Review of discrete fracture network characterization for geothermal energy extraction

Abstract

Geothermal reservoirs are highly anisotropic and heterogeneous, and thus require a variety of structural geology, geomechanical, remote sensing, geophysical and hydraulic techniques to inform Discrete Fracture Network flow models. Following the Paris Agreement on reduction of carbon emissions, such reservoirs have received more attention and new techniques that support Discrete Fracture Network models were developed. A comprehensive review is therefore needed to merge innovative and traditional technical approaches into a coherent framework to enhance the extraction of geothermal energy from the deep subsurface. Traditionally, statistics extracted from structural scanlines and unmanned aerial vehicle surveys on analogues represent optimum ways to constrain the length of joints, bedding planes, and faults, thereby generating a model of the network of fractures. Combining borehole images with seismic attributes has also proven to be an excellent approach that supports the stochastic generation of Discrete Fracture Network models by detecting the orientation, density, and dominant trends of the fractures in the reservoirs. However, to move forward to flow modelling, computation of transmissivities from pumping tests, and the determination of hydraulically active fractures allow the computation of the hydraulic aperture in permeable sedimentary rocks. The latter parameter is fundamental to simulating flow in a network of discrete fractures. The mechanical aperture can also be estimated based on the characterization of geomechanical parameters (Poisson’s ratio, and Young’s modulus) in Hot Dry Rocks of igneous-metamorphic origin. Compared with previous review studies, this paper will be the first to describe all the geological and hydro-geophysical techniques that inform Discrete Fracture Network development in geothermal frameworks. We therefore envisage that this paper represents a useful and holistic guide for future projects on preparing DFN models.