Fault Diagnosis of Ship Main Power System Based on Multi-Layer Fuzzy Neural Network

Artificial neural network (ANN) has been successfully applied to fault diagnosis systems in real-world applications. But only single network is used for diagnosis, which is not good at handling expert knowledge. Multiple faults in complex systems occur commonly in practice. When the single network is used to deal with complicated problems of fault diagnosis, it'll be so gigantic that a series of difficulties will be brought to network training. Based on the analysis of hierarchical classified diagnostic model, a multi-layer fuzzy neural network (MFNN) is presented for fault diagnosis of ship main power system. The diagnostic result indicates that the model is feasible and valid. With good generalization performance, the network has significantly improved the diagnostic precision