Insight into safety and economic trade-offs for high-pressure PEMEC operation using Bayesian optimization

As a pivotal technology for clean hydrogen production, proton exchange membrane electrolysis cells (PEMEC) have garnered widespread industrial adoption. The implementation of high-pressure operation in PEMEC systems enhances hydrogen output pressure, thereby reducing compression demands and associated costs in downstream transportation. However, this strategy potentially introduces critical safety challenges associated with gas crossover, particularly hydrogen-in‑oxygen (HTO) enrichment. To systematically address the inherent trade-offs between economic performance and safety risks, this study employs a novel proposed multi-objective optimization framework based on Bayesian optimization. Our methodology is designed to achieve explicit system-level objectives, maintaining the HTO concentration constraint below the safety threshold of 2%, while concurrently minimizing the levelized cost of hydrogen (LCOH) and the inherent safety index (ISI). First, we established an integrated process model by combining first-principles unit models with process simulation to assess system performance and safety. Second, we implemented Bayesian optimization to navigate the multi-objective trade-offs. This method is well-suited for this problem due to its ability to efficiently handle high-dimensional, non-linear constrained optimization. Within this framework, the multi-objective optimization process simultaneously quantifies gas crossover dynamics for each set of evaluated parameters. The optimization results identify an optimal operational window, an operating pressure of 25–50 bar, a current density of 1.5–2.0 A/cm2, and a membrane thickness below 100 μm. This window provides a scientifically rigorous foundation for balancing economic and safety objectives in industrial-scale PEMEC applications.