Application of a novel RBF algorithm to fault diagnosis of hydro-turbine generating unit

This paper presents a novel RBF neural network model.In this model,the immune genetic principles are introduced to select the number and positions of hidden layer RBF centers,and the output layer weights are decided with the recursive least squares algorithm.This is an effective model for fault diagnosis of hydro turbine unit.The application to a practical generator set shows advantages of the model,swiftness,high precision and easiness in generalization.This approach is provided for monitoring and fault diagnosis of hydro-turbine unit.