基于季节调整和回归分析法的月售电量预测方法研究

Study of Monthly Electricity Sales Forecasting Based on Seasonal Adjustment and Multi Factor orrection Method

准确的月售电量预测,对电力公司的业绩考核、控制利润平衡以及电力营销工作等都有着极其重要的作用。通过对天津市的月度售电量历史数据进行全面分析和深入研究,探索其增长变化规律及周期性波动特征。首先提出使用Holt-Winters指数平滑法对天津市月度售电量进行季节调整,以刻画其趋势性和季节性;再根据天津的实际数据展开月售电量主要影响因素分析,重点分析考虑经济、温度、春节分布与月售电量之间的关系,并进一步对季节调整后的数据进行修正,建立改进后的预测模型。结果表明,引入多种影响因子后的改进方法充分结合了季节调整对趋势时间序列精度预测的优点,又有利于提高月售电量预测精度。

To explore the growth regularity and periodic fluctuation characteristics,this paper makes a comprehensive analysis and in-depth study on the historical data of monthly electricity sales in Tianjin and proposes a method for forecasting monthly electricity sales. First,the seasonal adjustment of monthly electricity sales is carried out by using Holt-Winters exponential smoothing method to describe the trend and seasonal; Then,according to the actual data of Tianjin launched a month the sale of electricity mainly focuses on the analysis of influence factor analysis,considering the temperature distribution and the Spring Festival,random events and relationship between the monthly consumption,further to the seasonally adjusted data correction,establish the improved prediction model. The results show that the improved method after introducing a variety of factors fully combines the advantages of seasonal adjustment to the prediction of the trend time series accuracy,and also helps to improve the prediction accuracy of the monthly electricity sales.