An evaluation of intelligent systems for fault diagnosis

This paper discusses the foundation of intelligent systems sufficient for studying Space station and/or utility power systems, where fault detection are of paramount importance. The functionality of these intelligent systems or tools, such as expert systems, artificial neural networks and fuzzy logic, for aiding in design, control and analysis of power systems diagnosis is emphasized. The argument for selecting them in terms of their abilities and platforms for using them are discussed. Finally research activities using these tools for both Space power systems (Space Station Freedom) and terrestrial distribution systems in a major utility are presented. The evaluation has shown that the intelligent systems hold significant promise for online fault analysis or diagnosis. Criteria for using these tools as embedded tools in classical methods or fault study are discussed. Intelligent systems are useful for managing input data and interpreting faulty results.<<ETX>>