Optimization of urban building patterns for pollution removal efficiency by assuming periodic dispersion

Air pollution is often a major problem in densely built cities, mainly due to the urban canopy consisting of high-rise buildings and other obstructions that hinders the transport of traffic-produced pollution. Ventilation characteristics principally depend on geometrical parameters such as the variations in building height and building density. The aim of this paper is to introduce an effective method for finding optimum building patterns via CFD modelling. In these models the urban canopy is approximated by an infinite 2D lattice of a given elementary building pattern, which also incorporates pollution sources. The surface mass transfer coefficients are computed for various wind directions and the non-dimensional form of the average value is used for characterizing ventilation intensity. The model does not require input data for boundary conditions, therefore the resulting ventilation intensity depends only on the geometrical configuration. Resultant velocity and turbulence profiles are found to be in close agreement with earlier models and wind tunnel measurements. The methodology of optimization is demonstrated on simple geometrical configurations and further results are shown for a more complex building arrangement characteristic of Central Europe. The model can be further developed to non-isothermal conditions, and supports the application of the Large Eddy Simulation.