Numerical study of supercritical carbon dioxide accidental releases in partially confined spaces

The objective of this thesis is to support a preliminary project for the reconversion of decommissioned offshore platforms by conducting Computational Fluid Dynamic (CFD) simulations for risk analysis. The focus of the project is to repurpose an oil platform into a carbon dioxide injection station for storage purposes, using Carbon Capture and Storage (CCS) technologies. However, due to the changes in platform layout and the presence of another hazardous substance, the previously conducted risk analyses for hydrocarbon extraction are no longer applicable. The primary focus of this thesis is to perform numerical modeling simulations of potential incidental scenarios related to the CO2 release that can occur on board the platform. These simulations are fundamental for risk analysis. Commercial models do not accurately reproduce these phenomena in the presence of obstacles (confined spaces), leading to inaccurate predictions of CO2 releases that can occur in gaseous, liquid, or supercritical phases. The research object of this study has allowed on the one hand to investigate the numerical models, available in the literature, as well as the existing experiments; the scope is to implement CO2 release simulations within confined spaces into SBAM, a Computational Fluid Dynamics (CFD) model specifically designed for analyzing incidental scenarios involving high-pressure fluid leakages in confined spaces, extending its range of applicability. This study has revealed that the application of the ideal gas equation of state (EOS) in the case of supercritical CO2 release simulations leads to inaccurate predictions of the results. Consequently, a real gas model is required to provide a more accurate estimation of the behavior of this substance. However, the implementation of real gas models in the ANSYS Fluent software posed several challenges. Various attempts to incorporate this real gas equation of state are discussed in this thesis to highlight the encountered issues.