Induced fault reactivation by thermal perturbation in enhanced geothermal systems

Highlights

• The non-uniformity of thermal drawdown along the fault damage zones significantly affects the magnitude and timing of induced earthquakes, by generating different magnitudes of perturbation on the stress state of the fault and surrounding rocks.

• A reduced injection rate delays the timing required to reach the moment of fault reactivation. The comparison of slip distance profiles suggests that the magnitude of induced earthquakes grows with the increase of injection rate.

• The overall slip distance increases with the decrease of injection water temperature. Moreover, the more pronounced thermal drawdown response tends to prompt the onset of seismic fault reactivation in an earlier stage.

• The transition period from aseismic slip to seismic slip also produces a substantial impact in the evolution of fault permeability. The evolution of fault permeability is subjected to the process of aseismic slip enhancement followed by an abrupt reduction at the post-failure stage, due to the joint stiffness is defined as more sensitive to normal compaction response.

Abstract

Heat extraction by circulating cold water through a geothermal reservoir could potentially induce earthquakes of large magnitudes. In this work, we explore the role of water injection inside a normal fault in triggering seismic slip, under different temperature and rate scenarios of fluid injection into fault damage zones. The resulted non-uniformity of thermal drawdown along the fault damage zones significantly affects the magnitude and timing of induced earthquakes. A non-dimensional expression integrating fault configuration (e.g. length and thickness) and injection condition (e.g. rate and temperature) is used to describe the relationship between the injection schedule and the resulting fault seismic slip event. As the dimensionless parameter $Q_D$ increases, suggesting a transition from heat conduction to convection, the dimensionless event timing also grows nonlinearly. The perturbation of fault stress field induced from localized thermal cooling process is pronounced, compared to decoupled hydro-mechanical scenarios. The stress field perturbation in the system due to thermal cooling is characterized through a Coulomb friction ratio analysis for evaluating the stress changes along the fault plane and a tensor-based stress perturbation analysis for quantifying the stress changes in the damage zones and host rocks. The thermal influence acting on local patches along the fault strike not only advances the timing of seismic slip, but also increases the magnitude of induced seismic events, by unloading the fault to prompt seismic rupture. The injection temperature has a significant impact on facilitating the onset of seismic slip, i.e. attempting to accelerate the timing and increase the magnitude of fault reactivation. The injection rate variation will affect the timing by changing the pore pressure field and heat transfer manner. Prior to the onset of fault reactivation, the thermal unloading response increases fault permeability by decreasing normal stress, such that more permeable channels in the fault allow fluid to diffuse. Produced plastic shear strain due to fault slip provide extra positive contributions to increased normal aperture through shear dilation, thereby the fault permeability increases significantly by around two orders of magnitude.

Introduction

Geothermal resources have been identified as a promising energy resource for relieving the intense energy demand, and therefore have been extensively exploited in many countries (Tester et al., 2006; Boyle, 2004). Especially, the Enhance Geothermal Systems (EGS) that have a significant potential to deliver extra energy supply such as electricity and heating have started to attract increased societal and industrial attentions. Given the nature of low virgin permeability of granitic rocks and poor geometric connectivity of transmissible natural fractures in EGS reservoirs, the most feasible practice of exploiting geothermal energy relies on injecting fluids without proppants, thereby enhancing the permeability through hydro-shearing (Rinaldi et al., 2015; McClure and Horne, 2011). The major challenge of this approach is the high likelihood of inducing earthquakes and microseismicity, which have already been observed in numerous geothermal power plants worldwide (Majer et al., 2007; Guglielmi et al., 2015). For example, the 2017 Pohang earthquake with a moment magnitude of 5.4 occurred two months after the injection of a large volume of fluids into a critically stressed fault zone (Kim et al., 2018). Similarly, an earthquake occurred with a magnitude of 3.4 following the hydraulic stimulation at the Basel EGS field (Bachmann et al., 2011; Mignan et al., 2015).

In geothermal reservoirs, it is critical to consider the coupled thermal-hydro-mechanical effects on modulating fault permeability (Gan and Elsworth, 2014a, 2014b), as it will directly determine the pressure diffusion and heat transfer rates along the fault. Furthermore, the seismic rupture in slip zones could grow beyond the pressurized regions, which is consistent with the field observations (Rivet et al., 2016; Cornet, 2016). In the past, extensive studies have examined the potential induced fault slip behavior in EGS reservoirs, either through physical experiments or numerical simulations. The classical rate- and state-dependent frictional law was used to characterize the evolution of the fault’s friction coefficient as a result of changes in slip velocity, which in turn affects the frictional strength of the fault (Dieterich, 1979; Ruina, 1983). Also, the fault slip behavior could also be investigated through the brittle response, by defining the plastic deformation in the weakening plane of the fault (Cappa and Rutqvist, 2011; Rutqvist et al., 2002). The coupled thermal-hydraulic-mechanical process generated significant perturbation by changing the effective stress state of fault that eventually leads to fault instability. There is a complex competition process in evolving the stress state of the fault, either pressurizing or depressurizing the fault zone. Since most fault slips occur below the water table where the formations are fully saturated by pore fluids, these fault slip movements generate heat in both matrix rocks and pore fluids. As the thermal expansion of water is much more significant than that of rocks, and the rock is much stiffer than water, consequently the fault zone is pressurized by trapping the pore fluids in undrained conditions, given if the fault zone is sufficiently impermeable. Such a response is defined as thermal pressurization (Sibson, 1973; Noda and Shimamoto, 2005; Rice, 2006; Rempel and Rice, 2006; Urpi et al., 2019). However, pore fluid depressurization due to fault zone transient dilation would arrest the nucleation of unstable slips, especially under fluid saturated conditions (Samuelson et al., 2009, 2011). Moreover, the distributions of mean and differential stresses in the fault also have strong impacts on determining the flow path and permeability enhancement, by affecting the seismic magnitude and permeability variation (Zhang et al., 2008; Fang et al., 2018). As the permeability of fault zones is mediated by the aperture of joints through the cubic law relationship, the evolution of stress state and associated plastic failure response along the fault zone will determine the conductivity of fault damage zones (Im et al., 2018), and thereby affect the outcomes of heat energy extraction. Depending on the magnitude of the initial permeability of the fault and reservoir, the evolution of rock permeability may be dominated by different physical processes (e.g. hydraulic, thermal and/or chemical processes) (Vilarrasa et al., 2017; Rutqvist et al., 2016). Recent work also revealed that the permeability enhancement by dilation through mix mechanism stimulation or cyclic injection stimulation could prompt the extraction of geothermal energy with less seismic responses (Rinaldi and Rutqvist, 2019; Norbeck et al., 2018; Rutqvist et al., 2016).

In the past, we have identified that the induced homogeneous thermal drawdown along the fault has a great likelihood in triggering fault reactivation, for the scenario of fluid injector and producer being placed at the opposite sides of the fault (Gan and Elsworth, 2014a, 2014b). It has been recognized that the local thermal gradient due to non-uniform heat transfer could generate substantial thermal stress, leading to significant alteration of the stress state of the fault. In this work, we propose to simulate another scenario of thermal drawdown by direct fluid injection in the fault damage zones of an EGS reservoir, in order to evaluate the role of localized thermal drawdown along the fault in triggering fault reactivation. We conduct numerical simulations based on a fully coupled thermal-hydro-mechanical simulator TOUGHREACT-FLAC3D (Taron and Elsworth, 2009), in which the mechanical plastic yield is governed by the Mohr-Coulomb failure criterion according to the mechanism of frictional weakening with an increment of plastic strain. We study the non-uniform thermal drawdown behavior by exploring a range of injection rate and temperature conditions, with a specific aim to elucidate the primary factors controlling the timing and magnitude of fault slip. The stress state perturbation in the fault zone induced by injection and cooling is characterized by a recently-developed tensor-based approach, which faithfully honors the tensorial nature of the stress field. By interpreting the stress data produced from numerical models, the perturbation around the fault will be quantified and visualized.