Atmospheric particulate dispersion over complex terrain: Plume trajectory and model optimization

Atmospheric dispersion behavior of radioactive particles from nuclear facilities has attracted widespread attention due to their profound impacts on the environment and public health. The spatial variation of radioactive particle distribution could be mainly attributed to complex terrain. A simplification of complex terrain into idealized two-dimensional ridges and three-dimensional hills is a common methodological approach for investigating terrain effects. Here, an integrated approach employing both wind tunnel experiments and Computational Fluid Dynamics (CFD) simulations was used to investigate radioactive particulate dispersion trajectories over these typical idealized terrains. Key parameters, such as dimensions, types of terrain, release heights and positions of pollutants, were systematically varied. A comprehensive wind tunnel dataset was established using laser-sheet visualization to enrich high-fidelity validation data for particle dispersion model. Furthermore, the wind tunnel results indicated that the two-dimensional ridge enhanced plume rise relative to flat terrain, with the magnitude of enhancement dependent on its location and height. Conversely, lateral flow around the three-dimensional hill mitigated its influence on the plume trajectory, leading to negligible sensitivity to the hill height. However, large hills significantly blocked particle dispersion, resulting in a concentration decrease in the downwind area. Additionally, a higher release height and greater distance from the source diminished the influence of the complex terrain. In the CFD simulations, discrepancy of particle concentration was observed over the two-dimensional ridge when using various turbulent and dispersion model constants derived for flat terrain. A set of turbulent and dispersion constants of the CFD model (C_1ε, C_2ε, σ_k, σ_ε, C_μ and c_L) was optimized for application to complex terrain through validation against wind tunnel experimental data. The optimized values are 1.5, 1.92, 1.67, 3.25, 0.09 and 0.2, respectively.