A rule-based exception reason diagnosis system for electric power data

In the power user electric energy data acquire system, due to the fault of electric energy metering device, the exception in grid system, the system file data errors, etc, it produces a variety of abnormal data. Analyzing the abnormal data rapidly, diagnosing the reason that leads to the generation of the abnormal data, will improve the application effect of the acquire system. In this paper, we present an exception reason diagnosis system for electric power data. Based on the expert experience, we abstract an abnormal data characteristic set and an exception reason set, set up a knowledge base of association rules to describe the relationship between them. Each rule contains a number of abnormal data characteristics as the antecedents and an exception reason R as the consequent. The support and confidence of the rule have been defined in the paper. On the basis of the knowledge base, we construct an exception reason diagnosis system for electric power data, using an association rule based classification algorithm to diagnose the exception reason for the abnormal data. The system contains the function to learn the real world experience to correct the knowledge base. At present, the exception reason diagnosis system is playing a role as a supplement for the power user electric energy data acquire system in Fujian province, and is online monitoring the abnormal data and automatic analyzing the reasons for the exceptions.