The impact of grain size of the porous media of the 3D images of a beadpack on the permeability values and non-Darcy flow behaviour is analyzed using pore-scale flow modelling utilizing the finite volume method. In this study, the sample used was a 3D image from random tight packing of spheres (beadpack) with a size of 250×250×250 voxels and a porosity of 0.36. Variation in grain size is carried out by upscaling the 3D image sample, so that this treatment affects the specific surface area value but does not change the porosity and tortuosity of the sample. Several variations of grain diameter include 100 μm, 125 μm, 150 μm, 175 μm, and 200 μm. In this study, we adopted PIMPLE algorithm, which combines SIMPLE and PISO algorithms to accomplish the Navier-Stokes equations of fluid dynamics in such porous media. The simulation results show that the permeability value exhibits an inverse relationship with the specific surface area. This correlation is in accordance with the Kozeny-Carman equation. In addition, the initiation of non-Darcy flow occurs at the reference length-based Reynolds number of 2.06 and the permeability-based Reynolds number of 0.049.