Analysis of CO2 dispersion in environment, Gaussian vs CFD modelling

Melis G;Fonzo A;Mecozzi E;Demofonti G;Bertoli M;TOLA, VITTORIO;CAMBULI, FRANCESCO;CAU, GIORGIO

2011-01-01

---

Abstract

CO2 release and dispersion to the environment, as a consequence of leakage or fracture in onshore/offshore pipelines, is a matter of great concern. This interest is due to the expected future deployment of a pipeline network as essential element in Carbon Capture and Storage (CCS) projects or EOR applications. Hazards posed by accidental or intentional releases of CO2 from pipelines are often not easy due to a complex thermodynamics in the CO2 escaping related to phase change followed by cold heavy gas dispersion. As a consequence, a risk assessment based on the study of the atmospheric dispersion should be performed in the pipeline design. An introduction to the CO2 release and dispersion occurrence is provided as a well as a review of the numerical models for dispersion prediction in the event of CO2 pipeline failure. Indeed, different mathematical models can be used for the simulation, mainly based on Gaussian codes or Computational Fluid Dynamics (CFD). The main features of the two different numerical approaches have been reviewed focusing on their ability to simulate possible hazard scenarios of CO2 atmospheric dispersion and to account for different parameters, such as atmospheric conditions, orography, gas release mode, etc. The comparison between Gaussian and CFD models has allowed highlighting differences that impact directly on the accuracy of the results and affect the conservatism level of a numerically-based risk assessment. The commercial CFD code Ansys Fluent and a freely available Gaussian code, ALOHA, have been used for the simulation of dispersion caused by the release from supercritical CO2 pipelines. The impact on the dispersion to the environment of different sources releasing CO2 has been studied and the results obtained have been processed for evaluating CO2 concentration evolution with respect to time and space.