Premesh CAD quality

Fridrik · January 10, 2018, 06:52

I have made a mesh of a very complex structure, and by using snappyHexMesh.

In the process there was a lot of failed meshes, and wondered if anybody has some kind of measure of the premesh quality of a CAD file?

I have an idea to a way of quantify this:
This is the length of the smallest line divided by the length of the average line
$\bar{S}=\frac{min(l_i)}{\frac{1}{n}\sum^n_{i=1}(l_i)}$

vishnupc · January 13, 2018, 11:05

Hi Fridrik,

When you say a 'failure' I would like to specifically know how ? I have not used snappyHexMesh and wanted to know if there are specific criteria setup to know that a mesh is a good one ?

Not sure if this answers your question but qualitatively I feel that the CAD is a good one if it is :

a. Connected as per the requirements of the simulation being performed (Based on which the geometry can have 1-sided or even 3+ sided edges)
b. Ideally should not have self intersections and coincident faces.

The main issue I have faced is that most geometries must be 'connected' as intended. This can be a little difficult when you are preparing a complex geometry prior to meshing. Once you DO do this check, your meshing procedure and subsequent cleanup should be not too difficult.

G'Luck !

January 10, 2018, 06:52	Premesh CAD quality	#1
Fridrik Member Fridrik Magnusson Join Date: Aug 2017 Location: Denmark Posts: 34 Rep Power: 8	I have made a mesh of a very complex structure, and by using snappyHexMesh. In the process there was a lot of failed meshes, and wondered if anybody has some kind of measure of the premesh quality of a CAD file? I have an idea to a way of quantify this: This is the length of the smallest line divided by the length of the average line $\bar{S}=\frac{min(l_i)}{\frac{1}{n}\sum^n_{i=1}(l_i)}$

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
[ICEM] Unexplained changes of mesh quality and blocking approach	salumi	ANSYS Meshing & Geometry	9	November 23, 2016 04:14
[ICEM] How to improve bad quality elements that are obtained after creating prism layers?	Rohith Giridhar	ANSYS Meshing & Geometry	2	July 11, 2015 18:04
[ICEM] Premesh quality criterion: "Angle" and "Min Angle"	saisanthoshm88	ANSYS Meshing & Geometry	3	February 26, 2014 10:23
[ICEM] Hex refinement negative quality afterwards	Anorky	ANSYS Meshing & Geometry	8	March 22, 2010 03:25
STAR-Works : Mainstream CAD with CFD	CD adapco Group Marketing	Siemens	0	February 13, 2002 12:23

January 13, 2018, 11:05		#2
vishnupc New Member Vishnu Join Date: Mar 2015 Posts: 3 Rep Power: 11	Hi Fridrik, When you say a 'failure' I would like to specifically know how ? I have not used snappyHexMesh and wanted to know if there are specific criteria setup to know that a mesh is a good one ? Not sure if this answers your question but qualitatively I feel that the CAD is a good one if it is : a. Connected as per the requirements of the simulation being performed (Based on which the geometry can have 1-sided or even 3+ sided edges) b. Ideally should not have self intersections and coincident faces. The main issue I have faced is that most geometries must be 'connected' as intended. This can be a little difficult when you are preparing a complex geometry prior to meshing. Once you DO do this check, your meshing procedure and subsequent cleanup should be not too difficult. G'Luck !