Metro Vehicle Auxiliary Inverter Fault Diagnosis System Research

Auxiliary inverter is the key point of city track traffic grid system. Itsworking quality determines the traffic security and guest comfortable directly. Inthis situation, it must impacts the efficiency of track traffic obviously. Accordingto our common sense, auxiliary inverter inspection is so important that we needto pay much attention on it.Auxiliary inverter inspection method influences velocity and accurately ofthe inspection. Then it is necessary to choose an efficiency inspection method.Based on the prior research and this863project, I try to combine the waveletanalysis and improved neural network to make trouble inspection effective. Thisthesis designs the city Auxiliary inverter inspection system and collects the dataof proper functions, voltage fluctuation, Pulse transient and frequency outagereport. After wavelet de-noising, it collects data and types in the improvedneural network. The fault was also diagnosed with traditional BP neural networkin order to verify the improved results of the neural network, and compared toprove the improved neural network more effectively on the auxiliary inverterfault diagnosis.