Session 3B: Enhanced (or Engineered) Geothermal Systems

Hydraulic fractures occur during the stimulation of Enhanced Geothermal Systems when fluid pressure exceeds the minimum principal stress. Fracturing could potentially lead to seismic events if hydraulic fractures connect to a nearby fault. Hydraulic fractures in brittle geothermal rocks can exhibit distinct hydraulic and mechanical characteristics implying different influence on induced seismicity. Upon hydraulic opening of a fracture, shear stress decreases to zero, perturbing stress on adjacent fractures and potentially inducing slip through elastic stress transfer. Moreover, the permeability of hydraulic fractures is strongly influenced by aperture size and influences pressure diffusion at distance. Developing a numerical model that incorporates these hydraulic fracture effects was useful to understand their influence on induced seismicity.

The model distinguished between hydraulic opening and closure under non-penetration constraints. Two friction models, rate and state model, and linear weakening model, were employed to solve shear traction and slip. The study investigated sensitivity of factors including in-situ stress, fault permeability, friction properties, and injection strategy on hydraulic aperture and slip development. The findings inform prediction of induced seismicity through measurable observations and design of injection strategy to mitigate induced seismicity.

In April 2022 and July 2023, a series of hydraulic stimulation and fluid circulation experiments were executed at the Utah FORGE injection well 16A(78)-32, accompanied by high-resolution seismic monitoring. These experiments have shown a consistent pattern of upward migration in seismic cloud development, indicating the ascendant propagation of both hydraulic fractures (HF) and natural fractures (NF) upon fluid injection. Building on our prior numerical simulations, the results have demonstrated that the upward growth of hydraulic fractures may interact with natural fractures leading to a more complex stimulated reservoir volume (SRV) consisting of a HF/NF network (Kumar, Liu, & Ghassemi, 2023). In this study, we performed a theoretical analysis to further analyze fracture propagation behavior in response to fluid injection in Utah FORGE. Leveraging rock properties derived from FORGE core samples, in-situ stress assessments from borehole image logs (Ye et al., 2022; Ye & Ghassemi, 2024), and stochastic fracture set orientations within the Utah FORGE context, our investigation aimed to analytically evaluate the growth of the SRV and fracture evolution. Our findings indicate that fracture set 2 (South-striking, moderately dipping West) and fracture set 4 (East-striking, sub-vertical dipping South) are likely to be among the first to propagate and grow during fluid injection. Additionally, our theoretical analysis reinforces the dominance of upward fracture migration as the primary propagation mechanism within the FORGE stimulation environment. We also discuss the results with reference to the hydroshearing conceptual model and its implications for SRV size and geometry.

In recent years, a number of innovative concepts have been proposed to extract geothermal energy from hot dry resources. These include connecting multilaterals between injector and producer (Holmes, et al., 2021), thermal reach enhancement (Moncarz and Suryanarayana, 2022), closed loop geothermal (Beckers, et al., 2022), down-borehole exchange (all of which are increasingly referred to as Advanced Geothermal Systems, or AGS), and the classical Enhanced Geothermal Systems (EGS) that connect two (or more) wells through some form of fracturing or stimulation. The EGS approach has historically received the most attention and funding, since its beginnings in 1970s in Fenton Hill (Brown et al., 2012). The recent successes of Fervo (Norbeck, et al, 2023) and Utah FORGE (Allis and Moore, 2019) have brought EGS closer to reality. Both these demonstration projects were at temperatures around 200°C. More recently, the idea of using EGS in superhot rock (with resource temperatures greater than 375°C) has emerged as a viable and scalable technique for the production of geothermal energy (US Department of Energy, 2024). Moncarz and Suryanarayana (2022) describe a novel AGS technique for exploiting superhot rock. The US DOE has recently increased its focus in this area by awarding a grant to Mazama Energy, to create EGS reservoirs in superhot rock.

All these methods involve the circulation of a working fluid through the fracture network in the resource to extract heat from the resource, which is then converted into electric power using an appropriate power cycle. The most commonly used working fluid is water.

Regardless of the concepts being proposed and tested, they all scale according to the parameters and properties governing heat transfer behavior in the resource. In this paper, we illustrate the heat transfer behavior of these systems using scaling parameters that govern the problem. It is shown that at pseudo-steady state, the non-dimensional temperature of the working fluid scales with the Biot number and a suitably defined Peclet number, as well as the ratio of the contact area to conduit area, thus encompassing all the parameters that govern heat transfer. The scaling curves provide a useful basis for comparing different heat extraction concepts, and for identifying improvements. It is shown that a single unified parameter known as the Number of Transfer Units, common in Heat Exchanger Analysis, can describe the thermal performance of practically any EGS or AGS.

One of the objectives of this study is to generalize the approach of the nearly 50-year-old study of Gringarten et al. (1975) who considered a network of very large (1 km2) infinite planar fractures, and define a non-dimensional time parameter. While perhaps not practical at the time of their study, their results indicate the potential for substantial power generation with a properly constructed network of fractures in a hot dry resource.

Further, when extracting heat from hot, dry resources, depending upon the technique being used, the thermal penetration depth, and therefore, rate of thermal decline, vary. The thermal penetration depth is analogous to the drainage radius in oil and gas production. In this paper, we present a semi-analytical approach to determine the non-dimensional penetration depth, as a function of non-dimensional time (the Fourier number). Combining this with the scaling analysis above, we then investigate the expected thermal decline rate and the pseudo-steady state energy recovery for a number of examples. The notion of thermal recovery efficiency (analogous to the recovery factor in hydrocarbon reservoirs) is discussed in the context of these examples.

The authors believe that this work serves as a useful, generalized basis for quickly evaluating different thermal recovery concepts, and examining the impact of different modifications or improvements to heat recovery and recovery efficiency. This work complements the more sophisticated, but time-consuming, numerical simulations that are often used in literature to model heat recovery assurance. While it does not replace these numerical methods, it provides valuable insight into the thermal performance of different AGS and EGS systems.

The experiences with the seismic risk in the EGS Basel project in 2006 were sobering and delayed the development of deep geothermal energy in Switzerland for almost 20 years. Since these dramatic experiences, there has been an increase in scientific progress as well as new setbacks, such as in the St. Gallen and Pohang projects. These failures have in turn provided many insights into mitigating seismic risks. Some of these lessons learned led to further developments that are now being tested, verified, and qualified in underground rock laboratories such as Bedretto (Switzerland) and in large-scale R&D field tests such at the Utah FORGE project (USA). These progresses include the field testing of multi-stage EGS stimulations, methodologies for risk studies, Advanced Traffic Light Systems, development of monitoring tools for high temperatures and their field verification and qualification. We will present an overview of the recent works and their implementation plans for the new EGS Haute-Sorne project in Switzerland.

In April of 2022, three stages of stimulation were carried out near the toe of well 16A(78)-32 at Utah FORGE site. Based on extent of the recorded microseismicity and modeling results, well 16B(78)-32 was drilled 100 m above well 16A. In July of 2023 circulation program was conducted between two wells. The aim was to implement low-rate injection to interrogate the reservoir between the injection and the production wells, 16A(78)-32 and 16B(78)-32, respectively, and to assess the effect of reservoir stimulation in April of 2022. The circulation testing was designed to use low injection rates so that hydraulic fractures would not be created or propagated. Back analysis of the circulation test was carried out using a coupled hydro-mechanical (HM) numerical model based on distinct element method (DEM). The objective of the modeling was to interpret the state of the reservoir by matching the recorded injection pressures and rates at both wells. The model includes an explicit representation of the DFN. Some DFN fractures were generated to match the observed microseismic event locations, while the majority were generated stochastically. The models indicate that the stimulated reservoir permeability is a function of distance from the injection well. Although seismicity recorded during stimulation extended 100 m above the injection well and circulation test confirmed connectivity between two wells, the back analysis indicated that two wells were not connected by the domain of increased permeability. Also, circulation at rates of 5 bpm resulted in additional fracture propagation manifested by decaying injection pressure history and microseismicity recorded during circulation.

The study explores the potential and value extraction through the fracture geometry calibration component of end-to-end lifecycle simulations for the 16A(78)-32 and 16B(78)-32 injector-producer pair of Utah FORGE. Sophisticated geomechanical, fracture propagation, slurry flow, models were used to predict the fracture network for different treatment scenarios.

The 3D geomechanical model was built from the structural geologic features and the thermal and mechanical properties profiles. As the first step of calibration, we used elastic properties calculated from openhole logs and empirical core test data. A high fidelity discrete fracture network (DFN) was used as an input in the 3D geo-model. The DFN was built based on data sets collected from five wells, microseismic monitoring during a 2022 stimulation campaign to constrain fracture orientations and intensity, and fracture characterization in 16B(78)-32. The simulations were conducted in a 3D non-homogenous mode with stress shadowing and fracture interactions. The viscous heating model as incorporated to account for the temperature and exposure time effect on fluid viscosity in the transient pumping phase. Another step of calibration was conducted using the microseismic data acquired in the previous stage 3 stimulation of well 16A(78)-32 in April 2022. The DFN characteristics, such as fracture length, orientation, anisotropy were adjusted to match the microseismic observations along with the magnitude of stress anisotropy. Connection efficiency in the doublet was also predicted based on the fracture hit observations.

Multiple sensitivity scenarios were conducted with different perforation strategies, pumping schedules and varying proppant and fluid properties, through calibration process. Fluid viscosity can be critical for upward proppant transport and creation of a reliable, sustainable fracture connection..