A New Approach for Fault Section Diagnosis of Distribution System Based on Data Mining Model
Zhi Liao · Journal of Tianjin University Science and Technology · 2002
Considering the working condition of feeder terminal units in practical application of fault section diagnosis for distribution networks,the damage of FTU elements and the distortion of information are unavoidable,and incorrect diagnosis may be caused by distorted fault patterns.This paper presents a data mining method,which is based on the combination of rough set(RS) theory and genetic algornhm(GA),to deal with distorted information and carry out the fault section diagnosis of distribution networks.In this approach,RS is used to analyze the generalized fault knowledge region and GA is used to mine the redundant relation and internal relevant rules among input information and fault section diagnosis results The high fault to1erance performance of the proposed approach is proved through comparison with that of feedforward neural networks model based fault section diagnosis.