Research On Fault Diagnosis Method Of Rolling Bearing Of Rocker Arm Transmission System Based On Convolutional Neural Network

As a key component in the rocker arm transmission system of the shearer,the rolling bearing's health status is related to the stable operation of the entire shearer.Rolling bearing failures often lead to a series of failures and even irreparable maj or property losses.Therefore,it is very important to study the fault diagnosis method of the rolling bearing of the shearer drive system of the shearer,which can play a role in ensuring the safety of the mining of the underground shearer and reducing economic losses.In the actual working conditions underground,the mining coal seam fluctuates due to changes in the mining coal seam.The shearer operation is often accompanied by noise interference and sudden changes in the working load,so the collected vibration signals are very complex and often doped The interference of the noise signal may even cause the background noise to overwhelm the normal vibration signal.How to perform feature extraction and signal analysis on the original signal under strong background noise has become the research difficulty of downhole equipment fault diagnosis.This paper summarizes the previous research work on deep learning and rolling bearing fault diagnosis,and establishes a fault diagnosis model of rolling bearings in rocker arm transmission system of shearer based on Convolutional Neural Network.This paper first summarizes the research status of rolling bearing fault diagnosis,and focuses on the rolling bearing fault diagnosis method based on convolutional neural network and image recognition.Then for the fault vibration signal collected by the rolling bearing in the rocker arm transmission system of the shearer under this working condition,it is difficult to efficiently extract and analyze the fault features,and a feature based on the deconvolution autoencoder is proposed.Extract the model.Then,for the problem that it is difficult to collect vibration signal data in coal mines and the amount of fault data is small,an improved VGG16 convolutional neural network model is proposed.Finally,the two are combined to establish the fault diagnosis model of the rolling bearing of the shearer drive system of the shearer,and the corresponding experiments are designed in each part to verify the effectiveness of the algorithm and the model.The experimental results prove that the average recognition accuracy of the fault diagnosis model of the deconvolution self-encoder combined with the improved VGG16 has reached 87.843%,and the highest recognition accuracy has reached 99.142%.From the experimental results,we can see the application effect of the rolling bearing fault diagnosis method based on the deconvolution autoencoder and the improved VGG16 in the rolling bearing fault diagnosis under variable speed,variable load and high noise conditions.This method not only eliminates the cumbersome and time-consuming signal preprocessing process in common fault diagnosis methods,reduces the diagnosis time,but also greatly improves the recognition accuracy,which effectively improves the fault diagnosis efficiency of the rolling bearing of the rocker arm transmission system of the shearer.Has certain practical application value.