Numerical study of near-field pollutant dispersion around a building complex emitted from a rooftop stack

The topic of environmental pollution is of special significance in the atmospheric boundary layer (ABL) especially in urban areas as it is one of the significant sources of poor indoor air quality due to contamination of fresh-air intakes. In city centres where external air pollution levels are relatively high, it is usually assumed that natural ventilation may not be able to provide adequate indoor air quality. Therefore mechanical ventilation and air-conditioning systems are thus being solicited to "clean" the incoming air (Kukadia and Palmer, 1998). There is evidence that such systems do not always provide clean fresh-air to the occupants of the building since several contaminants from nearby outside sources exist (e.g. vehicle exhaust, rooftop stack exhaust, wind-blown dust). Control of the pollutant sources and understanding the dispersion mechanisms, therefore, shall be considered as the first alternative to evaluate better these harmful phenomena.;This thesis focuses on dispersion and transportation of pollutant emissions from a building rooftop stack situated in the wake of a neighbouring tower using numerical simulation approach. The main objective of this work is to contribute to the "best-practice" of numerical modelling for dispersion studies. For that, wind tunnel tests as well as full-scale experiments are numerically reproduced to shed light on the uncertainties related to the complex dispersion phenomenon when using CFD simulations.;In the first study of this thesis, the behaviour of the flow and pollutant concentration fields around the two-building configuration are investigated by means of various k - epsilon turbulence models (i.e. standard, re-normalization group (RNG) and realizable k - epsilon models). The results show that the realizable k - epsilon model yields the best agreement with wind tunnel experimental data for lower stack height and smaller momentum ratio, while the RNG k - epsilon model performs best for taller stacks. Despite an overestimation of concentrations using the realizable k - epsilon model, it remains the only model that provides the correct trend of concentration distribution in the lower region between the two buildings. Based on this finding, the second study deals with the ability of CFD to simulate controlled (wind tunnel scale) and non-controlled (fullscale) environments using realizable k - epsilon model. This study details also the main steps for conducting consistent and reliable numerical simulations for dispersion studies. Additionally, CFD is shown to simulate better controlled environments than non-controlled environments.;The third study investigates the influence of two important parameters related to the pollutant exhaust source, i.e. stack height and pollutant exhaust velocity, on the concentration fields measured in the wind tunnel. The results show that increasing the stack height has an effect that is similar to increasing the pollutant exhaust velocity on the concentration distributions and that such effect depends upon the wall of the building under consideration. In addition,recommendations on fresh-air intake locations for the two buildings are provided. In the final study, an unsteady turbulence model (i.e. detached-eddy simulation) is tested to evaluate the flow-field and the dispersion field around the two-building configuration. The results show that the flow fluctuation capture is crucial to address better the dispersion in the wake of buildings. Consequently, the strengths of using an unsteady approach are compared to RANS methodology which provides however good results far from the exhaust source. The results of this extensive research support the use of an unsteady methodology in future works.