基于诱导有序加权平均算子和马尔可夫链的中长期电力负荷组合预测模型.pdfVIP

第 34 卷 第 3 期 电 网 技 术 Vol. 34 No. 3 2010 年 3 月 Power System Technology Mar. 2010 文章编号：1000-3673 （2010 ）03-0150-07 中图分类号：TM 715 文献标志码：A 学科代码：470·4051 基于诱导有序加权平均算子和马尔可夫链的 中长期电力负荷组合预测模型 龙瑞华，毛弋，毛李帆，孙东杰，张芳明，张婷，陈宇哲 （湖南大学电气与信息工程学院，湖南省 长沙市 410082 ） A Combination Model for Medium- and Long-Term Load Forecasting Based on Induced Ordered Weighted Averaging Operator and Markov Chain LONG Rui-hua, MAO Yi, MAO Li-fan, SUN Dong-jie, ZHANG Fang-ming, ZHANG Ting, CHEN Yu-zhe (College of Electrical and Information Engineering，Hunan University, Changsha 410082, Hunan Province, China) ABSTRACT: To remedy the defects in traditional combination 负荷预测 method for medium- and long-term load forecasting, a method, 0 引言 which combines the induced ordered weighted averaging operator (IOWA) with Markov chain (MC) to construct an 中长期负荷预测是电网规划的基本工作，其精 IOWA-MC combinative forecasting model, is proposed. 度的高低直接影响电网规划的质量。但是，由于 According to the order of fitting accuracy of each forecasting 影响负荷预测的直接和间接因素较多、随机性大， method at each time point the proposed model endows them 单一的负荷预测模型难以准确地描述电力负荷变 with weights to ensure the correlativity of weight coefficients 化规律。组合预测通过整合多种预测模型包含的 and fitting accuracy in combinative forecasting model at any time point; meanwhile, by use of MC the forecasted accuracy 信息，能有效改善模型预测精度[1-4] 。目前，大多 condition of each forecasting method at the forecasted time 数组合预测模型基于不变权组合理论，即通过最 point can be qualitatively surmised, thus its weight coefficient 优化模型或其它方法计算出各个单项预测方法在 at the forecasted time point can be determined. Calculation 组合预测中的权系数，然后假定权系数不变进行 result