Feature Enhancement And Extraction Of Early Weak Fault Signal Of Rolling Bearing

As one of the key components widely used in mechanical equipment,rolling bearings directly affect the working performance of the entire mechanical equipment.Therefore,studying the state monitoring and fault diagnosis methods of rolling bearings has important theoretical value and practical significance.This paper takes rolling bearings as the research object,and conducts related research work on the feature enhancement and extraction of early weak fault signals.The research content of this article includes:（1）First,the relevant theories are explained on the structural composition,failure types and failure causes of rolling bearings;Then according to different types of failures,the vibration characteristics of rolling bearings are analyzed;finally,the theoretical formulas of the natural vibration frequency and failure characteristic frequency of rolling bearings are introduced,which provide theoretical support for the enhancement and extraction of fault characteristics.（2）A fault diagnosis method of Stochastic Resonance（SR）rolling bearing based on Adaptive Bat Algorithm（ABA）is proposed.The method is based on the bistable stochastic resonance system model,combined with the idea of re-sampling frequency conversion,and optimizes the structural parameters of the bistable stochastic resonance through the adaptive bat algorithm,so that the fault characteristics are enhanced,and the fault information of the rolling bearing can be judged more accurately.（3）Aiming at the difficulty in feature extraction of early weak fault signals,a feature extraction and diagnosis method based on Multiscale Sample Entropy（MSE）and Convolution Neural Network（CNN）is proposed.The method first uses the multi-scale sample entropy algorithm to analyze the complexity of the time series of fault vibration signals,and obtains feature vectors at different scales;then use the convolutional neural network to classify the generated feature vectors of different scales,so as to distinguish the signal characteristics in four working states of rolling bearing normal operation,rolling element failure,inner ring failure and outer ring failure;Finally,the feature extraction and diagnosis of the early weak fault signals of rolling bearings are realized.