Simulating Pore-Scale Chemical Transport

Overview

Image-based modelling can be used to analyse mass transfer phenomena through porous media, with particular applications to reservoir rock pore-throats networks. The aim of these analyses is to improve our understanding and characterization of the way fluids move through variable pore-scales.

Using real geometries from micro-CT, this project involved generating a 3D model in Simpleware software for visualisation and processing prior to meshing and export to COMSOL Multiphysics® for studying chemical transport mechanisms.

Characteristics:

  • Micro-CT data of real rock geometries obtained from open source library
  • Image processing and segmentation in Simpleware ScanIP
  • Robust multiphase meshing of pore structure in Simpleware +FE
  • Pore-scale chemical transport simulation in COMSOL Multiphysics®

Thanks to

Memorial University of Newfoundland: M. Mahmoodi

Image Processing

RAW image files of pore space and microstructure were obtained using micro-CT data from an open-source rock CT-image library provided by The Imperial College Consortium on Pore-Scale Modelling (PERM). Simpleware ScanIP was used to threshold the data to obtain rock and pore phases, and to crop the 3D model to a suitable dimension for meshing. The ScanIP flood fill algorithm was also used to ensure high-quality segmented images for meshing.

Meshing

A very robust CFD mesh of the multiphase model was generated using Simpleware module +FE and exported directly to COMSOL for solving a Navier-Stokes equation and to calculate basic parameters such as absolute permeability. This workflow allowed for modelling of the transportation of chemical components in the real pore-level geometry.

Conclusions

In this project a high-quality 3D model and robust mesh was generated in Simpleware software in order to simulate pore-scale chemical transport in porous media. The straightforward workflow sets up the potential for new rock physics studies into dispersion and diffusion, as well as multi-phase flow, depending on available computational resources.