Numerical benchmarking of GREAT cell experiments: insights into the impact of polyaxial stresses on fluid flow in fractured rock

A systematic benchmark suite is conducted to evaluate numerical methods in hydro-mechanical (HM) fracture mechanics. These benchmarks draw inspiration from experimental data collected using the GREAT cell at the University of Edinburgh—an advanced equipment designed to analyze fractured rocks under rotating stress conditions that simulate real-world subsurface environments. Given the inherent complexity of the GREAT cell experiments, the benchmarks have been simplified to replicate key behaviors while remaining manageable for computational modeling. This approach allows researchers to assess and compare numerical methods for simulating fracture propagation and hydro-mechanical interactions.

Two numerical approaches were utilized to perform these simulations: the variational phase field (VPF) method and the lower interface element (LIE) method. The VPF method employs a diffuse fracture representation, which enables it to model dynamic fracture propagation like branching and merging without the need for predefined paths. In contrast, the LIE method uses a discrete fracture representation, where fractures are explicitly embedded as interfaces within the computational mesh. While the LIE method is computationally efficient for stationary or pre-existing fractures, it lacks the inherent capability to simulate propagating fractures. By comparing these complementary approaches, the study highlights their respective strengths and limitations, providing valuable insights into fracture behavior under diverse hydro-mechanical conditions.

All numerical implementations and benchmarks are available in the OpenGeoSys platform, ensuring accessibility and reproducibility. This research contributes to the DECOVALEX 2027 providing tools for robust numerical code evaluation.