A maintenance and fault-diagnosis expert system integrated with artificial neural network for Wire-EDM

Commercial wire electrical discharge machining (Wire-EDM) machines possess high degree of automation and are quite robust. Nevertheless, some faults such as wire-breaking and unsatisfactory accuracy may still occur due to improper operations or inappropriate machine maintenance. A maintenance and fault-diagnosis system which integrates artificial neural network (ANN) and expert system (ES) is developed. It is time-saving in knowledge acquisition, is easy to maintain and is capable of self-learning. The occasions which call for machine maintenance are advised automatically. Suggestions to eliminate faults are proposed sequentially according to the inferred priority once a fault is taking place. Moreover, it can provide explanations.