A New Technique for the Modelling of Gas Production from Shale Gas Reservoirs

Abstract

The physics of hydrocarbon storage in shale and mechanisms involved in the production of hydrocarbons from shale are different from those for conventional reservoirs. Recent experimental studies and numerical investigations have demonstrated fundamental differences even between different shale plays. This, together with the combination of multiple porosities, gas desorption and slippage effects, represents a huge challenge for simulators. However, with reasonable assumptions one can develop mass conservation and flow equations at the nano-scale. In our research we implemented the most relevant physical mechanisms for fluid flow in shales into our in-house Automated Differentiation based General Purpose Research Simulator (AD-GPRS). Here we propose and test a three-step up-scaling technique that we believe is adequate for incorporating essential physics for modeling production from shale gas reservoirs. The effect of micro- and macro- fractures are included in this process. Three scales considered are core sample scale, reservoir gridblock scale and reservoir scale. We start from the core sample scale using mathematical models for shale gas physics, which are valid for this scale, introduce different sets of micro fractures into the domain. Then we compare production from the fractured domains against an equivalent homogeneous domain. Different combinations of fracture density and ratio of fracture/matrix permeability are analyzed for the fractured cases. The procedure allows us to include the effect of micro-fractures into a scaling function of permeability. The scaling function behaves linearly with increase in micro-fracture density and follow a power law distribution with increase in fracture/matrix permeability ratio. An analytical form of the scaling function is proposed. To account for the macro-fractures at the reservoir gridblock scale we develop a dual porosity model where the effect of micro-fractures is incorporated into the matrix domain through the proposed scaling function. Our results show that fracture network significantly increases rates for the initial years of production, whereas slippage and desorption show increase of 10-15% in cumulative production.