Electronic Thesis and Dissertation Repository

Degree

Doctor of Philosophy

Program

Civil and Environmental Engineering

Supervisor

Hanping Hong

Abstract

Approaches to simulate single- and multiple-component multiple-station ground motion records with target spatial coherency and spatial correlation structures are developed for scenario events in this study. To develop the approaches, spatial correlation of the Fourier amplitude spectrum for a random orientation and for two orthogonal directions is assessed, and empirical equations are suggested. Moreover, spatial coherency of ground motions for two orthogonal horizontal directions is estimated using actual records from seven seismic events. Empirical coherency function is suggested for the components of records in two orthogonal horizontal directions at single and multiple recording stations. It was also found that the coherency for the records along the major and minor principal axes at a recording station is similar to that for two randomly oriented orthogonal directions.

Based on the proposed approaches in this thesis, spatial correlated and coherent ground motions can be simulated for randomly oriented uni-directional excitations at considered sites for a scenario event. For the simulation, it is considered that the reference Fourier amplitude spectrum for scenario events can be defined by using the stochastic point-source method or the stochastic finite-fault method. It is shown that the estimated spatial correlation and coherency from the simulated records match well the target spatial correlation and coherency. Furthermore, the application of the simulated records for seismic risk assessment of a group of buildings is presented. The results indicate that the spatial correlation of Fourier amplitude spectrum must be considered in estimating the distribution of the aggregated seismic loss of spatially distributed group of buildings.

Spatial correlated and coherent ground motions are simulated for two orthogonal horizontal directions at considered sites for a scenario event. Again, it is shown that the estimated spatial correlation and coherency from the simulated records adequately match the target spatial correlation and coherency.

Share

COinS