3D Image Processing Based Reservoir Rock Core Analysis

  • Haiyang Zhang

    Student thesis: Master's Thesis

    Abstract

    Core analysis is one of the cornerstone techniques to acquire petrophysical data for reservoir development and evaluation. Identifying the pore structure and understanding flow and transport behaviors in sub-surface geological rocks is nontrivial task. Core analysis in laboratory are costly and time-consuming. The alternative nowadays with the advancement of imaging techniques and computation capability, the properties of rock matrix can be determined by numerical modeling. Pore network modeling is one of the numerical modeling methods. This consists of high-resolution digital images of pores and grains structures and evaluation of the rock properties using numerical simulation at the pore-scale. The objective of this research is to conduct numerical core analysis. This includes two-phase flow simulation through pore networks extracted from the 3D micro-CT images of three different carbonate samples from Abu Dhabi reservoirs. Pore network properties include the pore size distribution, the topological structure, the pore-space development uniformity, and the pore-throat shape factor. The results of the network method are the connected porosity, total porosity, and absolute permeability. These were compared and found similar to the values measured in the laboratory. Two-phase flow simulations were on the extracted pore networks. These simulations lead to an estimation of the capillary pressure in the primary drainage process. The obtained estimated capillary pressure agreed well with the experimental mercury injection capillary pressure (MICP). The waterflooding oil-water relative permeability was then estimated. The effect of wettability on the relative permeability in the uniform-wet system and the mixed-wet system was also investigated. In addition, a regression-based renormalization method was proposed to upscale the absolute permeability. The performance of this method was validated on different rock samples. The results showed this method could significantly increase the upscaling accuracy than the conventional renormalization method. This method also worked well for both homogenous and heterogeneous rock samples. Another upscaling method using image registration was also presented and tested on the Silurian Dolomite rock sample. A good agreement with the experimental data demonstrated the potential of the application of multi-scale images on permeability upscaling.
    Date of AwardJul 2021
    Original languageAmerican English

    Keywords

    • Digital Rock Physics
    • Pore Network Modeling
    • Core Analysis
    • Relative Permeability Prediction
    • Absolute Permeability Upscaling.

    Cite this

    '