基于支持向量机的负荷非线性组合预测研究

A New Load Nonlinear Combination Forecasting Method Based on Support Vector Machines

短期负荷预测是电力系统调度管理部门制定开停机计划及在线安全分析的基础,为了提高电力行业经济效益和社会效益,精确的负荷预测是最重要的工作内容。近年来的研究表明,组合预测比单项预测具有更高的精度。为了提高短期负荷的预测精度,提出一种基于支持向量机的负荷非线性组合预测方法,并与BP神经网络组合预测相比较,测试结果表明了该方法的有效性与优越性。

Period power system load forecasting is scheduling downtime to open the development of management plans and on-line security analysis. In order to improve the economic and social benefits of electricity industry, Accurate load forecasting is the most important work. It has been shown that combining forecasts may produce more accurate forecasts than individual ones in recent years. In order to improvethe accuracy of short-term load forecasting, this paper presents a new nonlinear composite forecasting method for load forecasting based on support vector machines. Analysis of the experimental results proved that the algorithm could achieve much effective than that of BP neutral network.