A SARIMAX coupled modelling applied to individual load curves intraday forecasting

被引:14
作者
Bercu, Sophie [1 ]
Proia, Frederic [2 ,3 ]
机构
[1] EDF Rech & Dev, Dept ICAME, F-92141 Clamart, France
[2] Univ Bordeaux 1, Inst Math Bordeaux, UMR 5251, F-33405 Talence, France
[3] Univ Bordeaux 1, INRIA Bordeaux, Team ALEA, F-33405 Talence, France
关键词
SARIMA(X) modelling; time series analysis; exogenous covariates; forecasting; seasonality; stationarity; individual load curve; TIME-SERIES PREDICTION; ELECTRICITY LOAD; PRICES;
D O I
10.1080/02664763.2013.785496
中图分类号
O21 [概率论与数理统计]; C8 [统计学];
学科分类号
020208 ; 070103 ; 0714 ;
摘要
A dynamic coupled modelling is investigated to take temperature into account in the individual energy consumption forecasting. The objective is both to avoid the inherent complexity of exhaustive SARIMAX models and to take advantage of the usual linear relation between energy consumption and temperature for thermosensitive customers. We first recall some issues related to individual load curves forecasting. Then, we propose and study the properties of a dynamic coupled modelling taking temperature into account as an exogenous contribution and its application to the intraday prediction of energy consumption. Finally, these theoretical results are illustrated on a real individual load curve. The authors discuss the relevance of such an approach and anticipate that it could form a substantial alternative to the commonly used methods for energy consumption forecasting of individual customers.
引用
收藏
页码:1333 / 1348
页数:16
相关论文
共 32 条
[1]  
[Anonymous], BIOMETRIKA
[2]  
[Anonymous], 1970, J AM STAT ASSOC, DOI DOI 10.2307/2284333
[3]  
[Anonymous], 2011, North American Power Symposium IEEE, DOI DOI 10.1109/NAPS.2011.6025124
[4]  
[Anonymous], 1993, Time Series Models
[5]   A functional wavelet-kernel approach for time series prediction [J].
Antoniadis, Anestis ;
Paparoditis, Efstathios ;
Sapatinas, Theofanis .
JOURNAL OF THE ROYAL STATISTICAL SOCIETY SERIES B-STATISTICAL METHODOLOGY, 2006, 68 :837-857
[6]   Bandwidth selection for functional time series prediction [J].
Antoniadis, Anestis ;
Paparoditis, Efstathios ;
Sapatinas, Theofanis .
STATISTICS & PROBABILITY LETTERS, 2009, 79 (06) :733-740
[7]  
Bondon P., 2009, IEEE WORKSH STAT SIG
[8]  
Box G. E. P., 1976, TIME SERIES ANAL F G
[9]  
Brockwell P.J., 1996, INTRO TIME SERIES FO
[10]  
Brockwell PJ., 1991, TIME SERIES THEORY M