基于EMD-SVD的矿山微震信号降噪方法及其应用OA北大核心CSTPCD

To enhance the accuracy of microseismic monitoring technology in the analysis and processing of microseismic signals,and to fully extract effective information from microseismic signal waveforms,a novel denoising method based on empirical mode decomposition(EMD)and singular value decomposition(SVD)is proposed for the non-stationary and nonlinear characteristics of mine microseismic signals.This method initially obtained the intrinsic mode function(IMF)components of the signal through EMD decomposition,and optimized the IMF components using correlation coefficients,variance contribution rates,and similari-ty.Subsequently,the selected IMF components were used to reconstruct the phase space data of one-dimen-sional microseismic signal time series.After decomposition by SVD,the SVD reconstruction order was es-tablished using the percentage of singular value energy,and the denoised one-dimensional microseismic time series was obtained based on the SVD restoration principle.Taking mine blasting in a mine in Shandong as an example,different denoising methods were applied to three types of typical microseismic signals,and their denoising effects were compared and analyzed.The results show that compared with traditional de-noising methods,the EMD-SVD denoising method improves the average signal-to-noise ratio by 35％and reduces the average mean square error by 50％,effectively eliminating noise components in the microseis-mic signal while preserving its characteristic information.This research is significant for analyzing mine mi-croseismic signals,locating microseismic events,and monitoring dynamic disasters in coal mines.