[gholamghar]

Hi All,
I'm trying to model the transition from laminar to turbulent flow on the suction side of a high pressure turbine blade. I'm using Transition SST model; After I run the code, transition happens near the trailing edge while in reality it should happen near leading edge.
What I did was to create the mesh in Gambitwith periodic B.C.s; I ran the code in fluent, density based, ideal gas, sutherland for viscosity, Transition SST, inlet pressure about 20atm, outlet pressure about 10 atm.
formulation:
implicit, all things are second order upwind except for Flow which is first order upwind( when I switch Flow to second order solution it does not converge)

B.C.:
Inlet ressure inlet
Outlet: pressure outlet
I have boundary layer mesh on suction and pressure side with growth ratio of 1.09 (which is less than 1.1)

after I iterated for about 2000 times I checked the Y+; suction side Y+ was Y+<1, pressure side Y+ was between 1 and 7, so I used the Adapt-Yplus/Ystar option in Fluent to refine the mesh on pressure side and now I have Y+<1 for pressure side.

Then I ran the code again; but still transition happens near trailing edge. I checked the intermittency contours at different times; after 1000 iterations when the solution is not converged yet, intermittency goes from 0 to 1 near the leading edge, as I let the solution to converge more, the location that intermittency suddenly increases from 0 to 1 on it,moves toward the trailing edge and after about 6000 iterations that solution becomes converged, my solution shows that transition is near trailing edge!
Any ideas what's going wrong here? I should get transition near the leading edge not trailing. any suggestions are highly appreciated.

By the way I changed the B.C.s a lot to just see what I get but no success; for example, I changed the inlet pressure to 15atm instead of 20atm, then there was no transition even at trailing edge. I increased it to 23 then solution didn't converge. Believe me I tried so many different and none works! Now I need an expert opinion!
Thanks

here is a link showing intermittency change close to leading edge on suction side :
http://i43.tinypic.com/2mydeuc.jpg

[Jim87]

Hi,

Your BC and Settings seam reasonable. The first thing I would check is if there is an comparable case in the tutorial guide or the fluent customers portal. These are often helpfull to get the first steps, maybe in your case to check the settings.

Maybe there is someone in this forum who is trained well in the simulation of a turbine. I can general say this to your unexpected results:

- Some cases are hard to get with the used numerical system. In a few applications a transient solver is a better idea then a steady solver.

- Check your mesh: How is the quality of your relevant blade area? Are there region with very high residuals or less cell quality?

- How did your residuals convergate? Is there high oscilation in any of your residulas? What is the final residual?

[gholamghar]

Originally Posted by Jim87

Hi,

Your BC and Settings seam reasonable. The first thing I would check is if there is an comparable case in the tutorial guide or the fluent customers portal. These are often helpfull to get the first steps, maybe in your case to check the settings.

Maybe there is someone in this forum who is trained well in the simulation of a turbine. I can general say this to your unexpected results:

- Some cases are hard to get with the used numerical system. In a few applications a transient solver is a better idea then a steady solver.

- Check your mesh: How is the quality of your relevant blade area? Are there region with very high residuals or less cell quality?

- How did your residuals convergate? Is there high oscilation in any of your residulas? What is the final residual?

thanks for the reply, I tried transient but it does nor converge, I'm using a BL mesh near the blade surface so the quality over there seems to be good..., the residuals converge up to 10-05 order, even more if I want to. This SST Transition problem is killing me, I have been working on this about 3 month and I haven't been able to move the transition close to the leading edge, it still stays there, I even looked at one of the fluent tutorials that they solve an airfoil with low pressures at th B.C.s and apparently transition is predicted very well over there, I changed my B.C.s and solver in accordance to that manual, but still for my case it happens near the trailing edge!!!

Is there anybody here that has solved a problem with the SST transition model, specially for a blade or an airfoil , and can tell me what how they set up their problem in Fluent exactly?
Thanks

[AUCH]

Hello,
As far as I understand if your transition is happening near the trailing edge, that means your freestream turbulence levels are low. In order to move the transition point near the leading edge, you need to specify a larger turbulence value (preferably the one that the experimental airfoil experiences). You can use the velocity inlet BC and specify the turbulence there. It has worked for me as far as transition is concerned. The bubble and transition points are all captured with reasonable accuracy. However, I am facing the problem that the lift coefficient is underpredicted, primarily because the simulation does not capture the peak suction pressure near the leading edge.
I hope this helps ..

[taxalian]

Hi,
I agree with the comments from AUCH, because in order to achieve good results with SST transition model. You need to specify appropriate turbulent levels i.e. turbulent intensity to achieve the desired transition in your case. I would recommend to have experimental turbulence levels if you have any, or try with different turbulent levels to achieve desired results.

[tom634]

Additionally, you can try to adjust Re theta and F_length values.