Optimizing non-isothermal turbulent inflow generation using precursor method: A sensitivity study

Abstract

Understanding airflow dynamics in urban environments is essential for enhancing urban planning and improving human comfort, particularly under the unstable conditions due to meteorological reasons and human activities. This study addresses the generation and application of turbulent inflow using the precursor method under unstable conditions, which is crucial for accurately capturing the complex flow characteristics. The precursor method successfully generated fully developed turbulent flow with uniform mean velocity and temperature profiles. A systematic analysis was conducted on the impact of various computational parameters, including mesh resolution, discretization schemes, Smagorinsky constant, and turbulent Prandtl number. The results indicate that finer streamwise mesh resolution may lead to overestimation of turbulent kinetic energy and temperature standard deviation due to model influence and experimental bias. It is recommended to use a blended scheme with over 90 % central differencing for velocity and total variation diminishing schemes for temperature to avoid oscillations. Increasing Smagorinsky constant simplifies vortex structures and dissipates small-scale vortices, while increasing turbulent Prandtl number enhances the coupling between velocity and temperature fluctuations; however, both parameters have minimal impact on the mean velocity and temperature results. Furthermore, applying the generated inflow to an isolated building case with a coarser mesh maintained the overall quality and accurately resolved the flow field around the building. This study provides valuable insights for generating non-isothermal turbulent inflow for LES and offering practical guidelines for CFD simulations in urban planning and environmental assessment.