convergence problems with simpleFoam

inf.vish · August 1, 2013, 08:34

Hi all,
I am trying to achieve a pressure driven steady state laminar flow of water through a pipe. (2D case). I am not able to achieve a convergence. But I solved the same case with velocity driven flow and the convergence reached in 105 iterations.
I am attaching my system directory as well as 0 directory.

looking forward to some solution.

0.zip

system.zip

constant.zip

gillimaniac · August 1, 2013, 10:28

Hey there,

first of all in "constant" you define the kinematic viscosity nu = 858e-9 but the kinematic viscosity of water at 20°C is 1e-6 m²/s i believe.

Additionally your applied pressure difference between inlet and outlet is very low. It calculates to:

1e-4 m²/s² * 1000 kg/m³ = 0,1 kg/(m*s²) = 0,1 Pa.

Cheers

inf.vish · August 1, 2013, 11:19

Oh my bad. Thanks for pointing out the mistake. I will try to rectify and run the simulation again. Thanks again

inf.vish · August 2, 2013, 05:29

The solution still won't converge. Can you help. I changed the value of nu to 1e-06 and pressure to 1 m^2/s^2

gillimaniac · August 2, 2013, 06:43

How does your case look after maybe 100 iterations?

Do you have perhaps inflow at your outlet?
In this case you can try using inletOutlet instead of zeroGradient for U @ outlet at first.

Perhaps you can plot your initial residuals and share the graph here.

gillimaniac · August 2, 2013, 06:45

Okay i took your case files, changed the nu and p values and the case converged after 224 iterations so there should be no problem.

please post your terminal content after some iterations.

inf.vish · August 2, 2013, 07:00

Yeah it converged after 224 iterations. You are absolutely correct about that. But did you look at the results? My results are very weird. The velocity distribution looks odd. It should have an entrance region and then gradually become parabolic.

inf.vish · August 2, 2013, 07:02

U.jpg

Just check the above image.

inf.vish · August 2, 2013, 08:36

Here is my test case again. Please tell me where am i going wrong. As far as i know the boundary conditions conform with those for Poiseuille flow.

2D_pipe_flow_pres_induced.zip

August 1, 2013, 08:34	convergence problems with simpleFoam	#1
inf.vish Member Vishal Join Date: Jul 2013 Posts: 73 Rep Power: 13	Hi all, I am trying to achieve a pressure driven steady state laminar flow of water through a pipe. (2D case). I am not able to achieve a convergence. But I solved the same case with velocity driven flow and the convergence reached in 105 iterations. I am attaching my system directory as well as 0 directory. looking forward to some solution. 0.zip system.zip constant.zip

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
SimpleFoam convergence problems	brahim	OpenFOAM Running, Solving & CFD	20	June 9, 2015 10:09
force convergence problems in CFX 6DOF rigid body solver	ajay_ks	CFX	8	March 25, 2013 05:02
Convergence and steady state using simpleFoam	sfigato	OpenFOAM Running, Solving & CFD	0	February 8, 2013 05:14
NACA0012 Convergence Problems	StudentAndrew	CFX	6	November 21, 2005 07:49
Convergence problems	Chetan	FLUENT	3	April 15, 2004 20:13

August 1, 2013, 10:28		#2
gillimaniac New Member Stefan Gaerling Join Date: Dec 2012 Posts: 22 Rep Power: 13	Hey there, first of all in "constant" you define the kinematic viscosity nu = 858e-9 but the kinematic viscosity of water at 20°C is 1e-6 m²/s i believe. Additionally your applied pressure difference between inlet and outlet is very low. It calculates to: 1e-4 m²/s² * 1000 kg/m³ = 0,1 kg/(m*s²) = 0,1 Pa. Cheers

August 1, 2013, 11:19		#3
inf.vish Member Vishal Join Date: Jul 2013 Posts: 73 Rep Power: 13	Oh my bad. Thanks for pointing out the mistake. I will try to rectify and run the simulation again. Thanks again

August 2, 2013, 05:29		#4
inf.vish Member Vishal Join Date: Jul 2013 Posts: 73 Rep Power: 13	The solution still won't converge. Can you help. I changed the value of nu to 1e-06 and pressure to 1 m^2/s^2

August 2, 2013, 06:43		#5
gillimaniac New Member Stefan Gaerling Join Date: Dec 2012 Posts: 22 Rep Power: 13	How does your case look after maybe 100 iterations? Do you have perhaps inflow at your outlet? In this case you can try using inletOutlet instead of zeroGradient for U @ outlet at first. Perhaps you can plot your initial residuals and share the graph here.

August 2, 2013, 06:45		#6
gillimaniac New Member Stefan Gaerling Join Date: Dec 2012 Posts: 22 Rep Power: 13	Okay i took your case files, changed the nu and p values and the case converged after 224 iterations so there should be no problem. please post your terminal content after some iterations.

August 2, 2013, 07:00		#7
inf.vish Member Vishal Join Date: Jul 2013 Posts: 73 Rep Power: 13	Yeah it converged after 224 iterations. You are absolutely correct about that. But did you look at the results? My results are very weird. The velocity distribution looks odd. It should have an entrance region and then gradually become parabolic.

August 2, 2013, 07:02		#8
inf.vish Member Vishal Join Date: Jul 2013 Posts: 73 Rep Power: 13	U.jpg Just check the above image.

August 2, 2013, 08:36		#9
inf.vish Member Vishal Join Date: Jul 2013 Posts: 73 Rep Power: 13	Here is my test case again. Please tell me where am i going wrong. As far as i know the boundary conditions conform with those for Poiseuille flow. 2D_pipe_flow_pres_induced.zip