A Correlation Analysis Model of Fault Location of Distribution System Based on RS-IA Data Mining
Zhang You Xu, Yi Fang Su, Pan Zhang, Rong Gang Ge, Fu Jian, Cong Liu, Shu Yan Yu, Lei Jiao Ge · Applied Mechanics and Materials · 2017
Fault location is an effective way to ensure quickly power recovery after a failure in power distribution system. Bad operation environment of FTU, damaged elements or lost information usually cause variations in fault information for correlation analysis in fault location of power distribution system. A data mining (DM) correlation analysis model based on rough sets (RS) theory and immune algorithm (IA) is proposed in this paper. Firstly, using RS theory to extract domain knowledge, a set of the given variant fault pattern is converted into a decision table in RS. Secondly, an attribute reduction of the decision table is made by using the IA theory, and the intrinsic correlation rules between input vector (condition attribute) and output vector (decision attribute) are mined. Then, the data mining method is used to deal with the distortion of FTU real-time input information. According to the current limit information sequence of section switches, line fault states in each section are judged to realize the fault location in power distribution network. Finally, the feasibility and effectiveness of the fault location model of distribution network based on RS-IA data mining model is verified by simulation.