time scale related to mesh size

yaseen wsu · May 6, 2016, 11:08

hello
I simulated free surface flow using coarse mesh with time scale 0.025 sec after about 600 iteration it reaches my residual (10e-4), but fir the same model with fine mesh and the same time scale, after 1500 iteration it does not reach the required residual. does physical time scale related to mesh size??
thanks

ghorrocks · May 7, 2016, 00:40

You usually need a smaller time step with finer meshes. It does not necessarily scale, but that is often a useful starting point. Use adaptive time stepping to find the necessary time step size and it takes care of this for you. Also, finer meshes usually are harder to converge as they have less numerical dissipation and are therefore less numerically stable. So you are getting the expected behaviour.

yaseen wsu · May 7, 2016, 02:30

Quote:

Originally Posted by ghorrocks

You usually need a smaller time step with finer meshes. It does not necessarily scale, but that is often a useful starting point. Use adaptive time stepping to find the necessary time step size and it takes care of this for you. Also, finer meshes usually are harder to converge as they have less numerical dissipation and are therefore less numerically stable. So you are getting the expected behaviour.

thanks a lot
but my case is steady state, is there time step adaption in steady state??? I only see time scale

ghorrocks · May 7, 2016, 06:01

In steady state your are free to adjust the time step size as required. Make it bigger if it is converging consistently but slowly, make it smaller if it is diverging or inconsistent convergence.

yaseen wsu · May 9, 2016, 03:55

dear glenn
I used automatic time scale with time scale factor 0.5 sec, but there are fluctuation in the equation residuals and not reach 10^-4 as shown in in fig. what does that reason?? because when I use 1 sec as time scale factor it gives error, is that due to small time or large tome scale, thanks

https://drive.google.com/open?id=0B3...HctaFA1TkwxeUE

ghorrocks · May 9, 2016, 08:01

I know you have said you have already read this, but this thread really is covered by this FAQ: http://www.cfd-online.com/Wiki/Ansys...gence_criteria

You really need to consider the issues it discusses.

Opaque · May 9, 2016, 16:53

Have you selected the option to output the equation residuals to the results file ?

One the equation residuals are in the results file, you can post-process them to find out where the maximum residual for each equation is located. Look carefully at that region of space for mesh quality issues, or specific flow features that may require a better mesh.

A bad mesh (coarse or fine) could prevent convergence of the iterative procedure.

Hope the above helps,

May 6, 2016, 11:08	time scale related to mesh size	#1
yaseen wsu Senior Member yaseen Join Date: Oct 2015 Location: Hawler Posts: 174 Rep Power: 11	hello I simulated free surface flow using coarse mesh with time scale 0.025 sec after about 600 iteration it reaches my residual (10e-4), but fir the same model with fine mesh and the same time scale, after 1500 iteration it does not reach the required residual. does physical time scale related to mesh size?? thanks

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
decomposePar problem: Cell 0contains face labels out of range	vaina74	OpenFOAM Pre-Processing	37	July 20, 2020 05:38
dynamic Mesh is faster than MRF????	sharonyue	OpenFOAM Running, Solving & CFD	14	August 26, 2013 07:47
OF 1.6 \| Ubuntu 9.10 (64bit) \| GLIBCXX_3.4.11 not found	piprus	OpenFOAM Installation	22	February 25, 2010 13:43
Modeling in micron scale using icoFoam	m9819348	OpenFOAM Running, Solving & CFD	7	October 27, 2007 00:36
IcoFoam parallel woes	msrinath80	OpenFOAM Running, Solving & CFD	9	July 22, 2007 02:58

May 7, 2016, 00:40		#2
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,943 Rep Power: 145	You usually need a smaller time step with finer meshes. It does not necessarily scale, but that is often a useful starting point. Use adaptive time stepping to find the necessary time step size and it takes care of this for you. Also, finer meshes usually are harder to converge as they have less numerical dissipation and are therefore less numerically stable. So you are getting the expected behaviour.

May 7, 2016, 06:01		#4
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,943 Rep Power: 145	In steady state your are free to adjust the time step size as required. Make it bigger if it is converging consistently but slowly, make it smaller if it is diverging or inconsistent convergence.

May 9, 2016, 03:55		#5
yaseen wsu Senior Member yaseen Join Date: Oct 2015 Location: Hawler Posts: 174 Rep Power: 11	dear glenn I used automatic time scale with time scale factor 0.5 sec, but there are fluctuation in the equation residuals and not reach 10^-4 as shown in in fig. what does that reason?? because when I use 1 sec as time scale factor it gives error, is that due to small time or large tome scale, thanks https://drive.google.com/open?id=0B3...HctaFA1TkwxeUE

May 9, 2016, 08:01		#6
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,943 Rep Power: 145	I know you have said you have already read this, but this thread really is covered by this FAQ: http://www.cfd-online.com/Wiki/Ansys...gence_criteria You really need to consider the issues it discusses.

May 9, 2016, 16:53		#7
Opaque Senior Member Join Date: Jun 2009 Posts: 1,901 Rep Power: 33	Have you selected the option to output the equation residuals to the results file ? One the equation residuals are in the results file, you can post-process them to find out where the maximum residual for each equation is located. Look carefully at that region of space for mesh quality issues, or specific flow features that may require a better mesh. A bad mesh (coarse or fine) could prevent convergence of the iterative procedure. Hope the above helps,