Fault diagnosis using a diagnostic shell and its verification results by connecting to an operator training simulator

This paper describes a fault diagnostic system using a diagnostic expert system shell, MELDASH, and results that confirm its effectiveness. The diagnostic shell that reflects and makes use of nature of model-based diagnosis is developed to overcome the drawbacks of methods that depend on operator knowledge. A high-performance fault diagnostic system is constructed by adding an application model to the diagnostic shell. A prototype system is verified by adding an application model to the diagnostic shell. A prototype system is verified by connecting it to an operator training simulator. It is able to make a proper diagnosis in 79 difficult cases. Verification results show that the prototype system has sufficient accuracy. The authors confirm the effectiveness of this fault diagnostic method for future energy management systems