| dc.description.abstract | Geological CO2 storage is an important initiative to mitigate climate change. Numerical models are valuable for estimating storage capacity and for risk assessment. Due to large scales involved, simplified models are desired. Vertical equilibrium (VE) models have shown orders-of-magnitude reduced runtime, but are challenged by formation heterogeneities. Herein, we extend a fully-implicit hybrid model developed in MRST with residual saturation and capillary exclusion. We relax the VE assumption (RVE) to model diffuse leakage of CO2 through sealing barriers. For comparisons, we extend a conventional trap analysis to be compatible with hybrid models. Our extended hybrid model agrees well with full-dimensional models, with runtimes being shorter. From the RVE assumption, it manages to trace the history of individual plumes spread among sealing barriers, but misses lateral spreading of CO2 beyond these. Stochastic simulations indicate that differences in depth of the CO2 plume between the models decrease for an increased number of sealing barriers, while differences in net CO2 trapping show an increasing trend for more sealing barriers. As a second purpose of the thesis, we develop a novel, semi-structured grid in MRST that conforms precisely to the facies of the FluidFlower rig, as part of the 11th SPE Comparative Solution Project. The grid is also compatible with hybrid modeling. Results from a full-dimensional compositional model correspond well with physical measurements, particularly adjacent to anticlines and faults, but also show grid-orientation effects that are subject to improvement. The hybrid model shows promising results on the 11th SPE CSP, despite convergence issues. Whether the FluidFlower rig is a good candidate for hybrid modeling remains an open question. The purpose herein is to demonstrate the versatility of our hybrid model for simulations on heterogeneous formations. We believe it is a promising candidate for history-matching and greater insight into numerical challenges for CO2 storage simulation in general, and for the 11th SPE CSP in particular. | eng |