Capturing transient effects in turbulent flows over complex urban areas with passive pollutants

Highlights: • A new variant of the hybrid T-RANS/VLES model is proposed for turbulent flows in complex urban areas. • Dynamical evolution of the T-RANS/VLES interface. • The model captures significantly larger amplitudes and frequencies of the transport variables in the wake regions. • Significant improvement in prediction of the turbulent kinetic energy. • Improved prediction of the passive pollutant dispersion and its variance. - Abstract: The present study addresses numerical modeling and simulations of turbulent flow and dispersion of a passive (non-reactive) pollutant in laboratory scale configurations that mimic complex urban areas. Two generic configurations are selected: a regular array of obstacles (of height 'h') with a central tower structure (of height '3h') and an array of randomly distributed obstacles (with heights of '1h', '2h' or '3h'). Both arrays exactly mimic the experimental conditions of Hilderman and Chong (2004) for water channel flow around simplified urban areas, with typical values of Reynolds and Schmidt numbers of 36,000 and 1920, respectively. Comparison between T-RANS and a new seamless hybrid T-RANS/VLES approach is performed. It is shown that the presented hybrid T-RANS/VLES approach is able to capture significantly larger amplitudes and frequencies of the transport variables in the wake regions. The long-term time-averaged statistics of the second-moments contain two parts: modeled (from the parent transport equations) and resolved contributions (from collecting the time-dependent evolution of transport variables). The resolved contributions prove to be the dominant mechanism in the wake regions, which result in significant improvement in predictions of the turbulent kinetic energy, and consequently, of the turbulent dispersion of the passive scalar. The version of the hybrid T-RANS/VLES approach proposed here, based on a dynamical evolution of the interface zone (seamless approach), proved to be a numerically efficient and robust method, which is recommended for future studies of turbulent flow and dispersion in complex urban areas