[ICEM] Improve hexa mesh quality and determinant

Ramon Quiñonez · January 29, 2019, 13:47

Hello,

I am currently working with the mesh of a flame stabiliser in a quarter of a combustion chamber. I am considering only hexahedral elements, and a annular feed at the outlet face of the contraction. I have been working with this geometry in the past producing sensible meshes, but this one has a simulated weld at the inlet of the contraction, so i just added more ogrids to the mesh, and the quality is very bad, and also the determinant. Could you please help me with some tips on how to improve my mesh?
I attach some photos, they show my current geometry and mesh, as you can see I have created some ogrids for the annular feed and the simulated welding.
1.jpg

2.jpg

6.jpg

7.jpg

9.jpg

Many thanks!

bluebase · January 30, 2019, 04:35

Hi José,

it seems you have merged some blocks/vertices. So you end up with degenerated blocks and negative cells.
See the min/max of the determinant. The minimum is actually negative, which tells there are cells with negative volume (respective to the convention).

If cell count is not a big issue here, i suggest the following.
Project the block topology in the middle section onto the domain's end faces. Do not merge any blocks, move all vertices which are in-line in the z direction to the the same respective x-y-coordinates. The edges approx. in z direction should end up parallel with the axis. Or with other words, looking onto the blocking in the x-y-plane should show an almost congruent blocking.

It might be not clear enough what i mean. If that's the case try to understand your blocking a bit better first:
Use the feature Scan Planes to figure out how the blocks are formed; where they end up. You can find this feature in the context menu of Premesh (in the project tree).

It allows to draw selected layers of the premesh, so you can visualize their texture.

Best regards,
Sebastian

January 29, 2019, 13:47	Improve hexa mesh quality and determinant	#1
Ramon Quiñonez New Member José Ramón Quiñonez Arce Join Date: Oct 2017 Location: Leeds, UK Posts: 17 Rep Power: 8	Hello, I am currently working with the mesh of a flame stabiliser in a quarter of a combustion chamber. I am considering only hexahedral elements, and a annular feed at the outlet face of the contraction. I have been working with this geometry in the past producing sensible meshes, but this one has a simulated weld at the inlet of the contraction, so i just added more ogrids to the mesh, and the quality is very bad, and also the determinant. Could you please help me with some tips on how to improve my mesh? I attach some photos, they show my current geometry and mesh, as you can see I have created some ogrids for the annular feed and the simulated welding. 1.jpg 2.jpg 6.jpg 7.jpg 9.jpg Many thanks!

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
[snappyHexMesh] sHM layer process keeps getting killed	MBttR	OpenFOAM Meshing & Mesh Conversion	4	August 15, 2016 03:21
[ICEM] Heart modeling - ICEM Hexa Mesh quality improvement	heart_modeller	ANSYS Meshing & Geometry	1	April 8, 2016 08:16
[snappyHexMesh] No layers in a small gap	bobburnquist	OpenFOAM Meshing & Mesh Conversion	6	August 26, 2015 09:38
snappyhexmesh remove blockmesh geometry	philipp1	OpenFOAM Running, Solving & CFD	2	December 12, 2014 10:58

January 30, 2019, 04:35		#2
bluebase Senior Member Sebastian Engel Join Date: Jun 2011 Location: Germany Posts: 566 Rep Power: 20	Hi José, it seems you have merged some blocks/vertices. So you end up with degenerated blocks and negative cells. See the min/max of the determinant. The minimum is actually negative, which tells there are cells with negative volume (respective to the convention). If cell count is not a big issue here, i suggest the following. Project the block topology in the middle section onto the domain's end faces. Do not merge any blocks, move all vertices which are in-line in the z direction to the the same respective x-y-coordinates. The edges approx. in z direction should end up parallel with the axis. Or with other words, looking onto the blocking in the x-y-plane should show an almost congruent blocking. It might be not clear enough what i mean. If that's the case try to understand your blocking a bit better first: Use the feature Scan Planes to figure out how the blocks are formed; where they end up. You can find this feature in the context menu of Premesh (in the project tree). It allows to draw selected layers of the premesh, so you can visualize their texture. Best regards, Sebastian