Abstract
In this study, we report comprehensive sensitivity analysis using artificial intelligence (AI) and machine learning techniques to determine the dominant operating and formation statistical parameters affecting the transient behavior of CO2 injection wellbore (i.e., wellbore pressure and temperature). This study focuses on the CO2 injection tests undertaken at the Field Research Station (FRS) injection site in Alberta, Canada. The CO2 wellbore transient behavior is simulated using a fully coupled nonisothermal transient multiphase wellbore-reservoir flow model (Jafari Raad et al., 2021). Radial Basis Function (RBF) Neutral Network and Polynomial Regressions are employed to characterize the dependency of the monitoring functions to the model parameters. The results show that the bottom-hole pressure (BHP) shows maximum sensitivity to the injection rate and minimum sensitivity to the injection pressure. The injection rate and the permeability of the target layer show the highest non-linearity effects. For a constant injection rate, the permeability of the target layer plays the most important controlling effect on the wellbore pressure. The results show that the injection pressure and permeability of the formation layer below the target formation represent the minimum non-linearity and interaction effects. This study provides an improved understanding of operating and reservoir parameters controlling the CO2 wellbore monitoring practices in CO2 injections.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have