利用经验模态分解及小波变换压制微震信号中的随机噪声

To suppress the random noise in microseismic signal by using empirical mode decomposition and wavelet transform

随机噪声的压制是微震信号分析过程中的重要环节,目前大多数降噪技术都不同程度的存在去噪效果差、易损伤有效信号等问题。针对微震信号的随机非平稳特征,提出一种联合经验模态分解(EMD)及小波阈值的降噪方法,压制微震信号中的随机噪声。该法首先使用EMD对微震信号进行自适应分解,得到有限个本征模态函数(IMF)。考虑到随机噪声主要集中在高频IMF分量中,基于噪声能量突变原则找出低频IMF分量与高频IMF分量的分界后,利用小波阈值方法对高频IMF进行降噪处理,最后将降噪后的高频IMF分量与剩余的低频IMF分量重构即可实现微震信号降噪。仿真分析及实验结果表明,该方法能充分保留微震信号的随机非平稳特征,较对比方法具有更好的降噪效果。

The suppression of random noise is an important step in the process of microseismic signal analysis.Nowadays,the most of noise reduction technologies have some problems.Owing to the stochastic non-stationarity of microseismic signal,this paper proposed a microseismic signal denoising method combined with empirical mode decomposition(EMD)and wavelet threshold denoising method to suppress random noise.The EMD can adaptively break signal down into a finite number of intrinsic mode functions(IMF)which are arranged according to the frequency from high to low order.After decomposition,the noise of microseismic signal is mainly concentrated in the higher frequency of the IMF component.Based on the principle of abrupt change of noise energy,the boundary of IMF component between high and low is found.The wavelet threshold method is used to process the high-frequency IMF.Finally,the high-frequency IMF components of denoised and the remaining low-frequency IMF components can be reconstructed to obtain the denoised microseismic signal.The simulation results show that the method can fully retain the transient non-stationary characteristics of the microseismic signal,and has a better denoising effect than that of the contrast method.