Redesign of Fault Diagnosis Expert System with Manual Intervention and Self-Learning Function

Introduce the manual intervention mechanism into the traditional fault diagnosis expert system, with which allowing the human experts to intervene the reasoning process and modify the results outputted by the expert system. Extract the pre-conditions and final results from the manual intervention behaviors, and store them in expert system knowledge base. Redesign the self-learning inference engine, making reasoning process suspending and debugging possible. And then develop a new matching algorithm with the help of statistical method, which can match the records in knowledge base and modify the processing work case automatically, and create new knowledge from manual intervention behaviors if matching failed. Thereby, more credible diagnosis results would be produced by the expert system newly designed.