A new semiparametric and EEMD based framework for mid-term electricity demand forecasting in China: Hidden characteristic extraction and probability density prediction. (December 2015)

• Record Type: Journal Article
• Authors: Shao, Zhen; Gao, Fei; Yang, Shan-Lin; Yu, Ben-gong
• Is Part Of: Renewable & sustainable energy reviews. Volume 52(2015:Dec.)
• Page Start: 876
• Page End: 889
• Publication Date: 2015-12
• Subjects: Mid-term electricity demand -- Forecasting -- Semi-parametric regression -- Ensemble Empirical Mode Decomposition -- Probability density forecasts; Renewable energy sources -- Periodicals; Power resources -- Periodicals; Énergies renouvelables -- Périodiques; Ressources énergétiques -- Périodiques; 333.794
• DOI: 10.1016/j.rser.2015.07.159
• Languages: English
• ISSNs: 1364-0321

Abstract

Abstract: One of the prime missions of the mid-long term electricity demand forecasting involves investigating the multidimensional fluctuation characteristics so that planners can sharpen their understanding of the intrinsic variation trend. To some extent, different facets of the actual fluctuation characteristics can be separated into components, and we can implement more targeted forecast by treating them separately and making more effective response to these characteristics. The purpose of this study is to present a new framework of mid-term demand forecasting along with the semi-parametric model and fluctuation feature decomposition technology, and to generate practical and reliable probability forecast through the application of measurable amount of external variables. To demonstrate the effectiveness, the framework is applied to the case study concerning the identification of potential volatility characteristic and long-term forecast (24-steps point forecasts and longer time scale probability forecasts up to January 2021) in Suzhou and Guangzhou, China. As expected, our proposed approach shows an outperformance result compare to the common decomposition forecast methods. The results also revealed that the extracted components present the opportunity to capture some of the hidden, but potentially important characteristics (e.g., climate fluctuation and economic development) from the original consumption data.