Gamma-theta turbulence model for transitional flow

kobe81 · October 10, 2016, 12:07

Hello,

I have to simulate a complex domain with both laminar and turbulent regions and have been suggested to use the gamma-theta turbulence model.
In order to understand its behavior I have conducted studies on simpler geometries before going to the real problem. In circular pipe flows the model has provided great results when compared to both analytical solutions for laminar flows and PIV experiments for turbulent ones, with mesh requirements of y+<1.

I have then tried to simulate a sudden contraction (5:1 diameter), to check the quality of the solution when the laminar and turbulent flows happened in the same domain, but I couldn't get the "right results". I have modelled two long tubes before and after the contraction to be able to compare the velocity profiles fully developed.
When the contraction has laminar-laminar or turbulent-turbulent reynolds it works well, but the results after the contraction are always wrong for laminar-turbulent flows (it seems like the turbulent side is still treated as laminar flow --> Vmax = 2Vave for example). The results before the contraction are fine.

So I tried to simulate the flow over a flat plate and compare the friction coefficient results to the ones showed in section 4.1.10 of the CFX manual, and once again my results are wrong. The image attached shows some of the results I obtained for the plate with different mesh conditions, but all with y+ = 0,05.

I haven't changed any constants on the gamma-theta model, does anyone know if its "standard formulation" has any limitations for this kind of simulation I am doing?

Thank you very much!

ghorrocks · October 10, 2016, 18:26

The gamma theta turbulence transition model is not a general turbulence transition model. It has been empirically tuned to work well on airfoils and turbomachinery blades. Any use outside that type of model is at your own risk.

So it does not surprise me that it does not handle a sudden contraction pipe flow to well. This is quite different to the airfoil condition so the model is probably not very appropriate.

I would think you should be able to get pretty good results on turbulence transition on flat plate however. This would be very sensitive to free stream turbulence conditions, the exact value of roughness on the plate and many other factors. Are you sure you have these issues correctly modelled?

danishrehman · May 22, 2017, 06:57

Hello Kobe81,

I am also trying to simulate transitional flow in pipes. Have you got any success with contraction at the inlet?

- When you say, it is providing excellent results if only pipe geometry is considered, which mesh are you considering? Structured or unstrutured?

- I have tried it with y+<1 structured grid, however it does not seem to give correct results (experimental data) of pressure loss in the pipe.

- I have also come across a numerical study "Breakdown of Laminar pipeflow into transitional intermittency and subsequent attainment of fully developed intermittent or Turbulent flow " Numerical Heat Transfer, Part B: Fundamentals. In this study author modifies gamma theta model coefficients in order to adopt the model to internal transitional flows. And results are quite good bridge between Poisuelle's law in laminar regime and Blausius equation in turbulent regime.

- Although I have modified the constants, its still not the same as of experimental data.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Issues on the simulation of high-speed compressible flow within turbomachinery	dowlee	OpenFOAM Running, Solving & CFD	11	August 6, 2021 07:40
Transitional flow (Gamma Theta model)	sharif88	CFX	3	March 6, 2014 16:59
SST turbulence with Gamma theta and Boussinesq model	kmgraju	CFX	2	June 11, 2013 01:46
Turbulence model for low Reynolds number flow?	Nokadu	Main CFD Forum	3	May 26, 2013 12:42
validating turbulence model on flow around a car	Pedro	CFX	1	February 20, 2008 17:32

October 10, 2016, 18:26		#2
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,871 Rep Power: 144	The gamma theta turbulence transition model is not a general turbulence transition model. It has been empirically tuned to work well on airfoils and turbomachinery blades. Any use outside that type of model is at your own risk. So it does not surprise me that it does not handle a sudden contraction pipe flow to well. This is quite different to the airfoil condition so the model is probably not very appropriate. I would think you should be able to get pretty good results on turbulence transition on flat plate however. This would be very sensitive to free stream turbulence conditions, the exact value of roughness on the plate and many other factors. Are you sure you have these issues correctly modelled?

May 22, 2017, 06:57		#3
danishrehman New Member Danish Join Date: Jan 2012 Posts: 4 Rep Power: 14	Hello Kobe81, I am also trying to simulate transitional flow in pipes. Have you got any success with contraction at the inlet? - When you say, it is providing excellent results if only pipe geometry is considered, which mesh are you considering? Structured or unstrutured? - I have tried it with y+<1 structured grid, however it does not seem to give correct results (experimental data) of pressure loss in the pipe. - I have also come across a numerical study "Breakdown of Laminar pipeflow into transitional intermittency and subsequent attainment of fully developed intermittent or Turbulent flow " Numerical Heat Transfer, Part B: Fundamentals. In this study author modifies gamma theta model coefficients in order to adopt the model to internal transitional flows. And results are quite good bridge between Poisuelle's law in laminar regime and Blausius equation in turbulent regime. - Although I have modified the constants, its still not the same as of experimental data.