Fault Diagnosis And Health Prediction System For Mine Hoists Based On Audio Signal

Mine hoist is known as the "throat" of coal mine production,and its reliable operation is directly related to coal mine production efficiency and personnel life safety.Therefore,it is important for the safe and reliable operation of mine hoist and efficient production of coal mine to detect hoist fault types and obtain hoist health level information in advance.In response to the problems of insufficient information features,narrow signal band,low accuracy of fault detection and health prediction system in current hoist monitoring technology,this paper designs an audio feature-based fault diagnosis and health prediction system for mine hoists.The system includes audio sensing and processing layer,data transmission layer and upper computer information interaction layer.The audio sensing layer is used for audio acquisition and processing;the data transmission layer is used to upload the signal processing results to the upper computer information interaction center,and the wired LAN communication method is used to meet the confidentiality and real-time requirements of coal mine data transmission;the upper computer information interaction layer is used to receive the signal processing results,and the analysis results are interacted with human-machine.Due to the complex working environment of the hoist,the audio signal is firstly processed for noise reduction in order to avoid the influence of environmental noise.In order to avoid the influence of environmental noise,the audio signal is firstly noise reduced to avoid the influence of environmental noise.In view of the problem that the audio signal feature information cannot be obtained comprehensively by using traditional time domain and frequency domain analysis methods,EEMD is proposed to decompose the audio signal and use the crag value as the preferential criterion of IMF,then use Hilbert transform to extract the features of different signals,and finally input the feature information to CNN network to realize the automatic detection of hoist motor fault.To address the problem that coal mines cannot obtain the health status information of hoist in advance,this paper firstly divides the health level according to the crag value and fault information of hoist operation,then uses GRU network to predict the crag value and IMF components respectively,and gets the health information of hoist according to the prediction result corresponding to the health level of hoist.The main shaft hoist of Pan San Mine of Huainan Mining Group is taken as the research object,and the results show that the combination of EEMD and Hilbert can extract different audio signal features well,and the use of CNN network improves the fault classification accuracy.The small deviation value of the prediction of each numerical component using GRU network significantly improves the accuracy of hoist health prediction,and the prediction results can provide a reliable reference for daily maintenance and health monitoring of hoists.Finally,the performance and function tests of the designed system are conducted.The results show that the audio feature-based fault diagnosis and health prediction system of mine hoist designed in this paper effectively assists hoist operation status monitoring and improves the accuracy of hoist fault detection and health level assessmentFigure[58]Table[14]Reference[63]...