Diagnosis Of Vibration Faults Of Locomotive Bearings Based On Multi-features

As an important rotating part in the mechanical equipment,the rolling bearing is one of the important fault sources of the mechanical equipment.Statistics show that in rolling machines with rolling bearings,about 70% of the failures are caused by rolling bearings.Therefore,the common fault diagnosis of rolling bearing is very important.With the development and application practice of rotating equipment condition monitoring and fault diagnosis technology,it is possible to use the vibration signal monitoring and analysis to determine the running status of rolling bearings.In view of the problem that the existing feature sets are not sufficient to fully characterize the essential features of the fault and according to the characteristics of autocorrelation function of vibration signal,a new feature of the fluctuation of autocorrelation function is proposed on the basis of the existing common feature sets.This topic further proposes a feature selection method based on intra-inter class standard deviation and weighted with the Analytic Hierarchy Process to screen sensitive feature sets.The Grey Correlation Analysis and Bias Discriminant Analysis are selected as classifier for the characteristics of few fault data in practical practice.Under the condition of few data,the new feature proposed by this topic can improve the accuracy of fault diagnosis by 6.21% through analyzing the rolling bearing data of Case Western Reserve University and the actual data of China Railway.And the correct rate of fault diagnosis is good when the sensitive feature set selected by the proposed method is used.In the data analysis,this topic uses the actual data of Chinese railway and achieves good results.Therefore,the method proposed in this paper can be applied in practice.