Refined Modeling of Effective Agglomerate Surface Area for Enhanced PEMFC Degradation Analysis

Understanding the degradation phenomena of proton exchange membrane fuel cells (PEMFC) is essential for their durability. Numerical simulation offers a powerful tool for analyzing and refining PEMFC design and operation. However, the computational demands of such simulations, particularly in assessing catalyst layer (CL) degradation of PEMFC under dynamic operating conditions, pose significant challenges. This study addresses these challenges by developing a multiobjective optimization algorithm (MOA) coupled with a model function capable of assessing cathode CL degradation to simulate the PEMFC dynamic behaviors based on durability test data. In the initial modeling phase, cathode CL properties were identified using a genetic algorithm to establish an extended model function describing the decay of effective agglomerate surface area profile over time. This function was subsequently incorporated into the MOA to validate its effectiveness in simulating PEMFC degradation performances via an electrochemical model. To further optimize computational efficiency, Gaussian process regression was employed to surrogate the electrochemical model and integrated with the MOA based on the limited size of real-world data. The integrated model produced both the effective agglomerate surface area profile and voltage signals for sustained PEMFC operation, achieving root-mean-square errors of less than 0.009 V with a simulation time of approximately 1 min. These findings underscore the central role of effective agglomerate surface area profile in intelligently interpreting the complex decay patterns of CL during PEMFC operation. This approach represents a substantial advancement in monitoring and forecasting of PEMFC degradation through the implementation of extended model functions in numerical simulations.