Recurrent attention encoder–decoder network for multi-step interval wind power prediction

In the context of large-scale wind power grid integration, accurate wind power forecasting is crucial for optimizing grid scheduling and ensuring safe grid connection. This study proposes a recurrent attention encoder–decoder network for multi-step interval wind power forecasting, combining Numerica...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Energy (Oxford) Ročník 315; s. 134317
Hlavní autoři: Ye, Xiaoling, Liu, Chengcheng, Xiong, Xiong, Qi, Yinyi
Médium: Journal Article
Jazyk:angličtina
Vydáno: Elsevier Ltd 15.01.2025
Témata:
Attention
Direct-prediction
energy
Fork-sequence
Moving-window
prediction
time series analysis
weather forecasting
wind farms
wind power
wind speed
Fork-sequence
Moving-window
Direct-prediction
Attention
ISSN:0360-5442
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:In the context of large-scale wind power grid integration, accurate wind power forecasting is crucial for optimizing grid scheduling and ensuring safe grid connection. This study proposes a recurrent attention encoder–decoder network for multi-step interval wind power forecasting, combining Numerical Weather Prediction (NWP) inputs with deep learning techniques. The approach leverages a sequence-to-sequence neural network and temporal attention mechanism, enabling better capture of latent patterns in historical data that are useful for future predictions, directly generating multi-step time series and final prediction intervals. Additionally, a moving window training scheme, integrating bifurcated sequences and hidden layers, is employed to organize historical data and improve the stability and performance of the sequence. Using offshore wind farm data, the wind speed and direction components (U, V) are decomposed, and experiments show that the proposed method outperforms existing methods in metrics such as a minimum PINAW of 0.119 and an average reduction of 19.37% in CWC. These results demonstrate high accuracy and reliability in interval forecasting, providing strong support for wind farm scheduling and grid optimization. [Display omitted] •Propose AMQRNN: Integrate NWP data to enhance wind power forecast accuracy.•Use Seq2Seq and Attention for direct multi-step time series prediction.•Innovative training: Forked sequences and moving windows improve model performance.•Model understanding: Context fusion boosts accuracy of historical data predictions.
Bibliografie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 23
ISSN:0360-5442
DOI:10.1016/j.energy.2024.134317