THREE ­DIMENSIONAL MODELING OF WOOD MOMENT CONNECTIONS

ABSTRACT

Small-scale wooden moment connections were tested under cyclic loading. These tests investigated the influence of the addition of fiber reinforcement and dowel spacing on the performance of the connections. A three-dimensional finite element model was developed as representative of these small-scale beam-column connections utilizing the finite element software ANSYS. Special emphasis was placed upon the definition of the wood material model utilized. An elasto-plastic anisotropic material model was defined, which was extended into transverse isotropy. Material testing in the form of dowel bearing, compression and tension tests was performed in order to determine the material parameters. The usage of a “foundation” model was assumed in order to more accurately represent the complex behavior of wood in compression. This “foundation” model was used to more accurately predict the non-recoverable deformation behavior, or “crushing” which occurs in the areas surrounding the dowels during loading. Material dowel bearing tests were used to determine the properties for this model. A general model utilizing the material properties determined from the material compression tests was used to define the material model in the remainder of the connection model. The three-dimensional finite element model was programmed to undergo the same cyclic loading as the original laboratory tests. Hysteretic loops and supporting data from the laboratory tests of the moment connections were used to validate the results of the computational finite element simulations. The results and shortcomings of this study are discussed in detail, and suggestions for improvements and further study presented.      

TABLE OF CONTENTS

List of Figures ……………………………………………………………………………………………………………… viii List of Tables ……………………………………………………………………………………………………………… xvii Acronymns ………………………………………………………………………………………………………………….. xix Nomenclature ……………………………………………………………………………………………………………….. xx Acknowledgements ……………………………………………………………………………………………………… xxv

1. Introduction …………………………………………………………………………………………………………….. 1
   • Background ………………………………………………………………………………………………………….. 1
   • Objective and Scope ……………………………………………………………………………………………… 2
2. Review of Background Theory and Existing Literature ………………………………………………… 3
   • Behavior of Wood Connections with Dowel-Type Fasteners ……………………………………… 3
   • Behavior under Cyclic Loading ………………………………………………………………………………. 5
   • Improving the Strength of Connections ……………………………………………………………………. 6
   • Modeling of Timber Connections ……………………………………………………………………………. 9
     • 2-Dimensional …………………………………………………………………………………………….. 9
     • 3-Dimensional …………………………………………………………………………………………… 11

3. Testing of Moment Connections ………………………………………………………………………………. 20
   • Test Methodology ……………………………………………………………………………………………….. 20
   • Analysis Methodology …………………………………………………………………………………………. 22
   • Results ……………………………………………………………………………………………………………….. 24
     • Hysteresis Loops ……………………………………………………………………………………….. 24
     • Cumulative Energy Dissipation …………………………………………………………………… 28
     • Failure ……………………………………………………………………………………………………… 30
     • Summary of Results …………………………………………………………………………………… 36

4. Testing of Material …………………………………………………………………………………………………. 39
   • Previously Completed Material Tests …………………………………………………………………….. 39
   • Further Material Tests ………………………………………………………………………………………….. 41
     • Tension Tests (Perpendicular to Grain) ………………………………………………………… 42
     • Compression Tests …………………………………………………………………………………….. 49
   • Results and Discussion ………………………………………………………………………………………… 54
     • Dowel Bearing Tests ………………………………………………………………………………….. 54
     • Tension Tests (Perpendicular to Grain) ………………………………………………………… 61
     • Compression Tests …………………………………………………………………………………….. 66

5. Computational Modeling Procedures ……………………………………………………………………….. 72
   • Modeling Procedure …………………………………………………………………………………………….. 73
     • …………………………………………………………………………………………………. 73
     • Contact …………………………………………………………………………………………………….. 76
     • Boundary Conditions …………………………………………………………………………………. 79
     • Loading ……………………………………………………………………………………………………. 81
   • Material Properties ………………………………………………………………………………………………. 82
     • Wood ……………………………………………………………………………………………………….. 82
     • ………………………………………………………………………………………………………… 82
     • Composite Fiber Reinforcement ………………………………………………………………….. 83
   • Previous Finite Element Model …………………………………………………………………………….. 83
   • Computational Methods ……………………………………………………………………………………….. 85
     • Nonlinear Static Analysis …………………………………………………………………………… 85
     • Nonlinear Transient Analysis ……………………………………………………………………… 87
     • Contact Algorithm …………………………………………………………………………………….. 88
   • Post-Processing of Model Solution ………………………………………………………………………… 88
     • Rotation ……………………………………………………………………………………………………. 89
     • Moment ……………………………………………………………………………………………………. 89
     • Energy Dissipation …………………………………………………………………………………….. 89
     • Failure Criteria ………………………………………………………………………………………….. 89

6. Development and Validation of Material Model ………………………………………………………… 91
   • Transverse Isotropic Model ………………………………………………………………………………….. 94
     • Anisotropic Plasticity …………………………………………………………………………………. 94
   • Material Parameters …………………………………………………………………………………………….. 97
     • General Material Model (GMM) ……………………………………………………………….. 101
     • Foundation Material Model (FMM) …………………………………………………………… 103
   • Validation …………………………………………………………………………………………………………. 106
     • Replication of Compression Tests ……………………………………………………………… 106
     • Replication of Dowel Bearing Tests …………………………………………………………… 110

7. Results and Discussion …………………………………………………………………………………………. 121
   • Further Validation Procedures …………………………………………………………………………….. 121
   • Model Results …………………………………………………………………………………………………… 126
     • Unreinforced 6.35 mm dowels using the FMM ……………………………………………. 126
     • Unreinforced 6.35 mm dowels using the GMM only ……………………………………. 139
     • Unreinforced 9.53 mm dowels …………………………………………………………………… 144
     • Reinforced Models (6.35 mm and 9.53 mm dowels)…………………………………….. 156
   • Factors Influencing Model Results ………………………………………………………………………. 156
     • Contact …………………………………………………………………………………………………… 156
     • Wood Material Model ………………………………………………………………………………. 158
     • ANSYS Solution Settings …………………………………………………………………………… 159
     • Boundary Conditions ……………………………………………………………………………….. 160
     • Applied Loading ……………………………………………………………………………………… 160
     • Potential Sources of Unquantifiable Error …………………………………………………… 162
     • Other ……………………………………………………………………………………………………… 163

8. Conclusions and Recommendations ……………………………………………………………………….. 164
   • Summary of Research ………………………………………………………………………………………… 164
   • Summary of Results …………………………………………………………………………………………… 164
     • Connection Tests ……………………………………………………………………………………… 164
     • Material Tests ………………………………………………………………………………………….. 165
     • 3D Finite Element Connection Model ………………………………………………………… 166
   • Scientific Contribution of Research ……………………………………………………………………… 167
   • Proposed Improvements and Future Research ……………………………………………………….. 168

Bibliography ………………………………………………………………………………………………………………. 170 Appendix A: Connection Specifications …………………………………………………………………………. 175 Appendix B: Connection Hysteresis Curves……………………………………………………………………. 178

1. INTRODUCTION

1.1.       Background

The need for improved performance of residential structures under seismic loading can be demonstrated by statistics set forth by Ayoub (2007). One of the most seismically volatile regions in North America is California, where 99% of residences are constructed of wood framing. After the 1994 Northridge earthquake, the Los Angeles housing department disclosed that 90% of housing in the affected areas, a figure of about 330 000, were damaged wood frame construction, with an estimated overall property loss of $20 million. It is also noted that the fraction of wood structures to total structures throughout the US is between 80 to 90%. Wood structures are particularly suited to earthquake loads due to the potential for high ductility of the frame connections, and a high strength to weight ratio as compared with concrete or steel construction. Wood frames can thus undergo significant deformations whilst retaining a large percentage of the original structural strength, and allow a considerable dissipation of energy. Heavy timber frames also have the benefit of a high fire resistance which is beneficial in areas of high seismicity. The influence of reinforcement on the performance of both wood connections and structures has been explored by many researchers, using several different techniques and variables. The main reasons for the use of reinforcement include the improvement of the connection strength and ductility, as well as the prevention of sudden brittle failure. These factors are all of high importance in the usage of wood structures in seismic areas. The current standards allow the reinforcement of connections using steel plates, which improve strength, but largely ignore factors such as ductility and moment carrying capacity, which are crucial in a seismic event. Recent research has focused upon the use of fiber textiles and fiber reinforced polymer rods for the purposes of reinforcement. However, there is no generally accepted usage of these materials in seismic areas. The use of finite element models in the prediction of connection behavior under cyclic loading is a method used in an attempt to better understand the action of the connections, and a move towards the eradication of impractical and expensive tests upon connections and structures of innumerable configurations. Again, several methods of modeling have been proposed, with varying results. The use of different factors in most of the models developed has meant that, once again, no generally accepted standard exists for the depiction of structural applications in wood. Furthermore, research on finite element models of wood connections in three dimensions is limited. There is no generally accepted method of modeling connections or of depicting material behavior. To the best of this author’s knowledge there are no recognized three-dimensional finite element studies of cyclic behavior upon wood connections to date.

1.2.       Objective and Scope

The objective of the this research is to update and improve upon existing finite element models of wood connections, and develop a more accurate representation of connection behavior under cyclic loading, in order to better predict the performance of the connections, and wood structures by extension. Previously completed tests upon small-scale wood connections using dowel-type fasteners, with select connections reinforced with glass fiber rods are used as a basis for the finite element models developed. These experiments were completed at The Pennsylvania State University in 2010, and consisted of two sets of connections of differing dowel diameter, with both reinforced and unreinforced configurations. Material tests were completed in order to define a working material model for application to the computational model. The results from the cyclic loading of these connections are used as a form of validation for the equivalent finite element models. The research herein aims to present a working three-dimensional finite element model of small-scale wood connections, with and without added reinforcement. The application and development of an accurate material model is another main focus of this research, as well as the usage of this material model in predicting load-displacement behavior under cyclic loading. THREE ­DIMENSIONAL MODELING OF WOOD MOMENT CONNECTIONS