基于SCE-UA支持向量机的短期电力负荷预测模型研究  

Research on short-term electricity load forecasting model using support vector machine based on SCE-UA algorithm

作　　者：李刚[1] 程春田[1] 曾筠[1] 林剑艺[1] 

机构地区：[1]大连理工大学水电与水信息研究所,辽宁大连116024

出　　处：《大连理工大学学报》2011年第2期263-268,共6页Journal of Dalian University of Technology

基　　金：国家自然科学基金资助项目(50909011);"九七三"国家重点基础研究发展计划资助项目(2009CB226111)

摘　　要：支持向量机(support vector machine,SVM)作为一种新颖的机器学习方法已成功应用于短期电力负荷预测,然而应用研究发现SVM算法性能参数的设置将直接影响负荷预测的精度.为此在对SVM参数性能分析的基础上,提出了SCE-UA(shuffled complex evolution-University of Arizona)支持向量机短期电力负荷预测模型建模的思路及关键参数的选取,在建模过程中引入了径向基核函数,简化了非线性问题的求解过程,并应用SCE-UA算法辨识SVM的参数.贵州电网日96点负荷曲线预测的实际算例表明,所提SCE-UA支持向量机模型不仅克服了SVM参数选择的盲目性,而且能提高预测准确率,是一种行之有效的短期电力负荷预测模型.Support vector machine（SVM） is a novel type of learning machine,which has been successfully applied to short-term electricity load forecasting.However,its application indicates that how to confirm the parameters of SVM algorithm directly affects forecasting precision.On the basis of analyzing the parameter performance of SVM for regression estimation,a short-term electricity load forecasting model SCE-UA（shuffled complex evolution-University of Arizona） based on SVM is presented.In the process of establishing the model,the radial basis kernel function is introduced,which simplifies the course of solving non-linear problems,and the SCE-UA algorithm is applied to identifying the parameters of SVM.The model is applied to short-term electricity load forecasting using the actual 96 points daily data from Guizhou power grid.The results show that the proposed model not only overcomes the blindness of identifying SVM parameters,but also increases forecasting precision.It is a feasible and effective short-term electricity load forecasting model.

关 键 词：负荷预测 支持向量机 SCE-UA 相似日 

分 类 号：TM74[电气工程—电力系统及自动化]