PROPAGATION OF UNCERTAINTY IN LIGHT-FRAME WOOD BUILDINGS

Light-frame wood construction is the most common form of construction for residential and low-rise commercial buildings (e.g. hotels and motels) in North America. In these buildings, shearwalls are the primary systems resisting lateral loads induced by earthquakes and winds. Past earthquake events have revealed that the structural performance of nominally identical light-frame wood buildings varied significantly. Although much research has been conducted investigating the uncertainty in the performance of wood shearwalls and buildings under earthquake loading, the focus of many of these studies was on the uncertainty due to the earthquake motions, referred as earthquake-to-earthquake uncertainty, the influences of variability in the lateral capacities of shearwalls on the seismic performance of light-frame wood buildings remain largely unexplored. Thus, primary focus of this thesis was to develop a framework to quantify the influence of wall-to-wall variability on the seismic performance of light-frame wood buildings. The inherent variability in the lateral capacity in light-frame wood shearwalls is largely attributed to the variability of the material properties of the wood and the fasteners or connections that connect parts of the walls together. To investigate the inherent uncertainty in light-frame wood buildings and shearwalls, a series of experimental connection tests on sheathing nails, framing nails and hold-downs were conducted. As part of this study, a new numerical model (M-CASHEW) which can be used to predict accurately the lateral responses of light-frame wood shearwalls was also developed. The data from the connection tests were utilized to model the various types of wood shearwall configurations commonly used in light-frame wood construction in the M-CASHEW program. The results were then compared with the experimental data from wood shearwall tests. Good agreements were observed between the model predictions and the test results. Two probabilistic connection models were developed and implemented in the M-CASHEW program to simulate and quantify the variability in the lateral responses of nominally identical wood shearwalls using a direct Monte Carlo simulation approach. In these two probabilistic connection models, the sheathing-to-framing, frame-to-frame and hold-down connection parameters were randomly generated based on the connection test data. The first probabilistic connection model considered no correlation among the connection parameters while the second specifically considered the correlation among the parameters using a distribution-free method that utilizes the Cholesky decomposition of the test data correlation matrix to simulate correlated connection parameters. Finally, based on the direct Monte Carlo simulation results, a new simplified shearwall simulation approach, which is more computationally efficient than the direct Monte Carlo simulation, was proposed. In the final phase of this study, the seismic response of a two-story light-frame wood building, the two-story CUREE shake table test structure, was modeled using a specialized structural analysis program developed for light-frame wood buildings. A methodology to simulate the stochastic response of light-frame wood buildings under earthquake loading was developed using the new simplified shearwall simulation model. This new stochastic building model can be used to determine the minimum number of realizations of a building needed in a dynamic time history analysis to obtain a reliable estimate of the peak drift distribution under earthquake loading