Online Fault Diagnosis of the Hybrid Electrical Multiple Unit Traction Converter

摘要

In this paper, the signatures and the diagnosis approach of the failure of switching devices in Hybrid Electrical Multiple Unit (HEMU) traction converter (TC) are developed. In this paper, the distorted voltage and current with such failures are treated as disturbance exerted over normal voltage and current without failures, and a special analytical failure model is built for the expression of voltage and current disturbance. Combined with the failure model, this paper proposes a novel reasoning process to locate malfunctioning switching device. The reasoning process is based on object-oriented colored Petri Net (OOCPN). Digitalized failure signatures are taken as inputs into the OOCPN reasoning machine, what stimulates the brain activities during fault diagnosis of an expert.