C3X Fully turbulent convergence problem

Mazze[ITA] · May 9, 2011, 04:44

Hi all, I'm doing a comparison with C3X 1988 blade test case between SST Fully turbulent, Gamma model (Reth=250) and Gamma-Theta model. Every case starts with High order advection scheme, High order turbulence, Local timescale factor 5.
During my runs I experimented very good convergence in second and third case, but in the first one the solution doesn't converge, with mom residuals between 10-3 and 10-4.
I tried increasing the time step, without any enhancement. The mesh has a good y+ quality, except for the trailing edge (y+ max approximately 5), but unfortunately i can't modify it.
Any suggestions to achieve convergence?

Mazze[ITA] · May 9, 2011, 17:50

ok, i decided the solution has converged. Although residuals didn't fall, the temperature of the solid is stable and also the imbalance.
Unfortunately I still haven't found an explanation for that behaviour.

ghorrocks · May 9, 2011, 19:54

http://www.cfd-online.com/Wiki/Ansys...gence_criteria

Also note you should use local time scale factor to convergence. You should switch to physical time scale for the final run to convergence.

Mazze[ITA] · May 10, 2011, 03:13

Thank you Glenn. I've already read the documentation. About your last statement: that step is recommended in every case? Because, except for the fully turbulent case, in the other cases all parameters are steady and residuals are below 10-6.

ghorrocks · May 10, 2011, 19:58

Yes, you should always use a physical time scale on the final run to convergence. You cannot judge convergence by residuals alone.

May 9, 2011, 04:44	C3X Fully turbulent convergence problem	#1
Mazze[ITA] Member Lorenzo Mazzei Join Date: Dec 2010 Posts: 60 Rep Power: 15	Hi all, I'm doing a comparison with C3X 1988 blade test case between SST Fully turbulent, Gamma model (Reth=250) and Gamma-Theta model. Every case starts with High order advection scheme, High order turbulence, Local timescale factor 5. During my runs I experimented very good convergence in second and third case, but in the first one the solution doesn't converge, with mom residuals between 10-3 and 10-4. I tried increasing the time step, without any enhancement. The mesh has a good y+ quality, except for the trailing edge (y+ max approximately 5), but unfortunately i can't modify it. Any suggestions to achieve convergence?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
how to define turbulent intensity in 2d problem	blueberry	FLUENT	1	March 13, 2013 10:09
Convergence problem for P1 & Energy	HP	FLUENT	5	May 21, 2005 16:01
convergence problem	Trushar	Phoenics	5	August 28, 2002 00:40
convergence problem with SIMPLER	NURAY KAYAKOL	Main CFD Forum	1	February 24, 1999 14:43
the turbulent problem	Daejin, Kang	Main CFD Forum	0	November 16, 1998 11:52

May 9, 2011, 17:50		#2
Mazze[ITA] Member Lorenzo Mazzei Join Date: Dec 2010 Posts: 60 Rep Power: 15	ok, i decided the solution has converged. Although residuals didn't fall, the temperature of the solid is stable and also the imbalance. Unfortunately I still haven't found an explanation for that behaviour.

May 9, 2011, 19:54		#3
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,854 Rep Power: 144	http://www.cfd-online.com/Wiki/Ansys...gence_criteria Also note you should use local time scale factor to convergence. You should switch to physical time scale for the final run to convergence.

May 10, 2011, 03:13		#4
Mazze[ITA] Member Lorenzo Mazzei Join Date: Dec 2010 Posts: 60 Rep Power: 15	Thank you Glenn. I've already read the documentation. About your last statement: that step is recommended in every case? Because, except for the fully turbulent case, in the other cases all parameters are steady and residuals are below 10-6.

May 10, 2011, 19:58		#5
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,854 Rep Power: 144	Yes, you should always use a physical time scale on the final run to convergence. You cannot judge convergence by residuals alone.