基于中、短时间尺度的春节负荷特征滚动预测

Spring Festival Load Characteristic Time-Process Forecasting Based on Medium and Short Time Scales

基于历史数据挖掘春节假日负荷规律,建立春节期间各日负荷特征的预测方法。首先,解耦周末假日的基础负荷,并通过负荷年增长率的识别预测春节假日的基础负荷。其次,定义气象特征,并以此为基础选取与春节假日气象相似的周末假日,基于相似周末拟合春节假日的中、短时间尺度负荷预测模型。最后,进一步建立春节假日负荷中、短时间尺度的综合预测模型,提高预测精度。采用某实际电网历史数据验证了该方法的有效性。

Based on historical data to explore the load law of Spring Festival,it establishes a method for predicting the load characteristics of each day during the Spring Festival.First,it decouples the basic load of weekends,and predicts the basic load of the Spring Festival through the identification of the annual growth rate of the load.Secondly,it defines the meteorological characteristics to select weekends which is similar to the Spring Festival based on this,and fits the mid-term and short-term load forecasting models of the Spring Festival based on similar weekends.Finally,it further establishes medium-term and short-term comprehensive forecasting model of the Spring Festival load to improve the forecasting accuracy.The validity of the method proposed is verified by the historical data of an actual power grid.