Model of selecting prediction window in ramps forecasting. (August 2017)

• Record Type: Journal Article
• Authors: Ouyang, Tinghui; Zha, Xiaoming; Qin, Liang; Xiong, Yi; Huang, Heming
• Is Part Of: Renewable energy. Volume 108(2017)
• Page Start: 98
• Page End: 107
• Publication Date: 2017-08
• Subjects: Wind power ramp events -- Ramp prediction -- Prediction time window -- Non-ramp data -- Genetic algorithm; Renewable energy sources -- Periodicals; Power resources -- Periodicals; Énergies renouvelables -- Périodiques; Ressources énergétiques -- Périodiques; 333.794
• DOI: 10.1016/j.renene.2017.02.035
• Languages: English
• ISSNs: 0960-1481

Abstract

Abstract: Prediction of wind power ramp events is important to the stability operation of power system, it is realized by combining wind power prediction of several continuous units for long-term power prediction, then using detecting algorithms to extract ramps. A prediction unit is a prediction time window. Its size impacts the accuracy of predicting ramps, and an optimization model is proposed to select the suitable window size. First, a swinging door algorithm is applied to extract ramp events from historical data. A model for optimizing the time window size is established based on the minimum non-ramp data in a ramp window. The solution of the proposed model is discussed, including the selection of variables, constraints and algorithm. The model presented in this paper is tested, and performance of selected time window is discussed. Computational analysis demonstrates the validity of the model. Highlights: Prediction time window affects the precision of wind power ramp prediction. Non-ramp data in ramp windows is used in modelling to select an optimal window size. A new method is proposed by using time windows as units in ramp prediction. Genetic algorithm is used to solve the optimization model, and SVM for prediction.