Abstract
Air pollution is a serious problem. Flemish governmental policies have been put into action to lower pollutant concentrations. Planned measures are e.g. lowering traffic emissions and innovative building configurations to enhance natural ventilation in urban environments. Since these measures are obviously very costly and the health cost of air pollution is enormous (estimated as € 8 billion per year in Belgium), assessing the effectiveness of the planned measures will result in an efficient allocation of our society's financial resources to lower these substantial
health costs. Mathematical modelling with computational fluid dynamics (CFD), allows quantifying urban pollutant concentrations and the effect of the proposed abatement strategies. However, concerns about the accuracy and computational cost of urban pollutant dispersion CFD models exist. To solve these problems, the following will be investigated: Combining different faster but less accurate Reynolds-averaged Navier Stokes (RANS) models into 1 model could increase the overall RANS performance. This strategy will be combined with the more accurate but slower large Eddy simulation (LES) models in a hybrid RANS/LES model, to speed up LES. In addition, a recently
developed uncertainty quantification method will be applied to identify as yet unknown relevant sources of uncertainty. Lastly, the developed methods and knowledge will be incorporated in a model of a real quarter in a Flemish city (Antwerp).
Researcher(s)
Research team(s)
Project type(s)