Analytical Framework for Evaluating Wind Hazard Vulnerability of Light Wood-Frame Buildings

Abstract

Light wood-frame construction constitutes most of the residential building inventory in the United States. The vulnerability of wood-frame buildings to wind hazards and a wide variability in observed performance have motivated research efforts toward understanding precisely how and why wood-frame structural failures occur and how wind damage can be effectively and economically mitigated in future events. Probabilistic analysis of wind performance has been an important strand within these efforts. The wind performance of a structure is essentially a function of structural resistance and wind load intensity, with sources of epistemic and stochastic uncertainty present in both resistance and load. This dissertation aims at a systematic examination of structural and aerodynamic factors that contribute uncertainty to wood-frame wind performance. The analysis framework is composed of two complementary models. (1) A structural vulnerability model evaluates wind uplift resistance by a Monte Carlo method in which probability distributions are assigned to factors like framing member spacing, wood species, and fastener measurements. (2) An aerodynamic vulnerability model, composed of regression neural networks trained on wind tunnel data, evaluates wind uplift intensity in terms of factors like roof shape, roof slope, and wind direction. Separately, these models provide a rational basis for comparing (1) system-level resistance across different wood-frame load paths and (2) expected wind uplift intensities across different low-rise geometries. The third element of the framework is a sensitivity analysis that operates on the models independently — to examine how structural and aerodynamic factors respectively influence resistance and load — as well as on a combined model that allows uncertainty in wind performance to be traced back to fundamental building characteristics. The framework lends itself to a range of applications, including wind-resistant design and wind speed estimation.