[ttu004]

Dear All

I am simulating a S814 aerofoil under Low Reynolds number. In wind tunnel experiment, the flow undergoes laminar-to-turbulent (LSB formation) via the separation-induced transition on both upper and lower surfaces.

For my CFX, I tried SST and SST Gamma theta models and revealed that only the SST Gamma Theta model has modeled the LSB on both upper and lower surfaces and is close to the experimental expectation. Comparison of the Turbulent Kinetic energy (TKE) and Velocity Vector plot are attached and has shown pronounced difference. The maximum y+ value for both simulations are approx. 0.55 on the leading edge tip. And the turbulence intensity were kept at approx. 0.07% around the foil.

For SST GammaTheta model, the TKE has started from near half of both surfaces and were matched with the location of LSB when superimposed with vector plot (location of LSB matched well with experiment). This indicates clearly an onset of transition in the boundary layer.
However, the TKE intended to start earlier when SST model was used.

My question are:
1. What is happening in the boundary layer for result from SST model?
2. Is this indicating that i am seeing a natural transition? (not separation-induced transition)
3. Can the SST model simulate cases with the laminar-to-turbulent transition flow (natural transition, separation-induced transition)

Very thankful for any thought and suggestion.

Cheers
Tom

[ghorrocks]

Q1: I do not understand your question. With the fully turbulent simulation the turbulent boundary layer builds from the start of the airfoil, with the transition model it builds from the transition point.
Q2: The laminar sep bubble is a natural transition. The laminar flow separates from the foil and straight away goes turbulent. The turbulent flow reattaches to the foil and that's the laminar separation bubble, and your model is capturing this.
Q3: SST has no transition model. The gamma-theta model is the transition model. If you use SST it will be assumed turbulent everywhere and this will over estimate the flow dissipation before transition.

Your models are showing the expected results: SST+gamma theta captures a transition and a laminar separation bubble, SST alone builds up too much turbulence and has not laminar separation bubble (and is wrong).

[ttu004]

Thanks for the reply. Glenn

The SST is not designed for predicting onset transition but only the SST transitional (Gamma-Theta) model does. That is why I am seeing reduced in lift when SST transitional model is used when the LSBs are modelled (especially, when modelling at low Reynolds number when LSB is enlarged and elongated...).

[ghorrocks]

That is correct.