Study on the fault diagnosis method of high-speed train auxiliary power supply system based on digital twin
With the booming development of railroad technology,high-speed trains have become the primary choice for people traveling medium and long distances.Meanwhile,its accompanying safety and comfort issues are gradually being emphasized.Auxiliary power supply system is an important guarantee for the normal operation of high-speed trains,and the system failure will lead to a series of problems such as the loss of passenger comfort,train travel disruption,and railroad schedule adjust-ment.To avoid such situations,this paper focuses on the fault diagnosis method of auxiliary power sup-ply system of high-speed trains.Based on the digital twin technology,the simulation model of auxiliary converter is established to simulate the real circuit operation.The acquired fault data set is then used as the analysis object to train various machine learning models on MATLAB R2018a platform.Mean-while,hybrid models are built in combination with optimization methods to compare the diagnostic per-formance.The experimental results show that digital twin technology can greatly reduce the difficulty of collecting data,and can quickly obtain a large number of rich categories of fault data.The hybrid model of BP(back propagation)neural network based on genetic algorithm has the best diagnostic classification of class faults with the average precision and accuracy of 86.7%and 95.6%.The diagno-sis efficiency and accuracy of machine learning model based on digital twin are high,which can be ap-plied in the fault diagnosis process of auxiliary power supply system of high-speed trains to ensure safe train operation.
fault diagnosishigh-speed trainauxiliary power supply systemmachine learninggenetic algorithm
NETL
NSTL
万方数据
维普
