基于MRF的配电网线损原因识别方法  

Cause identification method for line loss ofdistribution network based on MRF

作　　者：拜润卿 何欣 陈仕彬 BAI Run-qing;HE Xin;CHEN Shi-bin(Electric Power Research Institute,State Grid Gansu Electric Power Company,Lanzhou 730070,China)

机构地区：[1]国网甘肃省电力公司电力科学研究院,兰州730070

出　　处：《沈阳工业大学学报》2021年第3期265-269,共5页Journal of Shenyang University of Technology

基　　金：国家电网公司科技项目(522722180003).

摘　　要：为了实现对配电网中高线损位置的有效控制和快速修复,确保配电网能够持续供电,提升电力系统运行的稳定性,提出一种基于MRF的配电网线损原因识别方法.结合状态机模型建立配电网线损原因识别模型,确定与配电网线损位置有关的先验条件,利用MRF的模糊性和不确定性对配电网线损原因进行识别.结果表明,所提出方法与经典的支持向量机和神经网络分类算法相比,样本训练时间、运算时间较短,线损的原因识别正确率可达到82.6%,具有运算速度快、操作简单和识别正确率高等优点,实际应用性能良好.In order to realize the effective control and rapid repair of high-line loss positions in the distribution network,ensure the continuous power supply of distribution network,and improve the stability of power system operation,a cause identification method for the line loss of distribution network based on MRF was proposed.A cause identification model for the line loss of distribution network was established in combination with a state machine model,and the prior conditions related to the line loss positions of distribution network were determined.The causes for the line loss of distribution network were identified by using the fuzziness and uncertainty of MRF.The results show that the as-proposed method has shorter training and operation time with an accuracy of 82.6%for the line loss cause identification,compared with the classical support vector machine and the neural network classification algorithms.This method possesses advantages,e.g.fast calculation speed,simple operation and high recognition accuracy,and exhibits good performance for practical application.

关 键 词：状态机模型 原因识别 配电网 先验条件 模糊性 不确定性 电力系统 稳定性 

分 类 号：TM711[电气工程—电力系统及自动化]