solver setting

eb.nabizadeh · July 26, 2012, 17:04

can any one help me please?
i have a 3D model(a house)that i want to check temp & flow in it.my model is transient & the volume of it is 400m3,walls are at initial temp 20 c and when wind blowing into it exchange it's heat with walls.but i have some problem in convergence.i want some one to say exact setting about my model.what under-relaxation factor is good for it,which viscose model is better.finally what must i do to converge my problem.thanks

eb.nabizadeh · July 26, 2012, 17:15

i don't have much time please help me early

Bionico · July 27, 2012, 02:12

Hello

If I were you I would try first with a k-epsilon model (Realizable o RNG).
Maybe it's better to reduce the URF (M=0.2 E=0.9 should be enough).
First order upwind for discretization and PRESTO! for pressure scheme.

Sixkillers · July 27, 2012, 03:55

Also It depends which model you use for air density. Are using ideal gas model or Boussinesq approximation? Generally speaking ideal gas model is harder to converge.

eb.nabizadeh · July 27, 2012, 09:50

i use Boussinesq model for density.i want to know what criterion(surface integrals) can i consider except residual?

eb.nabizadeh · July 28, 2012, 04:16

dear Bionico
when in do what you said i facing this error:floating point exception!
and when i change my turbulence model it didn't converge.
can you help me more?

Bionico · July 30, 2012, 02:42

ok, it's a convergence problem...
since this is a transient simulation, try reducing the Courant number under 1

eb.nabizadeh · July 30, 2012, 17:45

you mean i use density based or pressure based?

Bionico · July 31, 2012, 02:01

No, but maybe (I don't remember very well) the Courant number can be controlled only in COUPLED problem (not if you have a SIMPLE one).
Therefore try reducing time step and increasing number of iteration/time step

July 26, 2012, 17:04	solver setting	#1
eb.nabizadeh New Member nabizadeh Join Date: May 2012 Location: iran Posts: 23 Rep Power: 13	can any one help me please? i have a 3D model(a house)that i want to check temp & flow in it.my model is transient & the volume of it is 400m3,walls are at initial temp 20 c and when wind blowing into it exchange it's heat with walls.but i have some problem in convergence.i want some one to say exact setting about my model.what under-relaxation factor is good for it,which viscose model is better.finally what must i do to converge my problem.thanks

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
[Other] A New Solver for Supersonic Combustion	nakul	OpenFOAM Community Contributions	20	February 22, 2019 09:08
Quarter Burner mesh with periosic condition	SamCanuck	FLUENT	2	August 31, 2011 11:34
Working directory via command line	Luiz	CFX	4	March 6, 2011 20:02
Cells with t below lower limit	Purushothama	Siemens	2	May 31, 2010 21:58
why the solver reject it? Anyone with experience?	bearcat	CFX	6	April 28, 2008 14:08

July 26, 2012, 17:15		#2
eb.nabizadeh New Member nabizadeh Join Date: May 2012 Location: iran Posts: 23 Rep Power: 13	i don't have much time please help me early

July 27, 2012, 02:12		#3
Bionico Senior Member Flavio Join Date: Sep 2011 Location: Brescia, Italy Posts: 181 Rep Power: 15	Hello If I were you I would try first with a k-epsilon model (Realizable o RNG). Maybe it's better to reduce the URF (M=0.2 E=0.9 should be enough). First order upwind for discretization and PRESTO! for pressure scheme.

July 27, 2012, 03:55		#4
Sixkillers Member Join Date: Nov 2011 Location: Czech Republic Posts: 97 Rep Power: 14	Also It depends which model you use for air density. Are using ideal gas model or Boussinesq approximation? Generally speaking ideal gas model is harder to converge.

July 27, 2012, 09:50		#5
eb.nabizadeh New Member nabizadeh Join Date: May 2012 Location: iran Posts: 23 Rep Power: 13	i use Boussinesq model for density.i want to know what criterion(surface integrals) can i consider except residual?

July 28, 2012, 04:16		#6
eb.nabizadeh New Member nabizadeh Join Date: May 2012 Location: iran Posts: 23 Rep Power: 13	dear Bionico when in do what you said i facing this error:floating point exception! and when i change my turbulence model it didn't converge. can you help me more?

July 30, 2012, 02:42		#7
Bionico Senior Member Flavio Join Date: Sep 2011 Location: Brescia, Italy Posts: 181 Rep Power: 15	ok, it's a convergence problem... since this is a transient simulation, try reducing the Courant number under 1

July 30, 2012, 17:45		#8
eb.nabizadeh New Member nabizadeh Join Date: May 2012 Location: iran Posts: 23 Rep Power: 13	you mean i use density based or pressure based?

July 31, 2012, 02:01		#9
Bionico Senior Member Flavio Join Date: Sep 2011 Location: Brescia, Italy Posts: 181 Rep Power: 15	No, but maybe (I don't remember very well) the Courant number can be controlled only in COUPLED problem (not if you have a SIMPLE one). Therefore try reducing time step and increasing number of iteration/time step