基于TOPSIS和递归等权法的中长期负荷组合预测

Medium and Long-term Load Combination Forecasting Based on TOPSIS and Recursive Equalization

中长期负荷预测是电力系统规划的重要环节。针对单一负荷预测模型预测精度不足、适用性有限等缺点,采用TOPSIS和递归等权法建立组合模型。以单一预测模型各个年份拟合的相对误差绝对值的倒数为属性集,单一预测模型的种类为方案集,通过TOPSIS法对单一模型的拟合精度进行排序和筛选,确定参与组合的模型;利用递归等权法确定参与组合模型的权重。通过对某省全社会用电量进行预测,并与单一预测模型进行误差比较,结果表明所提组合模型具有较高的实用价值。

Medium and long-term power load is a key link of power system planning.In view of low accuracy and limited applicability of single load forecasting model,technique for order preference by similarity to an ideal solution(TOPSIS)and recursive equalization are used to build combination model.Reciprocal of fit rel-ative error absolute value in each year of various single forecasting models are taken as an attribute set and categories of various single forecasting models are taken as a scheme set.TOPSIS is used to sort and filter for prediction accuracy of single forecasting models,and the models of participation are identified.Then,the re-cursive equalization is use to determine the weights of the various single models.The high practical value of combination model is verified through social power consumption forecasting in a province and comparison of error with single forecasting model.