Enhancing electrical conductivity and mechanical strength of gas diffusion layers through multi-objective optimization

<p dir="ltr">Proton exchange membrane fuel cells (PEMFCs) are key to decarbonizing industries, with composite gas diffusion layers (CGDLs) playing a crucial role in their performance. A longstanding challenge in this field is achieving a balance between mechanical strength (MS) and effective transport properties (e.g., electrical-thermal conductivity, permeability and diffusivity) while minimizing reliance on</p><p dir="ltr">resource-intensive experimental optimization. Current study, by tuning theoretical macroscopic stress–strain behaviour of a composite gas diffusion layer (CGDL) to our liking, specifically focuses on developing two constitutive relations of effective electrical conductivity (EEC) and MS models in CGDLs, and implements two multi-objective optimization frameworks that simultaneously enhance key parameters of CGDL microstructures, including porosity, thickness, diameter and orientation of fibers, saturation, and temperature. The most important key mechanism enabling simultaneous enhancement lies in the strategic balancing of fiber orientation and porosity; optimized orientation improves load distribution and mechanical resilience, while refined porosity and fiber alignment facilitate more efficient conductive pathways with minimal structural compromise. Utilizing the non-dominated sorting genetic algorithm-III (NSGAIII) and a hybrid deep learning-based (DL-based) surrogate model coupled with a grid search (GS) optimizer, we achieved significant improvements of 59.98%, 35%, and 40.18% in through-plane EEC, in-plane EEC, and MS, respectively, for optimized CGDL compared to primary case. The findings provide a foundation for designing more efficient CGDLs lead to durable PEMFCs.</p>