Process monitoring and safety‐informed control of a proton exchange membrane water electrolysis system

Abstract

This study provides an experimental validation of a multiple‐input multiple‐output (MIMO) model predictive control (MPC) strategy, coupled with dynamic risk modeling, to address two critical aspects of proton exchange membrane water electrolysis (PEMWE) operation: (i) process safety, by mitigating temperature imbalances, and (ii) system performance, through precise hydrogen production control. A cyber‐physical platform was developed for real‐time monitoring, state‐space modeling and validation, risk metrics analysis, control implementation, and visualization. Open‐loop experiments revealed limitations in managing thermal gradients, underscoring the need for feedback operating strategies. The proposed closed‐loop MPC approach achieved precise tracking of hydrogen production while maintaining safety by ensuring temperature stability. Moreover, the dynamic risk metrics show how thermal risk evolves with temperature and offer guidance for decision‐making. These findings demonstrate the effectiveness of MIMO MPC in enhancing the operational safety and efficiency of PEMWE systems, providing a foundation for scalable and sustainable hydrogen production.