P/Z for CO2 Injection into Depleted Gas Reservoirs

Abstract

Abstract Pressure data can be used to identify possible leaks in a CO2 storage site. However, conventional diagnostic plots used during gas production cannot directly be used to interpret the data, because CO2 properties have a different pressure dependency compared to the hydrocarbon gas initially in place. This paper investigates the possibility of adjusting p/Z plots for CO2 storage into depleted gas reservoirs to make them suitable for monitoring during injection. The p/Z method is explained analytically for both gas production and CO2 injection. A method is proposed to graphically ensure that the p/Z plot during injection overlies the p/Z plot during depletion – the injection line follows the p/Z plot backwards. The method proposes to use a volume-weighted Z-factor to scale both p/Z and injected volume. This method is verified with compositional thermal simulations in a homogeneous model. The uncertainty of both pressure and Z-factor is investigated and the implications of these uncertainties are discussed. Additionally, the proposed method is tested for reservoirs that do not show volumetric depletion behavior. For a volumetric gas reservoir (straight-line p/Z) without near-wellbore cooling it is possible to follow the original p/Z plot backwards with the proposed method. Cooling of the near-wellbore area causes a temperature induced increase of density. The impact of this phenomenon is less than 4% on the p/Z estimate in the described case. With an analytical estimate of the extent of the cooling area, the p/Z value can be corrected to an error of less than 1.5%. Uncertainty in Z for natural gas and CO2 at reservoir conditions can be significant, but generally result in a low impact on the accuracy of the method. The uncertainty in static pressure is larger with CO2 injection due to the higher fluid viscosity and more uncertain downhole density. For reservoirs with ‘slow gas’ behavior, the proposed method does not result in an overlap with the depletion p/Z plot. However, the method does reproduce the path that methane injection follows. The proposed methodology indicates to what extent the storage is filled and has the potential to identify large leakage of CO2 (>1% of GIIP) for volumetric gas reservoirs. To verify the observed potential leak, it is recommended to perform full field simulation.