Research On Fault Diagnosis And Early Warning Algorithm For Traction Power Supply System

The stable reliability of traction power supply system as an important component of electrified railway directly affects the safe running of trains. Once a fault happened on a traction power supply system, it not only caused the power cut affected the operation of trains of the supplied, but also may lead to more congestion, causing inconveniences and threats to passengers and lossing to the national economy. Therefore, the research on the issue of fault diagnosis and early warning for traction power supply system quickly and accurately is of great significance.In this paper, the fault diagnosis algorithm of transformer and overhead lines has been researched, which is the key equipments of the traction power supply system, including fault diagnosis of internal components of the traction transformer and fault location of the catenary.Firstly, basic ideas and principles of the model-based diagnosis methods and the fault tree analysis method has been introduced, then the advantages and disadvantages of each other were analyzed, the link existences between the two methods was revealed, and on this basis a fault tree diagnosis algorithm of model-based was proposed and the basic thoughts and implementation process were expounded. Then, studied the fault mechanism of traction transformer and catenary, established the knowledge representation library of the fault tree of traction transformer and catenary, applied the proposed algorithm above into the traction transformer diagnosis, analyzed its superiority in theory, and achieve fault location of the catenary fastly combined with traveling wave method. Finally, designed a software system for traction transformers and overhead lines for fault diagnosis and early-warning based on Labwindows software. By designing the data input module, fault diagnosis module, fault localization module and early warning analysis module, the fault diagnosis functionality for internal components of traction transformer and catenary fault location functionality have been implemented. Through the instance data, the validity of the software system was verified and proved the feasibility of the proposed algorithm.