Interdependence between damage indices and ground motion parameters based on Hilbert–Huang transform

Abstract

Feature extraction from seismic accelerograms is a key issue in characterization of earthquake damage in structures. Until today, a number of effective classical parameters such as peak ground acceleration (PGA) and Arias intensity have been proposed for analyzing the earthquake motion records. The aim of this paper is to search for new crucial characteristic seismic parameters which provide information pertinent to the damage indicators of the structures. The first proposed parameter is the maximum amplitude (AHHT_max) and the second is the mean amplitude (AHHT_mean). Emphasis of our work has been placed on the use of the Hilbert–Huang transform (HHT). A set of 13 natural accelerograms from worldwide well-known sites with strong seismic activity have been used. The HHT has been applied to the nonlinear and non-stationary data (earthquake recordings). Each complex seismic accelerogram is decomposed into several simple components called intrinsic mode functions (IMFs). Using the IMFs a three-dimensional time–frequency distribution of earthquake excitation is computed and two new seismic parameters are proposed and evaluated. After the numerical computation of all the seismic parameters (classical and proposed), nonlinear dynamic analysis is carried out to provide the post-seismic damage status of the structure under study. Two structural damage indices are utilized and the degree of interrelation among them and the seismic parameters is provided by correlation coefficients. Furthermore, two different reinforced concrete structures are examined. Results indicate the high correlation of the new seismic parameters (AHHT_max, AHHT_mean) with the damage indices and confirm that HHT is a promising tool for extracting information to characterize damage in structures.