A novel hybrid approach based on cuckoo search optimization algorithm for short‐term wind speed forecasting

• Published in: Environmental progress & sustainable energy Vol. 36; no. 3; pp. 943 - 952
• Main Authors: Zhang, Kequan, Qu, Zongxi, Wang, Jianzhou, Zhang, Wenyu, Yang, Feiyue
• Format: Journal Article
• Language: English
• Published: 01.05.2017
• Subjects: cuckoo search optimization algorithm; hybrid model; neural network; wind speed prediction
• ISSN: 1944-7442; 1944-7450
• DOI: 10.1002/ep.12533

A type of renewable energy, wind power, which has a large generating capacity, has increasingly captured the world's attention. Forecasting wind speed is of great significance in wind‐related engineering studies, the planning and designing of wind packs, wind farm management and the integration of wind power into electricity grids. Because of the chaotic nature and intrinsic complexity of wind speed, reducing forecasting errors related to wind speed has been an important research subject. This article proposes a hybrid forecasting model based on the wavelet neural network (WNN) method and ensemble empirical mode decomposition (EEMD) de‐noise technology. Moreover, the cuckoo search optimization (CSO) algorithm is employed to optimize the parameters of the WNN model to avoid the over‐fitting problem and improve fitting accuracy. A case study is performed to forecast 10‐min wind speed data for wind farms in the coastal areas of China, and a comparison of the forecasting results with other models demonstrates that the proposed model can provide desirable forecasting results and improve forecasting accuracy. © 2017 American Institute of Chemical Engineers Environ Prog, 36: 943–952, 2017