Predictive capability of coupled rock behaviour – development of an experimentally based benchmark for numerical quality assurance (BeNuQuA)
Public PublishedSimulation tasks and numerical codes have become increasingly complex in the context of temporal and spatial upscaling for understanding and predicting subsurface rock mass mechanics, for example associated with radioactive waste disposal. Benchmark procedures based on laboratory datasets are required to verify the predictive capabilities of computational approaches. In this study, a high-quality laboratory dataset of conventional geomechanical experiments on granite was generated. Numerical simulations of the experiments were performed as proof-of-concept using COMSOL Multiphysics and RocScience RS to derive a benchmark procedure for numerical quality assurance that can be applied to computational approaches in the context of nuclear waste disposal and beyond. The laboratory schedule was specifically developed for the numerical simulation of time-dependent deformation characteristics in granite. Splitting tensile strength, uniaxial compressive strength, and triaxial compressive strength of carefully characterised specimens were investigated at different strain rates covering five orders of magnitude. Strength decreased with decreasing strain rate, and the presence of water decreased strength significantly. A significant contribution of end face friction in the experimental setup to the results of strength tests was verified and recommendations for preparational and experimental procedures in deformation experiments on granite were derived. Based on the laboratory dataset, 2D numerical simulations with RocScience RS2 successfully reproduced the effect of different lubricants to modify end face friction on strength, and COMSOL Multiphysics was able to reproduce the time-dependent deformation characteristics observed for granite. Using crack phase field damage modelling, COMSOL Multiphysics predicted triaxial compressive strength from uniaxial compressive strength by adjusting nothing but the boundary conditions. In both approaches, the adaptation of microstructural properties was required to successfully simulate the experimental findings pointing to a distinct need to further improve the understanding of microstructural processes causing the time-dependent deformation characteristics and to evaluate the potential for temporal upscaling to long-term processes exceeding those covered by laboratory experiments. The results of this study will significantly contribute to gaining more confidence in the predictive capabilities of numerical codes and identify code-specific parameters that are critical for successful prediction.
When referencing this report, please use the following citation: Witte L.C., Asghari Chehreh H., Backers T., Duda M., Aydin M. & Parvin S., 2024. Predictive capability of coupled rock behaviour – development of an experimentally based benchmark for numerical quality assurance (BeNuQuA). ReSeeD Research Data Repository, Ruhr University Bochum.
- Title
- Predictive capability of coupled rock behaviour – development of an experimentally based benchmark for numerical quality assurance (BeNuQuA)
- Creator or Contributor
- Resource type
- Keyword
- License
- Alternative title
- Witte L.C., Asghari Chehreh H., Backers T., Duda M., Aydin M. & Parvin S., 2024. Predictive capability of coupled rock behaviour – development of an experimentally based benchmark for numerical quality assurance (BeNuQuA). ReSeeD Research Data Repository, Ruhr University Bochum.
- Description
- Simulation tasks and numerical codes have become increasingly complex in the context of temporal and spatial upscaling for understanding and predicting subsurface rock mass mechanics, for example associated with radioactive waste disposal. Benchmark procedures based on laboratory datasets are required to verify the predictive capabilities of computational approaches. In this study, a high-quality laboratory dataset of conventional geomechanical experiments on granite was generated. Numerical simulations of the experiments were performed as proof-of-concept using COMSOL Multiphysics and RocScience RS to derive a benchmark procedure for numerical quality assurance that can be applied to computational approaches in the context of nuclear waste disposal and beyond. The laboratory schedule was specifically developed for the numerical simulation of time-dependent deformation characteristics in granite. Splitting tensile strength, uniaxial compressive strength, and triaxial compressive strength of carefully characterised specimens were investigated at different strain rates covering five orders of magnitude. Strength decreased with decreasing strain rate, and the presence of water decreased strength significantly. A significant contribution of end face friction in the experimental setup to the results of strength tests was verified and recommendations for preparational and experimental procedures in deformation experiments on granite were derived. Based on the laboratory dataset, 2D numerical simulations with RocScience RS2 successfully reproduced the effect of different lubricants to modify end face friction on strength, and COMSOL Multiphysics was able to reproduce the time-dependent deformation characteristics observed for granite. Using crack phase field damage modelling, COMSOL Multiphysics predicted triaxial compressive strength from uniaxial compressive strength by adjusting nothing but the boundary conditions. In both approaches, the adaptation of microstructural properties was required to successfully simulate the experimental findings pointing to a distinct need to further improve the understanding of microstructural processes causing the time-dependent deformation characteristics and to evaluate the potential for temporal upscaling to long-term processes exceeding those covered by laboratory experiments. The results of this study will significantly contribute to gaining more confidence in the predictive capabilities of numerical codes and identify code-specific parameters that are critical for successful prediction.
- When referencing this report, please use the following citation: Witte L.C., Asghari Chehreh H., Backers T., Duda M., Aydin M. & Parvin S., 2024. Predictive capability of coupled rock behaviour – development of an experimentally based benchmark for numerical quality assurance (BeNuQuA). ReSeeD Research Data Repository, Ruhr University Bochum.
- Date
- Subject
- Language
- Software version
- Rocscience RS2 11.018
- Comsol Multiphysics 6.2
- Related item
- Identifier
- Funding reference
- Publisher
- Last modified
- 2025-06-05
Relations
Items
| Thumbnail | Title | Date Uploaded | Visibility | Actions |
|---|---|---|---|---|
|
241010_Benuqua_Final_Report.pdf | 2025-06-02 | Public | Download |