Numerical modelling of fault reactivation in carbonate rocks under fluid depletion conditions – 2D generic models with a small isolated fault

Highlights

• Numerically modelling the reactivation of a critically-stressed fault in a carbonate reservoir for fluid depletion condition.

• The model is based on field and laboratory experimental data.

• Horizontal stress decrease and fault shear-stress increase trigger fault reactivation during depletion.

• Normal-faulting downthrows and strain localization along the fault results from fault reactivation.

• Fault and damage zone permeability changes and fluid flow pattern varies during and after reactivation.

Abstract

This generic 2D elastic-plastic modelling investigated the reactivation of a small isolated and critically-stressed fault in carbonate rocks at a reservoir depth level for fluid depletion and normal-faulting stress conditions. The model properties and boundary conditions are based on field and laboratory experimental data from a carbonate reservoir. The results show that a pore pressure perturbation of −25 MPa by depletion can lead to the reactivation of the fault and parts of the surrounding damage zones, producing normal-faulting downthrows and strain localization. The mechanism triggering fault reactivation in a carbonate field is the increase of shear stresses with pore-pressure reduction, due to the decrease of the absolute horizontal stress, which leads to an expanded Mohr's circle and mechanical failure, consistent with the predictions of previous poroelastic models. Two scenarios for fault and damage-zone permeability development are explored: (1) large permeability enhancement of a sealing fault upon reactivation, and (2) fault and damage zone permeability development governed by effective mean stress. In the first scenario, the fault becomes highly permeable to across- and along-fault fluid transport, removing local pore pressure highs/lows arising from the presence of the initially sealing fault. In the second scenario, reactivation induces small permeability enhancement in the fault and parts of damage zones, followed by small post-reactivation permeability reduction. Such permeability changes do not appear to change the original flow capacity of the fault or modify the fluid flow velocity fields dramatically.

Introduction

Fault reactivation is a widely observed structural phenomenon important for both mineral and petroleum systems. For mineral systems, fault reactivation commonly results in the localization of deformation and dilation, permeability enhancement, fluid focusing and hence, favourable conditions for mineralization (e.g. Schaubs and Wilson, 2002, Sorjonen-Ward et al., 2002, Jolley et al., 2004, Zhang et al., 2008). For petroleum systems, fault reactivation and seal breach over geological time can reduce the integrity of hydrocarbon traps, leading to leakage of the accumulation and hence, represents a major risk for hydrocarbon exploration in regions experiencing tectonic events (e.g. Anderson et al., 1994, O'Brien et al., 1999, Gartrell et al., 2006, Langhi et al., 2010).

Most previous studies about fault reactivation concern fault structural behaviour under the active tectonics or during laboratory deformation. In a broad context, Sibson (1985) noted that fault reactivation accommodated the bulk of intra-plate deformation in frictional seismogenic regimes and he described the stress conditions for fault reactivation. Sibson's view is supported by extensive observations of fault reactivation in mineralized terrains or petroleum basins. Some examples include: (1) Bull and Scrutton (1990) reported fault reactivation in the Indian Ocean related to the rheology of oceanic lithosphere; (2) Cox (1995) showed that the reverse-fault reactivation for small shear stress and a large fluid pressure regime led to multiple episodes of fluid flow through the fault; (3) McClay and Bonora (2001) illustrated the formation of restraining step-over fault structures as the result of strike-slip reactivation of pre-existing fault systems, and compared similarities between the patterns from analogue models and several hydrocarbon fields/basins. Also, a large number of numerical modelling studies have simulated fault reactivation for convergent or extensional deformation conditions, such as: (1) Walsh et al. (2001) simulating the development of a growth fault for comparison with data from seismic interpretations; (2) McLellan et al. (2004) numerically modelling deformation localization and fluid flow around a normal fault in an extensional basin; (3) Zhang et al. (2009) modelling the distribution of down-throw, strain and fluid flow patterns around pre-existing faults for extensional conditions; (4) Buchmann and Connolly (2007) numerically simulating stress distribution and fault reactivation in the Upper Rhine Graben for transtensional conditions; and (5) Zhang et al. (2012) modelling the fault reactivation and fluid flow around a step-over “flower” fault structure for the Laverton gold region in the Yilgarn Craton of Western Australia for transpressional deformation conditions.

There is also another type of fault reactivation not directly related to active tectonics (i.e. under static conditions of the far-field stresses), which can be initiated by fluid injection (e.g. Van Ruth et al., 2006, Rutqvist et al., 2015) or depletion (e.g. Zoback et al., 2001, Safari et al., 2013), predominantly in sedimentary basins or the immediately underlying basement rocks. Fault reactivation as the result of fluid injection in a tectonically-static regime is easy to understand on the basis of conventional geomechanics, because the elevation of fluid pore pressure will decrease effective stresses and shift the Mohr's circle of stress for a fault toward the failure envelope (Safari et al., 2013). However, the geomechanical explanation for fault reactivation under fluid depletion conditions is less intuitive, because fluid pore pressure reduction will increase effective stresses and shift the stress Mohr-circle of the fault further away from the failure envelope. Segall (1989) attributed fault reactivation and seismicity during production in oil and gas fields to the development of pore pressure reduction and local contraction of the rock mass in response to pore fluid extraction, changes of poroelastic stresses and resultant local rock deformation. A number of more recent poroelastic studies (e.g. Zoback et al., 2001, Streit and Hillis, 2002, Hawkes et al., 2005, Nacht et al., 2010, Safari et al., 2013) proposed a geomechanical theory governing fault reactivation during reservoir depletion under a normal-faulting stress setting. Assuming the presence of stable vertical stresses (the maximum principal stresses) in reservoirs, these studies show that in poroelastic materials, decrease in the minimum horizontal stress accompanying a decrease in fluid pressure during depletion triggers fault reactivation. The ratio (A) between the minimum horizontal stress decrease and pore pressure decrease is a critical parameter that is determined by Poisson's ratio and Biot's coefficient (Zoback et al., 2001, Hawkes et al., 2005). For example, Zoback et al. (2001) showed that for the friction coefficient of 0.6 (friction angle = ∼31°), A ≥ 0.67 will trigger fault reactivation, but different critical ratio ranges are reported for different parameters and data (Streit and Hillis, 2002, Hawkes et al., 2005).

This study uses a coupled elastic-plasticity and fluid-flow numerical modelling approach to simulate the reactivation of a small isolated fault in a carbonate rock reservoir undergoing fluid depletion. The models are based on a set of specific laboratory experimental data for a deep carbonate reservoir. Our focus is on the geomechanical and hydrological impacts of fault reactivation in carbonate rocks including: 1) the patterns of strain and stress perturbation; 2) reactivation-associated fault and/or damage zone permeability development; and 3) effects on fluid flow patterns. This modelling study is part of a 3-year research program investigating fault reactivation in carbonate reservoirs in Brazil. As such, the generic models presented here reflect the relevant architecture and depth case of the reservoir, and model property parameters for laminated continental carbonate (travertine), fault core and damage zones are derived from laboratory work of this research program (Delle Piane et al., 2016, Giwelli et al., 2016a, Giwelli et al., 2016b).