Wind power forecasting considering data privacy protection: A federated deep reinforcement learning approach

In a modern power system with an increasing proportion of renewable energy, wind power prediction is crucial to the arrangement of power grid dispatching plans due to the volatility of wind power. However, traditional centralized forecasting methods raise concerns regarding data privacy-preserving a...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Applied energy Ročník 329; s. 120291
Hlavní autoři: Li, Yang, Wang, Ruinong, Li, Yuanzheng, Zhang, Meng, Long, Chao
Médium: Journal Article
Jazyk:angličtina
Vydáno: Elsevier Ltd 01.01.2023
Témata:
algorithms
Data openness and sharing
Deep reinforcement learning
energy
Federated learning
issues and policy
prediction
Privacy protection
Uncertainty modeling
wind power
Wind power forecasting
Data openness and sharing
Privacy protection
Deep reinforcement learning
Federated learning
Wind power forecasting
Uncertainty modeling
ISSN:0306-2619, 1872-9118
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:In a modern power system with an increasing proportion of renewable energy, wind power prediction is crucial to the arrangement of power grid dispatching plans due to the volatility of wind power. However, traditional centralized forecasting methods raise concerns regarding data privacy-preserving and data islands problem. To handle the data privacy and openness, we propose a forecasting scheme that combines federated learning and deep reinforcement learning (DRL) for ultra-short-term wind power forecasting, called federated deep reinforcement learning (FedDRL). Firstly, this paper uses the deep deterministic policy gradient (DDPG) algorithm as the basic forecasting model to improve prediction accuracy. Secondly, we integrate the DDPG forecasting model into the framework of federated learning. The designed FedDRL can obtain an accurate prediction model in a decentralized way by sharing model parameters instead of sharing private data which can avoid sensitive privacy issues. The simulation results show that the proposed FedDRL outperforms the traditional prediction methods in terms of forecasting accuracy. More importantly, while ensuring the forecasting performance, FedDRL can effectively protect the data privacy and relieve the communication pressure compared with the traditional centralized forecasting method. In addition, a simulation with different federated learning parameters is conducted to confirm the robustness of the proposed scheme. •A DDPG-based prediction model is built for ultra-short-term wind power forecasting.•Propose a FedDRL forecasting scheme to handle the data privacy and openness.•Combine automatic machine learning with DRL for hyperparameters selection.•Simulation tests were carried out on real-world historical data to examine the proposal.
Bibliografie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 23
ISSN:0306-2619
1872-9118
DOI:10.1016/j.apenergy.2022.120291