Digital Rock Analysis: A Morphological Approach from Micro- to Meso-Scale in Petrophysics

Constructing 3D digital rock models is essential for accurately estimating petrophysical properties using porescale modelling. These models represent rock microstructures and capture complex features such as porosity, pore geometry, connectivity, and grain size distribution that are vital for computing petrophysical rock properties such as permeability. Reconstructing 3D models from two-dimensional images or statistical methods joins the micro and core scales, enabling applications in contexts lacking physical samples.

In this paper, we adopt the morphological approach (MA) to construct 3D digital rock models using 2D SEM and micro-CT images in tarmat-bearing formation. This methodology provides accurate, physics-based insights into porosity, permeability, and their relationship, enabling more reliable predictions.

Machine learning (ML) techniques, such as SliceGAN, have been conducted to reconstruct realistic 3D models from 3D images, hence tackling difficulties in regions like “tarmat”, where direct 3D imaging poses difficulty. The integration of machine learning into the 3D reconstruction process with MA offers a practical approach for checking the reliability of real rock structures in the 3D reconstruction process by using MA, hence enabling advancements in numerical modelling at the mesa and pore scale.