Research on intelligent operation and maintenance fault diagnosis and knowledge reasoning based on knowledge graph

Aiming at the problems of difficult multi-source data fusion and low fault knowledge reuse rate in the operation and maintenance of smart grid power communication equipment, a knowledge graph based intelligent operation and maintenance fault elimination framework is proposed. Construct a hierarchical knowledge graph containing device entities, and combine the WBLA model and TFIDF cos algorithm to achieve fault severity recognition and intelligent diagnosis. In the experiment, there were 3696 fault datasets, and the accuracy of severity recognition was 95.98%, which is 16.2% higher than traditional methods; In 1000 cases, the accuracy rates of fault location and disposal measures were 98.4% and 97.5%, respectively, which improved operation and maintenance efficiency and decision-making accuracy.