基于滤波技术的上海日最大电力负荷气象预报模型

电力负荷与气象条件密切相关，为建立上海市日最大电力负荷的预报模型，利用2010—2013年上海市日最大电力负荷数据及同期气象资料，分析日最大电力负荷的时间变化特征及其与气象因子的相关性，并基于滤波技术将日最大电力负荷分离为时间趋势项和逐日变化项，用逐步回归方法针对冬季和夏季分别建立预测模型。结果表明：①上海日最大电力负荷的各个节假日效应存在差异，春节节假日效应持续时间最长，影响最大，国庆节假期前半段节假日效应明显大于后半段。夏季的周末效应最强。②采用逐步回归方法建立的气象预报模型效果较好，回代年和预测年的平均预测相对误差均小于5%。

Meteorological Forecast Model of Daily Maximum Electrical Load in Shanghai Based on Filter Technique

Electrical load is significantly impacted by weather conditions. Therefore, the seasonal variation, holiday effects, weekend effects, and the weather dependency of daily maximum electrical load are investigated by using the daily maximum electrical load data in Shanghai and meteorological observations from the Xujiahui weather station from May 2010 to December 2013. The stepwise regression method and filter technique are employed to build the meteorological forecast model in winter and summer. The model consists of two components: regular change component and weather sensitive component. The results show that the effects of various holidays are different. The Spring Festival has the longest and greatest influence on daily maximum power load. The holiday effect is more pronounced during the first several days of the National Day than in the other days. The weekend effect is greater in summer than in other seasons. The best model is able to explain most of variability in daily maximum electrical load, with the mean prediction relative error less than 5%.