基于模糊RBF神经网络的短路故障区域检测  

Short-circuit fault area detection based on fuzzy RBF neural network

作　　者：宋新甫 高明 王云飞 罗锐 时光远 白宇 SONG Xin-fu;GAO Ming;WANG Yun-fei;LUO Rui;SHI Guang-yuan;BAI Yu(State Grid Xinjiang Economic Research Institute,Urumqi 830000,China;State Grid Economic and Technological Research Institute Co.,Ltd.,Beijing 102209,China)

机构地区：[1]国网新疆经研院,乌鲁木齐830000 [2]国网经济技术研究院有限公司,北京102209

出　　处：《信息技术》2023年第12期184-190,共7页Information Technology

摘　　要：为有效判定配电网短路故障区域,基于模糊RBF神经网络进行配电网短路故障区域检测研究。通过假设配电网正方向规定短路故障信息向量,应用模糊RBF神经网络算法划分聚类输入向量,利用隶属度函数描述各维分类情况,通过模糊规则处理隶属度值得到所输出分类情况,参照短路故障信息向量搜索短路故障检测点,通过区域始末端关联矩阵监测点的定向连接关系,实现配电网短路故障区域的判定。实验结果表明:该模型故障识别精度较高,抗噪性能较强,诊断准确率较高,可进行配电网短路故障区域判定。In order to effectively determine the short-circuit fault area of the distribution network,based on the fuzzy RBF neural network,the detection of the short-circuit fault area of the distribution network is carried out.By assuming that the short-circuit fault information vector is specified in the positive direction of the distribution network,the fuzzy RBF neural network algorithm is used to divide the clustering input vector,the membership function is used to describe the classification of each dimension,and the output classification is obtained through the membership value processed by the fuzzy rules.Then,the short-circuit fault detection point is searched with reference to the short-circuit fault information vector,and the determination of the short-circuit fault area of the distribution network is realized through the directional connection relationship of the monitoring points of the association matrix at the beginning and end of the area.The experiment results show that the model has high fault identification accuracy,strong anti-noise performance and high diagnostic accuracy,which can be used to determine the short-circuit fault area of the distribution network.

关 键 词：模糊RBF神经网络 配电网 短路故障 模糊聚类 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]