Research On The Methods For Extracting Acoustic Emission Signals Of Coal Rock Burst Based On The Blind Separation

When the coal rock burst under the action of stress, the energy is released in the form of acoustic emission signal. Through the analysis of characteristics of fracture of coal rock acoustic emission signal, can predict the coal rock stress, coal and gas outburst dangerous degree. In addition, through the location of the source of the coal rock acoustic emission, prediction of the prominent position, can forecast the disaster in advance, therefore, the acoustic emission signal can be used as the research object of mine safety warning. However, there is not only a large number of noise in the mine environment, but also some acoustic emission signals generated by the excavation and mining. The characteristics of this kind of signal are very similar to the characteristics of the acoustic emission signals used for disaster warning and monitoring, and their spectrum is overlapping. Therefore, for the acquisition of acoustic emission signal, first of all to remove background noise interference, and further from the hybrid tunneling mining cracking noise emission signals accurately extract the effective force of coal and rock of AE signal, it is particularly important for the follow-up work. Blind source separation method is a natural choice when the source signal and the mixed mode of different kinds of jamming signal are unknown. The research contents of this thesis are as follows:(1) The mechanism, propagation characteristics and time-frequency characteristics of the acoustic emission signal of the coal rock burst are analyzed, and the sources and characteristics of the noise in the mine environment are analyzed.(2) Focuses on the independent component analysis(ICA) of this commonly used blind source separation method. The data preprocessing methods, the independent criteria and the optimization algorithm and the evaluation index of the separation performance are described. Finally, it makes a simulation analysis on the separation performance of three typical ICA algorithm(Fast ICA, SOBI and JADE), results show that, Fast ICA algorithm is best, more suitable for monitoring and early warning for the effective coal rock rupture of accurate extraction of acoustic emission signal.(3) In view of the disadvantage of the traditional ICA algorithm for the poor separation performance of the noisy mixed signal, the wavelet denoising method is adopted in the ICA algorithm. In order to restore the source signal to the maximum extent and not distortion, in accordance with the first wavelet pre denoising, and then ICA separation, finally the wavelet denoising steps. Through simulation experiments, a reasonable choice in the wavelet function and decomposition scale parameters, signal separation performance improved. Therefore, the ICA method combined with wavelet threshold denoising can improve the extraction performance of acoustic emission signals.(4) Wavelet denoising effect is affected by wavelet basis function and decomposition scale and other factors, which determines the final separation performance directly. Based on ICA sparse code shrinkage algorithm(SCS) in denoising, the direct use of signal characteristic basis function is used to reduce the noise, less destruction of the signal, to avoid the constraints of the wavelet basis function, especially in the low signal to noise ratio(SNR), the superiority of the SCS noise reduction algorithm is highlighted. In addition, the algorithm is suitable for the extraction of the source signal from the single channel observation signal. The experimental results show that the SCS based ICA algorithm is more suitable for the extraction of the acoustic emission signals of coal and rock burst in the strong background noise environment compared to the ICA algorithm combined with the wavelet denoising algorithm.