Assessment of RANS turbulence models for numerical study of laminar-turbulent transition in convection heat transfer

In this study, the prediction of different turbulence models for simulation of transition heat transfer in forced convection over flat plate and natural/mixed convection between two flat plates was investigated. These turbulence models included Spalart-Allmaras (S-A), Low-Re k - epsilon, Re-Normalisation Group (RNG) k - epsilon, Low-Re k - omega Shear Stress Transport (k - omega SST), k - k(L) - omega, gamma - Re-theta and V-SA transition models. In forced convection flow over horizontal flat plate, variation of skin friction coefficient and Stanton number with local Reynolds number, Re-x, were studied. The simulations showed that in forced convection V-SA, gamma - Re-theta and k - k(L) - omega transition models can closely predict the critical distance of flow transition from laminar to turbulence regime comparing to other turbulence models, while in natural convection k - k(L) - omega and k - epsilon models provided better prediction. In the case of mixed convection, the effects of different parameters such as heated wall temperature, inlet velocity, inclination angle and Richardson number on fluid flow and heat transfer were investigated in details. It was found that the changes in the transition point depends on the relative strength of the natural to forced convective flow and heat transfer. The transition point was delayed when the inclination angle of the channel was increased while increase in Richardson number promoted the transition to turbulence. In addition, Low-Re k - omega SST and RNG k - epsilon predicted a shorter transition region, while gamma - Re-theta model also predicted a delayed transition point. (C) 2017 Elsevier Ltd. All rights reserved.