Simulation Of Pollutant Dispersion Around Buildings Group

This paper is directed towards a deeper understanding of pollutant dispersion with respect to various simple building configurations by using the commercial CFD package-FLUENT.Taking an isolated street canyon (open country roughness and urban country roughness) as an example, a two-dimensional numerical model is developed to study the three basic problems in the simulation of the flow and pollutant dispersion within buildings. The three basic problems are the determination of suitable computational domain, the optimal k-εturbulence model and the choice of structure grid and unstructured grid. The RNG k-εturbulence model is found to be the optimal turbulence model in this kind of approach.A two-dimensional numerical model with RNG k-εturbulence model is developed to simulate the wind flow and pollution concentration around single building and parallel building column configurations. It is found out that the pollutant diffusion is determined by the wind speed and local turbulent diffusivity. By arranging the position of buildings, the area of high pollutant concentration can be reduced efficiently. An optimal numerical simulation strategy based on the two-dimensional model is proposed to replace the three-dimensional numerical model as a prepare approach in some conditions. After the two-dimensional model obtains the optimal plan, the three-dimensional model is used to check up the above optimal plan of two-dimensional results and determines it finally.A two-dimensional numerical model considering line pollutant source is adopted to simulate the pollutant diffusion within three positions of nine buildings configuration. The optimized building configurations for pollutant diffusion are found in this research.The population dynamic model of virus infection is introduced in the air quality model as an unsteadily pollutant source for the first time. The virus dispersion and possible infection area are obtained. A virus alarm system is established. An inverse mathematical method and graphics recognition technology are proposed to determine the parameters of non-linear population dynamic model with known experimental data.