Efficiency improvement of PEM Fuel Cells using Reinforcement Learning

This research focuses on designing and implementing, machine learning (ML)based control strategies for proton exchange membrane (PEM) fuel cells (FC) stacks will be investigated for the efficient production of electricity while considering lifespan constraints. ML will be used in three areas: ML based system modeling, Reinforcement Learning (RL) and ML-based model predictive control (MLMPC), with a primary focus on controller development for real-time implementation. This project specifically investigates using RL to improve the efficiency of a Horzion H-500 XP fuel cell. Previous studies by the Mechanical Engineering Energy Control Lab (MEECL) using RL on Hydrogen Diesel Dual Fuel combustion engines have been quite successful. The initial research goal is to implement RL control on the fuel cell with similar approachs to these studies.