Empowering Enhanced Geothermal Systems: Integrating Modelling and Laboratory Experiments

Enhanced geothermal systems (EGS) have emerged as a key technology for harnessing high-temperature energy from hot dry rock (HDR) at depths of ~3-10 km. Rapid progress in horizontal drilling and multi-stage hydraulic fracturing has enabled creation of large stimulated reservoir volumes, yet major challenges remain - fracture-flow short-circuiting, limited ability to characterize permeability evolution during stimulation, and injection-induced seismicity.

We illustrate and address some of these challenges through coupled thermo-hydro-mechanical (THM) simulations of a partially bridging multi-stage hydraulic-fracture well design that mitigates short-circuiting and delays thermal breakthrough relative to conventional fully bridging designs (Yu et al., 2021a). Building on this framework, we quantify permeability evolution and evaluate induced seismicity driven by long-term thermo-poroelastic stressing during production (Yu et al., 2023). By separating thermoelastic, poroelastic, and combined thermo-poroelastic contributions, we show that thermoelastic stressing dominates stress redistribution and seismicity rate in the stimulated reservoir, while poroelastic effects tend to stabilize and delay failure by moderating stress changes.

During shear stimulation, microearthquakes (MEQs) carry information about hydraulic rock properties, particularly permeability. We develop a mechanistic scaling that links seismic moment as diagnostic of incremental permeability creation during shear stimulation, and validate the relationship using laboratory fault-reactivation experiments with absolute constraints on seismic moment, demonstrating proportionality between permeability change and seismic moment (Yu et al., 2026). This MEQ-permeability linkage is applied to field data from the 2021 Gonghe EGS stimulations to infer the evolving 3D permeability distribution by estimating source parameters (e.g., source radius and stress drop) from raw waveforms. The resulting workflow enables near-real-time reservoir state estimation to support operational decision-making during stimulation.

Speaker: Pengliang Yu, Pennsylvania State University