TY - JOUR

T1 - Comparison of optimal basis function for the underground microseismic wave processing in wavelet packet transform

AU - Zhang, Mingwei

AU - Sasaoka, Takashi

AU - Shimada, Hideki

AU - Matsui, Kikuo

PY - 2013/12

Y1 - 2013/12

N2 - To improve the accuracy of signal analysis and processing for the underground microseismic waves, an optimal basis function is indispensable in the wavelet packet transform (WPT). Based upon the microseismic wave groups monitored in a deep coal mine, wavelet bases in the Daubechies, Symlets and Coiflets families were screened, and the optimal wavelet packet basis was strictly determined by its reconstruction capability on the original wave and its conservation capability on its characteristic components. Signal reconstruction and conservation capabilities were evaluated by two parameters, root mean square error and correlation coefficient. The energy reserving capability of the optimal basis function was finally discussed to verify its superiority. The results turn out that the wavelet bases db1, sym4, sym5 and sym8 are more appropriate for the microseismic wave compared to others as their better signal reconstruction capability. Among them, basis sym5 is the optimal wavelet basis function for the microseismic wave. When processed by the wavelet basis sym5, the maximum energy components of waves are effectively reserved and the reconstructed characteristic components have the highest relevancy with that of the original wave.

AB - To improve the accuracy of signal analysis and processing for the underground microseismic waves, an optimal basis function is indispensable in the wavelet packet transform (WPT). Based upon the microseismic wave groups monitored in a deep coal mine, wavelet bases in the Daubechies, Symlets and Coiflets families were screened, and the optimal wavelet packet basis was strictly determined by its reconstruction capability on the original wave and its conservation capability on its characteristic components. Signal reconstruction and conservation capabilities were evaluated by two parameters, root mean square error and correlation coefficient. The energy reserving capability of the optimal basis function was finally discussed to verify its superiority. The results turn out that the wavelet bases db1, sym4, sym5 and sym8 are more appropriate for the microseismic wave compared to others as their better signal reconstruction capability. Among them, basis sym5 is the optimal wavelet basis function for the microseismic wave. When processed by the wavelet basis sym5, the maximum energy components of waves are effectively reserved and the reconstructed characteristic components have the highest relevancy with that of the original wave.

UR - http://www.scopus.com/inward/record.url?scp=84891399415&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84891399415&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:84891399415

VL - 73

SP - 71

EP - 85

JO - Memoirs of the Graduate School of Engineering, Kyushu University

JF - Memoirs of the Graduate School of Engineering, Kyushu University

SN - 1345-868X

IS - 3

ER -