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 (A HHT max ) and the second is the mean amplitude (A HHT 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 (A HHT max , A HHT mean ) with the damage indices and confirm that HHT is a promising tool for extracting information to characterize damage in structures.

No abstract

The Hilbert–Huang transform is used to generate artificial seismic signals compatible with the acceleration spectra of natural seismic records. Artificial spectrum-compatible accelerograms are utilized instead of natural earthquake records for the dynamic response analysis of many critical structures such as hospitals, bridges, and power plants. The realistic estimation of the seismic response of structures involves nonlinear dynamic analysis. Moreover, it requires seismic accelerograms representative of the actual ground acceleration time histories expected at the site of interest. Unfortunately, not many actual records of different seismic intensities are available for many regions. In addition, a large number of seismic accelerograms are required to perform a series of nonlinear dynamic analyses for a reliable statistical investigation of the structural behavior under earthquake excitation. These are the main motivations for generating artificial spectrum-compatible seismic accelerograms and could be useful in earthquake engineering for dynamic analysis and design of buildings. According to the proposed method, a single natural earthquake record is deconstructed into amplitude and frequency components using the Hilbert–Huang transform. The proposed method is illustrated by studying 20 natural seismic records with different characteristics such as different frequency content, amplitude, and duration. Experimental results reveal the efficiency of the proposed method in comparison with well-established and industrial methods in the literature.

No abstract

No abstract