Convex hull approach for determining rock representative elementary volume for multiple petrophysical parameters using pore-scale imaging and Lattice–Boltzmann modelling. (June 2017)

• Record Type: Journal Article
• Authors: Shah, S.M.; Crawshaw, J.P.; Gray, F.; Yang, J.; Boek, E.S.
• Is Part Of: Advances in water resources. Volume 104(2017)
• Page Start: 65
• Page End: 75
• Publication Date: 2017-06
• Subjects: Representative element volume -- Porosity -- Single phase permeability -- Pore-scale -- Convex hull; Hydrology -- Periodicals; Hydrodynamics -- Periodicals; Hydraulic engineering -- Periodicals; 551.48
• DOI: 10.1016/j.advwatres.2017.03.008
• Languages: English
• ISSNs: 0309-1708

Abstract

Highlights: Qualitative and quantitative method to capture heterogeneity at pore-scale. Multiple petrophysical parameters to determine representative element volume. Enhancing computational efficiency to calculate petrophysical properties. Abstract: In the last decade, the study of fluid flow in porous media has developed considerably due to the combination of X-ray Micro Computed Tomography (micro-CT) and advances in computational methods for solving complex fluid flow equations directly or indirectly on reconstructed three-dimensional pore space images. In this study, we calculate porosity and single phase permeability using micro-CT imaging and Lattice Boltzmann (LB) simulations for 8 different porous media: beadpacks (with bead sizes 50 µm and 350 µm), sandpacks (LV60 and HST95), sandstones (Berea, Clashach and Doddington) and a carbonate (Ketton). Combining the observed porosity and calculated single phase permeability, we shed new light on the existence and size of the Representative Element of Volume (REV) capturing the different scales of heterogeneity from the pore-scale imaging. Our study applies the concept of the 'Convex Hull' to calculate the REV by considering the two main macroscopic petrophysical parameters, porosity and single phase permeability, simultaneously. The shape of the hull can be used to identify strong correlation between the parameters or greatly differing convergence rates. To further enhance computational efficiency we note that the area of the convex hull (for well-chosen parameters such as the log of the permeability and the porosity) decays exponentially with sub-sample size so that only a few small simulations are needed to determine the system size needed to calculate the parameters to high accuracy (small convex hull area). Finally we propose using a characteristic length such as the pore size to choose an efficient absolute voxel size for the numerical rock.