基于温度近因效应的多元线性回归电力负荷预测

Load Forecasting of Multiple Linear Regression Based on Temperature Recency Effect

为分析负荷与温度的关系,充分利用有用的气象信息,以结构简单、解释能力强的回归预测方法为基础,提出基于温度近因效应的多元线性回归负荷预测模型。在构建回归模型时,引入虚拟变量描述负荷在年、周和日周期上的周期性变化规律,对周分类时兼顾考虑节假日负荷特殊性及与休息日负荷的相似性,对节假日及其邻近日做相应转换;在近因效应方面,采用滞后时温度和24h移动平均温度。研究结果表明,考虑温度近因效应可较大程度提升负荷预测精度。

In order to analyze the relationship between load and temperature and make full use of meteorological information,a temperature recency effect load forecasting method based multiple linear regression with simple structure and strong explanatory performance was proposed.Dummy variables were introduced to the regression model for describing the periodic rules of the load in year,week and day.Due to the similarity of holidays and rest days,the specificity of holidays load was considered and coped with to improve the practicability of model.The lag temperature and 24 hours moving average temperature were adopted to handle the of temperature recency effect.The results show that the prediction accuracy can be greatly improved by considering the temperature recency effect.