The complex fluid dynamics, microphysics and chemistry governing the transport and transformation of pollutants emitted directly in the urban canopy, necessitate the adoption of sophisticated numerical methods for studying air quality in urban areas. Since its foundation in 1990, LHTEE/AUT has undertaken manifold activities related to the atmospheric environment and will contribute to all the modelling activities of TRAPOS with advanced numerical models.
As the geometrical details of street canyons are of vital importance to the transformation and transport of pollutants, several street canyon configurations will be studied. More specifically, various height to width aspect ratios will be considered both for two-dimensional and three-dimensional, symmetric and asymmetric street canyon configurations. Slanted roofs will be taken into account with the aid of terrain following coordinates. Three-dimensional unsteady effects will also be studied and compared to experiments. Space periodic boundary conditions applied to typical aggregates of buildings, will enable modelling parts of a city.
Vehicle induced turbulence, both of mechanical and thermal origin, is a very important agent, as it determines the initial dispersion of pollutants in the streets. These effects will be taken into account with the aid of boundary conditions of rough moving walls at a temperature comparable with the exhaust gas temperature.
For low wind speeds over an urban area, buoyant effects, due to the differential heating of the street canyon walls caused by radiation, become dominant and drive the air within the street canyons. Such effects will be taken into account for typical conditions and configurations.
Respirable suspended particulate matter emitted by buses and heavy duty tracks will be treated by a Lagrangian particle tracking module.
As a model of turbulence, the two equation k-e model has served as a compromise between accuracy and complexity and has found the widest application in modelling fluid conditions in street canyons. The k-e model has proved to give poor results for modelling flows characterised by anisotropic turbulence, as well as for flows near stagnation regions. In order to improve turbulence modelling for flows around buildings, an anisotropic k-e model and a realizability correction for cm will be implemented.
Chemical reactions with small time scales of the order of a few minutes can have a decisive effect on the composition of the emitted pollutants. Fast NO-NO2-O3 cycles, for example, can lead to substantial production of NO2, before scales are reached which are comparable to the resolution of mesoscale models. Microscale models, taking into account such chemical processes, can provide mesoscale models with more accurate emission data. Microscale models, on the other hand, need boundary conditions from larger scale models. For more accurate source receptor relationships, a multiscale model is therefore needed, starting at regional scale and going down to microscale (street level). Within such a model cascade, each scale provides the next smaller one with appropriate boundary conditions for momentum, heat and pollutant concentrations at the lateral boundaries. At the small scale end of the cascade, microscale parametric computations provide improved momentum, heat and pollutant vertical fluxes to be used as input to the larger scales. Within the frame of TRAPOS, the basis of such a model system will be set, using the mesoscale model MEMO with nesting capabilities to treat the large scales and the microscale model MIMO for the small scales.
All cases to be studied will be defined through the co-operation with the groups in the network. The results of the simulations will be compared with the results of relevant models and measurements within the network. Modelling work at LHTEE/AUT will be conducted with two CFD software packages: The MEMO/MIMO model system, developed in collaboration with the Institute for Technical Thermodynamics of the University of Karlsruhe and the CFD software package TASCflow developed by the Advanced Scientific Computing (ASC) Ltd., Ontario, Canada. These packages are capable of simulating fluid flow and heat and mass transfer in complex situations from regional scale (MEMO) to microscale (MIMO and TASCflow). MIMO will be used as the basic microscale solver, while TASCflow will form a basis for fast developments and comparisons with the results of MIMO.