From invasion percolation to flow in rock fracture networks

TL;DR

This paper develops an advanced invasion percolation model to simulate immiscible two-phase flow in complex 3D fracture networks, incorporating fracture inclinations, intersections, and trapping effects, validated against real geological data.

Contribution

It introduces modifications to invasion percolation to better model flow in anisotropic 3D fracture networks, including trapping and geometric considerations, with validation on real data.

Findings

Enhanced invasion percolation model captures fracture inclination effects.

Flow and saturation behaviors depend on network anisotropy and size.

Model results align with observed geological fracture network data.

Abstract

The main purpose of this work is to simulate two-phase flow in the form of immiscible displacement through anisotropic, three-dimensional (3D) discrete fracture networks (DFN). The considered DFNs are artificially generated, based on a general distribution function or are conditioned on measured data from deep geological investigations. We introduce several modifications to the invasion percolation (MIP) to incorporate fracture inclinations, intersection lines, as well as the hydraulic path length inside the fractures. Additionally a trapping algorithm is implemented that forbids any advance of the invading fluid into a region, where the defending fluid is completely encircled by the invader and has no escape route. We study invasion, saturation, and flow through artificial fracture networks, with varying anisotropy and size and finally compare our findings to well studied, conditioned fracture networks.