Performance comparison of dispersion and computational fluid dynamics models in the dispersion of gaseous stack emissions

Abstract

Estimation of air pollution is crucial especially for regulatory purposes. It is from one of the primary goals of air pollution scientists to make air quality estimations with high precision. In this study, the dispersion of SO2 from a coal-fired power plant was simulated over real terrain, using dispersion and computational fluid dynamic models that have different calculation approaches for the prediction of pollutant concentrations. This study aimed to improve the estimation capacity of computational fluid dynamic calculation by using 3D terrain in air quality studies. The numerical results from these two approaches were compared with the observed concentrations, and model performances were assessed. The Eulerian-based computational fluid dynamic model performed better than the Gaussian-based dispersion model with high R2 values of 0.99, and it solved the over-prediction problem of the dispersion model at low wind speeds. Although the Eulerian-based computational fluid dynamic model was computationally more demanding, it can be used as an effective method for assessing the dispersion of gaseous pollutants and evaluating local scale air pollution in complex urban terrain with the use of appropriate boundary conditions and actual geometry situations.