Digital Rock Physics: Mechanical properties from nano-tomography to whole core scale analysis

Abstract

Digital Rock Physics (DRP) is an emerging technology that has potential to generate fast and cost effective special core analysis (SCAL) properties compared to conventional experimental techniques and modeling techniques. The primary workflow of DRP consists of three elements: 1) image the rock sample using high resolution 3D scanning techniques (e.g. X-ray CT, FIB/SEM), 2) process and digitize the images by segmenting the pore and matrix phases 3) simulate the desired physical properties of the rocks such as elastic moduli and velocities of wave propagation. While DRP based mechanical property predictions have been successful for homogeneous sandstones, for carbonates the success has been limited due to their complex heterogeneity at multiple length scales from tens of nanometers to several centimeters. With the advent of high resolution 3D imaging capability it is now possible to characterize actual rock microstructure and heterogeneity in 3D at the micron scale. We have used a finite element method (FEM) that works directly on the digital images by treating each voxel as an element eliminating the meshing step. DRP studies that used single-scale image data for simulations noticed a constant overestimation of numerical elastic modulus compared to experimental results regardless of the computational approach used. We worked towards complementing the previous studies in this field and increasing the accuracy of numerical elastic property predictions. We addressed several key components that serve as sources of uncertainties and integrated them to develop a DRP workflow through following steps: 1) establishing the importance of rigorous representative volume element (RVE) determination in DRP, 2) ensuring comparison of experimental vs. numerical data at the same length scale and for the same carbonate samples, 3) resolving complex carbonate pore structures down to 1 µm (using X-ray CT) and around 50 nm (using FIB/SEM) through multi-scale imaging protocol, 4) estimating more accurate input moduli for simulations through local modulus characterization using nanoindentation technique and 5) integrating nanoindentation moduli plus five-scale image data to upscale effective elastic properties in two real reservoir carbonate rocks. Integration of nanoindentation based input alone caused a minimum of 15% decrease in simulated bulk and shear modulus overestimation.

Date of Award	May 2017
Original language	American English
Supervisor	Mohamed Sassi (Supervisor)