A novel automata and neural network based fault diagnosis system for PLC controlled manufacturing systems

•DFA based control process model is utilized to detect and isolate faults online.•Remaining faults are detected by analyzing the power usage profile.•A set of neural networks is used to detect abnormalities in the power usage profile.•The combination of above two methods always yields a very high FDI accuracy.•Can also detect analog sensor and loosely-coupled actuator faults effectively.