Pb in conv. k-e 3D Help

fluser · October 30, 2009, 06:27

Hello,
I am running a 3d simulation using the standard K-E model.

after 10 iterations "sliding mesh simulation" (in steady state flow) values of k & epsilon are diverging. I have the message: "Turbulent viscosity limited to viscosity ratio of 1e5 in cells".
I used pressure based solver.

Can anyone help with this please? Thanks.

CFD_4 · October 30, 2009, 06:28

pl. check mesh quality.

fluser · October 30, 2009, 06:44

I do grid--check, everything is good!

fluser · October 30, 2009, 07:18

and if I make simulation in laminar flow, solution converge!

tez · October 30, 2009, 16:51

Try with very small under-relaxation numbers and gradually increase them to default values. That might help.

fluser · November 2, 2009, 07:16

Hi Tez,
I divided UR factors by 100, results are better . but there are fluctuations of the epsilon parameter. After 100 iterations the residuals of k and e parameters are about 5.e-2.
tell me plz. how to increase gradually these parameters? by stopping and restart calculation and increasing URP without initializing?

greeneye · November 3, 2009, 07:42

I think it is better to give a inviscid solution and then set it as a intial condition for turbulent solution.another way is to set the "turbulent insity" and "turbulent viscousity ratio" to 1 at velocity inlet.

fluser · November 4, 2009, 06:26

Hi greeneye,
I apply your second proposition but k & epsilon residuals remind at too high values (about 1.e-1).
I have a simple question for you: how to use a solution as an initial condition in Fluent?
by iterating without initilize solution?
thanks

October 30, 2009, 06:27	Pb in conv. k-e 3D Help	#1
fluser New Member Mines Join Date: Oct 2009 Posts: 5 Rep Power: 17	Hello, I am running a 3d simulation using the standard K-E model. after 10 iterations "sliding mesh simulation" (in steady state flow) values of k & epsilon are diverging. I have the message: "Turbulent viscosity limited to viscosity ratio of 1e5 in cells". I used pressure based solver. Can anyone help with this please? Thanks.

October 30, 2009, 06:28	mesh	#2
CFD_4 New Member wayusen Join Date: Oct 2009 Location: Pune Posts: 6 Rep Power: 17	pl. check mesh quality.

November 3, 2009, 07:42	hello	#7
greeneye New Member reza Join Date: Nov 2009 Posts: 5 Rep Power: 17	I think it is better to give a inviscid solution and then set it as a intial condition for turbulent solution.another way is to set the "turbulent insity" and "turbulent viscousity ratio" to 1 at velocity inlet.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Problems with the RSM in simpleFoam	sberg	OpenFOAM Running, Solving & CFD	10	February 25, 2014 20:39
[Commercial meshers] Converting Gambit Neutral Meshes	plmauk	OpenFOAM Meshing & Mesh Conversion	2	July 20, 2007 08:08
FvVectorMatrixresidual before solving	cosimobianchini	OpenFOAM Running, Solving & CFD	0	January 28, 2007 12:02
[Gmsh] Gmsh with physicals	Rasmus Gjesing (Gjesing)	OpenFOAM Meshing & Mesh Conversion	9	February 2, 2006 15:36
conv variable in cfx4.4	Tad	CFX	1	March 12, 2005 12:42

October 30, 2009, 06:44		#3
fluser New Member Mines Join Date: Oct 2009 Posts: 5 Rep Power: 17	I do grid--check, everything is good!

October 30, 2009, 07:18		#4
fluser New Member Mines Join Date: Oct 2009 Posts: 5 Rep Power: 17	and if I make simulation in laminar flow, solution converge!

October 30, 2009, 16:51		#5
tez New Member Tez Join Date: Oct 2009 Posts: 12 Rep Power: 17	Try with very small under-relaxation numbers and gradually increase them to default values. That might help.

November 2, 2009, 07:16		#6
fluser New Member Mines Join Date: Oct 2009 Posts: 5 Rep Power: 17	Hi Tez, I divided UR factors by 100, results are better . but there are fluctuations of the epsilon parameter. After 100 iterations the residuals of k and e parameters are about 5.e-2. tell me plz. how to increase gradually these parameters? by stopping and restart calculation and increasing URP without initializing?

November 4, 2009, 06:26		#8
fluser New Member Mines Join Date: Oct 2009 Posts: 5 Rep Power: 17	Hi greeneye, I apply your second proposition but k & epsilon residuals remind at too high values (about 1.e-1). I have a simple question for you: how to use a solution as an initial condition in Fluent? by iterating without initilize solution? thanks