Supervisors: Helge K. Dahle
Short descirption of the thesis:
This thesis investigates how vertical grid resolution and permeability averaging affect water injectivity in reservoir simulations. Discrepancies between predicted and actual injection rates in early injection stages often result from simplifications in reservoir models, such as the use of coarse grids. Based on the principle that fluids follow the path of least resistance, an analytical expression is derived relating injectivity to resistance. Simulations using Eclipse 100 are conducted for single-phase water injection into homogeneous and heterogeneous reservoirs, while varying vertical grid resolutions and permeability distributions. Injectivity is given by the pressure difference between an injector and a producer at steady state.
Results indicate that coarse grid resolution overestimates resistance, while averaging permeability underestimates it. The latter has a greater impact, suggesting both simplifications may contribute to overestimated injectivity. These effects are interpreted using the relation linking injectivity to resistance, showing that changes in injectivity when upscaling are likely due to variations in resistance arising from differences in flow path distribution.
This thesis also explores how reduced injectivity affects power demand. Calculations across a range of skin factors, injectivity indices, and flow rates show that power demand increases significantly with high skin factors and low injectivity indices, particularly at high injection rates.
These findings improve the understanding of how modeling assumptions influence injectivity estimates and may ultimately influence decision-making in injection projects. Furthermore, the greatest potential for energy savings occurs in cases of low injectivity and high injection rates, offering useful input for improving the energy efficiency of injection operations.