配电网故障定位容错算法  

Fault-tolerant algorithm for fault location in distribution network

作　　者：王艳松 宗雪莹 衣京波 WANG Yansong1, ZONG Xueying1 , YI Jingbo2(1. College of Information and Control Engineering, China University of Petroleum, Qingdao 266580, China; 2. Shengli Petroleum Administration Bureau Shengli Power Plant,Dongying 257087,Chin)

机构地区：[1]中国石油大学信息与控制工程学院,山东青岛266580 [2]胜利石油管理局胜利发电厂,山东东营257087

出　　处：《电力自动化设备》2018年第4期9-15,共7页Electric Power Automation Equipment

基　　金：国家自然科学基金资助项目(51477184);山东省自然科学基金资助项目(ZR2012EEL20)~~

摘　　要：非健全故障信息下故障区段的快速准确定位对于提高配电网供电可靠性具有重要作用。分析首端电压、电流量和短路回路等值电抗的关系,提出基于径向基函数(RBF)神经网络的短路回路等值电抗估计方法,仿真分析表明短路回路等值电抗估计结果受故障距离、过渡电阻的影响较小。然后,以馈线终端设备(FTU)故障信息和短路回路等值电抗为故障特征,应用改进的BP神经网络构建故障区段定位模型。对大量测试样本的分析表明,改进的BP神经网络建立的故障区段定位模型比极限学习机网络算法的定位精度高、泛化能力好,短路回路等值电抗能够辅助修正FTU故障信息的畸变,提高BP神经网络故障定位的容错性。Fast and accuracy fault location for distribution network with incomplete information plays an important role in power supply reliability improvement. The relationship between short circuit equivalent reactance and voltage and current in head end is analyzed, then the method of short circuit equivalent reactance estimation is proposed based on RBF（ Radial Basis Function） neural network. The simulative results show that the result of short circuit equivalent reactance estimation is slightly ildluenced by fault position and transition resistance. And then with the fault information of FTU and short circuit equivalent reactance as fault characteristics, a fault section locating model is built by using improved BP （Back Propagation） neural network. The analysis of a large amount of test samples shows that fault section locating model built by using improved BP neural network has higher fault locating accuracy and better generalization ability than that built by ELM （Extreme Learning Machine ） algorithm, and short circuit equivalent reactance can help to modify the distorted fault information of FTU, which greatly improves the fault-tole- rance ability of BP neural network fault Locating model.

关 键 词：配电网 短路回路等值电抗 故障定位 BP神经网络 极限学习机网络 

分 类 号：TM76[电气工程—电力系统及自动化]