基于模糊PSO-BP神经网络的配电网故障选线方法  

Fuzzy PSO-BP Neural Network Method of Distribution Network Fault Line

作　　者：郝静[1] 朱锋[1] 王韵然 兰华[1] 许云仕[1] 李琳琳[3] 

机构地区：[1]东北电力大学电气工程学院,吉林吉林132012 [2]中石油北京长城钻探公司,北京100020 [3]辽宁省抚顺供电公司保护自动化所,辽宁抚顺113008

出　　处：《电测与仪表》2011年第12期53-57,共5页Electrical Measurement & Instrumentation

摘　　要：将模糊粒子群优化BP神经网络的方法引入配电网单相接地故障选线中。通过matlab仿真试验,得出各线路的零序电流信号,并将多种提取的特征量进行信息融合。利用模糊粒子群优化BP神经网络进行故障选线。测试结果表明,与传统模糊BP神经网络选线方法相比,该方法具有训练速度快、误判率低的优点,能够满足电力系统对精度和准确性的要求。This paper proposes fuzzy particle swarm optimization method of BP neural network into a single distribution network in the phase to ground fault line.The matlab simmulation tests show that it obtained the line of zero sequence current signals and a variety of extract characteristic information fusion.We use particle swarm optimization method of fuzzy BP neural network for fault line.The results show that this method has more training speed and lower false positives than the traditional fuzzy BP neural network method,and the system can meet the power requirements for precision and accuracy.

关 键 词：粒子群优化算法 故障选线 BP神经网络 零序电流信号 

分 类 号：TM711[电气工程—电力系统及自动化]