Geomechanically coupled modelling of fluid flow partitioning in fractured porous media.

Abstract

Naturally fractured reservoirs are characterised with complex hydro-mechanical dynamics. In these reservoirs, hydrocarbons can be stored and produced from the rock matrix, the fracture network, or both. Normally the fracture network is depleted much faster than the matrix blocks due to its increased hydraulic conductivity; consequently, the recovery factor is low for these reservoirs. Additionally, the in-situ stress profile changes with reservoir depletion and affects fluid flow dynamics of the fractured reservoir. Therefore, dynamic characterisation of fractured reservoirs is considered a challenging task, responsible for inefficient exploitation of their reserves. This dissertation focuses on characterising matrix-fracture fluid flow partitioning subjected to variable overburden stress loading. Understanding of the matrix-fracture hydro-mechanical interaction would assist in developing optimum production plans to maximize recovery from fractured reservoirs. Initially, three different fracture implementation techniques - (1) simulating fracture as an equivalent porous medium; (2) implementing it as a sub-dimensional feature within the porous matrix; and (3) considering fracture domain as an open channel - were evaluated using a set of published laboratory core flooding data. The best fracture simulation approach was identified to be fracture implementation as an open channel interacting with matrix block. This approach takes into consideration the coupling of Darcy flow equation in the matrix domain to Navier-Stokes flow formulation in the fracture. The efficiency of this fracture simulation approach was significantly enhanced when coupled further with poro-elasticity physics and stress dependent permeability. In the next step, the coupled open channel fracture simulation approach was applied to perform a sensitivity analysis on the effect of all parameters of the governing equations on fracture and matrix flow. The results of this analysis were statistically analysed, with specific attention to the analytical formulation of the governing equations, to develop coupled empirical flow models for fracture and matrix. These empirical models incorporate both flow physics of matrix and fracture, as well as mechanical loading impacts. An analysed multiphase flow scenario demonstrated the compatibility of the coupled simulation approach with multiphase flow investigations in fractured porous media. A novel core flooding set-up, capable of separated fracture and matrix flow measurement, was designed and built to enable laboratory evaluation of the developed empirical models. This set-up enabled monitoring of pressure front within matrix and fracture, taking the advantages of several differential pressure transducers along the core plug length. Variation of the matrix and fracture flow in response to different stress loading scenarios was investigated in the laboratory. Furthermore, laboratory validation indicated that the matrix flow model is capable of predicting laboratory measurements with an acceptable accuracy; however, the fracture flow model seemed to need more improvement. Probable factors that could have caused inaccuracy in the fracture flow model were discussed and actions for improving it were recommended as an extension to this research. Application of the empirical models in fractured porous medium characterisation simulations reduces the coupling-related numerical complexities. The coupled empirical models can predict flow dynamics of fractured reservoirs under various stress regimes. They demand much less computational effort and, as they incorporate geometrical factors, they can be up-scaled conveniently. In terms of production planning for fractured reservoirs, the empirical models can assist engineers to manage matrix and fracture production efficiently based on overburden stress variations.