Effect of grayscale threshold on X-ray computed tomography reconstruction of gas diffusion layers in polymer electrolyte membrane fuel cells

In X-ray computed tomography (CT) reconstructions of gas diffusion layers (GDLs), grayscale threshold selection is a critical issue. Although various selection methods exist, they all have their own drawbacks. This study investigates the influence of grayscale threshold on GDL properties and compares Otsu and porosity-adaptive thresholds. We utilized X-ray CT to reconstruct a Toray carbon paper sample (TGP-H-060) at a resolution of 2 μm. Using reconstructed 3D models generated under different grayscale thresholds, we performed structural analysis, computational fluid dynamics simulation, and compression simulation. We subsequently calculated porosity, tortuosity, permeability, and macroscopic stress-strain relationships, quantitatively analyzing the sensitivity of these parameters to the change of grayscale threshold. The results indicated that small change in the grayscale threshold can significantly impact the transport and mechanical properties of reconstructed GDLs. The difference between Otsu and porosity-adaptive thresholds is notable, and the porosity-adaptive threshold appears to be less accurate than the Otsu threshold.