RANS and LES models comparison for a cross-shaped jet flow with application in personalized ventilation

The need for thermal comfort and clean air for occupants in buildings or vehicles is vital since we spend more than 90% of our time inside these enclosed environments. Worldwide, current directions of the leading powers are oriented towards the reduction of the energy consumptions and HVAC systems make no exception. Personalized Ventilation (PV) applied to buildings may represent a solution to this problem. The main idea of PV is to provide clean air close to the face of each occupant and to improve thermal comfort in his microenvironment.
This study is a part of a larger research work related to the optimization of PV air diffuser for HVAC systems in terms of passive control of jets flows in order to control jet development before it impinges on the occupant’s face.
The flow field of an elementary lobed jet from a cross-shaped orifice with straight edges is investigated numerically using Large Eddy Simulation (LES) model and the results are compared with those from seven RANS (Reynolds Averaged Navier-Stokes) turbulence models and with experimental results. The RANS models used are the RNG k-ε turbulence model, k-ε standard, k-ε realizable, the Shear Stress Transport (SST) k-ω, the standard k-ω, the Spalart-Allmaras turbulence models and the Reynolds Stress turbulence Model (RSM). The initial Reynolds number based on the jet centreline exit streamwise velocity and on the equivalent diameter is around 4000. Numerical results are analysed based on PIV measurements performed in the same flow.
The objective is to assess the capability and limitations of the studied viscous models to predict the significant features of the cross-shaped air jet by numerical simulation.
The study revealed that none of the turbulence models was able to predict well all jet characteristics in the same time. For instance, the Reynolds Stress turbulence Model (RSM) predicted better the local jet flow expansion in the longitudinal minor and major planes, whereas global flow expansion and ambient air induction are better predicted by the Shear Stress Transport (SST) k-ω turbulence model. Furthermore the RANS models are not able to capture the flow’s temporal dynamics.
Also, the LES model allows obtaining global mean quantities as RANS models but delivers supplementary information about the temporal vortex dynamics.