Research On Fault Diagnosis Method Of High-speed Train Bogie Based On Data Drive

High speed train bogie is an important part of high speed train traction system.In the process of fault diagnosis,a data-driven method is proposed to solve the problems of multiple data sources,high data dimensions and complex actual conditions.Considering the characteristics of the trajectory of the sample point in the actual running data sample,the feature extraction is carried out,the fault diagnosis algorithm which can lock the fault location is proposed,and the fault diagnosis model which is suitable for the current environment is established.The main research contents of this paper include the following three aspects.(1)The research foundation of this paper.Introduce and extract data feature description of high speed train bogie.As an important part of high speed train,this paper introduces the maintenance goal of high speed train bogie,the current method of fault determination,and the hidden danger of minor faults.At the same time,the characteristics of the collected experimental data are introduced.(2)A fault detection method for bogie of LOF(Local Outlier Factor)high-speed train based on variance between variables was proposed.First,considering the problem of high computational complexity of space-time in big data,IVV(inter-variable variance)algorithm is used to integrate the data tracks with similar running states for calculation,so as to reduce the computational complexity.Secondly,in the case that the data are more affected by external factors,the nonlinear threshold value is used to replace the linear threshold value as the fault detection standard for the horizontal comparison of other fault-free components of the same high-speed train.Finally,in the case that the location of fault components of high-speed trains is initially locked,the LOF(Local Outlier Factor)algorithm is used to deeply detect the Outlier degree of a single measuring point,and to lock the precise location of fault,so as to improve the accuracy of the algorithm.A simulation experiment is carried out with the actual operating data of a high-speed train,and the effectiveness of the algorithm in fault detection of bogies of high-speed trains is verified.(3)A feature extraction method based on DTW(dynamic time warping)is proposed for fault diagnosis of high-speed train bogies.By taking the dynamic evolution feature and the traditional static feature as the input of fault diagnosis model,the running state of high-speed train is comprehensively considered.First of all,the sampling points over a period of time sampling data together as a dynamic trajectory,using the DTW algorithm is used to draw a portion of the adjacent sampling point data of trajectory similarity matrix.Secondly,a new feature "deviation degree" is defined.The deviation degree of the component is calculated on the basis of similarity matrix by referring to the calculation method of K-L divergence.This method also reduces the dimension and computational complexity of the data to a certain extent on the basis of considering the time-varying evolution of data samples.Finally,a simulation experiment is carried out with the actual running data of a high-speed train,and the deviation degree and static characteristics are input into the fault diagnosis model,which verifies the effectiveness of the algorithm in fault diagnosis of bogie of high-speed train.