Analysis of Power Grid Fault Diagnosis Based on Association Rules and its Evaluation Index Optimization Method

Aiming at multiple solutions and misdiagnoses for fault diagnoses based on the power grid analytic model, a method of power system fault diagnoses based on association rules is proposed. The complete fault diagnoses are obtained by analyzing the uncertainty about the protective relay action, circuit breaker action and the alarm information. The analytic model is given, combined with the probability of various uncertainty events to construct the evaluation index of the diagnosis. These uncertainties are used to classify alarm levels. At the same time, the evaluation index is determined by using comentropy to classify the weight of the uncertainty. Finally, the optimal diagnosis is obtained, and verified the effectiveness of the method by an experiment example.