Wavelet analysis techniques for ground motion simulation and modification

Graduate researcher: Yoshi Yamamoto

Software relating to this project is provided here.

The goal of this research is to develop a stochastic method to generate artificial ground motions with time and frequency non-stationarities using wavelet analysis. These artificial ground motions are modeled by two types of probability distributions of wavelet packet coefficients, which correspond to large- and small-amplitude portions of the ground motions. The parameters of this model are then estimated using regression analysis as a function of magnitude, distance, and site conditions. Using this regression model, we can generate the artificial ground motions for arbitrary conditions. Further, because this simulation procedure is so computationally inexpensive, it can be combined with a seismic source model to provide a direct simulation method that circumvents traditional Probabilistic Seismic Hazard Analysis and selection of recorded ground motions.

Selected publications:

Yamamoto, Y. (2011). Stochastic model for earthquake ground motion using wavelet packets. Ph.D. Thesis. Dept. of Civil and Environmental Engineering, Stanford University, Stanford, CA. 295p.