ROTATING MACHINERY MONITORING AND FAULT DIAGNOSIS WITH NEURAL NETWORK ENHANCED FUZZY LOGIC EXPERT SYSTEM

This paper presents an intelligent monitoring and fault diagnosis approach for rotating machinery by utilizing artificial neural networks and fuzzy logic expert systems (FLES). A recurrent neural network (RNN) is introduced to filter the input signal before they are forwarded to the expert system. The RAIN is trained based on existing operational data so that it can adapt to a specific machine's configurations and conditions. The RNN is able to generate proper baseline signal even if the machine is not under the exact same condition. A fuzzy logical expert system is then used for diagnosis based on the residual signal generated by the RAW. The system is incorporated into an existing comprehensive roto-dynamics software package named LVTRC.