Research On Recognition Method Of Microseismic Signal In Coal Mine

The microseismic signal of coal mine contains abundant coal and rock rupture information,which can provide basis for monitoring and early warning of coal mine dynamic disasters such as rockburst.However,due to the influence of some factors as human and environment,the collected signals are often doped with blasting signal,mechanical vibration signal and other background noises.These will directly affect the results of microseismic monitoring and source location.Therefore,it is of great significance to study the identification method of coal mines microseismic signals.Aiming at the problem of low accuracy of the existing microseismic signal recognition method,this paper focuses on the improvement of noise reduction and feature extraction in microseismic signal recognition.The main research contents and achievements of this paper are as follows:(1)In the aspect of microseismic signal denoising,aiming at the main problems of wavelet analysis and empirical mode decomposition denoising,the local mean decomposition(LMD)method and the wavelet packet denoising method are combined in this paper,and an improved microseismic signal denoising method is proposed.In this method,the LMD is used to decompose the microseismic signal to obtain the product function components,and the difference between the noise and the dominant component of the signal is found by using the correlation coefficient between the original signal and the product function components.Then the wavelet packet is used to denoise the product components with noise.Through simulation experiments on microseismic data from coal mines and the experimental results are compared with the LMD denoising method and the LMD-based wavelet threshold denoising method.The experimental results show that the denoising method proposed in this paper has higher signal-to-noise ratio and lower root-mean-square error than the contrast method,and the noise reduction effect is ideal.(2)In the aspect of feature extraction,aiming at the problems of the fractal box dimension can only describe the microseismic signal as a whole and ignore the local feature,Combined with the multifractal spectrum features,an improved feature extraction method of microseismic signal is proposed in this paper.This method mainly studied the fractal box dimension and multifractal spectrum characteristics of coal mine microseismic,blasting and mechanical vibration signals,and extracted 7 parameters with great difference as signal characteristics.According to the established 7-dimensional feature vector,support vector machine is used to train and classify 100 groups of coal mine microseismic,blasting and mechanical vibration signals after noise reduction.The experimental results show that this method can correctly identify the coal mine microseismic signal,and the recognition accuracy is 94%,compared with the existing two recognition methods were increased 4%and 2.33%.The improved method improves the accuracy of microseismic signal recognition and provides a new idea for coal mine microseismic signal recognition in this paper.It has important guiding significance and application value for the accurate location of microseismic events in coal mines and the early warning of coal-rock dynamic disasters such as rockburst.