Research On Rotating Machinery Fault Diagnosis Method Based On Multi-information Fusion Of Convolutional Neural Network

With the rapid development of automation technology and information technology,the complexity of industrial automation production equipment is increasing.In order to ensure the normal operation of the equipment,it is necessary to establish a sound health detection system.Due to the increasing number of detection points in the industrial process and the improvement of the sampling frequency of various sensors,the detection data of the whole industrial production equipment is increasing.Compared with the traditional spectrum analysis method,the new intelligent algorithm can make better use of the large amount of data generated.In this paper,on the basis of summarizing the previous research on deep learning and mechanical fault diagnosis,a fault diagnosis model of rotating machinery based on convolution neural network and multi information fusion is established.This paper first designs and implements a parallel convolutional neural network bearing diagnosis model based on spark.The model uses spark technology to train the weak learning machine,and uses bagging algorithm to collect the calculation results of the weak learning machine to complete the parallelization of network training.On the basis of the above model,multi information fusion technology is used to improve the accuracy of the whole network model.Experiments show that the training speed of the model is 3.38 times faster than that of the traditional method in the cluster composed of four computers.Finally,a bearing diagnosis simulation system including the bearing diagnosis program and the diagnosis result display terminal is built by using the above network model.