Evaluation of fast fluid dynamics with different turbulence models for predicting outdoor airflow and pollutant dispersion

Fast fluid dynamics (FFD) could provide informative and efficient airflow and concentration simulation. The commonly used turbulence model in FFD was Re-Normalization Group (RNG) k-epsilon turbulence model which solved two transport equations to obtain eddy viscosity. To reduce this part of time and further improve computing speed, this investigation implemented no turbulence model, Smagorinsky model and dynamic Smagorinsky model which calculated eddy viscosity without solving equation in FFD in an open-source program, OpenFOAM. By simulating several outdoor cases of varying complexity and comparing with experiment and CFD, this study assessed the accuracy and computing efficiency of FFD with four turbulence models. Compared with CFD, FFD greatly improved the computing speed without reducing accuracy. The simulation of FFD without turbulence model was fast but inaccurate. FFD with Smagorinsky model increased the computing speed while ensuring the same accuracy as RNG k-epsilon turbulence model. FFD with dynamic Smagorinsky model provided accurate results with high efficiency. Computation errors arose mainly from inaccurate prediction of turbulence dispersion. The computing cost was associated with the number of transport equations and calculation method of model coefficient. This investigation recommended the use of FFD with dynamic Smagorinsky model for outdoor airflow and pollutant dispersion studies.

Automated summary

This study implemented fast fluid dynamics (FFD) with different turbulence models, including no turbulence model, Smagorinsky model, and dynamic Smagorinsky model, in the open-source program OpenFOAM. By simulating various outdoor cases and comparing with experiment and computational fluid dynamics (CFD), the accuracy and computing efficiency of FFD with different turbulence models were assessed. FFD greatly improved computing speed without sacrificing accuracy compared to CFD. Dynamic Smagorinsky model provided accurate results with high efficiency.