Fault analysis of condenser based on RBF network and D-S evidence theory

Abstract

A novel Information fusion fault diagnosis method is proposed for condenser fault analysis. Condenser fault diagnoses were analyzed by two algorithms of Radical Basis Function (RBF) neural network. And then the method of information fusion diagnosis was used for improving the results form the two networks. This method has both advantages of the simple features of neural networks and the uncertainty capabilities of information fusion in the application. Through the condenser fault simulation test, it can be verified to improve the accuracy of fault diagnosis, while reducing the complexity of the algorithm. © 2012 Springer-Verlag.