A Fault Diagnosis Method of Transmission Network Based on Bayesian Network and Fault Decision Table

In view of the complexity and low accuracy of existing fault diagnosis methods for transmission network, this paper proposes a novel transmission network fault diagnosis method combining simplified Bayesian network and fault decision table. First, using the information from Supervisory Control and Data Acquisition (SCADA) system, there develop a fault area identification method, which utilizes circuit breaker (CB) information to isolate the components to build a fault area. In fault area, there creates the simplified Bayesian network to associate components with CBs. Further, the paper proposes a calculation method to calculate the fault probability of components in fault area, which is able to determine the suspicious component set. According to the sequence of relay actions, a fault decision table for the suspicious component set can be established. Finally, the fault condition of components, CBs and protection devices are diagnosed by comparing the local fault decision table. The above method greatly simplifies the complexity of Bayesian network, and the test results show that the speed and accuracy of its fault diagnosis have been significantly improved.