Research On Fault Diagnosis Method Of Rolling Bearing Based On Entropy Theory

The running state of rolling bearings affects the stability and safety of the whole equipment,which can be utilized to characterize the fault monitoring and performance degradation evaluation of rolling bearings.In this paper,a feature extraction method based on improved multiscale sample entropy(IMSE)is proposed to identify the fault of rolling bearing combining with machine learning theory.Meanwhile,a performance degradation index of improved information entropy(?E)is proposed to evaluate the performance degradation trend of rolling bearings.During the research process of bearing fault diagnosis,an IMSE algorithm is proposed for feature extraction aiming to solve the problem of missing information in vector similarity calculation based on sample entropy method.The algor:ithm is based on the similarity between the pythagorean theorem and the vector.The improved sample entropy(ISE)is multi-scaled to obtain the IMSE,and the fault type of rolling bearing is identified by support vector machine(SVM)classifier.The experimental results show that this method can improve the accuracy of rolling bearing fault type identification compared with those of multi-scale entropy and sample entropy methods.In addition,a nonlinear analysis method based on HE is proposed to evaluate the performance degradation of rolling bearings.Based on the construction of a special probabilistic mass function,the information entropy is calculated,and the information entropy is standardized to obtain the ?E parameter,which reflects the degradation process of bearing performance.This parameter can reflect the complexity of one-dimensional time series and have high sensitivity to signal changes.Moreover,using the ?E values can distinguish between different types of signals.The characteristics make the parameter HE available for the detection of early faults in rolling bearings.The experimental results show that the method can effectively extract the sensitive characteristics of the bearing running state,and has fast operation time and minimal parameter requirements.Based on the bearing experimental data of Case Western Reserve University(CWRU),the rolling bearing life-cycle vibration data of the Intelligent Maintenance Center of Cincinnati University(IMCCU),and the rolling bearing fatigue experimental data collected by our laboratory,the effectiveness of the proposed methods are verified.Experimental results show that the sensitive features reflecting the running state of rolling bearings can be extracted effectively based on the IMSE algorithm,and the fault information in the vibration signals can be identified.The evaluation method of bearing performance degradation based on ?E characteristic parameters provides a new research idea for predicting the degradation trend of rolling bearing life-cycle performance.