Evaluation of Various Turbulence Models in Predicting Airflow and Turbulence in Enclosed Environments by CFD: Part 2-Comparison with Experimental Data from Literature

• Published in: HVAC&R research Vol. 13; no. 6; pp. 871 - 886
• Main Authors: Zhang, Zhao, Zhang, Wei, Zhai, Zhiqiang John, Chen, Qingyan Yan
• Format: Journal Article
• Language: English
• Published: Atlanta Taylor & Francis Group 01-11-2007; Taylor & Francis Ltd
• ISSN: 1078-9669; 2374-4731; 1938-5587; 2374-474X
• DOI: 10.1080/10789669.2007.10391460

Numerous turbulence models have been developed in the past two decades, and many of them can be used in predicting airflows and turbulence in enclosed environments. It is important to evaluate the generality and robustness of the turbulence models for various indoor airflow scenarios. This study evaluated the performance of eight turbulence models, potentially suitable for indoor airflow, in terms of accuracy and computing cost. These models cover a wide range of computational fluid dynamics (CFD) approaches, including Reynolds averaged Navier-Stokes (RANS) modeling, hybrid RANS and large-eddy simulation (or detached-eddy simulation [DES]), and large-eddy simulation (LES). The RANS turbulence models tested include the indoor zero-equation model, three two-equation models (the RNG k-∊, low Reynolds number k-∊, and SST k-ω models), a three-equation model (
model), and a Reynolds-stress model (RSM). The investigation tested these models for representative airflows in enclosed environments, such as forced convection and mixed convection in ventilated spaces, natural convection with medium temperature gradient in a tall cavity, and natural convection with large temperature gradient in a model fire room. The air velocity, air temperature, Reynolds stresses, and turbulent heat fluxes predicted by the models were compared against the experimental data from the literature. The study also compared the computing time used by each model for all cases. The results reveal that LES provides the most detailed flow features, while the computing time is much higher than for RANS models, and the accuracy may not always be the highest. Among the RANS models studied, the RNG k-ω and a modified
model perform the best overall in four cases studied. Meanwhile, the other models have superior performance only in particular cases. While each turbulence model has good accuracy in certain flow categories, each flow type favors different turbulence models. Therefore, we summarize in the conclusions and recommendations both the performance of each particular model in different flows and the best suited turbulence models for each flow category.