基于灰色理论和季节指数的月度用电量预测模型  

Monthly electricity consumption prediction model based on grey theory and seasonal index model

作　　者：晏伟宸 YAN Weichen(Power Service Center,Shandong University of Political Science and Law,Jinan 255014,China)

机构地区：[1]山东政法学院动力服务中心,济南255014

出　　处：《黑龙江电力》2024年第4期306-310,共5页Heilongjiang Electric Power

摘　　要：传统季节指数预测模型采用平均数作为季节指数,存在精度不足的问题。为获得简便而准确的月度用电量预测模型,在传统季节指数预测模型的基础上,引入非线性趋势函数,并使用灰色预测模型GM(1,1)增强对随机性的预测,得到改进的季节指数预测模型。通过对中国第一、二产业各4个周期44组月度用电量数据的计算分析,验证改进算法的有效性。计算结果表明,改进算法可以更准确地预测月度用电量,较传统季节指数模型和传统GM(1,1)模型的MAPE指标分别降低了2%~21%和4%~10%,RMSE指标分别降低了39%~67%和22%~76%,是一种简单、有效的算法。Traditional seasonal index prediction model uses average value as seasonal index,so there exists inaccuracy problem.In order to predict monthly electricity consumption accurately and easily,introduces non-linear trend function based on traditional seasonal index algorithm and grey prediction model GM(1,1),aiming to better predict randomness.Improved seasonal index algorithm is produced.Through separate calculation of 44 groups of data in 4 periods in China s first sector and second sector,effectiveness of the improved algorithm is[JP2]proved.Calculation results show that,compared with traditional seasonal index algorithm and traditional GM(1,1),[JP]the improved algorithm s MAPE reduces 2%~21%and 4%~10%,and the improved algorithm s RMSE reduces 39%~67%and 22%~76%.To sum up,the improved algorithm is an effective and simple method.

关 键 词：灰色理论 季节指数模型 月度用电量 预测模型 MAPE RMSE 

分 类 号：TM715[电气工程—电力系统及自动化]