Abstract
Accurate wind speed prediction has been becoming an indispensable technology in system security, wind energy utilization, and power grid dispatching in recent years. However, it is an arduous task to predict wind speed due to its variable and random characteristics. For the objective to enhance the performance of forecasting short-term wind speed, this work puts forward a hybrid deep learning model mixing time series decomposition algorithm and gated recurrent unit (GRU). The time series decomposition algorithm combines the following two parts: (1) the complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN), and (2) wavelet packet decomposition (WPD). Firstly, the normalized wind speed time series (WSTS) are handled by CEEMDAN to gain pure fixed-frequency components and a residual signal. The WPD algorithm conducts the second-order decomposition to the first component that contains complex and high frequency signal of raw WSTS. Finally, GRU networks are established for all the relevant components of the signals, and the predicted wind speeds are obtained by superimposing the prediction of each component. Results from two case studies, adopting wind data from laboratory and wind farm, respectively, suggest that the related trend of the WSTS can be separated effectively by the proposed time series decomposition algorithm, and the accuracy of short-time wind speed prediction can be heightened significantly mixing the time series decomposition algorithm and GRU networks.
Highlights
Wind power has been vigorously developed because it is clean and renewable
For the objective to enhance the forecasting performance of short-term wind speed, a hybrid deep learning model is proposed. This model integrates a time series decomposition algorithm and gated recurrent unit (GRU) networks, and has the following characteristics: (1) Several intrinsic mode functions (IMFs) and a residue can be obtained by employing CEEMDAN to decompose the normalized wind speed time series (WSTS); (2) wavelet packet decomposition (WPD) is adopted to further decompose the first IMF that contains the complex and high-frequency signal of the raw data; (3) GRU networks are established for all the resulting component signals, including training samples and predicting the outputs; (4) the eventual predicted values are calculated by superimposing the prediction results of all components
As a deep learning algorithm, GRU can effectively capture the nonlinear fluctuation of WSTS, can significantly enhance the forecasting performance aiming at short-term wind speed
Summary
Wind power has been vigorously developed because it is clean and renewable. A great number of wind plants have been constructed, and offshore wind power has received plenty of attention. For the objective to enhance the forecasting performance of short-term wind speed, a hybrid deep learning model is proposed This model integrates a time series decomposition algorithm and GRU networks, and has the following characteristics: (1) Several IMFs and a residue can be obtained by employing CEEMDAN to decompose the normalized WSTS; (2) WPD is adopted to further decompose the first IMF that contains the complex and high-frequency signal of the raw data; (3) GRU networks are established for all the resulting component signals, including training samples and predicting the outputs; (4) the eventual predicted values are calculated by superimposing the prediction results of all components. As a deep learning algorithm, GRU can effectively capture the nonlinear fluctuation of WSTS, can significantly enhance the forecasting performance aiming at short-term wind speed.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have