Study of impinging turbulent jet flows using the isotropic low-Reynolds number and the algebraic stress methods

Abstract

The performance of two-equation turbulence models (such as the low Reynolds number k–ɛ model of Launder and Sharma 1974) is evaluated versus the algebraic stress model (ASM) (Rodi 1976) for high Reynolds number (Re=9×104) jet flows with strong streamline curvature due to impingement onto a flat plate. The partial differential equations for the conservation of mass and momentum are solved using a finite volume method and the predicted velocities are compared to experimental data by Myszko (1997). The paper demonstrates that in the free-jet region both models over-predict the thickness of the jet. The ASM predicts faster jet growth rate and smaller jet thickness than the low Reynolds number model resulting to closer agreement with the experiments. As a consequence of the better performace of the ASM in the free-jet region, predictions in the wall-jet region showed that despite the use of the logarithmic law-of-the-wall function, the ASM results are closer to the experimental points than the predictions obtained with the two-equation model. However, the rate of peak velocity decay is far higher than the experimental one with both turbulence models. Again, the decay rate predicted with the ASM fits better the exprimental data. The implementation of the ASM exhibited convergence problems most of which were atributed to the cross-derivative terms in the k and ɛ equations and were treated using a linear under-relaxation technique. In general the ASM predictions were more accurate than the low Reynolds number k–ɛ model, with an extra computational cost of less than 25%, which makes the model very attractive for the prediction of turbulence characteristics of high Reynolds number flows with strong stramline curvature.