Transient run taking too long time

Alexander Tham Keat Fu · June 19, 2012, 04:30

Hi all.

My name is Alexander. I am running transient run for solid-liquid flow in a horizontal annulus.

I am facing a problem, which is my run is taking too long time. It sums up to 3 days and still running.

Previously, I have make changes in my setup as I read the related posts in this forum.

My current time step is of 0.001 s and my run is computed over a period of 100s. My model length is only 2 ft!!!

May I know what are the potential problems that I may have not identified in my setup?

Thank you very much.

Regards,
Alexander Tham
tham.keatfu@gmail.com

oj.bulmer · June 19, 2012, 15:48

Have you considered using adaptive time-stepping?

In case if you choose to use that, make sure you select a reasonable residual target and convergence criteria. If you select too stringent and aggressive convergence target (i.e. small), adaptive time-stepping takes very small time-steps, and convergence takes like aeons.

Also, running local parallel may help in speeding things up.

Regards
OJ

Alexander Tham Keat Fu · June 19, 2012, 22:11

Thanks, OJ.

I will try using adaptive time step and let you know how is the progress. In addition to that, may I know in a normal transient simulation, how long would it takes?

Thanks. Appreciate your feedback.

Regards,
Alex

mauricio · June 20, 2012, 07:53

How long it'd take may depend on several things, the models that determine the amount of equations solved, the timestep size, the convergence criteria, the mesh resolution... Did you do a mesh sensitivity analysis to check you're not using a mesh that is too fine for this purpose?

June 19, 2012, 04:30	Transient run taking too long time	#1
Alexander Tham Keat Fu New Member Tham Keat Fu Join Date: Jun 2012 Posts: 2 Rep Power: 0	Hi all. My name is Alexander. I am running transient run for solid-liquid flow in a horizontal annulus. I am facing a problem, which is my run is taking too long time. It sums up to 3 days and still running. Previously, I have make changes in my setup as I read the related posts in this forum. My current time step is of 0.001 s and my run is computed over a period of 100s. My model length is only 2 ft!!! May I know what are the potential problems that I may have not identified in my setup? Thank you very much. Regards, Alexander Tham tham.keatfu@gmail.com

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Problem: Very long "write" time (~2h-3h) for results and transient results	Shawn_A	CFX	16	April 12, 2016 21:49
Superlinear speedup in OpenFOAM 13	msrinath80	OpenFOAM Running, Solving & CFD	18	March 3, 2015 06:36
Time step in transient simulation	shib	FLUENT	0	June 17, 2010 14:07
Modeling in micron scale using icoFoam	m9819348	OpenFOAM Running, Solving & CFD	7	October 27, 2007 01:36
Could anybody help me see this error and give help	liugx212	OpenFOAM Running, Solving & CFD	3	January 4, 2006 19:07

June 19, 2012, 15:48		#2
oj.bulmer Senior Member OJ Join Date: Apr 2012 Location: United Kindom Posts: 473 Rep Power: 20	Have you considered using adaptive time-stepping? In case if you choose to use that, make sure you select a reasonable residual target and convergence criteria. If you select too stringent and aggressive convergence target (i.e. small), adaptive time-stepping takes very small time-steps, and convergence takes like aeons. Also, running local parallel may help in speeding things up. Regards OJ

June 19, 2012, 22:11		#3
Alexander Tham Keat Fu New Member Tham Keat Fu Join Date: Jun 2012 Posts: 2 Rep Power: 0	Thanks, OJ. I will try using adaptive time step and let you know how is the progress. In addition to that, may I know in a normal transient simulation, how long would it takes? Thanks. Appreciate your feedback. Regards, Alex

June 20, 2012, 07:53		#4
mauricio Member Mauricio Labarca Join Date: Aug 2009 Posts: 35 Rep Power: 17	How long it'd take may depend on several things, the models that determine the amount of equations solved, the timestep size, the convergence criteria, the mesh resolution... Did you do a mesh sensitivity analysis to check you're not using a mesh that is too fine for this purpose?