[Other] Non-convergence for smaller mesh spacings

fferroni · January 11, 2012, 04:22

Hello.

I'm running some simulations of MHD duct flow using mhdFoam, and I ran some a case with a square duct cross section mesh of 20x20 and 40x40 elements and they both converged. I tried running a 100x100 case and now it doesn't converge! How is this possible? I swear I haven't changed anything else. I ran checkMesh and everything was good, and it looks alright on Paraview too...

Anyone wish to point out possible reasons?

Regards,

Fran

akidess · January 13, 2012, 03:07

You have to keep an eye on the Courant number. If you refine your mesh without reducing the time step, you are increasing the Courant number. If it gets too large, the solution algorithm will become unstable.

fferroni · January 13, 2012, 06:42

Ah ok. So for an evenly spaced mesh, the time-step needs to be reduced proportionally to the decrease in mesh spacing?
Is this the condition you are referring to? http://en.wikipedia.org/wiki/Courant...Lewy_condition

Thank you. I will see if it works!

Regards,

Fran

akidess · January 13, 2012, 08:18

Yes, except that in this case the limitation is not due to an explicit time integration scheme, but to maintain pressure-velocity coupling with the PISO-algorithm.

January 11, 2012, 04:22	Non-convergence for smaller mesh spacings	#1
fferroni Member Join Date: Nov 2011 Posts: 44 Rep Power: 15	Hello. I'm running some simulations of MHD duct flow using mhdFoam, and I ran some a case with a square duct cross section mesh of 20x20 and 40x40 elements and they both converged. I tried running a 100x100 case and now it doesn't converge! How is this possible? I swear I haven't changed anything else. I ran checkMesh and everything was good, and it looks alright on Paraview too... Anyone wish to point out possible reasons? Regards, Fran

January 13, 2012, 03:07		#2
akidess Senior Member Anton Kidess Join Date: May 2009 Location: Germany Posts: 1,377 Rep Power: 30	You have to keep an eye on the Courant number. If you refine your mesh without reducing the time step, you are increasing the Courant number. If it gets too large, the solution algorithm will become unstable. __________________ On twitter @akidTwit Spend as much time formulating your questions as you expect people to spend on their answer.

January 13, 2012, 08:18		#4
akidess Senior Member Anton Kidess Join Date: May 2009 Location: Germany Posts: 1,377 Rep Power: 30	Yes, except that in this case the limitation is not due to an explicit time integration scheme, but to maintain pressure-velocity coupling with the PISO-algorithm. __________________ On twitter @akidTwit Spend as much time formulating your questions as you expect people to spend on their answer.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Convergence	Centurion2011	FLUENT	48	June 15, 2022 00:29
sliding mesh problem in CFX	Saima	CFX	46	September 11, 2021 08:38
Convergence problem with mesh?	tareqkh	Main CFD Forum	0	July 13, 2016 18:38
[mesh manipulation] Importing Multiple Meshes	thomasnwalshiii	OpenFOAM Meshing & Mesh Conversion	18	December 19, 2015 19:57
Influence of mesh refinement on the convergence	saisanthoshm88	CFX	6	November 26, 2010 07:58

January 13, 2012, 06:42		#3
fferroni Member Join Date: Nov 2011 Posts: 44 Rep Power: 15	Ah ok. So for an evenly spaced mesh, the time-step needs to be reduced proportionally to the decrease in mesh spacing? Is this the condition you are referring to? http://en.wikipedia.org/wiki/Courant...Lewy_condition Thank you. I will see if it works! Regards, Fran