| Microseismic monitoring technology is one of the effective means in the field of mine safety monitoring,which collects rock rupture signals in the monitoring area through reasonable arrangement of vibration pickups,and uses inversion means to obtain the source and energy information of rock rupture signals,so as to make early warning for dangerous areas in time.The installation of vibration pickup is the premise of microseismic monitoring,and the intelligent identification and de-noising of effective microseismic signal is an important part of microseismic monitoring technology,which can effectively improve the efficiency and reliability of microseismicity monitoring.The mine environment is complex and changeable,and many interference factors make the collected signals not only have low signal-to-noise ratio,but also have many and miscellaneous.Therefore,it is particularly important to identify rock rupture signals from a large number of microseismic events and effectively denoise it.In this thesis,based on the actual mine microseismic data,the intelligent identification and denoising of effective microseismic signals and the influence of sensor installation on monitoring results are studied.It mainly completes the following tasks:(1)According to the characteristics of microseismic monitoring system,a method of intelligent identification of effective microseismic signals based on random forest is proposed.Firstly,the STA/LTA method,FFT transform and CEEMDAN decomposition are used to extract the characteristics of microseismic signals in many aspects,and the characteristics of each type of signal are marked,and then the random forest classifier model is trained by using the characteristics to accurately identify the rock fracture signal.The test results show that the random forest method can effectively identify rock rupture signals,and the classification accuracy of microseismic characteristic samples is 98.4%.(2)Aiming at the problem of low signal-to-noise ratio of microseismic signal,wavelet threshold denoising algorithm is used to denoise the microseismic signal.Based on the Garrote threshold function,an improved threshold function is proposed to solve the problems of soft and hard threshold functions in signal processing.It not only preserves the continuity of the soft-thresholding function,but also avoids the oscillation phenomenon at the transformation point due to the reconstruction of the hardthresholding function.At the same time,in order to meet the characteristic that the noise wavelet coefficients will decrease with the deepening of th decomposition level,the logarithmic function is introduced to optimize the threshold rule.The experimental results show that compared with the classical wavelet threshold denoising algorithm,the proposed algorithm can retain more detailed information in the signal while removing the noise components,and is closer to the pure rock fracture signal.The denoising effect is evaluated by the evaluation index,and compared with the traditional threshold denoising algorithm,the average signal-to-noise ratio is improved by 1.3dB,and the minimum average root mean square error is 0.0089.(3)In order to make the signal waveform characteristics of the microseismic monitoring system good and less affected by environmental factors,a buried installation method and three external installation methods are selected for comparative test.By analyzing the signals collected by the four installation methods,a more suitable sensor installation method for microseismic monitoring is selected.(4)Taking the actual mine microseismic detection data as an example,the proposed two methods of microseismic effective signal intelligent identification and denoising are verified,and the research method is visualized by using MATLAB GUI tool. |
PDF Full Text Request