Fuzzy fault diagnosis method based on particle swarm optimization algorithm

A new fuzzy fault diagnosis method using particle swarm optimization (PSO) algorithm to determine its membership function is proposed in view of weights training technology of neural network. The brief introduction to fuzzy fault diagnosis method based on fuzzy classification concept is described at first. Then the process of obtaining membership function using PSO algorithm is demonstrated. A fitness function for fault diagnosis is presented. Finally, a numerical simulation for fault diagnosis of steam turbine-generator sets is given to verify the effectiveness of the proposed method.