基于深度强化学习的绿氢生产过程调度优化

刘京鑫; 赵亮

doi:10.11949/0438-1157.20260420

您当前的位置：

首页 >

文章列表页 >

基于深度强化学习的绿氢生产过程调度优化

智能过程工程 | 更新时间：2026-05-09

- 基于深度强化学习的绿氢生产过程调度优化
- Deep reinforcement learning for scheduling of green hydrogen production process
- 化工学报 2026年
- 作者机构：
  
  华东理工大学信息科学与工程学院，上海 200237
- 作者简介：
  
  刘京鑫（2000—），男，硕士研究生，1453597365@qq.com
  赵亮（1979—），男，博士，教授，lzhao@ecust.edu.cn
- 基金信息：
  
  国家自然科学基金面上项目(62373154;22178103)
- DOI：10.11949/0438-1157.20260420
  中图分类号： TP 273.1
- 收稿：2026-03-28，
  
  修回：2026-05-09，
  
  录用：2026-05-09，
- 稿件说明：
移动端阅览
刘京鑫, 赵亮. 基于深度强化学习的绿氢生产过程调度优化[J/OL]. 化工学报, 2026.

LIU Jingxin, ZHAO Liang. Deep reinforcement learning for scheduling of green hydrogen production process[J/OL]. CIESC Journal, 2026.
刘京鑫, 赵亮. 基于深度强化学习的绿氢生产过程调度优化[J/OL]. 化工学报, 2026. DOI： 10.11949/0438-1157.20260420.

LIU Jingxin, ZHAO Liang. Deep reinforcement learning for scheduling of green hydrogen production process[J/OL]. CIESC Journal, 2026. DOI： 10.11949/0438-1157.20260420.

摘要

针对风光发电波动导致电解槽负载频繁变化、弃电与购电并存等问题，本文建立了光伏—风电—储能—电解槽耦合制氢过程模型，并将在线调度表述为马尔可夫决策过程；在多目标奖励中考虑产氢收益、购电成本、弃电抑制、负载平滑及储能荷电状态（SOC）安全约束，采用软演员-评论家（SAC）算法获得最优调度策略。基于NASA气象数据与电价序列的算例表明，所提方法在多场景下比专家规则与模型预测控制取得更高产氢量与可再生能源利用率，并显著降低弃电，且单步推理开销可满足5 min调度周期的实时应用需求。

Abstract

To address the issues of frequent load fluctuations of electrolyzers caused by volatile wind and solar power

as well as the coexistence of renewable energy curtailment and electricity purchase costs

this paper establishes a coupled photovoltaic-wind-energy storage-electrolyzer hydrogen production system model

and formulates the online scheduling problem as a Markov decision process. The multi-objective reward function comprehensively considers hydrogen production benefits

electricity purchase costs

renewable energy curtailment mitigation

load smoothing

and state-of-charge (SOC) security constraints

and the Soft Actor-Critic (SAC) algorithm is adopted to obtain the optimal scheduling strategy. Case studies based on NASA meteorological data and electricity price sequences demonstrate that the proposed method achieves higher hydrogen production and renewable energy utilization efficiency than expert rules and model predictive control under various scenarios

significantly reduces renewable energy curtailment

and the single-step inference cost can meet the real-time application requirements of a 5-minute scheduling cycle.

关键词

Keywords

references

国家发展和改革委员会，国家能源局 . 氢能产业发展中长期规划（ 2021 - 2035 年）[EB/OL ] . 2022 . https://zfxxgk.nea.gov.cn/2022-03/23/c_1310525630.htm https://zfxxgk.nea.gov.cn/2022-03/23/c_1310525630.htm .

IEA . Global Hydrogen Review 2024 [R ] . Paris : International Energy Agency , 2024 . https://www.iea.org/reports/global-hydrogen-review-2024.

Hydrogen Council . Hydrogen Scaling Up: A Sustainable Pathway for the Global Energy Transition [R ] . Brussels : Hydrogen Council , 2017 . https://hydrogencouncil.com/en/newsroom/.

Aminaho E N , Aminaho N S , Aminaho F . Techno-economic assessments of electrolyzers for hydrogen production [J ] . Applied Energy , 2025 , 399 : 126515 .

Sayed-Ahmed H , Toldy Á I , Santasalo-Aarnio A . Dynamic operation of proton exchange membrane electrolyzers: Critical review [J ] . Renewable and Sustainable Energy Reviews , 2024 , 189 : 113883 .

Nguyen E , Olivier P , Péra M C , et al . Impact of short-term intermittent operation on experimental industrial PEM and alkaline electrolyzers [J ] . International Journal of Hydrogen Energy , 2025 , 126 : 516 - 530 .

Minnah P , Pope K . Degradation modelling and the impact of intermittent operation on proton exchange membrane electrolyzers [J ] . International Journal of Hydrogen Energy , 2026 , 202 : 152962 .

Johnsen A G , Mitridati L , Zarrilli D , et al . The value of ancillary services for electrolyzers [J ] . Computers & Chemical Engineering , 2026 , 204 : 109360 .

Zeng A D , Wang Z R , Wang J W , et al . Optimal scheduling of electricity–hydrogen–heat integrated energy system considering electrolyzer dynamic temperature and pipeline network heat storage [J ] . International Transactions on Electrical Energy Systems , 2025 , 2025 : 9478361 .

Zhang L Q , Dai W Z , Zhao B , et al . Multi-time-scale economic scheduling method for electro-hydrogen integrated energy system based on day-ahead long-time-scale and intra-day MPC hierarchical rolling optimization [J ] . Frontiers in Energy Research , 2023 , 11 : 1132005 .

Amores E , Rodríguez J , Carreras C . Influence of operation parameters in the modeling of alkaline water electrolyzers for hydrogen production [J ] . International Journal of Hydrogen Energy , 2014 , 39 ( 25 ): 13063 - 13078 .

Masampally V S , Mitra K , Runkana V . Modeling and simulation of hydrogen production in a PEM water electrolyzer at high current densities [J ] . Energy Conversion and Management , 2026 , 352 : 121110 .

Zhang Y T , Zhang T C , Liu S W , et al . Multi-energy complementary optimal scheduling based on hydrogen gas turbine considering the flexibility of electrolyser [J ] . IET Renewable Power Generation , 2024 , 18 ( 12 ): 1972 - 1985 .

Cardona P , Costa-Castelló R , Roda V , et al . Model predictive control of an on-site green hydrogen production and refuelling station [J ] . International Journal of Hydrogen Energy , 2023 , 48 ( 47 ): 17995 - 18010 .

Ulleberg Ø . Modeling of advanced alkaline electrolyzers: a system simulation approach [J ] . International Journal of Hydrogen Energy , 2003 , 28 ( 1 ): 21 - 33 .

Abdin Z , Webb C J , Gray E M . Modelling and simulation of an alkaline electrolyser cell [J ] . Energy , 2017 , 138 : 316 - 331 .

Ashraf H , Kawai E , Mae M , et al . Advancing electrochemical modelling of PEM electrolyzers through robust parameter estimation with the weighted mean of vectors algorithm [J ] . Multiscale and Multidisciplinary Modeling, Experiments and Design , 2025 , 8 ( 9 ): 397 .

Mößle P , Herrmannsdörfer T , Welzl M , et al . Economic optimization for the dynamic operation of a grid connected and battery-supported electrolyzer [J ] . International Journal of Hydrogen Energy , 2025 , 100 : 749 - 759 .

Zheng Y , You S , Bindner H W , et al . Optimal day-ahead dispatch of an alkaline electrolyser system concerning thermal–electric properties and state-transitional dynamics [J ] . Applied Energy , 2022 , 307 : 118091 .

Voulkopoulos X , Dimitriadis C N , Georgiadis M C . Optimal scheduling of a RES–Electrolyzer aggregator in electricity, hydrogen and green certificates markets [J ] . International Journal of Hydrogen Energy , 2024 , 78 : 1099 - 1107 .

Wang J B , Wen J F , Wang J R , et al . Coordinated scheduling of wind-solar-hydrogen-battery storage system for techno-economic-environmental optimization of hydrogen production [J ] . Energy Conversion and Management , 2024 , 314 : 118695 .

Haarnoja T , Zhou A , Abbeel P , et al . Soft actor-critic: Off-policy maximum entropy deep reinforcement learning with a stochastic actor [EB/OL ] . 2018 : arXiv : 1801 . 01290 . https://arxiv.org/abs/1801.01290 https://arxiv.org/abs/1801.01290 .

Haarnoja T , Zhou A , Hartikainen K , et al . Soft actor-critic algorithms and applications [EB/OL ] . 2018 : arXiv : 1812 . 05905 . https://arxiv.org/abs/1812.05905 https://arxiv.org/abs/1812.05905 .

Bäumler A , Benterki A , Meng J W , et al . Energy management strategies based on soft actor critic reinforcement learning with a proper reward function design based on battery state of charge constraints [J ] . Journal of Energy Storage , 2024 , 90 : 111797 .

Zhao Y J , Lin F , Yang Z P . Soft actor-critic-based energy management strategy for hybrid energy storage system in urban rail transit DC traction power supply system [J ] . Electric Power Systems Research , 2026 , 252 : 112372 .

Lu W C , Gao Y , Sun Z , et al . An Improved soft actor–critic framework for cooperative energy management in the building cluster [J ] . Applied Sciences , 2025 , 15 ( 16 ): 8966 .

Van Quang T , Doan D T . Online transfer learning (OTL) for accelerating deep reinforcement learning (DRL) for building energy management [J ] . Journal of Building Performance Simulation , 2025 : 1 - 20 .

RMI . 2024 China Power Market Outlook: 10 Key Trends for Market Players [R ] . RMI , 2024 . https://rmi.org/wp-content/uploads/dlm_uploads/2025/01/final-2024-China-Power-Market-Outlook-10-Key-Trends-for-Market-Players-2501117.pdf.

Li D F , Li Z H , Li X J , et al . Optimization method of time-of-use electricity price for the cost savings of power grid investment [J ] . Frontiers in Energy Research , 2024 , 12 : 1361116 .

Dong L , Huang Y , Xu X , et al . Research on priority scheduling strategy for smoothing power fluctuations of microgrid Tie-lines based on PER-DDPG algorithm [J ] . IET Generation , Transmission & Distribution, 2024 , 18 ( 20 ): 3221 - 3233 .

Tang D G , Ge P X , Yuan C Q , et al . Optimal management of coupled hydrogen-electricity energy systems at ports by multi-time scale scheduling [J ] . Applied Energy , 2025 , 391 : 125885 .

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

基于ANN-GA集成的ORC混合工质智能筛选与性能优化

基于深度强化学习的某异构生产线工厂生产智能排产

配备电锅炉和蓄热装置的热电联产系统性能分析

考虑不均匀性的SOEC电堆3D建模优化分析

多目标优化驱动的PEM电解槽性能研究