Quantitative classification of near-fault ground motions selected by energy indicators. (January 2022)

• Record Type: Journal Article
• Authors: Zhao, Dahai; Wang, Huiwei; Wang, Ding; Zhu, Ruiguang; Zhang, Jinghui
• Is Part Of: Structures. Volume 35(2022)
• Page Start: 780
• Page End: 791
• Publication Date: 2022-01
• Subjects: Pulse indicator -- Energy indexes -- Earthquake engineering -- Wavelet analysis; Structural engineering -- Periodicals; 624.1
• DOI: 10.1016/j.istruc.2021.11.032
• Languages: English
• ISSNs: 2352-0124

Abstract

Abstract: The pulse-like ground motions usually lead to great inter-story displacement and permanent deformation of the structure. In this paper, a new method based on coarse screening of seismic ground motion energy is proposed to identify the pulse-like ground motions. In general, the pulse-like ground motions have a large portion of output energy during the pulse period. Therefore, the energy of the pulse is large enough to be selected as an energy indicator for the initial data. The peak energy in the original record is often the largest, so the energy indicator of the peak is firstly used to select the data. Then, the wavelet packet transform (WPT) is used to decompose the signal and to extract the coherent velocity pulses by the first set of sub-signals. Here, the peak point method is adopted to determine the pulse-starting ( t s ) and pulse-ending ( t e ) time instants in the velocity time history. After acquiring the energy of the velocity-pulse (EVP) of the signal by t s and t e, the initial coarse screening is completed. Finally, the method proposed by Baker and Shahi in 2014 is used to determine the presence of the pulse. The energy matrices is used to obtain more pulse-like ground motions than the identification results of the previous rough indicator Peak Ground Velocity (PGV). A set of pulse-like ground motions were identified from the data training set. The linear fitting between the energy indicators and PGV was performed based on the results of the identification. The fitting results show that some pulse-like ground motions are lost directly after filtering the data with PGV 30 cm/s.