A Power Network Fault Diagnosis Method Based On Association Rules And Augmented Naive Bayes

Power network fault diagnosis is system-level fault diagnosis in the power dispatching centers. Its analysis depends on information of protect devices, breakers'movements as well as electrical measurement such as electric current and voltage. Then can predicate the possible fault sections according to the logic of protect movement information and the operators'work experience.The goal of researching dispatcher on the ways of grid fault diagnosis is to judge the fault by computer instead and other units, using computer rapid calculation ability, provides the analysis result occur. accurately fast, so that can assist dispatcher to make rational operation decision after faultWith the enlarging of power system scale and the increasing of power network automatization level, more and more comprehensive information is transmitted to the power control centers. Power system response is being complication while there is fault in the power grid, especially complex faults, the abnormal movements of automechanism, circuit breakers or protective relays and uncertainties such as losing information such as channel disturbance which will make trouble for power system fault diagnosis. That will lead to serious results such as fault expanding if the fault was misjudged. Hence it is quite necessary to research and develop a fault diagnosis method which not only can deal with the great amount of datum at the first stage,but alsot could improves the fault-tolerance performance of fault diagnosis system so that dispatchers could identify the fault rapidly and e power system could operate safe and stably.Therefore, the authors propose a new approach of Augmented Naive Bayesian Network based on Association Rules data mining to diagnose faults in power network. At first, the protections and circuit breakers are taken as conditional attributes and faulty region as decision-making attribute, various faults are investigated and decision table is established. Then by use of attribute reducing method based on data mining association rules and modifying thresholds based on interactive data mining, the optimal attribute reduction combination is directly extracted. Finally, by means of choosing some specific 2-frequent item sets, the Augmented Naive Bayesian Network model is built and the nodal probability is trained.Considering the fault information being both localized and connected, this paper attempts the idea of power network faults diagnosis. The fault diagnosis software which based on Association Rules data mining and Augmented Naive Bayesian Network is programmed by C# programming language. Results of calculation complex examples demonstrated that the proposed method is superior to some traditional artificial intelligence ones. so the approach of Augmented Naive Bayesian Network based on Association Rules data mining to diagnose faults in power network is available.