Fault Diagnosis Method of Power System Based on Bayesian Network

Fault diagnosis is to estimate the fault state of the system through the observed information. The accuracy of diagnosis results depends on the recognition degree of the diagnosis method to the system fault. In order to deal with the uncertainty issue in fault diagnosis, many methods which combine probability emerge as the times require. Bayesian network is an uncertain reasoning method proposed in the 1980s, which offers a natural and intuitive expression for causal relationship and dependency relationship. It is characterized by strong probabilistic reasoning ability, clear semantics, easy to understand, etc., which has become one of the hot topics in the field of artificial intelligence research. This work first described the basic knowledge from Bayesian theorem to Bayesian network, then extracted a case for analysis, and finally established a corresponding Bayesian network for power system fault diagnosis.