Treatment of geoseismic data as a non-stationary process

摘要

The traditional Fourier-spectra-based tools do not adequately capture the evolutionary and localized features of the natural-systems responses. When these systems are subjected to seismic loads the Fourier analysis may misinterpret the information due to the time-variation of frequency characteristics in non-stationary processes. This study explores the use of the Hilbert-Huang Transform for analyzing earthquake recordings and the associated dynamic-soil behavior. The HHT, integrated by the Empirical Mode Decomposition and the Hilbert Transformation, is an empirical based data-analysis method with an adaptive basis of expansion that can produce physically meaningful representations of data from nonlinear and non-stationary processes. Hilbert-Huang Transform enables engineers to analyze non-stationary oscillation systems and to obtain more detailed intensity descriptions on time-varying frequency diagrams. HHT is used in this work to examine responses of soft-soils deposits in Mexico City. The results indicate that the proposed methodology is able to extract some motion characteristics useful in geoseismic studies which are not properly seen when are employed conventional data processing techniques.