Miranda, and 4 co-authors, reported in early 2023 that understanding the natural fracture network is essential for geothermal-related investigations. However, the geometrical attributes depend on the scale of observation. Therefore, a multiscale characterization of the fracture network is essential to ensure that forward heat and flow simulations are based on stochastically generated discrete fracture network models representative of the natural fracture system observed. This was the goal of this work.
Fracture data was collected from satellite imagery, outcrops and well cores to evaluate the scale effect and to study the impact of fracture size and density on the performance of engineered geothermal systems by numerical modeling. The numerical simulations highlighted that networks made of small fractures (0.08 to 27 m) tend to decrease the performance of the system compared to a network made of large fractures (22 to 1,437 m).
However, thermal short-circuiting is easily reached in the latter scenario. Thus, the simulations suggest that the best-case network is made of fractures ranging between 1.57 to 135 m with fractures spaced by 5 m. This scenario provides the best compromise between heat extraction, water losses, hydraulic impedance and thermal drawdown. Despite the uncertainties, the fracture data used highlights the importance of multiscale fracture analysis for heat-flow simulations of geothermal reservoirs.
The authors concluded that despite the scale effect being known, it is still a challenging aspect to consider when analyzing fracture data with the goal of generating stochastic DFNs and under-stand the deep-seated structures.
The results of this work highlight that orientation, size and connectivity are greatly dependent on the scale of observation. Despite existent, the scale effect on fracture spacing and linear frequency is less significant. Outcrop data alone suggest one main fracture set while the combination of satellite image and outcrop data indicates five main families. Furthermore, outcrop data indicates fractures smaller than 30 m while structural lineaments of up to near 1,500 m were identified in the satellite image. A power-law distribution provides the best fit for the fracture size. Additionally, core data reveals more widely spaced fractures, and consequently lower linear fracture frequency, than outcrop data. The data also suggests that fracture spacing can be fit to an exponential distribution. Fracture aperture is found to obey a power-law distribution. Finally, connectivity analysis suggests that the small-scale outcrops tend to have greater connectivity than large scale (i.e., satellite imagery scale) outcrops.
The numerical simulations of EGS performance suggest that a network composed by larger fractures (satellite imagery size; 22 to 1,437 m) performs better than a network composed by smaller fractures (outcrop size; 0.08 to 28 m). One reason is the lower shear stiffness of larger fractures. However, a network made exclusively of larger fractures tend to increase the probability of reservoir thermal short-circuiting. The simulations also illustrate that a small fracture size network needs greater fracture frequency and/or greater stimulation and circulation pressures to achieve similar flow rates to a large fracture size network.
Despite being a case study, the work described here aims at acquiring new knowledge on the geometry and connectivity of natural fracture networks. This is essential to proceed with exploration steps and evaluate the technical feasibility of EGS in Precambrian basement rocks. This represents a valuable contribution for the geo-thermal community assessing the potential of geothermal technologies in northern environments and dealing with important data gaps. However, the authors are well aware that deep geothermal exploration boreholes are needed to validate any estimates based on outcrop analogues. Nevertheless, to raise the interest for geothermal exploration drilling, research needs to be undertaken using data that can be economically collected. This is where this contribution gains its scientific value.
Managed & Built by I2M Consulting, LLC | © 2010-
