Enhancing fracture-network characterization and discrete-fracture-network simulation with high-resolution surveys using unmanned aerial vehicles

Abstract

A workflow is presented that integrates unmanned aerial vehicle (UAV) imagery with discrete fracture network (DFN) geometric characterization and quantification of fluid flow. The DFN analysis allows for reliable characterization and reproduction of the most relevant features of fracture networks, including: identification of orientation sets and their characteristics (mean orientation, dispersion, and prior probability); scale invariance in distributions of fracture length and spatial location/clustering; and the distribution of aperture values used to compute network-scale equivalent permeability. A two-dimensional DFN-generation approach honors field data by explicitly reproducing observed multi-scale fracture clustering using a multiplicative cascade process and power law distribution of fracture length. The influence of aperture on network-scale equivalent permeability is investigated using comparisons between a sublinear aperture-to-length relationship and constant aperture. To assess the applicability of the developed methodology, DFN flow simulations are calibrated to pumping test data. Results suggest that even at small scales, UAV surveys capture the essential geometrical properties required for fluid flow characterization. Both the constant and sublinear aperture scaling approaches provide good matches to the pumping test results with only minimal calibration, indicating that the reproduced networks sufficiently capture the geometric and connectivity properties characteristic of the granitic rocks at the study site. The sublinear aperture scaling case honors the directions of dominant fractures that play a critical role in connecting fracture clusters and provides a realistic representation of network permeability.