Study on environment-concerned short-term load forecasting model for wind power based on feature extraction and tree regression. (10th August 2020)

• Record Type: Journal Article
• Authors: Liu, Jicheng; Li, Yinghuan
• Is Part Of: Journal of cleaner production. Volume 264(2020)
• Publication Date: 2020-08-10
• Subjects: Short-term load forecasting -- Wind power -- Feature extraction -- Tree regression -- Interpretative structural modeling -- Random forest; Factory and trade waste -- Management -- Periodicals; Manufactures -- Environmental aspects -- Periodicals; Déchets industriels -- Gestion -- Périodiques; Usines -- Aspect de l'environnement -- Périodiques; 628.5
• DOI: 10.1016/j.jclepro.2020.121505
• Languages: English
• ISSNs: 0959-6526

Abstract

Abstract: With the increasingly severe situation of environment and gradually depleted energy, wind power has been put on the schedule and comes to an efficient way of power supply and sustainable development. However, wind power generation depends on weather to a great extent, throwing great threats to demand side and supply side. To ensure the normal operation of power system, load forecasting especially for the short term has become a necessary phase. Considering from external environment, this paper aims to identify the influence factors of wind power load and conduct short-term load forecasting with random forest coupling interpretative structural modeling, so as to realize a more accurate and reliable prediction. The data containing information about climate and power load is gathered and put into the calculation of case study. Results show that proposed model performs well, which can be verified and validated through scenario analysis, sensitivity analysis, and comparison analysis. Furthermore, according to the results of factors identification and load forecasting, some suggestions are put forward to deal with the environmental impact on wind power. Graphical abstract: Image 1 Highlights: A novel prediction model combining interpretative structural modeling with random forest is constructed for load forecasting. Interpretative structural modeling gets improvements by transformation from qualitative to quantitative. The environmental influence factors of wind power are identified and the hierarchy structures between them are divided. The proposed model is verified by scenario analysis, sensitivity analysis, and comparison analysis. Suggestions about how to deal with the impact of external environment on wind power are proposed based on the results.