Comparative Assessment of RANS and DNS for Different Indoor Ventilation Regimes

Although Reynolds-Averaged Navier–Stokes (RANS) models are widely employed, there is a need to understand their limitations and accuracy for predicting the complex turbulent flow phenomena in indoor ventilation environments. The objective of this study is to assess RANS models directly by conducting a comparative analysis of RANS and Direct Numerical Simulations (DNS). We study two canonical cases that represent basic ventilation problems: the flow within a room containing a constant temperature source that represents a person, and an empty room with an exponential heat load commonly used to represent a group of people. The variation on the relation between the convection and ventilation speed will be used as a parameter to change the importance of the drivers of the flow. We demonstrate that RANS simulations provide an adequate prediction of the flow kinematics and the convection process for the case of an isolated constant-temperature source, with minor discrepancies in temperature prediction. On the contrary, the results observed in the exponentially heated volume were different for DNS and RANS, where neither the kinematics nor the heat transport was properly represented.