摘要
为提高配电网理论线损计算的效率和精度,提出了基于系统聚类和果蝇优化支持向量回归机的配电网理论线损计算方法。先对样本数据进行系统聚类,将样本分成几个内部数据比较相似的群,每群分别建立支持向量回归模型,并用果蝇优化算法动态寻找最优的训练参数,克服了支持向量回归机训练参数选取的盲目性,提高了支持向量回归机的学习效率和计算精度。对两个不同规模的实际配电网进行测算仿真,结果表明,所提方法具有较高的计算精度和实用价值,可以为供电企业快速评估配电网线损提供科学辅助工具。
To improve the efficiency and accuracy of distribution network line loss calculation, proposing theoretical line loss calculation method based on system cluster and fruit fly optimization algorithm SVR for distribution network. First systematic cluster sample data,divide the sample data into several internal similar data groups,build support vector regression model for each group separately,and fruit fly optimization algorithm to search the optimal training parameters for SVR,overcome the blindness of selecting the SVR training parameters,improving the learning efficiency and the accuracy of SVR. Calculating two practical line data of different sizes of distribution network,the results show that the proposed method has higher accuracy and practical value,providing the power companies scientific support tools for evaluating distribution network line loss quickly.
出处
《煤矿机电》
2016年第3期11-14,共4页
Colliery Mechanical & Electrical Technology
关键词
配电网
理论线损
系统聚类
果蝇优化
支持向量回归机(SVR)
distribution network
theoretical line loss
system cluster
fruit fly optimization
support vector regression(SVR)