International Journal of Smart Grid and Clean Energy Study on Bayesian Network Parameters Learning of Power System Component Fault Diagnosis Based on Particle Swarm Optimization

Power system component fault diagnosis problem is a key issue in case of the failure of the power system. A Bayesian network, in which the network parameters are learnt by a particle swarm optimization algorithm, is proposed in this paper to establish the statistical diagnosis model. The Noisy-Or and Noisy-And structure are employed to construct the framework of the model, where the 4-level Bayesian network makes the fault prediction with properly given parameters. In order to verify the performance of our proposed method, a typical power system component fault diagnosis problem is used for empirical case study, and the result demonstrates the effectiveness of the proposed method.