Mesh Independence - Valid Strategy?

Ianto · August 27, 2009, 09:04

Dear All,

I'm running a mesh independance study - solve same problem on coarse, medium & fine meshes & examining solution from each, in the hope my parameter (head loss) converges towards the analytical solution as the mesh is refined & gives an indication of the mesh size necessary for acceptably small error. I have a convoluted passage which is my geometry of interest, to which I've added artificial inlet and outlet extensions. At the end of these I apply boundary conditions.

My question is with an eye on maintaining manageable mesh sizes - am I correct in thinking I can leave areas of the mesh which aren't of interest (extended inlet and outlets) unrefined, and only refine the areas I am studying? Is this still a valid mesh dependance study?

Thanks!

Ianto.

SKK · August 27, 2009, 09:17

In any kind of meshing, it is necessarry to have smooth transition between mesh sizes. So, if you are refining your mesh in the region of importance, you would have to refine your mesh in the extruded regions (especially near the interface of your interior to boundary region) aswell so the gradients etc are calculated properly and so their effect on the interior mesh. The mesh in outlet region will perhaps have more effect in the solution than that at the inlet owing to the effect of boundary values on the calculation of the interior cells.

shyamdsundar · August 31, 2009, 05:36

If I am right, what you intended to do is local refinement. It could give you a (grid) converged solution. But, what you are trying to do is to remove the errors only in that region. You still need to be sure that the errors generated in the rest of the region is smaller than the region of interest (which you refined). To check it, you can probably do two grid refinements (coarse to medium) with first one refined locally, and second one refined globally.

August 27, 2009, 09:04	Mesh Independence - Valid Strategy?	#1
Ianto New Member Join Date: Mar 2009 Posts: 9 Rep Power: 17	Dear All, I'm running a mesh independance study - solve same problem on coarse, medium & fine meshes & examining solution from each, in the hope my parameter (head loss) converges towards the analytical solution as the mesh is refined & gives an indication of the mesh size necessary for acceptably small error. I have a convoluted passage which is my geometry of interest, to which I've added artificial inlet and outlet extensions. At the end of these I apply boundary conditions. My question is with an eye on maintaining manageable mesh sizes - am I correct in thinking I can leave areas of the mesh which aren't of interest (extended inlet and outlets) unrefined, and only refine the areas I am studying? Is this still a valid mesh dependance study? Thanks! Ianto.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
[Other] vtk mesh or Abaqus mesh to OpenFOAM	bigphil	OpenFOAM Meshing & Mesh Conversion	27	November 23, 2015 18:31
fluent add additional zones for the mesh file	SSL	FLUENT	2	January 26, 2008 12:55
basic of mesh refinement	arya	CFX	4	June 19, 2007 13:21
unstructured mesh grid independence for Fluent	Shane Schouten	FLUENT	0	October 11, 2006 17:50
Mesh Independence and Y +	sam	FLUENT	1	September 13, 2006 06:50

August 27, 2009, 09:17		#2
SKK Member Join Date: Mar 2009 Posts: 55 Rep Power: 17	In any kind of meshing, it is necessarry to have smooth transition between mesh sizes. So, if you are refining your mesh in the region of importance, you would have to refine your mesh in the extruded regions (especially near the interface of your interior to boundary region) aswell so the gradients etc are calculated properly and so their effect on the interior mesh. The mesh in outlet region will perhaps have more effect in the solution than that at the inlet owing to the effect of boundary values on the calculation of the interior cells.

August 31, 2009, 05:36		#3
shyamdsundar New Member Shyam Join Date: Apr 2009 Posts: 29 Rep Power: 17	If I am right, what you intended to do is local refinement. It could give you a (grid) converged solution. But, what you are trying to do is to remove the errors only in that region. You still need to be sure that the errors generated in the rest of the region is smaller than the region of interest (which you refined). To check it, you can probably do two grid refinements (coarse to medium) with first one refined locally, and second one refined globally.