Fault diagnosis for power system based on a special Bayesian network

To deal with a lot of uncertain information existing in the grid when faulted, this article uses a special Bayesian network-Noisy-or, Noisy-and nodes model to build object-oriented Bayesian networks. An error back propagation algorithm is used to train the Bayesian network so as to update the network´s parameters. Finally a case simulation is presented to prove the accuracy and validity of this method for power system fault diagnosis.