Bedrikovetski, PavelZeinijahromi, AbbasBorazjani, SaraShokrollahi, Amin2025-07-042025-07-042025https://hdl.handle.net/2440/145686The urgent need to decarbonise energy and heavy industries is driving research into underground carbon dioxide storage. These technologies play a crucial role in reducing emissions and facilitating the use of sustainable alternatives. Deep saline aquifers and depleted hydrocarbon reservoirs possess the capacity and containment efficiency required to store significant volumes of gas. However, challenges during injection and storage include: (i) a decline in well injectivity—defined as the ability to inject a specified volume per unit of time—due to fines migration and other formation damage mechanisms, and (ii) poor sweep efficiency, which refers to the proportion of the porous volume filled by the injected gas during one pore volume injection. In highly heterogeneous formations, this can result in limited storage capacity. Poor injectivity and poor sweep efficiency can constrain storage capacity and increase operational costs, thereby hindering the widespread deployment of storage projects globally. High uncertainty in geological information during underground gas storage necessitates fast predictive modelling, which facilitates risk analysis and multivariate simulation runs to account for these uncertainties. A promising approach for rapid predictive modelling that incorporates key physical mechanisms is analytical modelling based on fundamental flow equations. Despite the significant advancement of analytical modelling in oil and gas production, its application to geological gas storage remains limited. Moreover, several critical aspects of two-phase gas-water flows in natural reservoirs have yet to be developed within the context of analytical modelling. These include the rate-dependency of phase permeabilities at both the core and reservoir scales, the non-Newtonian properties of high-rate gas flows (the inertial Forchheimer effect), and heterogeneity along flow paths in streamline simulators. This thesis addresses these gaps. Fines particles within pore spaces can be mobilised during gas injection, leading to clogged pore throats and reduced injectivity. While fines mobilisation during hydrocarbon production is controlled by drag forces, during gas injection, it is primarily governed by capillary forces, which are influenced by aqueous-phase saturation that decreases due to displacement and evaporation. Despite extensive research on formation damage control in hydrocarbon production, there remains a gap in analytical modelling of formation damage during underground gas storage in heterogeneous reservoirs. This gap is also addressed in the present thesis. The aforementioned analytical models for underground gas storage in heterogeneous reservoirs require fluid property inputs. A promising methodology for obtaining these data is advanced computational data-driven modelling approach. In the present thesis, we employ these methods to predict the necessary properties of gas-brine/hydrocarbon systems for modelling purposes.enUnderground Gas StorageAquifersFormation DamageEnvironmentColloidal transportExact solutionFractional flow theoryThesis