[kael]

Hi!
I'm doing the analysis of 2D airfoil in low-Re regime (around 10^5 - 10^6) which the transitional B.L phenomenon is expected. (e.g., laminar separation bubble). I've done some literature study based on this problem. For turbulence model, Transition SST (gamma-Re, theta) seems give better prediction to the aerodynamics forces and the flow physics compared to other turbulence models.

I tried to model the problems in Fluent and compare the results (Cl, Cd, and Cp distribution along chord airfoil) with the exp data. At low aoa, the results match accurately with the exp data but at higher aoa the drag values are way-off.

Based on this, I have some question regarding the correct modelling in Fluent:

1. Steady or Transient? (I got converged solution for the low aoa with steady state solver but at higher aoa the convergence is hard to get, and Cd’s discrepancy got lot bigger. My doubt is if there's some unsteady behavior (e.g., separation B.L) is it better to use transient solver?

2. Incompressible or Compressible? (I know for this low re, it can be assumed that the flow is incompressible. My question is should I stick with constant properties or set the ideal-gas properties and let the solver develop itself?

3. Decay of Turbulent Intensity (Tu). Fluent Theory Guide for Transition SST turbulence model give an explanation about the decaying Tu from inlet to the body with respect to streamwise distance downstream of the inlet. Basically saying, I need to set the Tu at inlet to corresponding value to get the actual Tu (exp Wind Tunnel Tu) at the body?

Best Regards,

[LoGaL]

To answer point 1 and 2

1) if the Cl and Cd do not stabilize, you can’t say the prediction is way off, it just didn’t stabilize. Switch to transient and get time average Cd, Cl

2) low re doesn’t mean incompressible. Low Mach means incompressible, and indeed you can have high Re, low Mach scenarios or viceversa. I would calculate your far field Mach number and see if you get something on the order of 0.1-0.2, if yes, then you can stick with incompressible

3 is not clear

[kael]

Originally Posted by LoGaL

To answer point 1 and 2

1) if the Cl and Cd do not stabilize, you can’t say the prediction is way off, it just didn’t stabilize. Switch to transient and get time average Cd, Cl

2) low re doesn’t mean incompressible. Low Mach means incompressible, and indeed you can have high Re, low Mach scenarios or viceversa. I would calculate your far field Mach number and see if you get something on the order of 0.1-0.2, if yes, then you can stick with incompressible

3 is not clear

Thank you for your responses LoGaL.

1. Cl and Cd values are stabilized, i set the convergence condition for these two parameters so the changes after 10 iterations get less than 0.0001.

2. It's low Mach flow, the freestream flow speed is 1.5m/s, equivalent to Mach 0.03.

[LuckyTran]

3. You need to provide inlet boundary conditions for all transport equations. You are using transitional SST, which means you need to provide profiles of k & omega/epsilon at the inlet just like you need to provide a velocity profile. Turbulence intensity is the term most people use for the sake of discussion, but technically it's k and omega that you need.


You need the k and omega profiles from the wind tunnel data. 99% of the time, that data was never measured. So people just tweak the value of the turbulence intensity at the inlet until their sought-after-parameter matches what was measured.

[kael]

Originally Posted by LuckyTran

3. You need to provide inlet boundary conditions for all transport equations. You are using transitional SST, which means you need to provide profiles of k & omega/epsilon at the inlet just like you need to provide a velocity profile. Turbulence intensity is the term most people use for the sake of discussion, but technically it's k and omega that you need.


You need the k and omega profiles from the wind tunnel data. 99% of the time, that data was never measured. So people just tweak the value of the turbulence intensity at the inlet until their sought-after-parameter matches what was measured.

So what do you mean is i should better specify corresponding values of k and omega at inlet rather than specified Tu and turbulent viscosity ratio? i assumed these values will be based on the eqn that fluent gives in fluent user guide right?

and by profile, did you mean that these values are not constant at inlet?