Intelligent Fault Diagnosis Method Of Rolling Bearing Based On Improved One Dimensional Convolutional Neural Network

Rolling bearings are the core components of rotating machinery,and the health condition of rolling bearings directly affects the overall performance of rotating machinery,which is of great significance to the safe operation of mechanical equipment.The bearing vibration signal contains the information about the running state of the bearing,and the health state of the bearing can be identified through analysis and processing.However,as the machinery fault diagnosis technology enters the era of big data,traditional feature extraction and fault diagnosis methods are difficult to mine the inhernet information of faults from a large number of fault data sets and accurately identify them.For this reason,the rolling bearing vibration signals are used as fault information carrier,and Convolutional Neural Network(CNN)is employed as the theoretical analysis tool to carry out rolling bearing intelligent fault diagnosis in this work.The main contents are as follows:(1)In order to solve the problem that the used intelligent diagnosis models usually have too many parameters and low recognition efficiency,a rolling bearing fault recognition method based on one-dimensional lightweight convolutional neural network is proposed in this paper.For one thing,the 1×1 convolution kernel is introduced to enhance the nonlinear expression ability of the one-dimensional convolutional neural network model;for another,the global average pooling layer is used to replace the fully connection layer in the traditional convolution neural network,so as to reduce the model parameters and the amount of calculation and prevent over fitting phenomenon.The experimental results show that the proposed method can accurately recognize different fault status of a real rolling bearing and has particular potential in engineering application.(2)Aiming at the problem that some deficiencies exist in fault diagnosis based on convolutional neural network because the sampled actual vibration data are often differently distributed and difficult to label.A fault diagnosis method of rolling bearing based on one-dimensional convolutional neural network with transfer learning is proposed.Firstly,a one-dimensional convolutional neural network model which can directly process vibration signals is established and pre-trained with the data in source domain.Then,maximum mean discrepancy(MMD)is used to measure the feature distribution distance between the source domain and the target domain in each layer of the pre-training model and determine whether the convolutional layers and fully-connected layers can be transferred,and after that the model is restructured through the initialization strategy.Finally,a small number of labeled data in target domain are used to train the model again,and then the fault data in target domain are classified.The experimental results show that the proposed method can realize the highly accurate fault classification of rolling bearings under variable operation conditions while there are only a few labeled data in the target domain.(3)In order to solve the problem that convolution neural network is difficult to play the best role when the vibration signal of rolling bearing is polluted by noise and the change of running load,a fault diagnosis method of rolling bearing under complex working conditions based on one-dimensional residual neural network is proposed.For one thing,the Batch Normalization and Instance Normalization simultaneously in the one-dimensional residual network is introduced to enhance the feature extraction ability of the model;for another,the Multiple Kernel Maximum Mean Discrepancy(MK-MMD)between the source domain and the target domain is introduced into the training process of the model,so as to improve the generalization performance of the model for the fault data sets with different distribution.The experimental results,obtained by using fault bearing data,show that the proposed method can achieve the highly accurate fault classification of rolling bearings while the fault signals are affected by both noise pollution and load change.