Fuzzy Information Approaches to Equipment Condition Monitoring and Diagnosis

Equipment condition monitoring plays a crucial role in the overall integrity of the power system. As a result, utilities invest significant time and finances into equipment monitoring and maintenance in order to anticipate failures or accelerated aging in power equipment. Such monitoring includes regular insulation condition tests for switching devices, reactors, power transformers, generator windings and so on. In general, many of the indicators of equipment condition are imprecise and/or unreliable. Engineers must have considerable experience with a particular test before that test becomes useful. Several utilities have developed expert systems to codify this experience and improve knowledge of the breakdown process. This work emphasizes the uncertainty modeling of diagnostic problems. Fuzzy information methods are employed to represent quantitatively the diagnostic capability of a system. Further, several methods are discussed for extracting information from test data and evaluating system performance. This research proposes that systematic representation of uncertainty can lead to significant improvements in diagnostic capability.