Abstract
Fine-grained sedimentary rocks, such as shales and mudstones, are crucial in various geoenergy applications, especially as geological barriers for underground storage of energy carriers (e.g., H2, CH4) and climate-relevant gases (in particular CO2). This thesis presents advancements in the characterization of these rocks, focusing on their pore structure, micromechanical properties, and compaction processes at microscale, using novel tools and approaches including high-speed nanoindentation mapping and machine learning-based image processing and data analyses.
The scientific work presented in this thesis is centered around three publications. Publication I assesses the impact of image processing methods on pore characterization in mudstones by comparing conventional thresholding with machine learning-based pixel classification, revealing comparable results between thresholding and fast pixel classification methods. However, the thresholding workflow results show significant variability in porosity estimations (a relative 38% to 62% data variability), with different threshold values influencing the resolved pore structure by connecting adjacent pores. The results highlight the sensitivity of the obtained porosity and pore structural data to the image processing method, suggesting the need for careful selection of image processing techniques to improve reproducibility and reliability. Publication II introduces a novel micromechanical profiling approach for mudstones that was developed using high-speed nanoindentation mapping combined with machine learning data analysis, enabling extraction of representative micromechanical parameters for the clay matrix. Despite the inherent geologic heterogeneity, the machine learning-based k-means clustering successfully classified the mechanical characteristics of the clay matrix, transitional phases (e.g., measurements on grain boundaries and structural discontinuities), and brittle minerals. High-resolution mechanical property maps were obtained for the clay matrix, showing average reduced elastic modulus (Er) and hardness (H) values of 16.2 ± 6.2 GPa and 0.5 ± 0.5 GPa, respectively. This approach facilitates a detailed understanding for the mechanical behavior of clay-rich matrix in mudstones under different experimental settings (e.g., indentation load and tip), demonstrating the feasibility of nanoindentation mapping combined with advanced data analysis for micromechanical characterization of complex geomaterials. Publication III covers investigations on compaction processes in mudstones from the Vienna Basin across varying depths (723.5 to 3213.5 meters) using the approach developed in Publication II, bridging a critical gap in understanding compaction processes at microscale. Compaction trends along the burial depth were obtained, with decreasing porosity (from 31 to 5 vol% for helium pycnometry) and enhanced micromechanical properties (Er and H increased from 6.8 ± 3.4 to 22.6 ± 7.5 GPa and from 0.2 ± 0.2 to 0.9 ± 0.2 GPa, respectively). A correlation coefficient matrix linking micromechanical properties, multiscale porosity measurements and mineralogical data was applied to identify the main factors controlling clay matrix compaction. It was found that mechanical compaction dominated over chemical diagenesis with respect to the micromechanical changes during burial diagenesis and resulting porosity reduction. Mechanical properties as a function of depth and porosity were described by empirical mathematical equations, which can serve as reference and input for future basin analysis and geomechanical modelling.
Collectively, this thesis establishes refined methodologies for the advanced characterization of pore structural and mechanical properties of mudstones. The established workflows and derived models are highly applicable to seal rock integrity assessment, and have great potential for supporting site screening and evaluation in underground storage applications regarding the quality of geological barriers.
The scientific work presented in this thesis is centered around three publications. Publication I assesses the impact of image processing methods on pore characterization in mudstones by comparing conventional thresholding with machine learning-based pixel classification, revealing comparable results between thresholding and fast pixel classification methods. However, the thresholding workflow results show significant variability in porosity estimations (a relative 38% to 62% data variability), with different threshold values influencing the resolved pore structure by connecting adjacent pores. The results highlight the sensitivity of the obtained porosity and pore structural data to the image processing method, suggesting the need for careful selection of image processing techniques to improve reproducibility and reliability. Publication II introduces a novel micromechanical profiling approach for mudstones that was developed using high-speed nanoindentation mapping combined with machine learning data analysis, enabling extraction of representative micromechanical parameters for the clay matrix. Despite the inherent geologic heterogeneity, the machine learning-based k-means clustering successfully classified the mechanical characteristics of the clay matrix, transitional phases (e.g., measurements on grain boundaries and structural discontinuities), and brittle minerals. High-resolution mechanical property maps were obtained for the clay matrix, showing average reduced elastic modulus (Er) and hardness (H) values of 16.2 ± 6.2 GPa and 0.5 ± 0.5 GPa, respectively. This approach facilitates a detailed understanding for the mechanical behavior of clay-rich matrix in mudstones under different experimental settings (e.g., indentation load and tip), demonstrating the feasibility of nanoindentation mapping combined with advanced data analysis for micromechanical characterization of complex geomaterials. Publication III covers investigations on compaction processes in mudstones from the Vienna Basin across varying depths (723.5 to 3213.5 meters) using the approach developed in Publication II, bridging a critical gap in understanding compaction processes at microscale. Compaction trends along the burial depth were obtained, with decreasing porosity (from 31 to 5 vol% for helium pycnometry) and enhanced micromechanical properties (Er and H increased from 6.8 ± 3.4 to 22.6 ± 7.5 GPa and from 0.2 ± 0.2 to 0.9 ± 0.2 GPa, respectively). A correlation coefficient matrix linking micromechanical properties, multiscale porosity measurements and mineralogical data was applied to identify the main factors controlling clay matrix compaction. It was found that mechanical compaction dominated over chemical diagenesis with respect to the micromechanical changes during burial diagenesis and resulting porosity reduction. Mechanical properties as a function of depth and porosity were described by empirical mathematical equations, which can serve as reference and input for future basin analysis and geomechanical modelling.
Collectively, this thesis establishes refined methodologies for the advanced characterization of pore structural and mechanical properties of mudstones. The established workflows and derived models are highly applicable to seal rock integrity assessment, and have great potential for supporting site screening and evaluation in underground storage applications regarding the quality of geological barriers.
| Translated title of the contribution | Fortgeschrittene bildbasierte und mikromechanische Charakterisierung feinkörniger Sedimentgesteine: Innovative Methoden zur Untersuchung geologischer Barrieregesteine |
|---|---|
| Original language | English |
| Qualification | Dr.mont. |
| Awarding Institution |
|
| Supervisors/Advisors |
|
| DOIs | |
| Publication status | Published - 2025 |
Bibliographical note
no embargoKeywords
- Shale
- mudstone
- geoenergy
- BIB-SEM
- nanoindentation
- machine learning
- image processing
- pore size distributions
- bootstrapping
- Vienna Basin
- seal rocks
- mudstone compaction
Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver