Doctor of Philosophy (PhD)


Civil and Environmental Engineering

Document Type



Light-frame wood buildings account for over 95% of all residential structures in the U.S, of which the majority are designed as low-rise buildings. These low-rise residential buildings in the U.S. have performed unsatisfactorily and are the largest source of the damage and fatality during the past extreme wind events. To deepen the understanding and reduce the vulnerability of the infrastructures, the accurate prediction of the hurricane loss has been an urgent need, and the hurricane catastrophe models are developed in response. However, the current hurricane catastrophe models are focused on the economy loss estimation rather than investigating the root causes of structural failures that only little or empirical structural analysis is involved. Thus, these models cannot reveal the realistic load paths nor the stage-wise damage propagation. This dissertation aims to develop a validated finite-element (FE) modeling frame work for predicting the system nonlinear performance of low-rise buildings under the spatiotemporally varying wind loads with the reasonable accuracy. This framework would serve for the successive damage prediction as a part of the risk assessment of low-rise buildings under extreme wind events.

To reach the final objective, a refined 3D modeling methodology is proposed first. This modeling methodology contributes to combine the strengths of each involved disciplines to achieve a desired resolution, i.e., the dynamic form of wind loads, the full-scale scope of modeling, and the extensive nonlinear representative of the critical components. It is validated by a large-scale wind test from the linear to the nonlinear range including the successive failure stages. This modeling methodology provides the foundation for the future research.

Secondly, a progressive failure prediction methodology aiming at finding the quantitative relationship between the wind intensity and the damage state of the building is well developed with an explicit explanation on the failure mode, the failure location, and the failure criteria. This methodology is also validated in the building scale and the individual connection scale by a corresponding destructive wind test with the agreement on the failure mode and sequence. Meanwhile, the database-assisted design (DAD) technique is extended from its original application on the linear prediction on the frames of the metal structure to the nonlinear modeling on the envelope of the wood structure in the current study.

This framework that consists of the building modeling and the failure prediction provides a guideline on the three crucial steps for a more accurate performance prediction: directly using the aerodynamic database derived from wind tests, applying the loads onto a refined building model considering the nonlinear behaviors of critical components, and conducting the analysis on the progressive failure process. Then, some attempts are made on the application of this framework, e.g., the effect of the geometric and loading forms on the load paths and structural failure is discussed, and the adequacy of the wind design by using the ASCE 7-10 wind provisions on residential buildings is evaluated.



Committee Chair

Cai, Steve