基于K-Means聚类和梯度提升树算法的配电网线损计算方法  

Calculation method of distribution network line loss based on improved k-means clustering and gradient boosting tree algorithm

作　　者：焦昊 王海林 陈锦铭 刘伟 JIAO Hao;WANG Hailin;CHEN Jinming;LIU Wei(Electric Power Research Institute,state Grid Jiangsu Electric Power Co.,Ltd.,Nanjing 21110,China;state Grid Jiangsu Electric Power Co.,Ltd.,Nanjing 210024,China)

机构地区：[1]国网江苏省电力有限公司电力科学研究院,南京211103 [2]国网江苏省电力有限公司,南京210024

出　　处：《自动化与仪器仪表》2022年第10期74-79,共6页Automation & Instrumentation

基　　金：国网江苏省电力公司科技项目(J2020097)。

摘　　要：针对传统配电网理论线损计算需要电气参量多、工作量大、计算结果准确率低等问题,提出一种基于改进K-Means聚类算法和GBDT(Gradient Boost Decision Tree,梯度提升树)算法的配电网线损计算的方法。先采用改进K-Means算法对配电网线损样本进行聚类分析,然后将聚类后的数据集作为GBDT算法的输入数据集训练模型,最后进行线损的计算。采用本算法与BP神经网络模型进行算例对比与分析,并利用扬州许方线路配电网实际线损值做实例验证。结果表明,所提算法具有计算快速、精度更高等优点。Aiming at the problems of traditional distribution network theoretical line loss calculations that require a lot of electrical parameters,a large workload,and low accuracy of calculation results,a method for calculating line loss of distribu-tion networks based on improved K-Means clustering algorithm and GBDT(Gradient Boost Decision Tree)algorithm is pro-posed.First,the improved K-Means algorithm is used to cluster analysis of the distribution network line loss samples,and then the clustered data set is used as the input data set training model of the GBDT algorithm,and finally the line loss is cal-culated.The calculation example is compared and analyzed using this algorithm and the BP neural network model,and the actual line loss value of the Yangzhou Xufang line distribution network is used for example verification.The results show that the algorithm proposed in this paper has the advantages of fast calculation and higher accuracy.

关 键 词：电力系统 配电网 梯度提升树 电网损耗 改进K-Means聚类算法 

分 类 号：TP202[自动化与计算机技术—检测技术与自动化装置] TM71[自动化与计算机技术—控制科学与工程]