The Research Of Type HXD1Electric Locomotive Traction Converter Fault Diagnosis

With the development of national economy and the policy tilt to the rail industry, heavy rail has developed rapidly in the recent years; and how to guarantee the safety operation of locomotives and trains has become the top priority of millions of railway workers. Traction converter is serving as a pivotal sub system for power conversion, and its operating status would directly impact the running quality of trains. Thus, the research on the safety operation of the traction converter on HXD1electric locomotive is discussed in this paper. This project is supported by the National Natural Science Fund "The Researches on Fault Diagnosis of the Electric Traction System on High-speed Trains"(No.61273158)This paper analyzes the failure characteristics of the traction converter of HXD1electric locomotive and how it works. The voltage signal was selected as the research object of inverter fault diagnosis, and different fault states of the traction inverter were simulated.This paper summarized the effects of converter device supports capacity on fault signal, and proposed a fault diagnosis method using DC voltage of four-quadrant rectifier and AC voltage of inverter.The voltage signal was decomposited with db3wavelet, and the energy spectrum of the signal was obtained. The feature vector of fault signal was constructed as a neural network input using the wavelets decomposition coefficients.To identify the fault feature vectors, BP neural network based on genetic algorithm was used to design a converter neural network fault diagnosis system. The results validate that neural network can effectively diagnosis inverter fault.On the basis of theoretical research, a HXD1locomotive traction converter diagnosis system was designed. The neural network weights were import to the system. The on-site test results verify the accuracy and reliability of these research methods. Figure78, Table6, references65.