Modelling reacting localized air pollution using Computational Fluid Dynamics (CFD)

A Computational Fluid Dynamics (CFD) approach is used to model reacting NO2 dispersion of vehicle pollutants released from a dual carriageway in Maidstone, UK. The simulations are carried out over the course of one full day during January, 2008.The developed CFD model utilizes a modified k-ɛ turbulence model and Arrhenius reaction kinetics with source terms for the reactions which include a photo-stationary set with peroxy radicals. An approach is taken whereby the reactions are solved specific to the rush hour period corresponding to the availability of certain hydrocarbons released from the vehicles.The results of the simulation are compared with field measurements taken at the site which is made up of several, different sized buildings on varying terrain in Maidstone UK. The predictions and field measurements are considered over a 12 h period with averaged hourly results.It was found that the reactive pollutant approach greatly improves the predictions as compared to the experiments. Furthermore the effect of peroxy radicals during rush hour periods is found to be a major disturbance to the photo-stationary set and its inclusion improved the predictions further.

Research highlights
► Comparisons of CFD and experimental pollutant concentrations.
► Reactive localized pollutants included in the CFD study.
► Photo-stationary reactions insufficient to predict NO2.
► Results significantly improved by modelling Peroxy Radicals.