Comparative Application of Time-Frequency Methods on Strong Motion Signals

Real-world physical signals are commonly nonstationary, and their frequency details change with time and do not remain constant. Fourier transform that uses infinite sine/cosine waves as basis functions represents frequency constituents of signals but does not show the variations of the signal frequ...

Full description

Saved in:
Bibliographic Details
Published in:Advances in civil engineering 2021, Vol.2021 (1)
Main Author: Ashtari Jafari, Mohammad
Format: Article
Language:English
Subjects:
Basis functions
Civil engineering
Continuous wavelet transform
Decomposition
Discrete Wavelet Transform
Earthquakes
Entropy (Information theory)
Filter banks
Fourier transforms
Methods
Seismic activity
Signal processing
Trigonometric functions
Wavelet transforms
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Real-world physical signals are commonly nonstationary, and their frequency details change with time and do not remain constant. Fourier transform that uses infinite sine/cosine waves as basis functions represents frequency constituents of signals but does not show the variations of the signal frequency contents over time. Multiresolution demonstration of the time-frequency domain may be achieved by the techniques that can support adjustable resolution in time and frequency. Earthquake strong motion signals are nonstationary and indicate time-varying frequency content due to the scattering from the source to the site. In this paper, we applied short-time Fourier transform, S-transform, continuous wavelet transform, fast discrete wavelet transform, synchrosqueezing transform, synchroextracting transform, continuous wavelet synchrosqueezing, filter bank synchrosqueezing, empirical mode decomposition, and Fourier decomposition methods on the near-source strong motion signals from the 7 May 2020 Mosha-Iran earthquake to study and compare the frequency content of this event estimated by these methods. According to the results that are examined by Renyi entropy and relative error, synchroextracting performed better in terms of energy concentration, and the Fourier decomposition method revealed the lowest difference between the original and reconstructed records.
ISSN:1687-8086
1687-8094
DOI:10.1155/2021/9933078