Please use this identifier to cite or link to this item:
Title: An artificial neural network hybrid with wavelet transform for short-term wind speed forecasting: A preliminary case study
Authors: Yousefi, M. 
Hooshyar, D. 
Khaksar, W. 
Sahari, K.S.M. 
Alnaimi, F.B.I. 
Issue Date: 2016
Abstract: Given the importance of an accurate wind speed forecasting for efficient utilization of wind farms, and the volatile nature of wind speed data including its non-linear and uncertain nature, the wind speed forecasting has remained an active field of research. In this study, the non-linearity of wind speed is tackled using artificial neural network and its uncertainty by wavelet transform. To avoid trial-and-error process for selection the ANN structure, the results of auto correlation factor (ACF) and partial auto correlation factor (PACF) on the historical wind speed data are employed. Instead of forecasting the time series directly, a set of better-behaved components of the data is achieved by decomposing the data using wavelet transform and are forecasted separately using a feedforward neural network. Finally, using an inverse wavelet transform, the future time series is reconstructed and the wind speed could be forecasted. The historical hourly wind speed from ABEI weather station in Idaho, United States is used for assessing the performance of the proposed algorithm. This data set is merely selected due to its availability. The data is divided to three parts of 50%, 25% and 25% for training, testing and validation respectively. The testing part of data set will be merely used for assessing the performance of the neural network which guarantees that only unseen data is used to evaluate the forecasting performance of the network. On the other hand, validation data could be used for parameter-setting of the network if required. The results shows that using wavelet transform can enhance the forecasting accuracy when it is compared with a regular neural network prediction algorithm. © 2015 IEEE.
Appears in Collections:COE Scholarly Publication

Show full item record

Google ScholarTM


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.