Caprock integrity evaluation for geosequestration of CO2 in depleted petroleum reservoirs.

Aminaho, Efenwengbe Nicholas

Abstract

The geological storage of carbon dioxide (CO2), also referred to as CO2 geosequestration, represents one of the most promising options for reducing greenhouse gases in the atmosphere. However, sometimes CO2 is captured with small amounts of other industrial gases such as sulphur dioxide (SO2) and hydrogen sulphide (H2S), incurring extra costs to separate these other acid gases before the CO2 can be stored in depleted petroleum reservoirs or aquifers - or, it might be necessary to inject CO2 with a small amount of these gas impurities to save cost. Moreover, during CO2 geosequestration in reservoirs, pressure variations during injection and storage could force some amount of CO2 into the caprock, thereby altering the petrophysical and geochemical properties of the caprock. Also, CO2-brine-rock interactions during CO2 geosequestration can impact the brittleness of the formations due to changes in their mineralogical compositions during the geochemical reactions. Therefore, studies on the co-injection of CO2 with other acid gases from industrial emissions and their impact on caprock integrity are paramount. In this study, numerical simulations were performed using TOUGHREACT codes, to investigate the injection of pure CO2 and CO2 co-injection with SO2 or H2S into carbonate and sandstone formations, and their migration to shale caprock. Mineralogical brittleness models were derived from existing models, and one of the models that accounts for the relative level of brittleness of minerals was applied to evaluate the mineralogical brittleness index of the rocks. In addition, a machine learning approach to predicting the brittleness index of rocks was adopted and an artificial neural network (ANN) model was developed to evaluate the brittleness index of rocks using data from the numerical simulations of CO2 geosequestration in sandstone and carbonate reservoirs, overlain by shale caprock. The ANN model was developed using Python programming language. The findings of the study revealed that SO2 gas dissolves faster in brine compared to H2S gas when co-injected with CO2 gas in reservoir rocks. Thus, the region in the reservoir contacted by SO2 gas is smaller compared to H2S gas. SO2 gas dissolves more in sandstone reservoirs compared to carbonate reservoirs, due to the availability of more Fe-bearing minerals in the sandstone formation. The findings of the study also indicate that porosity and permeability increase for the CO2 alone and CO2-H2S injection cases, in both the carbonate and shale rocks; while for the CO2-SO2 injection case, porosity and permeability increase in the shale rock and carbonate rock (initially composed of calcite and dolomite) and decrease in the carbonate (pure and impure limestone) and sandstone rocks, due to anhydrite precipitation from the injection zone to the reservoir-caprock interface. During cyclic injection and withdrawal of CO2, for all the injection cases, the porosity and permeability of the sandstone reservoir decreased in a few layers directly below the perforation interval of the production zone. In all the sequestration cases and techniques, the brittleness of the shale and sandstone rocks decreases, while the change in the brittleness of the carbonate rocks varies depending on calcite precipitation or dissolution. The brittleness of the carbonate rock (initially made up of calcite and dolomite minerals) decreased only in a small region in the lower part of the reservoir. Therefore, carbonate reservoirs with similar mineralogy may be suitable for SO2 gas co-injection with CO2. Based on the mineralogical composition of the formations in this study, the injection of pure CO2 or CO2 co-injection with H2S or SO2 decreased the brittleness index of the clay-rich shale caprock slightly. The brittleness index of carbonate-rich shale caprock might increase during pure CO2 geosequestration. Therefore, a proper understanding of the mineralogical composition of formations before CO2 geosequestration is vital. The ANN model developed in this study predicted the brittleness index of rocks with R2 value greater than 99% and mean absolute percentage error (MAPE) of less than 0.7%. Hence, the ANN model predicts the brittleness index of the formations accurately.

Citation

AMINAHO, E.N. 2024. Caprock integrity evaluation for geosequestration of CO2 in depleted petroleum reservoirs. Robert Gordon University, PhD thesis. Hosted on OpenAIR [online]. Available from: https://doi.org/10.48526/rgu-wt-2445726

• Thesis Type: Thesis
• Deposit Date: Aug 26, 2024
• Publicly Available Date: Aug 26, 2024
• DOI: https://doi.org/10.48526/rgu-wt-2445726
• Keywords: Carbon capture; Geosequestration; Petrology; Petroleum reservoirs; Machine learning
• Public URL: https://rgu-repository.worktribe.com/output/2445726
• Award Date: May 31, 2024