A comparison of turbulent CFD with Gaussian dispersion models on a methane emission test site

Abstract

This article investigates the influence of structures on the atmospheric transport of methane in an outdoor industrial environment, and provides a comparison in the predictive capability of Gaussian dispersion models against a turbulent computational fluid dynamics (CFD) model, implemented in OpenFOAM. Direct atmospheric measurements from the Methane Emissions Technology Evaluation Center (METEC), as well as a detailed computational mesh of on-site structures, are used to calibrate the turbulent closure model. By comparing the CFD model with and without the computational mesh of structures, it is shown that structures on METEC exhibit only a small effect on concentrations predicted by the CFD model. The calibrated CFD model is then used to assess the fidelity of the commonly employed Gaussian puff model that ignores the effect of any structures or topography. Despite the presence of structures, the Gaussian puff model is in consistent agreement with predictions from the CFD model, however both models fail to capture certain trends in the measurement data. To show that one cannot conclude, in general, that methane transport is unaffected by structures, a simulation study is performed with an emission source placed upwind of a large structure. In this case, it is found that predictions from the Gaussian puff model can deviate significantly from the CFD model due the Gaussian model’s inability to resolve spatially inhomogeneous wind fields caused by structures. This finding highlights the continued importance of CFD modeling for evaluating atmospheric dispersion models in environments with complex structures and topography.