A hybrid model based on data preprocessing for electrical power forecasting. (January 2015)

• Record Type: Journal Article
• Authors: Xiao, Liye; Wang, Jianzhou; Yang, Xuesong; Xiao, Liyang
• Is Part Of: International journal of electrical power & energy systems. Volume 64(2015:Jan.)
• Page Start: 311
• Page End: 327
• Publication Date: 2015-01
• Subjects: Electrical power forecasting -- Hybrid model -- Forecasting accuracy -- Model selection; Electrical engineering -- Periodicals; Electric power systems -- Periodicals; Électrotechnique -- Périodiques; Réseaux électriques (Énergie) -- Périodiques; Electric power systems; Electrical engineering; Periodicals; 621.3
• DOI: 10.1016/j.ijepes.2014.07.029
• Languages: English
• ISSNs: 0142-0615

Abstract

Highlights: Improve the accuracy and stability significantly of electrical power forecasting. Propose a hybrid model based on several artificial neural networks. Use model selection in process of hybrid model. Apply longitudinal data selection to conduct the input data. Abstract: Electrical power forecasting plays a vital role in power system administration and planning. Inaccurate forecasting can lead to the waste of scarce energy resources, electricity shortages, and even power grid collapses. On the other hand, accurate electricity power forecasting can enable reliable guidance for the planning of power production and the operation of a power system, which is also important for the continued development of the electrical power industry. Although thousands of scientific papers address electricity power forecasting each year, only a small number are devoted to developing a general model for electricity power prediction that improves performance in different cases. This paper proposes a hybrid forecasting model for electrical power prediction that incorporates several artificial neural networks and model selection. To evaluate the forecasting performance of the proposed model, this paper uses half-hourly electrical power data of the State of Victoria and New South Wales of Australia as a case study. The experimental results clearly indicate that for this particular dataset, the forecasting performance of the proposed hybrid model is outstanding compared to that of the single forecasting model.