Machine learning-driven optimization of gas diffusion layer microstructure for PEM fuel cells

Abstract

The Gas Diffusion Layer (GDL) is a vital component within Proton Exchange Membrane Fuel Cells (PEMFCs), playing a crucial role in mass and heat transport. Enhancing GDL microstructures directly improves transport properties, thereby leading to more efficient and durable PEMFCs. In this study, we developed a novel machine learning methodology to optimize the GDL microstructure and properties. The developed optimization framework demonstrated high efficacy, with an R² score ∼95 % in 6 out of 7 properties and a R² score ∼90 % for the GDL-Micro-Porous Layer (MPL) contact resistance. We validated our machine learning approach by comparing the predicted GDL properties to those calculated through digital characterization using physics-based methods from the stochastically generated GDL, using the optimal manufacturing parameters identified by the optimizer. Our machine learning model predicted accurately 7 GDL properties decreasing the computational cost from ∼3 to 4 h wall time (physical model) to ∼3 s wall time. Results show that low fiber concentration accompanied by low compression ratio achieve maximum diffusivity and minimum GDL-MPL contact resistance. Furthermore, prioritizing maximum electrical and/or thermal conductivities while minimizing GDL-MPL contact resistance require high fiber concentration with high compression ratio. This optimization strategy shows significant potential for improving gas transport, water management, efficient current collection, and thermal regulation within PEMFCs.