Fault Diagnosis Of Traction Power Supply System And Its Key Equipment Research

The stability of electrical power is crucial to the operation of the entire high-speed railway.When any fault occurs to the traction power supply system(TPSS),it may course the outage of the power and shut down the high-speed railway electric locomotive.The transporation plans of railway may be affected.It may bring inconvenience to people's life and great damage to economy.Thus,on the one hand,it is essential to find the fault in the system.On the other hand,traction transformer is the most improtant equipment,fault diagnosis and prognosis of which is of great significance to the whole system.Against this background,this paper studies fault diagnosis of TPSS and fault diagnosis and prognosis of traction transformer.This paper starts with the TPSS.The Bayesian network is established according to relay protection principle and the topology of TPSS.The fault rate of relay protection and circuit breaker is applied in parameter learning of Bayesian network to reduce the uncertainty.Finally,the fault diagnosis of TPSS is implemented by Bayesian reasoning.As for fault diagnosis of traction transformer,this paper analyze the dissolved gas in transformer oil firstly.The three-ratio method codes and the percentage of dissolved gases in oil are used as input of neural network.The neural network is trained by particle swarm algorithm.Finally,the fault diagnosis of traction transformer is based on output of neural network.As for fault prognosis of traction transformer,firstly,we get a list of fitting values of dissolved gases in oil by modified gray model(MGM).Finally,the predicted data is entered in neural network and the output of neural network is used to calculate the possibility.The diagnosis model of TPSS considers the uncertainty of information under different weather condition.The diagnosis model of traction transformer combine neural network with three-ratio method.The fault prognosis model of traction transformer is based on the diagnosis model of traction transformer,using the MGM to predicte the data of dissolved gases in oil.