Research On Fault Diagnosis Design Of Traction Network Based On Cloud Platform

Affected by the complicated and varied natural conditions along the high-speed railway,the traction network transmission line is prone to fault trip,by which the safe and reliable operation of the power supply equipment and the transmission network are directly affected.In order to quickly and effectively restore the stable operation of the traction network,it is necessary to accurately identify the type of fault,and then different measures must be taken to restore the railway for different types of faults,which is of great significance for maintaining the safe and stable operation of the high-speed railway.In this paper,the traction network line fault feature extraction and pattern recognition algorithm and the application of fault diagnosis method of traction network based on cloud platform is studied:with sample data of traction network substation statistically analysis,the fault type of the traction network is divided,and eight main failure modes such as foreign objects,electric locomotives,weather effects,equipment anomalies,overload,protection misoperation,external network and lightning strikes are studied,and the research direction of traction network fault diagnosis technology based on frequency domain method is determined;In order to determine the failure mode identification of the traction network,the characteristics of non-stationary transient signals of fault traveling waves is aimed at and a main feature extraction algorithm for fault signals of traction network is proposed,which is based on wavelet packet energy spectrum method and principal component analysis method;The GA-BP neural network algorithm model which is used to complete the diagnosis of the fault mode of the traction network,the main energy features obtained from the known fault traveling waves are learned as the training samples of the algorithm model,and the fault pattern recognition algorithm is designed and the recognition accuracy is realized,whose recognition accuracy is above 90%;The application system research of cloud server technology and traction network fault diagnosis method is proposed,and according to the user,function and performance requirements analysis,the design of the fault diagnosis system based on cloud platform for traction network is completed.The experimental results of the data processing system based on cloud platform architecture and fault pattern recognition for traction network show that the system has strong robustness,fast response and high recognition in the fault diagnosis of traction network,which the application requirements of fault mode identification of traction network transmission lines can be meet by.