互信息在电力系统中长期负荷预测中的应用  

Application of mutual information to power system medium and long-term load forecasting

作　　者：原媛[1] 顾洁[1] 黄薇[2] 陈春琴[3] 高宇博[3] 

机构地区：[1]上海交通大学电气工程系,上海200240 [2]上海市电力公司,上海200122 [3]上海市电力公司市区供电公司,上海200080

出　　处：《华东电力》2009年第2期236-239,共4页East China Electric Power

摘　　要：传统电力负荷预测方法常以电力负荷自身的历史序列特征外推预测未来负荷,或者用确定性相关关系来模拟少数几种电力负荷关联因素的影响,导致对负荷变化规律的把握存在缺陷。信息领域中的互信息理论用于分析两个随机变量间的相互依赖程度,其中包含了线性和非线性关联关系,用以描述电力负荷及其影响因素的关联程度,为进一步的预测工作提供依据。应用互信息理论,建立了电力负荷预测的互信息网络模型,根据各影响因素与电力负荷间的互信息对其量化、比较和逐层筛选,在得到各经济社会指标与负荷之间的关联程度的基础上,对未来负荷进行预测。将本模型应用于实际电力系统预测,得到了令人满意的预测结果。The traditional load forecasting methods are often based on the historical sequence characteristics of the load itself, or use definitude relationship to simulate influences of several related factors, which leads to defects in load prediction. The mutual information （MI） which is used to analyze the dependence degree of two random variables, contains linear correlation and non-linear correlation and can be used to describe the relationship of load and its relative elements. According to MI, the MINI for load forecasting was built. By using the model, the relative elements could be quantified, compared, and selected, and the relationship between each economy index and the load was obtained. Practice proved the model effective.

关 键 词：互信息 中长期负荷预测 关联关系 决策属性 

分 类 号：TM714[电气工程—电力系统及自动化]