Fault Detection of Rolling Bearing Based on Improved Residual Neural Network
In view of the problem that the traditional depth learning algorithm can not complete the on-site detec-tion due to the large amount of computation when facing the complex algorithm with a large amount of computation un-der special environments such as mines,this paper proposes a rolling bearing fault detection method based on im-proved residual neural network.This method reduces the loss of information as much as possible by adding jump con-nections in convolutional residual blocks and identical residual blocks,and replaces the ordinary convolution in partial residual blocks with depth separable convolution,which greatly reduces the computational complexity.The experiment shows that the improved residual neural network can effectively extract the feature information of data,improve the speed of operation,and greatly improve the accuracy of rolling bearing fault detection,with an accuracy of 99.97%,while solving the problem that large amounts of data are difficult to operate on-site in harsh environments.
Rolling bearingResidual neural networkTroubleshootingDepth separable convolution
NETL
NSTL
万方数据
