Hybrid Reinforcement Learning Algorithm Improves Proton Exchange Membrane Electrolyzer Performance

Strategic Insights -

A study published in the International Journal of Hydrogen Energy has developed a novel hybrid reinforcement learning (RL) framework to optimize the operation of proton exchange membrane electrolyzers (PEMELs). The research aimed to identify the exact parameters of PEMELs, essential for their large-scale applications. The hybrid RL method was shown to be more accurate and robust than existing algorithms, reducing the sum of absolute voltage errors by 0.53 V, 0.65 V, and 0.41 V compared to particle swarm optimization (PSO), honey badger algorithm (HBA), and feedforward neural network (FNN), respectively.

Key Takeaways:

  • A novel hybrid reinforcement learning (RL) framework is proposed for parameter identification of proton exchange membrane electrolyzers (PEMELs).
  • The hybrid RL method uses a deep reinforcement learning agent to eliminate prediction error and has excellent fitting performance.
  • The accuracy of the hybrid RL method is validated through comparisons with honey badger algorithm (HBA), particle swarm optimization (PSO), and feedforward neural network (FNN).
  • The robustness of the hybrid RL method is also validated under different operation conditions.
  • The hybrid RL parameter identification method has high accuracy with robust performance, reducing the sum of absolute voltage errors by 0.53 V, 0.65 V, and 0.41 V compared to PSO, HBA, and FNN, respectively.
  • The research was conducted by Seunghun Jung, Hangyu Cheng, Jiahui Chen, Shengnan Liu, and Young-Bae Kim from Chonnam National University, with financial support from the National Research Foundation of Korea.
  • The study has been peer-reviewed and published in the International Journal of Hydrogen Energy.

Statistics:

  • The hybrid RL method reduces the sum of absolute voltage errors by 0.53 V, 0.65 V, and 0.41 V compared to PSO, HBA, and FNN, respectively.
  • The research was conducted by a team of 5 authors from Chonnam National University.
  • The study received financial support from the National Research Foundation of Korea.
  • The research has been peer-reviewed and published in the International Journal of Hydrogen Energy.

Sources:

  • NewsRx. Findings on Mathematics Reported by Investigators at Chonnam National University (Parameter Identification With Hybrid Reinforcement Learning Algorithm of Proton Exchange Membrane Electrolyzer). Journal of Engineering. October 13, 2025; p 1053.
  • International Journal of Hydrogen Energy, 2025;174.