Reliability-based design optimization methodology for enhancing performance and efficiency in catalyst manufacturing for polymer electrolyte membrane fuel cells

Polymer Electrolyte Membrane Fuel Cells (PEMFCs) have emerged as an alternative to internal combustion engines, offering higher efficiency and reduced environmental impact. High component costs, particularly the cathode catalyst layer (CL), pose a significant barrier to widespread commercialization. This study addresses this challenge by optimizing the material and design of the cathode CL. Traditional deterministic design optimization (DDO) approaches have been explored but often fail to account for inherent uncertainties introduced during the manufacturing process. To overcome this limitation, we propose a data-driven reliability-based design optimization (RBDO) approach to optimize key CL design parameters, including the weight ratio of ionomer to carbon (wt_I/C), the weight ratio of Pt to carbon (wt_Pt/C), the porosity of cathode CL (ε_cCL), and platinum loading (L_Pt), to maximize cell voltage (V_cell), considering CL manufacturing costs (Cost_CL) and power density of membrane electrode assembly (Ṗ_MEA) as performance constraints. Compared to a typical PEMFC stack with L_Pt aligned to Department of Energy targets of 0.125 mg/cm², results of the DDO show that V_cell is improved by 12 mV, with a reliability of 49 % for Cost_CL and 99.97 % for Ṗ_MEA, respectively. In contrast, the RBDO approach provides a reliability of 95 % for Cost_CL and 97.95 % for Ṗ_MEA, at the expense of a 32 mV drop in V_cell.