Atmospheric dispersion downstream a two-dimensional obstacle: experimental evaluation of turbulence closure models

This study investigates the turbulent dispersion of pollutants in the wake of a two-dimensional square obstacle, focusing on assessing the reliability of turbulence closure models used in atmospheric pollutant dispersion. Utilizing Laser Doppler Anemometry (LDA) and Particle Image Velocimetry (PIV), we characterized the flow dynamics, identifying a recirculation zone with an average length of 6H downstream of the obstacle. This zone, characterized by high shear and increased turbulent viscosity, plays a crucial role in turbulent momentum exchange. We evaluated the turbulent kinetic energy (t.k.e.) budget, estimating its dissipation rate ($\varepsilon$), and found traditional isotropy and Taylor hypothesis methods inadequate within the wake region. Furthermore, we explored pollutant dispersion from a linear source located downstream. Analysis of mean concentration and variance revealed that the log-normal distribution is most effective for modeling concentrations within the recirculating region, while the Gamma distribution suits areas outside it. Testing various closure models for turbulent mass fluxes highlighted the limitations of the Simplified Gradient Diffusion Hypothesis model, favoring more complex closure models for longitudinal trends, though these still faced challenges with intensity estimation. The Simplified Gradient Diffusion Hypothesis model proved robust for vertical mass fluxes, with satisfactory results in turbulent diffusivity and turbulent Schmidt number calculations. Our findings underscore the necessity for improved spatial resolution in measurements near pollutant sources, particularly between 0.5H and 2H downstream. The experimental results allowed us to assess the reliability of a wide variety of parameterizations adopted in Reynolds-Averaged Navier-Stokes (RANS) formulations and within pollutant dispersion models. This database serves as a benchmark to test the accuracy of more sophisticated modeling approaches such as Large Eddy Simulations (LES) and Direct Numerical Simulations (DNS).