Inter-Model Flow Topology Comparison of Simple CFD, RANS and LES Simulations for Complex Street Canyons

Urban microclimate plays an important role in the usage of water and energy resources, transport of pollutants, and the general comfort of urban inhabitants. Street canyons are critical components of urban infrastructure which strongly contribute to the complexities of the flow structures in urban areas. For example, non-linearities associated with interacting wakes and recirculation zones produce non-intuitive flow topologies that affect momentum and scalar transport. In this study, we asses and compare flow topology features from three computational fluid dynamics (CFD) methods with different levels of physics representations against experimental data for step-down (tall building followed by short building) and step-up (short building followed by tall building) street canyons. The first method is a “simple CFD” approach, which uses the Quick Urban Industrial Complex (QUIC-URB) wind model in which empirical flow parameterizations are applied, the second method uses a Reynolds Averaged Navier-Stokes (RANS) model, and the third utilizes a fully-coupled fluid-structure interaction (FSI) Large-Eddy-Simulation (LES) model. The models are particularly examined on their capability of predicting major flow topology features. For the step-down street canyon these features include: a primary vortex core between the buildings, a secondary recirculation zone, a saddle point and a wall-originating ground shear layer. For the step-up street canyons, the two key topological features are a counter-rotating vortex core and a stagnation point. The disparity amongst the results is evident in the models' ability to predict the wall-originating ground shear layer and the counter-rotating vortex pair in step-down and step-up street canyons, respectively. Moreover, the ability to track the dynamics of the flow topology features as a function of the down-wind-to-upwind building height ratio and the canyon width-to-length ratio is found as another important distinction among different CFD models.