Dynamic Analysis Method On Small Current Fault Of Distribution Network Based On Distributed Computing

The idea of smart grid have promoted the development of distributionautomation. Large scale of resonance grounding devices were deployed indistribution network. Single-phase ground fault occurrence probability is high, andquickly locating the fault zone is difficult. An effective method to analyze failuredata, locate the fault zone, identify fault type is urgently needed. This paper studiedlocation method of small current fault by using distribution computing, which issignificant for identifying line-fault.This paper combined the feature of small current fault, proposed a locationmethod using comparison of zero-sequence current. Proposed master-slave zonefault location method based on differential evolution (DE) algorithm. Against thedeficiencies of DE algorithm, optimized it by introducing hybrid encoding, dualdifferences strategy and individual disaster strategy. According to the principle ofresonance grounding system fault location, defined rules to divide line.combinedwith the improved differential evolution algorithm to establish unified fault locationevaluation model. Improved differential evolution effectively solved the problemcaused by failed data produced by dynamic changing of monitoring devices. Thispaper studied the technology of coefficients of wavelet packet decomposition toreconstruct signal and fault classification method of dynamic center point.Simulation demonstrate the effectiveness and accuracy of the classificationalgorithm. Finally，designed and implemented a small current fault location systembased on distribution computing, elaborated details of functions of every layers inthe location system, given the implementation of location algorithm and faultclassification algorithm in the location system.