Time filter

Source Type

Hao G.,Wuhan University | Hao G.,Chongqing Three Gorges University | Gong T.,Wuhan University | Dong H.,Wuhan University | And 3 more authors.
Earth Science Frontiers | Year: 2015

Aiming at non-stationary and nonlinear signals like the Earth's natural pulse electromagnetic fields, this paper adopted a clustering algorithm based on ensemble empirical mode decomposition(EEMD), which effectively extracted the instantaneous energy spectrum, energy concentration distribution bands, time and frequency corresponding to the maximum amplitudes and other time-frequency characteristics before the Lushan MS 7.0 earthquake. Compared with the Hilbert-Huang transform (HHT) method in the decomposition of empirical mode decomposition (EMD), EEMD method can effectively suppress the mode mixing effect existing in EMD decomposition process. The paper also made a comparative study among this method and Fourier transform and wavelet analysis. The result shows that using HHT method based on EEMD decomposition to the unsteady data of the Earth's natural pulse electromagnetic fields is better for representing inherent characteristics of the original data, and thus EEMD method owns higher validity. ©, 2015, Chinese Academy of Forestry. All right reserved.


Hao G.,Wuhan University | Hao G.,Chongqing Three Gorges University | Chen Z.,Wuhan University | Zhao J.,Wuhan University | And 6 more authors.
Earth Science Frontiers | Year: 2016

The present paper aims at the non-stationary characteristics of Earth's natural pulse electromagnetic field (ENPEMF) signals by using the normalized STFT-WVD (NSTFT-WVD) transformation, and the main analysis focuses on time-frequency characteristics of the ENPEMF signal before Lushan MS 7.0 earthquake. By comparing short-time Fourier transform (STFT)and Wigner-Ville distribution (WVD) transformation with the linear frequency modulation signal, this paper analyzes their advantages and disadvantages of both time and frequency focusing performance. NSTFT-WVD transform method can give us the better time-frequency aggregation and inhibits the cross-terms. The results show that NSTFT-WVD transform can reflect the real ENPEMF signal time-frequency-energy spectrum distribution before and after the earthquake, which could render more obvious silent state in entire frequency and sustained 1-2 days before the earthquake. The time-frequency representation of data channels 2 and 3 are basically consistent with this characteristic, which well represents the feature of impending earthquake precursors. © 2016, Editorial Office of Earth Science Frontiers. All right reserved.

Loading Environmental Center for Resources Conservation collaborators
Loading Environmental Center for Resources Conservation collaborators