基于CEEMD-LMD-SCN的集合型配电网故障选线方法

Line Selection Method of Integrated Distribution Network Fault Based on CEEMD-LMD-SCN

为解决线路发生单相接地故障时,过渡电阻较高导致故障特征不明显,以及噪声干扰情况下难以对故障线路进行准确识别的问题,提出一种基于互补集合经验模态分解(complementary ensemble empirical mode decomposition,CEEMD)、局部均值分解(local mean decomposition,LMD)和随机配置网络(stochastic configuration network,SCN)的故障选线方法。首先采用CEEMD与LMD对零序电流进行分解,分别计算其对应分量的能量熵;然后将求解出的能量熵值组合形成组合特征向量,利用SMOTE(synthetic minority oversampling technique)算法扩充数据,获得训练及测试数据;采用泛化能力较强的SCN网络建立配电网故障选线模型。仿真结果表明:本文故障选线方法在不同故障距离、不同接地电阻和不同故障初始角度的情况下能有效实现故障线路的选择,在高阻以及噪声干扰情况下,该方法适应性依然良好。

In order to solve the problem of single-phase ground fault,the characteristics of fault are not obvious when the transition resistance is high,and it is difficult to accurately identify the fault line in the case of noise interference.A fault line selection method based on complementary ensemble empirical mode decomposition(CEEMD),local mean decomposition(LMD)and stochastic confi-guration network(SCN)was proposed.First,the zero-order currents were decomposed by using CEEMD and LMD,the energy entropy of its corresponding components were computed separately,then the solved energy entropy was combined to form a combined eigenvector,the data was augmented by using the synthetic minority oversampling technique(SMOTE)algorithm,data was divided into a training set and a test set,and the SCN network with strong generalization ability was used to establish the distribution network fault line selection model.The simulation results show that the line selection method can effectively select the proper fault line under different fault distances,different grounding resistance,and different initial fault angles.This method has strong tolerance to high resistance and a relatively good anti-noise capability.