Monitoring Injectivity for CO2 Injection in Depleted Gas Reservoirs

Abstract

Abstract Determining the injectivity of CO2 injection wells in depleted gas reservoirs is important, because injectivity directly impacts the operating envelope of the injector. Injectivity evolves over time due to changing fluid properties, but it can also change by processes like hydrate formation or fracturing (summarized in the skin parameter). This paper investigates how accurate injectivity monitoring is, based on coupled well-reservoir simulation and identifies threshold values for identifying injectivity loss or gain. The relation between bottomhole pressure (BHP) and rate is a function of several reservoir properties and fluid properties, each carrying its own uncertainty. This paper quantifies the impact of these uncertainties on injectivity for CO2 injection into depleted gas reservoirs using a coupled wellbore and reservoir simulator. The analysis is performed for three scenarios with varying pipeline conditions and reservoir properties, representative for CCS projects in the Netherlands. The resulting BHP uncertainty is translated to equivalent skin values. This is the threshold value above or below which the observed bottomhole pressure data conclusively suggests injectivity gain or loss. A good estimate of the fluid properties is important to accurately determine injectivity (changes). Although the uncertainty in density is small when using the appropriate Equation of State (like Span & Wagner for pure CO2), the deviation when using a standard Equation of State (like Peng-Robinson) is significant. The parameter having the largest impact is reservoir pressure for high injectivity reservoirs and relative permeability for low injectivity reservoirs. The combined effect of all parameters (with the exception of reservoir pressure) is equivalent to a skin value of approximately 3 for the selected cases. Without reducing the uncertainty, it is not possible to distinguish between uncertainty in injectivity and actual wellbore damage of that magnitude. There are ways to decrease the uncertainty in both estimated average reservoir pressure and fluid properties, like the collection of bottomhole temperature data. Shut-in pressure data will further reduce the average reservoir pressure estimate. However, the presence of skin will be difficult to assess from pressure transient data due to wellbore storage and near-wellbore cooling. Rate dependent skin can be assessed when there is rate variation. Permeability and relative permeability changes in the near-wellbore will be difficult to distinguish from wellbore damage. History matching the observed data by a coupled wellbore-reservoir simulator will further decrease the uncertainties. Monitoring tools for CO2 injection in depleted reservoirs are yet to be developed. This paper improves the understanding of the uncertainties related to injectivity monitoring, which haven’t been quantified so far. Also, the value of a coupled wellbore and reservoir simulator is demonstrated in this paper.