Short-Term Wind Power Forecasting Based on Spatiotemporal Correlations
- https://doi.org/10.2991/meees-18.2018.3How to use a DOI?
- Short-Term wind power forecasting, Deep Neural Network, Convolutional Neural Network, spatiotemporal correlations.
Wind power forecasting is of great significance for promoting the new energy resources accommodation and the economic efficiency, security and stability of power grid operation. The traditional wind power forecasting model has monotonous elements for input and demonstrates a lack of spatiotemporal correlation between elements. Therefore, a deep neural network model considering space-time correlation is proposed. Firstly, a multi-dimensional space-time data input modeling method is proposed based on meshed numerical weather forecasting. Then, a variety of deep neural network models for wind power prediction are established. Multi-layer CNN (Convolutional Neural Networks) are utilized for feature extraction, meanwhile LSTM (Long Short-Term Memory) networks are used for pattern memory. Finally, the prediction of single station and regional wind power is carried out respectively, which proves the effectiveness and feasibility of the method. The results show that, compared with the traditional single-layer neural network, the deep neural network in this paper can effectively mine the spatiotemporal correlation between data and improve the prediction accuracy of single-station wind power forecasting.
- © 2018, the Authors. Published by Atlantis Press.
- Open Access
- This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).
Cite this article
TY - CONF AU - Jinli Dou AU - Chun Liu AU - Bo Wang PY - 2018/05 DA - 2018/05 TI - Short-Term Wind Power Forecasting Based on Spatiotemporal Correlations BT - Proceedings of the 2018 International Conference on Mechanical, Electrical, Electronic Engineering & Science (MEEES 2018) PB - Atlantis Press SP - 12 EP - 15 SN - 2352-5401 UR - https://doi.org/10.2991/meees-18.2018.3 DO - https://doi.org/10.2991/meees-18.2018.3 ID - Dou2018/05 ER -