Comprehensive evaluation of constant and variable turbulent Schmidt numbers for CFD simulation of near-field air pollutant dispersion

Turbulent Schmidt number (Sc_t) is an essential parameter of the standard gradient diffusion hypothesis that is used for the Reynolds-Averaged Navier-Stokes (RANS) equation-based computational fluid dynamics (CFD) simulation of air pollutant dispersion. Many studies have attempted to choose the optimum Sc_t without considering its variability with local turbulent flow properties and height dependency. Recently, there has been a renewed interest in formulating variable Sc_t to model near-field air pollutant dispersion. However, these formulations are not widely employed due to a lack of knowledge concerning their use and advantages. This study aims to fill this knowledge gap by applying three variable Sc_t formulations for simulating pollutant dispersion around an isolated CEDVAL A1-5 building. CFD simulations were conducted using constant and variable Sc_t and two sets of inflow boundary conditions. The results suggested that the variable Sc_t formulation directly depending on flow parameters and physical properties, such as Sct_R2, display accurate prediction and is compatible with all inflow boundary conditions, differently from Sct_G and Sct_R1, which depend on model parameters, i.e. C_µ. Moreover, the three variable Sc_t formulations modeled similar spatial pollutant dispersion, but they modeled dispersion patterns that were noticeably different from that of constant Sc_t. A statistical analysis shows that the three variable Sc_t estimated the mean Sc_t in the range 0.6–0.7, with 50–60% occurrence probabilities, despite their values fluctuating between 0.11 and 1.85 in the area affected by the building.