Speaker
Description
Digital rock physics (DRP) is widely used to predict petrophysical properties from pore-scale images, yet its application to low-porosity crystalline rocks remains limited. In granites, low connected porosity, complex mineral intergrowth, fine inclusions, and alteration textures challenge conventional grayscale-based phase identification (segmentation), reducing the reliability of predicted effective elastic and transport properties. Property estimates are known to be sensitive to subtle microstructural features, and thus these difficulties in phase identification are particularly relevant for property prediction based on cores from crystalline geothermal reservoirs and crystalline-hosted mineral deposits.
We developed a geology-driven DRP workflow constrained by independent geological and laboratory observations for a granitoid sample from the Frontier Observatory for Research in Geothermal Energy (FORGE), Utah, USA. High-resolution X-ray computed tomography (XRCT) was used to acquire three-dimensional images of the rock microstructure at a resolution of 6.9 µm/voxel. Multiphase segmentation combines grayscale normalization, histogram-based thresholding, targeted morphological operations (isolated voxel removal, boundary smoothing), and watershed algorithms. These steps were guided by thin-section petrography, scanning electron microscopy (SEM) observations, and laboratory measurements, including bulk density, connected porosity for grayscale calibration, and phase volume validation.
The resulting segmented volumes distinguish quartz, feldspar, ferromagnesian minerals, including amphiboles, accessory phases, and pore space. Finite-difference simulations of elastic wave propagation were performed on segmented subvolumes with an edge length of 400 voxels (0.00276 m $\times$ 0.00276 m $\times$ 0.00276), with phase-specific elastic properties assigned to each mineral and dry pore phase, based on literature values for 100% intact crystals. Computed P- and S-wave velocities show good overall agreement with laboratory ultrasonic measurements but systematically predict higher effective stiffness than experimental data (ultrasonic measurements at 1 MHz on 0.04 m diameter samples). This discrepancy of approximately 20-30% indicates that unresolved microporosity, insufficient grain-to-grain contact stiffness modeling, and limited representative elementary volume (REV) size remain critical sources of uncertainty in pore-scale elastic modeling.
Our results emphasize the need to extend this segmentation workflow to multiscale imaging and upscaling strategies (XRCT + FIB-SEM) that better capture grain contacts and sub-resolution porosity, as well as pressure-dependent measurements to account for in-situ stress effects. The presented approach contributes to the construction of geologically consistent digital twins of crystalline Enhanced Geothermal Systems (EGS) and provides pore-scale insights relevant for geothermal reservoir characterization and mineral exploration beyond sedimentary systems.
| Country | Germany |
|---|---|
| Student Awards | I would like to submit this presentation into both awards |
| Acceptance of the Terms & Conditions | Click here to agree |
