Abstract

An incompressible, high-Reynolds number flow (slightly less then 1 Mio. per chord) over a smoothly contoured, asymmetric, wall-mounted hump was computationally studied using the LES (large eddy simulation) and DES (detached eddy simulation) methods. In addition, several second-moment and eddy-viscosity closures within the RANS (Reynolds-averaged Navier–Stokes) framework were tested. The focus of the investigation was on the effects of local perturbation of the hump boundary layer introduced by spatially uniform (in the spanwise direction) steady suction and oscillatory suction/blowing through a narrow opening (1.7 mm) situated at the hump crest immediately upstream of the natural separation point. Reference experiments have shown that both flow control mechanisms cause a shortening of the recirculation bubble relative to the baseline configuration with no flow control. All statistical turbulence models used in the RANS framework resulted in a substantially larger recirculation zone independent of the modelling level, being a consequence of a too low turbulence level in the separated shear layer. Accordingly, the effect of the steady suction, namely the reduction of the reattachment length, was underpredicted. The LES method, despite a relatively coarse mesh (with a total of 4 Mio. cells) for such a high-Reynolds number, wall-bounded flow, was capable of capturing important effects of the flow control qualitatively and quantitatively. DES failed to do so in the suction case, despite good results in the baseline and oscillatory blowing/suction cases, indicating that a shallow separation from curved surfaces poses a challenge to this hybrid RANS/LES strategy. A sensitivity study of the RANS/LES interface position within the DES approach shows that a RANS region chosen too thin (with the interface situated at the very beginning of the logarithmic layer) can lead to a strong reduction of the turbulent viscosity causing a low turbulence level within the shear layer region aligned with the recirculation zone, which in turn leads to a larger separation bubble.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call