Fault diagnosis model based on Bayesian network considering information uncertainty and its application in traction power supply system

Abstract

Fault diagnosis is of great significance in maintaining the safe and stable operation of a system. Fast and precise location of faults is important for restoring power supply systems. In this paper, considering the interference of information uncertainty to fault diagnosis under different weather conditions, a fault diagnosis model based on Bayesian network is proposed. The model inherits the capability of processing uncertain information of a Bayesian network and divides the information into historical information and evidence information. The uncertainty of historical information is reduced by modifying the parameters through improving parameter learning formulas, while that of evidence information can be reduced by logical judgment. A fault simulation model based on Monte Carlo method is used to generate simulation data for parameter learning, and a case study confirms the correctness of the proposed fault diagnosis and fault simulation models. © 2018 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.