Modeling of a Heavy Gas Turbulent Dispersion in Simplified and Real-Scale Urban Environments

We report recent advances in the development of a computationally efficient and accurate time-dependent RANS-based (T-RANS) method for predicting turbulent heavy-gas dispersion in complex urban environments. Two test cases were analyzed: (i) the simplified Kit Fox field-scale setup, and (ii) a realistic urban neighborhood in Beijing. In the first case, the T-RANS approach outperformed the simplified atmospheric dispersion model (SLAB), showing significantly better agreement with field measurements due to its ability to capture local flow and turbulence features within obstructed regions. The numerical robustness and computational efficiency of the method were further validated in the real-scale Beijing neighborhood, incorporating detailed building geometries and a mesh of over 100 million control volumes. By utilizing a passive-element approach to resolve complex building geometries and accounting for the buoyancy-driven effects of local gas concentrations, this method provides a high-fidelity, computationally efficient framework for predicting the temporal evolution of turbulent heavy-gas dispersion in complex built environments.