Teta term in Menter's Momentum thickness Reynolds number equation for transition

Adrian281 · June 28, 2012, 18:56

Hi, I'm currently undertaking a task of implementing the gamma-Re transition formulation for turbulence modelling by way of a UDF in Fluent.
I've come across a problem.
There's lambda term in used for prediction of transition onset, denoted lambda underscore theta. In the formulation of this term, a parameter given by a capital theta is required. (equation 28 in Langtry & Menter, 2009, 'Correlation-based transition modelling for unstructured parallelized computations CFD')
The problem is, that in all the literature I have looked in, no correlation or formulation for this term is given.
As you would expect, this has brough my progress to a standstill.
I would be grateful if anyone could suggest where I might be able to look for an answer.
Many thanks.

mb.pejvak · June 28, 2012, 20:09

it should be calculated iteratively. you should take following step:
1-guess teta (teta=0)
2-find F(lateta) and Rethetat (from form. 35-38)
3-find theta from Retheta (Retheta=ru*theta*U/mu - you can see it in Nomenclature)
4-go to step 2 and iterate until it converge.

I hope I can explain it well.

Adrian281 · June 29, 2012, 09:53

That makes it much clearer. Thanks very much.

Adrian281 · June 29, 2012, 10:02

Just one thing, to calculate theta in step three, which value for Retheta do I use? Do I make it equal to Rethetat?

mb.pejvak · June 29, 2012, 12:10

I am sorry, i made a mistake. In fact, Rethetat is true in step three instead of Retheta.

Adrian281 · June 29, 2012, 12:35

But then Rethetat is calculated using thetat, whereas lambdatheta is calculated using just theta.

mb.pejvak · June 29, 2012, 22:35

yes that is true, but in calculation of landa it was recommended to Theta= thatat. In fact, I have never understand why and in all paper and thesis I have read, never describe its reason.
if I complete my former reply, following steps can explain it better:

(a) guess the initial value for Thetat (=0 was recommended)
(b) calculate the Landa by setting Theta= thatat
(c) calculate the Flanda
(d) calculate the Rethetat
(e) calculate the t from the definition of Thetat =(no/U)Ret.
(f) go to step 2 and iterate until it converge.
The calculation may converge in a few iterations.

Adrian281 · July 1, 2012, 12:31

I'll try that out. Many thanks for your help.

Adrian281 · July 14, 2012, 15:29

A friend of mine is also implementing the Transition formulation of Menter for the Spalart Allmaras turbulence model based on the paper by Medida and Baeder 'Application of the Correlation based gamma-retet Transition Model to the Spalart-Allmaras Turbulence model' and has encountered a problem with the blending function.
In the paper it is specified that the blending function is used to distinguish a wall boundary layer and a free shear layer to deactivate the production term in the wall boundary layer only. However, this function depends on the dissipation rate, which is not available in the S-A model. The paper says that the blending function has been turned off manually, but does not explicitly specify how this is done.
Can anyone suggest how this is done? Any help would be greatly appreciated.

TedBrogan · November 30, 2012, 19:30

Adrian,

I am trying to put in the SA transition model currently also. I ran into the same problem with how to handle the F_theta blending function. Do you know how your friend ended up resolving that?

June 28, 2012, 18:56	Teta term in Menter's Momentum thickness Reynolds number equation for transition	#1
Adrian281 New Member Adrian Join Date: Mar 2012 Posts: 8 Rep Power: 14	Hi, I'm currently undertaking a task of implementing the gamma-Re transition formulation for turbulence modelling by way of a UDF in Fluent. I've come across a problem. There's lambda term in used for prediction of transition onset, denoted lambda underscore theta. In the formulation of this term, a parameter given by a capital theta is required. (equation 28 in Langtry & Menter, 2009, 'Correlation-based transition modelling for unstructured parallelized computations CFD') The problem is, that in all the literature I have looked in, no correlation or formulation for this term is given. As you would expect, this has brough my progress to a standstill. I would be grateful if anyone could suggest where I might be able to look for an answer. Many thanks. aero_head likes this.

June 28, 2012, 20:09		#2
mb.pejvak Senior Member Mehdi Baba Mehdi Join Date: Jan 2011 Posts: 158 Rep Power: 15	it should be calculated iteratively. you should take following step: 1-guess teta (teta=0) 2-find F(lateta) and Rethetat (from form. 35-38) 3-find theta from Retheta (Retheta=ruthetaU/mu - you can see it in Nomenclature) 4-go to step 2 and iterate until it converge. I hope I can explain it well. aero_head likes this.

June 29, 2012, 22:35		#7
mb.pejvak Senior Member Mehdi Baba Mehdi Join Date: Jan 2011 Posts: 158 Rep Power: 15	yes that is true, but in calculation of landa it was recommended to Theta= thatat. In fact, I have never understand why and in all paper and thesis I have read, never describe its reason. if I complete my former reply, following steps can explain it better: (a) guess the initial value for Thetat (=0 was recommended) (b) calculate the Landa by setting Theta= thatat (c) calculate the Flanda (d) calculate the Rethetat (e) calculate the t from the definition of Thetat =(no/U)Ret. (f) go to step 2 and iterate until it converge. The calculation may converge in a few iterations. Last edited by mb.pejvak; June 30, 2012 at 01:11.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Cell Reynolds Number	laliong	Main CFD Forum	12	September 17, 2019 03:18
[snappyHexMesh] No layers in a small gap	bobburnquist	OpenFOAM Meshing & Mesh Conversion	6	August 26, 2015 09:38
Constant velocity of the material	Sas	CFX	15	July 13, 2010 08:56
how to integrate this x momentum equation term	Prakash	Main CFD Forum	0	June 13, 2008 10:45

June 29, 2012, 09:53		#3
Adrian281 New Member Adrian Join Date: Mar 2012 Posts: 8 Rep Power: 14	That makes it much clearer. Thanks very much.

June 29, 2012, 10:02		#4
Adrian281 New Member Adrian Join Date: Mar 2012 Posts: 8 Rep Power: 14	Just one thing, to calculate theta in step three, which value for Retheta do I use? Do I make it equal to Rethetat?

June 29, 2012, 12:10		#5
mb.pejvak Senior Member Mehdi Baba Mehdi Join Date: Jan 2011 Posts: 158 Rep Power: 15	I am sorry, i made a mistake. In fact, Rethetat is true in step three instead of Retheta.

June 29, 2012, 12:35		#6
Adrian281 New Member Adrian Join Date: Mar 2012 Posts: 8 Rep Power: 14	But then Rethetat is calculated using thetat, whereas lambdatheta is calculated using just theta.

July 1, 2012, 12:31		#8
Adrian281 New Member Adrian Join Date: Mar 2012 Posts: 8 Rep Power: 14	I'll try that out. Many thanks for your help.

July 14, 2012, 15:29		#9
Adrian281 New Member Adrian Join Date: Mar 2012 Posts: 8 Rep Power: 14	A friend of mine is also implementing the Transition formulation of Menter for the Spalart Allmaras turbulence model based on the paper by Medida and Baeder 'Application of the Correlation based gamma-retet Transition Model to the Spalart-Allmaras Turbulence model' and has encountered a problem with the blending function. In the paper it is specified that the blending function is used to distinguish a wall boundary layer and a free shear layer to deactivate the production term in the wall boundary layer only. However, this function depends on the dissipation rate, which is not available in the S-A model. The paper says that the blending function has been turned off manually, but does not explicitly specify how this is done. Can anyone suggest how this is done? Any help would be greatly appreciated.

November 30, 2012, 19:30		#10
TedBrogan New Member patty Join Date: Nov 2010 Posts: 21 Rep Power: 15	Adrian, I am trying to put in the SA transition model currently also. I ran into the same problem with how to handle the F_theta blending function. Do you know how your friend ended up resolving that?