基于Kohonen聚类和SVM组合算法的电网日最大负荷预测  

Forecasting of the Daily Maximum Load Based on a Combined Model of Kohonen Clustering and SVM

作　　者：马小慧[1] 阳育德[1,2] 龚利武 

机构地区：[1]广西大学电力系统最优化研究所,广西南宁530004 [2]广西电力系统最优化与节能技术重点实验室,广西南宁530004

出　　处：《电网与清洁能源》2014年第2期7-11,共5页Power System and Clean Energy

基　　金：广西理工科学实验中心重点项目(LGZX201209);广西教育厅高校科研项目(201203YB011);广西大学科研基金项目(XBZ120037)~~

摘　　要：电力系统负荷受到气象等多种因素的共同影响,为提高电力系统短期预测的准确度,提出了一种将Kohonen聚类和SVM结合的预测算法。该方法考虑到电力系统变化周期性的特点,通过对学习样本进行聚类,选出具有相同特征的历史相似日,构造相似日训练样本并利用SVM模型进行预测。样本特征向量中同时包含了多种气象因子及节假日等因素,经过仿真实验,验证了该方法的有效性。Power system load is affected by a variety of factors including meteorological conditions. To improve the power system short-term prediction accuracy, this paper pre- sents a prediction algorithm combined with Kohonen clustering and SVM. The methodology takes into consideration of the characteristics of cyclical changes in power system, and chooses the similar days with the same characteristics in history by clustering the leaning sample and constitutes the training samples for the similar days, and prediction is conducted by the SVM model. The sample feature vector contains a variety of factors such as meteorological factors and holidays. The simula- tion has verified the effectiveness of the algorithm.

关 键 词：Kohonen聚类 支持向量机 负荷预测 

分 类 号：TM715[电气工程—电力系统及自动化]