Study On Health Management Technology For Auxiliary Power Supply System On Subway Vehicles

Subway vehicles are the most popular urban travelers because they are convenient,on-time and eco-friendly.At present,China's urban rail transit construction is vigorously carried out,and the level of scientific research and technology related to subway vehicles has been greatly improved.Because of the complex logic of the vehicle electrical system and work in adverse conditions of high temperature shocks at long times,so there are mostly electrical fault in the subway vehicle fault.Most of the electrical faults can not be solved in the planning,it is unrealistic to completely avoid the occurrence of online faults.How to detect the early failure of a vehicle in time and eliminate the hidden trouble of a fault in time to detect the fault timely and accurately when the fault occurs and to quickly repair the fault is great significance to improving the usability of the vehicle and reducing the operating cost.Especially for the subway vehicle auxiliary inverter system,its good condition leads directly to the ride comfort of passengers and the safe operation of the train.Its high failure rate requires that the system be in good condition before the failure occurs monitoring and diagnosing the faults correctly and quickly to ensure the safety and reliability of the system,which are the focuses of the research on urban rail vehicles in the present.In this paper,used the pearl-line auxiliary inverter system of Shanghai urban rail transit as the research object,analyzed and studied the structure and principle of its auxiliary inverter system.The main circuit of the auxiliary inverter is constructed based on the actual auxiliary inverse parameters through the Matlab Simulink environment Simulation circuit model,respectively,the circuit component level hardware and software failure simulation,used wavelet transform to extract the feature vector and used the Mahalanobis distance and support vector machines to achieve classification.The main works in research are as follows:(1)Construct the main circuit simulation model of Auxiliary Inverter for subwayvehicles with the parameters according to the actual pearl-line Auxiliary Inverter components parameters,obtain the three-phase AC output waveform,then verify the accuracy of the simulation model and analyze and simulate Circuit status,extract the three-phase AC output waveform for the study signal.Focusing on soft and hard fault circuit status sampling the circuit state from a single or multiple component fault perspective.(2)The application of wavelet packet decomposition and reconstruction on three-phase current signal,contrasted with the wavelet decomposition and studied the superiority of wavelet packet,the principle of wavelet packet decomposition and determined the wavelet function and decomposition layer number.After decomposition,completed and accurated the feature vectors exactly.Then according to different needs in the feature dimensionality reduction,used the different methods for dimensionality reduction to reduce sample dimensions.(3)Application of main circuit state recognition algorithm,studied the basic principles of Mahalanobis distance and SVM respectively.According to different algorithms,determined the sampling plan and applied.After the test proved,that the two methods had some different advantages and disadvantages,can achieve good single or multiple components to recognize the soft and hard faults.It is proved that this method can identify the hidden state when there is no failure,and achieve the monitoring of the health status of auxiliary power supply system in subway vehicles.