Faculty of Civil Engineering and Geosciences Department of Geosciences and Engineering
Master of ScienceHierarchical coarsening of simulation model for in-situ upgrading process by Raul Fucinos M.Oil shales are sedimentary rocks containing organic matter in the form of kerogen which accounts for more than 5 trillion barrels of oil in place according to Birol, 2010; therefore, oil shales represents a plausible solution for the constantly increasing demand for hydrocarbons. Oil shale production is currently done using two different techniques, surface retorting and in-situ retorting. The last one being the focus of this study. During this process, the sedimentary rock containing the kerogen is brought into a high-temperature environment with oxygen deficit. At this stage, the organic matter is subject to a thermo-chemical decomposition that finally releases the hydrocarbon in liquid and gas forms. This process is also known as pyrolysis. During this process, solid and fluid components experience compositional and physical changes, which requires complex chemical models represented by multiple species and several governing relations.In this work, we first developed a numerical solver for closed systems with simple kinetics models. This initial work allowed us to analyze the dynamic behavior of each component during the chemical decomposition of the kerogen and its impact in the porosity of the system. Then, we described an accurate base model for chemical decomposition of kerogen. This model was then implemented in our inhouse simulator ADGPRS. The model is based on the most recent understanding of pyrolysis process, and it incorporates coupling of chemical kinetics to heat and mass transport. Due to the high number of species, variations of porosity as consequence of the transformation of solid species into fluid products and complex multi-scale structure of porous media, the simulation performance of the high-fidelity model is limited. Therefore, in the second step of this work, we introduce a hierarchy of coarser models to improve the run-time of forward solution without significant reduction in accuracy. We applied coarsening in time, space, and chemical representation, and quantify errors introduced at each coarsening level. In conclusion, we provided recommendations for large-scale modeling of in-situ upgrading process. v