| dc.description.abstract | Organic-rich shales have been recognized as one of the most important energy resources in the world due to their ubiquitous presence. However, there are numerous engineering challenges serving as obstacles for exploiting these geo-materials with multiscale microstructure. This work addresses an important aspect of engineering challenges in understanding the complex behavior of organic-rich source rocks, namely their anisotropic poroelastic behavior at multiple scales.
To this end, we utilize a framework obtained by combining experimental characterization, physically-based modeling and uncertainty quantification that spans and integrates scales from nanoscale to macroscale. The multiscale models play a crucial role in predicting macroscale mechanical properties of organic-rich shales based on the available information on poromechanical properties in microscale. Recently a three-level multiscale model has been developed that spans from the nanometer length scale of organic-rich shales to the scale of macroscopic composite. This approach is powerful in capturing the homogenized/effective properties/behavior of these geomaterials. However, this model ignores the fluctuation/uncertainty in mechanical and compositional model parameters. As such the robustness and reliability of these estimates can be questioned in view of different sources of uncertainty, which in turn affect the requisite information based on which the models are constructed. In this research, we aim to develop a framework to systematically incorporate the main sources of uncertainty in modeling the multiscale behavior of organic-rich shales, and thus take the existing model one step forward. Particularly, we identify and model the uncertainty in main model parameters at each scale such as porosity and elastic properties. To that end, maximum entropy principle and random matrix theory are utilized to construct probabilistic descriptions of model parameters based on available information. Then, to propagate uncertainty across different scales the Monte Carlo simulation is carried out and consequently probabilistic descriptions of macro-scale properties are constructed. Furthermore, a global sensitivity analysis is carried out to characterize the contribution of each source of uncertainty on the overall response. Finally, methodological developments will be validated by both simulation and experimental test database. | en |
