user defined turbulence model

manuutin · July 31, 2009, 07:13

Hi! I'm trying to implement a user defined turbulence model in star-cd 3.26 through subroutines vistur.f and sorkep.f, but it seems that the dissipation equation can't be activated when turbulent viscosity is user defined (vistur.f). Does anyone know a way around this (a switch maybe)?

Another thing: one can't modify the source terms in the wall adjacent cells through sorkep.f at least when standard k-eps model is used. Is this possible if turbulent viscosity is user defined (vistur.f) or does star force the no-good wall functions in that case too

?

Regards,
Mika Nuutinen

sdwijeya · September 30, 2009, 22:11

How do you know that the dissipation equation is not working? I tried a case to incorporate streamline curveture effect into k-epsilon equation but did not get physically realistic results. The flow was accelerating and I couldnot figure out whats going on.
When I tried restricting vistur.f to only where the streamlines curve using cell sets, the code diverges.

manuutin · October 1, 2009, 05:08

Just by activating sorkep.f and modifying the sources only in the wall adjacent cells (through cell sets) changes nothing because star computes these sources internally based on the wall functions. I learned that Switch 12 allows to use modify k and eps through sorkep.f when vistur.f is activated. Other tricks are still required to modify the wall adjacent source terms....

A_Lyaskin · October 12, 2009, 08:59

Quote:

Originally Posted by sdwijeya

How do you know that the dissipation equation is not working?

You can see it in the run summary or at the Equation Status tab in Primary Variables - when you turn on VISTUR subroutine both k and eps will be off. You can turn k back on with SOLVE command but this doesn't work for eps.

Actually I'm facing the similar problem as manuutin - I want to use standard k-eps equations but calculate turbulent viscosity in a non-standard way. I was thinking about defining a passive scalar and treating it as eps (adding the appropriate source term for its transport equation and etc.) but it doesn't look like as reasonable solution...

SKK · October 13, 2009, 12:15

If I remeber correctly, there is a trick to modify vist via posdat.f. It has been ages since I've used STAR CD so I do not remeber the details properly. But if you make a search on this forum around the period of 2004-5 you might be able to find something. Using posdat you can modify data at different stages of solution, ie what goes into different equations. I think it was something like If (level.eq.1) then... using this you can change your VIST = VIST * correction or however you want to calculate it. This lets you use STAR's in build k-e solver but allowes modification ti VIST. You might have to under-relax a lot in order to avoid divergance at the initial few iterations. Hope this helps.

A_Lyaskin · October 14, 2009, 06:29

Thanks, SKK! I'll try the posdat hack.

July 31, 2009, 07:13	user defined turbulence model	#1
manuutin New Member Mika Nuutinen Join Date: Jul 2009 Posts: 7 Rep Power: 17	Hi! I'm trying to implement a user defined turbulence model in star-cd 3.26 through subroutines vistur.f and sorkep.f, but it seems that the dissipation equation can't be activated when turbulent viscosity is user defined (vistur.f). Does anyone know a way around this (a switch maybe)? Another thing: one can't modify the source terms in the wall adjacent cells through sorkep.f at least when standard k-eps model is used. Is this possible if turbulent viscosity is user defined (vistur.f) or does star force the no-good wall functions in that case too ? Regards, Mika Nuutinen

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Modified k-e turbulence model UDF	Travis	Fluent UDF and Scheme Programming	7	November 11, 2018 21:21
Fan heater model: what turbulence source to use?	andy20	CFX	7	March 3, 2008 17:42
validating turbulence model on flow around a car	Pedro	CFX	1	February 20, 2008 17:32
turbulence model question	Jason Wei	Main CFD Forum	1	May 6, 2003 01:45

September 30, 2009, 22:11		#2
sdwijeya New Member Sameera Wijeyakulasuriya Join Date: Aug 2009 Posts: 12 Rep Power: 17	How do you know that the dissipation equation is not working? I tried a case to incorporate streamline curveture effect into k-epsilon equation but did not get physically realistic results. The flow was accelerating and I couldnot figure out whats going on. When I tried restricting vistur.f to only where the streamlines curve using cell sets, the code diverges.

October 1, 2009, 05:08		#3
manuutin New Member Mika Nuutinen Join Date: Jul 2009 Posts: 7 Rep Power: 17	Just by activating sorkep.f and modifying the sources only in the wall adjacent cells (through cell sets) changes nothing because star computes these sources internally based on the wall functions. I learned that Switch 12 allows to use modify k and eps through sorkep.f when vistur.f is activated. Other tricks are still required to modify the wall adjacent source terms....

October 13, 2009, 12:15		#5
SKK Member Join Date: Mar 2009 Posts: 55 Rep Power: 17	If I remeber correctly, there is a trick to modify vist via posdat.f. It has been ages since I've used STAR CD so I do not remeber the details properly. But if you make a search on this forum around the period of 2004-5 you might be able to find something. Using posdat you can modify data at different stages of solution, ie what goes into different equations. I think it was something like If (level.eq.1) then... using this you can change your VIST = VIST * correction or however you want to calculate it. This lets you use STAR's in build k-e solver but allowes modification ti VIST. You might have to under-relax a lot in order to avoid divergance at the initial few iterations. Hope this helps.

October 14, 2009, 06:29		#6
A_Lyaskin Member Anton Lyaskin Join Date: Mar 2009 Location: Samara, Russia Posts: 66 Rep Power: 17	Thanks, SKK! I'll try the posdat hack.