Combining Different Intelligent Reasoning Techniques for Explaining Faults on Turbine Generators

Abstract

Expert systems have been developed for the diagnosis of turbine generators, and are heuristic-based systems. The level of diagnostic explanations they provide to their end users is questionable. This paper presents and discusses the results of case- and model-based reasoning techniques for diagnosing and explaining faults of turbine generators. The techniques provide explanations that match user's expectations and reflect less ‘human expert rules’. This also led to the design of a new generation of smart systems which will explain and persuade better plant operators that their diagnosis was correct.