Abstract
Ensuring the efficient scheduling of power systems and enhancing the grid's renewable energy integration efficiency heavily relies on the precision and dependability of wind power prediction. Based on this, an ultra-short-term wind power point-interval prediction framework is proposed. Firstly, the wind power data is preprocessed. Secondly, Random Forest(RF) is used to filter the factors affecting wind power data to eliminate redundant features. Third, using improved sparrow search algorithm(ISSA) to ascertain the optimal parameters in variational mode decomposition(VMD), the wind power sequence is decomposed to obtain a more regular sequence. Then, combining ISSA, bidirectional gated recurrent unit(BiGRU) and Attention, the ISSA-BiGRU-Attention model is constructed for the prediction of wind power subsequences. Finally, using the kernel density estimation(KDE) of Grid Search(GS)-Cross-Validation(CV), prediction intervals for wind power at varying confidence levels are calculated. The experimental results show that the RF-ISSA-VMD-ISSA-BiGRU-Attention prediction model has the great prediction accuracy compared with the comparison models. In dataset 1, the model possesses a good prediction result with an R2 of 0.997408 which is an improvement of 7.98 % compared with LSTM. In addition, the GS-CV-KDE interval prediction model enhances the practicability of prediction outcomes and offers more efficient prediction information for the assurance of power system safety and reliability.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.