Impacts of natural fracture geometries on heat production in enhanced geothermal systems: from 3D discrete fracture models to uncertainty analysis

Fracture geometry is critical for Enhanced Geothermal Systems (EGS) performance. This study uses 3D ensemble simulations to investigate how natural fracture characteristics affect long-term heat production. Discrete fracture models are built stochastically using Monte Carlo simulations with power-law length distributions, Fisher orientation distributions, and power-law aperture-length scaling relationships, based on data from the Soultz EGS site. Coupled thermo-hydraulic simulations are performed using the finite element method (FEM) to solve fluid flow and heat transport equations. An analytical solution for doublet heat extraction in idealized fracture-matrix systems verifies the numerical model accuracy. Results from fifty realizations show that fracture geometries significantly influence production temperature and hydraulic-head differences. Temperature and geothermal energy production increase slightly with fracture length and density but decrease with aperture-length correlation exponent. Increasing the aperture-length correlation exponent from 0 to 2 reduces mean temperature from 170 °C to 119 °C, decreases energy production by approximately 37 % (from 20.3 PJ to 12.8 PJ over 60.6 years), and conversely increases hydraulic-head differences, demonstrating that fracture heterogeneity significantly impacts EGS performance through promoting preferential flow channeling while reducing connectivity. An exponential function best fits both relationships. The uncertainty of heat extraction performance decreases with increasing fracture length and density. While parameters are based on Soultz site characterization, the exponential relationship between fracture heterogeneity and thermal performance is expected to hold qualitatively in other fractured geothermal reservoirs.