Vibration Fault Diagnosis of Steam Turbine Shafting Based on Probability Neural Networks

Information entropy is an effective description for the uncertainty of a system, and could be used for the symptom to detect the vibration changes of steam turbine. Based on the faulty signals collected from rotor test rig, three information entropy: singular spectrum entropy, power spectrum entropy, wavelet energy spectrum entropy were calculated as information entropy data. Probability neural networks(PNNs) was explored to fuse the three information entropy. Research shows that with the advantages of Bayes classifier and neural networks, PNNs has good classification ability to typical vibration faults of turbine, the classification accuracy is 100% for training data, 80% for unseen data. Compared with the classification accuracy of minimum distance classifier(MDC) and improved MDC, PNNs has higher classification accuracy. It can be deduced that PNNs is a practical fusion diagnosis method for typical fault identification of turbine rotor.