Research On Bearing Performance Degradation Evaluation And Remaining Life Prediction Based On PHM

Rolling bearing is one of the important components of rotating machinery,which is widely used in wind turbines and other complex equipment.Through prognostics and health management(PHM),the health monitoring of rolling bearings in use can predict the occurrence of faults in advance and maintain them actively,as to reduce costs and avoid safety problems.In this paper,based on PHM technology,aiming at bearing degradation assessment and residual life prediction,combined with bearing life data,the research is carried out gradually.According to the fault characteristics extracted from historical data,through artificial intelligence algorithms and deep learning methods,the degradation assessment,state recognition and residual life prediction of rolling bearings during operation are realized.To solve the problem of the difficult detection of early performance degradation point of rolling bearing,a performance degradation evaluation method of rolling bearings combining improved variational mode decomposition and comprehensive feature index is proposed.The adaptive variational mode decomposition method is used to process the original vibration signal of the bearing.The singular value feature of the effective component is extracted and combined with the entropy energy ratio and the confidence value feature to form the comprehensive feature matrix of the rolling bearing degradation.Based on this,the SVDD degradation assessment model is constructed to realize the early weak fault detection and performance degradation assessment of the rolling bearing.The validity of the method is verified by the bearing life test data.The detection results of the early performance degradation points are earlier than other degradation assessment methods,which provides a new idea for the performance degradation assessment of rolling bearings.To solve the problem of bearing degradation state recognition,a bearing degradation recognition method based on degradation index and Dark Net model is proposed.Based on the performance degradation index,the degradation stage of the bearing is divided,and then the bearing state is further identified by constructing a deep convolutional neural network model.Finally,it is verified on the published IMS bearing data set and applied to the full life data of the high-speed shaft of the wind turbine.The results prove the effectiveness and accuracy of the proposed method and have engineering practical value in state recognition.To solve the problem of insufficient generalization for common health indicators in the bearing remaining useful life(RUL)prediction,a model of the RUL prediction based on a dual-input deep convolutional neural network(DUALCNN)is proposed.Firstly,the bearing signals are processed by an adaptive maximum correlation kurtosis deconvolution(MCKD),and the time series features can be obtained by fusing features.The vibration signal is processed by continuous wavelet transform(CWT)to obtain the degraded image features.Then a dual-input deep convolutional neural network model is used to extract feature maps from image features and time series features to construct health indicator(HI)for representing the degradation state of the bearing.Finally,the predicted HI is combined with a gated recurrent unit(GRU)network to predict the remaining useful life of the bearing.The proposed method is validated on the publicly available XJTU bearing datasets and is applied to the historical monitoring data from high-speed shaft bearings in a wind turbine.The experimental results show that the proposed method significantly improves the generalization performance of the HI.Meanwhile,the proposed method has an excellent performance in terms of the accuracy of remaining useful life prediction.