PNNL

Abstract

Numerical models are essential tools for CO2 sequestration projects and should be included in the life cycle of a project. Common practice involves modeling the behavior of CO2 during and after injection using site-specific reservoir and caprock properties. Little has been done to systematically evaluate and compare the effects of a broad but realistic range of reservoir and caprock properties on potential CO2 leakage through caprock. Broad-based research addressing the impacts of caprock properties and their heterogeneity on seal permeation is absent. Efforts along this direction require obtaining information about the physically reasonable range of caprock and reservoir properties, effectively sampling the parameter space to fully explore the range of these properties, and performing flow and transport calculations using reliable numerical simulators. In this study, we identify the most important factors affecting CO2 leakage through intact caprock and try to understand the underlying mechanisms. We use caprock and reservoir properties from various field sites and literature data to identify the range of caprock thickness, permeability, and porosity that might occur. We use a quasi Monte Carlo sampling approach to ensure that the full range of caprock and seal properties is evaluated without bias. For each set of sampled properties, the migration of injected CO2 is simulated for up to 200 years using the water-salt-CO2 operational mode of the STOMP simulator. Preliminary results show that critical factors determining CO2 leakage rate through intact caprock are, in decreasing order of significance, the caprock thickness, caprock permeability, reservoir permeability, caprock porosity, and reservoir porosity. This study provides a function for prediction of potential CO2 leakage risk due to permeation of intact caprock, and identifies a range of acceptable seal thicknesses and permeability for sequestration projects. As a byproduct, the dependence of CO2 injectivity on reservoir properties is also evaluated.