A fuzzy logic expert system for fault diagnosis and security assessment of power transformers

A fuzzy logic expert system for diagnosis and security assessment of power transformers (FDSAPT) is presented. A prototype of this expert-system has been accomplished with the AI-language Turbo-Prolog on a microcomputer and satisfactory application results were obtained. In building a sophisticated expert system for FD-SAPT, the main difficulties arise from the inexactness or impreciseness of the fault information, the uncertainty of the inference rules, the complexity of representing experts´ knowledge, and the inference process. To solve these problems, the test information was represented by fuzzy sets obtained by the operation of predicate rules. A production rule with a certainty factor was adopted to represent the knowledge of experts, and a fuzzy inference approach based on fuzzy synthetic judgment theory was introduced into the inexact reasoning of the expert system. A complex inference mechanism with depth-first search strategy was successfully used in this system. The FDSAPT system has been applied to Beijing electric power system and satisfactory results were obtained