摘要
在对主变负荷进行预测时,通常需要采用不同的预测方法,而仅得到不同方法下的预测结果是不够的,还需知道对各主变而言哪种预测方法最适宜、相应的预测结果最可靠。提出了一种最佳算法选择的决策树法,该方法在对已有的5种负荷预测方法的特性进行分析后,选出6个关键特征属性,利用历年负荷预测结果数据建立一个由特征属性值确立最佳预测方法的决策树,最后将所得的决策树应用于目标年最佳预测方法的选取。针对上海市松江区110(35)kV主变作实例分析,结果表明决策树方法能够较准确地判别出适应各个主变的最佳算法。
Different forecasting methods are used for forecasting the load of main transformers. Merely obtaining the prediction result is not enough,the most appropriate method should be selected for each transformer based on its own conditions. This paper proposes a method to select the optimal forecasting method by using the decision tree. This method uses the load data of previous years to establish a decision tree,which contains six characteristic properties,and then applies the decision tree to the selection of optimal forecasting method for the target year. Tests have been done to the 110(35) kV main transformers of Shanghai Songjiang District,the result indicates that the method could determine the optimal forecasting method adapted to the transformers with good accuracy.
出处
《电网与清洁能源》
2014年第3期93-97,共5页
Power System and Clean Energy
关键词
主变负荷预测
决策树
特征属性
最佳算法选择
main transformer load forecasting
decision tree
characteristic property
optimal method selection
