采用支持向量机和模拟退火算法的中长期负荷预测方法  

Mid-long Term Load Forecasting Based on Simulated Annealing and SVM Algorithm

作　　者：李瑾[1] 刘金朋[2] 王建军[2] 

机构地区：[1]北京科技大学经济管理学院,北京市海淀区100083 [2]华北电力大学经济与管理学院,北京市昌平区102206

出　　处：《中国电机工程学报》2011年第16期63-66,共4页Proceedings of the CSEE

基　　金：中电联预研基金项目(92103804)~~

摘　　要：准确的中长期负荷预测能够提高电力系统的经济效益和社会效益。分析了支持向量机(support vector machine,SVM)模型,并针对利用支持向量机进行负荷预测需要人为地确定相关参数的不足,提出了利用支持向量机进行中长期预测的新方法。该方法利用模拟退火(simulated annealing,SA)算法自动优化参数。实例验证结果表明,所提出的方法可以有效地选取支持向量机模型的参数,降低支持向量机的建模误差和测试误差,该方法与利用默认参数支持向量机进行预测的方法相比,有效地提高了负荷预测精度。Accurate mid-long term load forecasting can improve the economic and social benefits of power system. The model of support vector machine （SVM） was analyzed, and a new mid-long term load forecasting method based SVM method was proposed in order to overcome the shortcoming that some parameters have to be determined by experience. The simulated annealing algorithm was applied to optimize the parameters in the method. An example proved that the proposed method can efficiently select the parameters of the SVM method. It can reduce the support vector machine modeling training error and testing error. Compared to the method using the default parameters SVM method, the accuracy of load forecasting is effectively improved.

关 键 词：电力系统 中长期负荷预测 模拟退火 支持向量机 

分 类 号：TM71[电气工程—电力系统及自动化]