Research On The Methods For Extracting Acoustic Emission Signals Of Coal Rock Burst Based On The Compressed Sensing

When the coal rock is deformed or fractured under the action of force,by monitoring and analyzing the changes of the acoustic emission signal that releases energy,the degree of destruction of the coal rock body can be inferred,the location of the outburst source can be predicted,and the early warning of coal rock dynamic disasters can be realized Therefore,the acoustic emission detection technology has become a research method to prevent the occurrence of mining disasters.However,the noisy operating environment under the mine results in the monitored disaster warning acoustic emission signals containing a large amount of noise signals,which reduces the accuracy of predicting coal and rock mass dynamic disasters.Therefore,this thesis studies the extraction method of acoustic emission signals from coal and rock mass under strong background noise.In view of the characteristics of the acoustic emission signals of coal and rock mass,which are random,non-stationary and susceptible to noise signal interference,a method for the acoustic emission signal noise reduction based on the theory of compressed sensing is proposed on the basis of wavelet transform noise reduction.This method avoids the influence of the basis function on the noise reduction effect in the wavelet transform filtering,effectively inhibits the interference of the noise signal,and improves the noise reduction accuracy.Considering the diversity of acoustic emission signals generated during coal rock rupture,this thesis conducts a blind source separation study on effective signals through three typical blind source separation algorithms.Comparison of separation efficiency evaluation,found that the PI signal separated by a relatively small Fast ICA algorithm,the similarity coefficient closer to 1,which best separation for accurate monitoring of the extracted effective rock bursting coal acoustic emission signals.To explore the limitations of the Fast ICA separation algorithm on the separation effect of noisy aliased signals,it laid the foundation for the combination of signal denoising and blind source separation algorithms.In order to improve the application effect of the blind source separation algorithm in the noisy aliased signal,through the simulation experiment,the signal after noise reduction and preprocessing of the signal was studied by Fast ICA blind source separation,and the comparative analysis of the evaluation indicators through wavelet transform,compressed sensing The separation effect of noise reduction preprocessing,the results show that the blind source separation effect based on compressed sensing is better than the separation effect under wavelet transform filtering.The mean square error(MSE),signal-to-noise ratio(SNR)and similaritycoefficient of the signal after noise reduction and separation are well improved and available extracting rock bursting coal acoustic emission signal in the background noise.