Personal tools
You are here: Home Forum Use Local Length works but Max Size does not for 1D mesh hypothesis

Local Length works but Max Size does not for 1D mesh hypothesis

Up to Use

Local Length works but Max Size does not for 1D mesh hypothesis

Posted by Anirudh Nehra at August 04. 2019

Hi to all,

I am doing a study where I want to mesh all bodies with this hypothesis:

For 1D I use either local length or max_size

For 2D, quadrangle mapping

Hexa(i,j,k) for 3D.

The geometry is partitioned orthogonally everywhere. Everywhere there are regular block/cubes so, I should not have problem with generating all Hex elements, whether I use Local length or Max_size for 1D.

However, for the partition named 'bc3_prep' and 'gnt_prep' and somewhat for 'bc5_prep', there are errors using Max_Size. but no errors using Local_length.

Even generating a default mesh over 'bc3_prep' using these parameters: Hexa (i,j,k)for 3D, Quadrangle for 2D, Max_Size for 1D(with propagation of nodes to opposite edges) causes error.

One other observation is that the above attempt at mesh generation succeeds if we use the default value(~0.6) that appears in Max_Size dialog box under 1D hypothesis.
I generate the geometry using script and most of partition planes(named by 'all_planes_x') are created by 'multi_translated_planes' feature. I did this kind of meshing before albeit not using translate_copy feature but creating planes manually(by picking point and normal) and did not have errors with using Max_size hypothesis.

I am attaching a link to the study. Could someone help me know why this problem occurs?

Link : https://www.dropbox.com/s/oryqozawly6eo3t/study1.tar.gz?dl=0

Thanks and Regards

Anirudh

Re: Local Length works but Max Size does not for 1D mesh hypothesis

Posted by Saint Michael at August 05. 2019

Hi Anirudh

I can't download the study. Only "Preview" action is available on it.

St.Michael

Re: Local Length works but Max Size does not for 1D mesh hypothesis

Posted by Anirudh Nehra at August 05. 2019

Re: Local Length works but Max Size does not for 1D mesh hypothesis

Posted by Saint Michael at August 09. 2019

Hi

Local length uses its precision when dividing edge. Due to that precision you get an expected result. Max_Size does not use precision, so you have to set it up more carefully. For example:

Local Length = 0.1 --> OK,
Max_Size = 0.1 --> KO,
Max_Size = 0.09 -->OK

St.Michael

Re: Local Length works but Max Size does not for 1D mesh hypothesis

Posted by Anirudh Nehra at August 09. 2019

Hi,

Thanks for the response.

However, I am not sure I understand it correctly. I can't see what difference does 0.1 or 0.09 make to an edge if all edges are divided in the same way. Since there is always an edge below another one, the max_size should propogate in a similar way.

I also believe is that maybe the edge orientation is not maintained in Max_size, so it starts cutting edge from other side for some thus edges leaving uneven number of cubes on top and bottom side.

 

Regards

Anirudh

Re: Local Length works but Max Size does not for 1D mesh hypothesis

Posted by Saint Michael at August 09. 2019

Consider two edges: of length 0.2 and 0.19999999999, Max_Size=0.1 will give different results on them while Max_Size=0.09, the same.

Since you constructed your edges not directly but by partitioning, there can be tiny fluctuations of length of result edges.

St.Michael

Re: Local Length works but Max Size does not for 1D mesh hypothesis

Posted by Anirudh Nehra at August 09. 2019

Hi,

Thanks again for the reply. I think I understand now.

Is there a partitioning tolerance that I can control so that it works the same way as Local Length hypothesis?

Also, is there a difference to creating a plane by picking a point and a normal, and by multi translation 1D? Is there a tolerance involved in multi translation, so the distance between multiple instances of multi translation varies a little from one instance to another?

Also, If I enquire using "what is" function of Salome about a partition, it says a compound.

Shouldn't a partition be CompSolid since faces are shared between the adjacent solids?

 

Regards

Re: Local Length works but Max Size does not for 1D mesh hypothesis

Posted by Saint Michael at August 09. 2019
There is no difference between Max_Size and Local length but tolerance usage. Just use Local length as it works for you and release your from care about the way of geometry construction.
CompSolid is rarely used, as it is just a particular case of compound.
 
St.Michael
 
Powered by Ploneboard
Document Actions