基于广域信息处理的配电网故障隔离技术研究  

Research on fault isolation technology of distribution network based on wide area information processing

作　　者：思勤[1] 郭杉[1] 贾俊青[1] SI Qin;GUO Shan;JIA Junqing(Inner Mongolia Power Research Institute,Hohhot 010020,China)

机构地区：[1]内蒙古电力科学研究院,内蒙古呼和浩特010020

出　　处：《电子设计工程》2024年第9期124-128,共5页Electronic Design Engineering

基　　金：内蒙古电力(集团)有限责任公司2021年科技项目(2021-52)。

摘　　要：针对分布式电源并入配电网后,传统算法进行故障检测时存在定位准确度偏低、反应速度较慢的问题,文中基于广域信息处理技术提出了一种配电网故障隔离方法。该方法采用模态分解算法将故障复杂信号分解为多种类基础小信号,使用支持向量机对这些小信号进行数据分类。但由于传统支持向量机的收敛速度较慢,因此通过引入粒子群算法对其参数加以优化,从而提升模型的运算速度。实验结果表明,在加入分布式电源的电网中,所提算法的故障定位准确率为96.7%,平均运行时间则为43.9 s,且这两项参数在对比算法中均为最优。由此证明,该算法可应用于实际工程中,为配电网故障隔离提供技术支撑。In response to the problems of low positioning accuracy and slow response speed in traditional algorithms for fault detection after the integration of distributed power sources into the distribution network,this paper proposes a fault isolation method for distribution networks based on wide area information processing technology.This method uses modal decomposition algorithm to decompose complex fault signals into multiple types of basic small signals,and uses support vector machine to classify these small signals into data.However,due to the slow convergence speed of traditional support vector machines,particle swarm optimization algorithm is also introduced to optimize its parameters,thereby improving the computational speed of the model.The experimental results show that in the power grid with distributed power sources,the proposed algorithm has a fault localization accuracy of 96.7%and an average operating time of 43.9 s,and these two parameters are the best in the comparison algorithm.This proves that the algorithm can be applied in practical engineering,providing technical support for fault isolation in distribution networks.

关 键 词：广域信息 故障隔离 模态分解法 支持向量机 粒子群优化 智能电网 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] TN99[自动化与计算机技术—计算机科学与技术]