event were similar. SADT's estimate of a timber shearwall's load-165 deformation behaviour was in agreement with experimental results. - FRICWALL's prediction of the seismic response of a conventional timber shearwall was in general agreement with experimental results. However,.it overestimated a f r i c t i o n damping device's a b i l i t y to improve the seismic response of a timber shearwall. These experiments divulged that the f r i c t i o n devices were unable to contribute to the energy dissipation of a shearwall un t i l wall deflections in the order of 25.4 mm (1.0 in) were reached. This delay was caused by bending in the framing members and play in the connections which transferred load between the devices and framing members. These problems offset the f u l l benefit of the devices, thus limiting their a b i l i t y to improve the seismic response of a timber shearwalls. 166 8 . CONCLUSION 8 . l Summary and Conclusions This thesis was intended to be a preliminary investigation into the potential application of the f r i c t i o n damping concept to wood structures for improving their seismic behaviour. The behaviour of a conventional and a f r i c t i o n damped timber shearwall under lateral load was discussed. Using two numerical models, SADT and FRICWALL, the racking and seismic performance of the two types of shearwalls were predicted. Prototypes of the f r i c t i o n damping devices were tested under cyclic loads to examine their hysteretic behaviour and i t s repeatability. Full scale versions of conventional and f r i c t i o n damped timber shearwalls were tested under racking, static c y c l i c and simulated dynamic earthquake loads on the shake table. Results from these tests were used to compare the racking, hysteretic and seismic behaviour of the two types of walls. The racking and static c y c l i c tests of the conventional walls were also used to establish the wall's hysteretic parameters for use in the FRICWALL program. Finally, the v a l i d i t y of the numerical models were checked through comparison with the experimental results. The findings made from this preliminary investigation are as follows: 1. The SADT analyses showed that the racking performance of a f r i c t i o n damped timber shearwall was superior to that of a conventional timber shearwall. It exhibited higher in-plane stiffness and sustained a higher racking load 167 before failure. 2. Cyclic loading of the f r i c t i o n damping devices under the sinusoidal frequencies found in typical earthquakes showed that they exhibited rectangular and very stable hysteretic behaviour .~ 3. The FRICWALL analyses showed that the optimal s l i p load for the f r i c t i o n damping devices considered in this investigation was 6.7 kN (1.5 kip). 4. The FRICWALL analyses showed that a f r i c t i o n damped timber shearwall could serve as a very practical and effective lateral load resisting system because i t s seismic response was superior over that of a conventional timber shearwall. 5. Full scale racking tests showed that the f r i c t i o n damped timber shearwalls were stronger and s t i f f e r than the conventional ones. These characteristics provided the walls with higher energy dissipation at failure. The failure was ductile and a sizeable portion of the racking resistance was retained at post-peak load deflections. 6. Pinched hysteretic behaviour was exhibited by both types of shearwalls. But the amount of pinching i s less in the f r i c t i o n damped walls, thus they were able to dissipate more energy. However, both types of walls were equally inefficient energy dissipating systems. 7. The seismic performance of f r i c t i o n damped timber shearwalls was only marginally better than that of the conventional walls. Other than a marginally lower peak 168 deflection, the response of the f r i c t i o n damped wall basically followed that of the conventional wall. 8. SADT provided a good prediction of the load-deformation behaviour of timber shearwalls. 9. FRICWALL provided a good estimate of the response time-history of a conventional timber shearwall, but i t overestimated the a b i l i t y of the f r i c t i o n damping devices to improve the seismic response of a timber shearwall. 8.2 Future Research The work presented in this research was carried out as a preliminary study into the potential application of the f r i c t i o n damping concept to timber structures. Although the f r i c t i o n damping devices did not improve the seismic response of a timber shearwall to the extend predicted by the numerical computations, they did offer some limited improvement. Through further research into the f r i c t i o n damping concept, and modifications of the f r i c t i o n devices, a f r i c t i o n damping system which may better improve the seismic performance of timber shearwalls could be developed. 169 BIBLIOGRAPHY ASTM, "ASTM E72-77 Standard Methods of Conducting Strength Tests of Panels for Building Construction". Annual Book of ASTM Standards: pp. 690-701, 1986. ASTM, "ASTM E564-76 Standard Methods of Static Load Test for Shear Resistance of Framed Walls for Buildings". Annual Book of ASTM Standards: pp. 1013-1017, 1986. Breyer, D.E., "Design of Wood Structures - Second Edition". McGraw-Hill Book Company, USA, 1988. CISC, "Handbook of Steel Construction". Universal Offset Limited, Markham, Canada, 1986. Clough, R.W. and Penzien, J., "Dynamics of Structures". McGraw-H i l l Book Company, USA, 1975. Canadian Wood Council, "Nails, Spikes and Staples". CWC Datafile WJ-2. Ottawa, Canada, 1982. Dean, J.A., Deam, B.L. and Buchanan, A.H., "Earthquake Resistance of Timber Structures". NZ Journal of Timber Construction, Vol. 5, No. 2, pp. 12-16, 1989. Dolan, J.D. "The Dynamic Response of Timber Shear Walls". Ph.D. Thesis, University of British Columbia, Department of C i v i l Engineering, Vancouver, B.C., Canada, 1989. Dowrick, D.J., "Earthquake Resistant Design for Engineers and Architects - Second Edition". John Wiley & Sons, Inc., 1987. Dowrick, D.J., "Hysteresis Loops for Timber Structures". Bulletin of the New Zealand National Society for Earthquake Engineering, Vol. 19, No. 2, pp. 143-152, 1986. Dowrick, D.J. and Smith, P.C., "Timber Sheathed Walls for Wind and Earthquake Resistance". Bulletin of the New Zealand National Society for Earthquake Engineering, Vol. 19, No. 2, pp. 123-134, 1986. F i l i a t r a u l t , A., "Analytical Predictions of the Seismic Response of Friction Damped Timber Shear Walls". Earthquake Engineering and Structural Dynamics, Vol. 19, pp. 259-273, 1990. F i l i a t r a u l t , A., " C i v i l 519 - Earthquake Engineering". C i v i l Engineering Course Notes, University of British Columbia, Department of C i v i l Engineering, Vancouver, B.C., Canada, 1989. 170 F i l i a t r a u l t , A., "Performance Evaluation of Friction Damped Braced Steel Frames Under Simulated Earthquake Loads". M.A.Sc. Thesis, University of British Columbia, Department of C i v i l Engineering, Vancouver, B.C., Canada, 1985. Foschi, R.O., "Analysis of Wood Diaphragms and Trusses. Part I: Diaphragms". Canadian Journal of C i v i l Engineering, Vol. 4, No. 3, pp. 345-352, 1977. Foschi, R.O., "Performance Evaluation of Shear Walls and Diaphragms with Waferboard Sheathing". Report to the Canadian Waferboard Association by Forintek Canada Corp., Vancouver, Canada, 1982. Foschi, R.O., "Performance Evaluation of Shear Walls with CDX Plywood Sheathing". Report to the Canadian Waferboard Association by Forintek Canada Corp., Vancouver, Canada, 1982. Foschi, R.O., Folz, B.R. and Yao, F.Z., "Reliability-Based Design of Wood Structures". F i r s t Folio Printing Corp. Ltd., Vancouver, Canada, 1989. Keenan, F.J., "Limit States Design of Wood Structures". Morrison Hershfield Limited, Toronto, Canada, pp. 329-339, 1986. National Research Council of Canada, "National Building Code of Canada", NRC No. 23174F, Ottawa, Canada, 1985. Smith, P.C, Dowrick, D.J. and Dean, J.A., "Horizontal Timber Diaphragms for Wind and Earthquake Resistance". Bulletin of the New Zealand National Society for Earthquake Engineering, Vol. 19, No. 2, pp. 135-142, 1986. Yung, W.C.W. and F i l i a t r a u l t , A. "Innovative Energy Dissipating System for Earthquake Design and Retrofit of Timber Structures". Proceedings from the 1989 CSCE Annual Conference, 1989. 

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:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0062628/manifest

Comment

Related Items