Research on a combined model based on linear and nonlinear features - A case study of wind speed forecasting

RENEWABLE ENERGY

卷:130

Article

Zhang, Kequan¹  Qu, Zongxi¹  Dong, Yunxuan²  Lu, Haiyan³  Leng, Wennan¹  Wang, Jianzhou⁴  Zhang, Wenyu¹ 

[1] Lanzhou Univ, Coll Atmospher Sci, Key Lab Arid Climat Change & Reducing Disaster Ga, Lanzhou 730000, Gansu, Peoples R China

[2] Lanzhou Univ, Sch Math & Stat, Lanzhou, Gansu, Peoples R China

[3] Univ Technol, Fac Engn & Informat Technol, Sydney, NSW, Australia

[4] Dongbei Univ Finance & Econ, Sch Stat, Dalian, Peoples R China

关键词: Wind speed forecasting;    Combined model;    Artificial fish swarm algorithm;    Ant colony optimization;   

DOI  :  10.1016/j.renene.2018.05.093

来源: Elsevier

【 摘 要 】

As one of the most promising sustainable energy sources, wind energy is being paid more attention by the researchers. Because of the volatility and instability of wind speed series, wind power integration faces a severe challenge; thus, an accurate wind energy forecasting plays a key role in smart grid planning and management. However, many traditional forecasting models do not consider the necessity and importance of data preprocessing and neglect the limitation of using a single forecasting model, which leads to poor forecasting accuracy. To solve these problems, a novel combined model based on two linear and four nonlinear forecasting algorithms is proposed to adapt both the linear and nonlinear characteristics of the wind energy time series. In addition, a modified Artificial Fish Swarm Algorithm and Ant Colony Optimization (AFSA-ACO) algorithm is proposed and employed to determine the optimal weight coefficients of the combined models. To verify the forecasting performance of the developed combined model, several experiments were implemented by using 10-min interval wind speed data in Shandong, China. Then, one-step (10-min), three-step (30-min) and five-step (50-min) predictions were conducted. The experimental results indicate that the developed combined model is remarkably superior to all benchmark models for the high precision and stability of wind-speed predictions. (C) 2018 Elsevier Ltd. All rights reserved.

【 授权许可】

Free