Short-term wind power prediction based on data mining technology and improved support vector machine method: A case study in Northwest China. (20th December 2018)

• Record Type: Journal Article
• Authors: Li, Cunbin; Lin, Shuaishuai; Xu, Fangqiu; Liu, Ding; Liu, Jicheng
• Is Part Of: Journal of cleaner production. Volume 205(2018)
• Page Start: 909
• Page End: 922
• Publication Date: 2018-12-20
• Subjects: Data mining -- Wavelet transform -- Support vector machine -- Wind power prediction -- Cuckoo search; 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.2018.09.143
• Languages: English
• ISSNs: 0959-6526

Abstract

Abstract: In recent years, wind power industry has been developing rapidly as the wind resources are clean, cheap and inexhaustible. However, it is difficult to supply steady wind power generation due to the strong randomness, volatility and uncontrollability of wind energy. Therefore, it is significant to propose an efficient wind power prediction model. In this paper, a short-term wind power prediction model is proposed based on data mining technology and improved support vector machine method. In this model, data mining is employed to investigate the relationship between wind speed and wind power output and then modify the invalid original data. Then, based on wavelet transform method, the high frequency parts of the original signal can be eliminated. Next, cuckoo search algorithm is used to optimize kernel function and penalty factor of support vector machine in order to improve the accuracy of the forecast result. Finally, a wind farm located in the Northwest China is selected to perform the case study. The results indicate that the proposed model has the best performance according to the values of several error assessment indexes, including mean absolute error, mean squared error and mean absolute percentage error. Highlights: A short-term wind power forecasting method is proposed. A case study in Northwestern China is performed. The prediction accuracy of the proposed method approaches to 90%. The proposed method has the smallest values of three evaluation indexes.