Deep Reinforcement Learning Based Optimal Operation of Low-Carbon Island Microgrid with High Renewables and Hybrid Hydrogen–Energy Storage System
Journal of Marine Science and Engineering · 25.01.2025 · 13 (2) · 225
DOI: 10.3390/jmse13020225 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Marine Science and Engineering, online erschienen am 25.01.2025. Verfasst von Wangwang Zhu und 5 weiteren Autor:innen. Die Arbeit wurde bisher 23-mal zitiert (Crossref) und nennt 42 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Natural Science Foundation of China, Postgraduate Research & Practice Innovation Program of Jiangsu Province, Shanghai Marine Equipment Research Institute.
Abstract (Originalsprache)
Hybrid hydrogen–energy storage systems play a significant role in the operation of islands microgrid with high renewable energy penetration: maintaining balance between the power supply and load demand. However, improper operation leads to undesirable costs and increases risks to voltage stability. Here, multi-time-scale scheduling is developed to reduce power costs and improve the operation performance of an island microgrid by integrating deep reinforcement learning with discrete wavelet transform to decompose and mitigate power fluctuations. Specifically, in the day-ahead stage, hydrogen production and the hydrogen blending ratio in gas turbines are optimized to minimize operational costs while satisfying the load demands of the island. In the first intraday stage, rolling adjustments are implemented to smooth renewable energy fluctuations and increase system stability by adjusting lithium battery and hydrogen production equipment operations. In the second intraday stage, real-time adjustments are applied to refine the first-stage plan and to compensate for real-time power imbalances. To verify the proposed multi-stage scheduling framework, real-world island data from Shanghai, China, are utilized in the case studies. The numerical simulation results demonstrate that the proposed innovative optimal operation strategy can simultaneously reduce both the costs and emissions of island microgrids.
Förderung
- National Natural Science Foundation of China
- Postgraduate Research & Practice Innovation Program of Jiangsu Province
- Shanghai Marine Equipment Research Institute
Zitieren
APA
Zhu, W., Wen, S., Zhao, Q., Zhang, B., Huang, Y., & Zhu, M. (2025). Deep Reinforcement Learning Based Optimal Operation of Low-Carbon Island Microgrid with High Renewables and Hybrid Hydrogen–Energy Storage System. Journal of Marine Science and Engineering, 13(2), 225. https://doi.org/10.3390/jmse13020225
Vancouver
Zhu W, Wen S, Zhao Q, Zhang B, Huang Y, Zhu M. Deep Reinforcement Learning Based Optimal Operation of Low-Carbon Island Microgrid with High Renewables and Hybrid Hydrogen–Energy Storage System. Journal of Marine Science and Engineering. 2025;13(2):225. doi:10.3390/jmse13020225
BibTeX
@article{Zhu2025,
author = {Zhu, Wangwang and Wen, Shuli and Zhao, Qiang and Zhang, Bing and Huang, Yuqing and Zhu, Miao},
title = {Deep Reinforcement Learning Based Optimal Operation of Low-Carbon Island Microgrid with High Renewables and Hybrid Hydrogen–Energy Storage System},
journal = {Journal of Marine Science and Engineering},
year = {2025},
volume = {13},
number = {2},
pages = {225},
doi = {10.3390/jmse13020225}
}
EUVIMED