CFD Online Discussion Forums

CFD Online Discussion Forums (
-   CFX (
-   -   anybody using transition model in CFX (

littlelz August 20, 2008 12:24

anybody using transition model in CFX
I am doing transition modelling for a simple flat plate model (gamma theta model). I check the user's guide and previous messages in this forum as well. still can't set correct turbulence intensity and viscosity ratio for inlet boundary. (running start only a few steps and then collapse with overflow error). the same model for fully turbulent simulation shows all right. any idea? please

Glenn Horrocks August 20, 2008 21:14

Re: anybody using transition model in CFX

Need more information. What do the residuals look like before it diverges? If it runs a few steps stop it before it crashes and look at the output - can you see what is going wrong? Outputting the residuals to the output file may also help.

Glenn Horrocks

littlelz August 21, 2008 07:09

Re: anybody using transition model in CFX
many thanks, Glenn.

I 've already modified the mesh and now it is converged, but only for velocity/turbulence variables. The turbulence transition variables, intermittency and retheta don't converge. max residual of intermit reach 4.e-2, and then get fluctuation between 1.e-1 and 4.e-2. max residual of rethetat reach 1.e-3, and then fluctuate between 1.e-3 and 2.e-3. What's your idea on it?

what's residual of transiton variables of your case?

I also want to know what it look like for the result of transition variables? any pictures?

thanks a lot again

Glenn Horrocks August 22, 2008 00:30

Re: anybody using transition model in CFX

Generally poor convergence of the transition variables is due to a laminar separation bubble moving about. Have a look at the overall lift/drag or whatever is important to determine whether it is significant of not - it generally is not.

Glenn Horrocks

iffy May 10, 2011 03:50

I am doing transition modelling for naca 0012 airfoil in cfx using Gamma theta model (Re number 5*10^ 5). At zero degree aoa, simulation results show laminar separation bubble at about 60% chord and then this bubble reattaches as tubulent boundary layer at about 90% chord. But this result is very different then what XFOIL shows. Xfoil shows no separation bubble and shows transition at about 80% chord. Also the residuals for during simulations for Intermitt do not converge (residuals drop down to 1.5*10^-4 and then oscillates).
Can anyone explain the reson behind the difference between xfoil and cfx results? I did not use Intermittency and Retheta coefficients in CFX-Pre, as I do not know what coefficients should I use, can this also be the reason?



ghorrocks May 10, 2011 18:52

CFX and XFoil use totally different approaches to modelling turbulence transition. There will be different inherent assumptions behind them which may well be difficult to root out, such as incoming turbulence levels, surface roughness etc.

Unless you propose to do turbulence transition research to tune the model to your situation I would just use the default Gamma Theta model, with default parameters.

iffy May 11, 2011 02:59

Glenn thank you for your reply.

Actually I have to compare my simulation results with some reference. The only reference I can find till now is XFOIL. I understand that the method to predict transition in XFOIL is completely different and xfoil can predict only natural transition. In CFX, I used low turbulence intenisty of 0.1%, inlet velocity is quite low (v=7.6 m/s) and I expect to get natural transition on airfoil but this did not happen.
Do you have idea what could go wrong with the simulation or could you recommend any literature from which I can compare the results.


ghorrocks May 11, 2011 06:48

All I can suggest is to make sure your simulation is accurate, both CFX and XFoil. Some tips for CFX is here:

XFoil does not have many setable parameters but regardless you should make sure it is as accurate as possible before making the comparison.

All times are GMT -4. The time now is 05:57.