Prediction System Of Insulator Leakage Current Based On Rough Set And Confidence

Insulator leakage current is closely related to the development process of flashover discharge,can comprehensively reflect the degree of contamination, humidity level, under-voltage of insulatorand insulator shape, and so on. Therefore, predicting the change of insulator leakage current value isof great significance for the power system increasing the ability of flashover prevention.This paper designs a predicting system based on rough sets and the confidence for insulatorleakage current, and gives the functional modules of the method. A method of knowledge acquisitionbasis of rough set theory is put forward, which guides the choice of conditions attributes of decisiontable, removes redundant condition attributes. Through attribute reduction for the decision table,gains the decision rules from the knowledge, establishes decision rules base which implies therelationship among data attributes. Calculates the confidence of the rules respectively, selects thelargest confidence rule of the decision rules for predicting, establishes a classifier which achieves thesystem data predict function, and assesses the accuracy of the classifier. Gives the analysiscombining with example, verifies the feasibility of this method.