TIFDES-an expert system tool for transformer impulse fault diagnosis

The objective of this paper is TIFDES-an expert system tool for impulse fault diagnosis in transformers. These important and complex decision tasks are, in general, performed by highly experienced testing personnel, whose knowledge can be expressed as a set of production rules. The knowledge base of this expert system is derived from IEEE and IEC standards, literature and expert human knowledge to include as many diagnosis rules as possible. To identify and locate a fault, an inference engine is developed to perform deductive reasoning based on the rules in the knowledge base and statistical techniques. The basic idea of TIFDES is to provide a nonexpert with the necessary information and interaction in order to make a fault diagnosis in a very friendly environment by a windowed display system.