Research On The Analysis Of Underground Mine Microseismic Monitoring Signal

Underground mine microseismic signal contains rich information of the internal state of rock mass change.Analyzing and identifying microseismic signal accurately and duly is vital for evaluation of rock mass stability and disaster forecasting.The microseismic signals are analyzed based on FFT,wavelet packet decomposition(WPD),empirical mode decomposition(EMD)and source characteristics,the results show that microseismic events can be divided into blasting vibration signal,rock fracture signal(large energy and small energy of rock fracture in ore)and interference signal(electrical and mechanical),preliminary seismic signal database is established;The energy of the blasting vibration signal is mainly concentrated in 13 to 16 frequency bands(375-500Hz),and the energy of the rock fracture signal and the mechanical interference signal is mainly concentrated in 1 to 4 frequency bands(0-125Hz),and partially concentrated in 7 frequency band(218.75-250Hz);The energy of the electrical interference signal is concentrated in 1 to 4 frequency bands(0-125Hz);The IMF1 and IMF2 components are the main frequency bands of the rock fracture signal and the electrical interference signal,and the IMF1-IMF3 are the main components of the mechanical interference signal,and the IMF1-IMF4 are the main components of the blasting vibration signal.Two methods for feature extraction and classification of microseismic signal are proposed based on wavelet packet decomposition and singular value decomposition(SVD),empirical mode decomposition and singular value decomposition(SVD).Taking the singular values as characteristic parameters,the recognition results of two methods are analyzed and recognition models are established based on linear discrimination analysis.The results show that the accuracy rate of WPD-SVD is generally higher when the type number of singular values is same;The accuracy rate of recognition models is 94.5%.To realize the goal of processing the microseismic signal in real time,this paper uses a combination of WPD,SVD and recognition models to develop a set of signal identification software based on Lab VIEW,which is able to satisfy the need of automatic classification.