Summary

The prediction of pollutant dispersion from rooftop emissions in an urban environment is an extremely complex phenomenon particularly in the vicinity of a group of buildings. The plume behaviour depends on the wind characteristics but is also severely affected by the surroundings. This causes effluents released from stacks located on one of the buildings to re-enter the same or an adjacent building, causing potential health problems to its occupants. The optimal design and placement of exhaust stacks to limit this re-ingestion on air intakes and other sensitive locations can be a considerable challenge. Unfortunately, the state-of-the-art of dispersion modeling, particularly around a non-isolated building configuration, is not sufficiently advanced for accurate predictions in order to avoid such situations. Therefore, there is a need to develop a new model or modify an existing model to take into account the effects of dispersion of effluents and in particular focus on the impact of buildings that are in close proximity of the source of pollutants. To address this issue, a collaborative research program between Concordia University and IRSST was elaborated relying both on experimental and numerical modeling. The experimental findings have been published in a companion report  (Stathopoulos et al. 2014) while the current report focuses on the numerical modeling phase of the research.

The purpose of this study was to contribute to a better understanding of air pollution aerodynamics in urban areas by focusing on the most representative non-isolated building configurations: a building located upstream of an emitting building, a building located downstream of an emitting building and two buildings, one located upstream and the other one located downstream of an emitting building. All these cases were compared with a reference case: an isolated emitting building. The effect of adjacent buildings on the near field of a pollutant source was analysed in terms of dilution distribution on the roof of an emitting building.

The current research methodology uses Computational Fluid Dynamics (CFD) to study pollutant dispersion in the vicinity of a cluster of buildings. This tool provides detailed information on flow pattern and concentration (or dilution) fields by solving the flow equations in the entire computational domain. Numerical simulation reliability is one of the main concerns of this study; therefore, validation of results through comparisons with wind tunnel data collected during the experimental phase conducted at Concordia University is included in this report. Mesh quality, boundary conditions, turbulence model choice, wall treatment and numerical parameters are some of the elements that can be calibrated through comparison with experimental data.

To achieve our objective, two steps are suggested; the first step is to generate sufficient information regarding the setup of CFD simulations for flow and dispersion in urban areas. Special attention is paid to transport processes in order to build the best numerical model possible for such applications. The second step is a parametric analysis for diverse cases of pollutant dispersion in an urban area. The results are presented in terms of normalized dilution at roof level of an emitting building, but also as iso-contour planes of dilution field and stream lines showing the airflow pattern of all the configurations analysed. From the first step it was observed that in general steady CFD simulations tend to underestimate dilution when comparing with wind tunnel results. This underestimation is probably caused by the inherent incapability of RANS to capture unsteadiness of the flow. An adjustment in the value of turbulent Schmidt number (Sct) permits to obtain a better agreement with experiment data. In fact, reducing Sct permits to increase turbulent diffusion and then increase dilution of the pollutant. The parametric analysis (second step) produced valuable information about scalars and velocity fields as well as about vortical structures formed in the leeward side and between buildings. Knowing how these flow characteristics interact with the surroundings is essential to improve the understanding of pollutant dispersion within an urban area.