基于K-Means聚类和Boosting算法的配电网线损计算方法  

Distribution Network Line Loss Calculation Method Based on K-Means Clustering and Boosting Algorithm

作　　者：马芳 张晨晖 MA Fang;ZHANG Chenhui(State Grid Shaanxi Electric Power Company Yulin Electric Power Supply Company,Yulin 719000,China)

机构地区：[1]国网陕西省电力公司榆林供电公司,陕西榆林719000

出　　处：《通信电源技术》2024年第1期1-3,共3页Telecom Power Technology

摘　　要：传统线损计算方法所需电气参数较多且计算过程烦琐,导致配电网线损计算结果精度较低,因此提出了一种基于K-Means聚类和Boosting算法的配电网线损计算方法。先采用K-Means聚类算法挖掘配电网的线路负荷有功电量、线路负荷无功电量、线路长度及线路负载率等电气特征指标,再将电气特征指标作为Boosting算法线损预测模型的输入数据,经过模型训练完成配电网线损的预测计算。实验结果表明,该设计方法的线损计算值与真实值之间的误差仅为4.27%,具有较高的配电网线损计算精度。Due to the large number of electrical parameters required by traditional line loss calculation methods and the cumbersome calculation process,the accuracy of distribution network line loss calculation results is low.A distribution network line loss calculation method based on K-Means clustering and Boosting algorithm is proposed.Firstly,the K-Means clustering algorithm is used to mine the electrical characteristic indicators of the distribution network,such as the active power and reactive power of the line load,the length of the line,and the load rate of the line.Then,the electrical characteristic indicators are used as input data for the Boosting algorithm based line loss prediction model.After model training,the prediction calculation of the distribution network line loss is completed.The experimental results show that the error between the calculated value of line loss under the design method and the actual value is only 4.27%,which has a high accuracy in the calculation of distribution network line loss.

关 键 词：K-MEANS聚类 BOOSTING算法 配电网线损 线损计算 

分 类 号：TM74[电气工程—电力系统及自动化]