Preliminary analysis on bizarre waveforms in strong motion records

B. Zhou,Lili Xie,Yu Haiying,Yongqiang Yang,H. Jinjun
2018-01-01
Abstract:One of the key questions in strong motion record processing is to identify bizarre waveforms and make suitable adjustments. Some typical strong motion records with strange waveforms in earthquakes of China were collected and analyzed in this paper. Five kinds of abnormal acceleration waveforms were revealed, including regular noise, spike, asymmetric waveform, obvious baseline drift and strong motion records packets separation. Firstly, the generation mechanism was analyzed preliminarily. It's found that the mechanism of spikes was rather complicated and mechanism of the asymmetric waveforms was of diversity. Secondly, reasonable processing approaches were offered. Regular noise can be reduced or even eliminated by filtering methods; spikes can be identified by ‘jerk’ method, ratio method and the consistency of the corresponding time of three-component PGA; the Butterworth low-pass filter can be applied to correct asymmetric waveforms; two pieces of strong motion record packets can be connected by searching for continuous and repeated data; accumulation can be used to find the clear baseline drift. Finally, the effects on time history, Fourier amplitude spectrum and response spectrum were studied before and after correction of strong motion records, and result shows that the abnormal waveform directly affects the characteristics of time history and frequency spectrum.
What problem does this paper attempt to address?