Personal tools
You are here: Home Forum Use Meshing adjacent geometries with huge dimension differences

Meshing adjacent geometries with huge dimension differences

Up to Use

Meshing adjacent geometries with huge dimension differences

Posted by A. Bhattarai at August 19. 2019

Dear Salome experts,

I need to connect two geometries with huge differences in their dimension. The geometries are available in link:

 

https://drive.google.com/file/d/1KA1_DOV-LJM0VuFv2TIMjOwQCUOHPNT1/view?usp=sharing

 

Is it possible to assign any proper hypothesis in salome to connect the nodes of these adjacent structures together?

Thank you very much in advance for the support.

 

Best regards,

Bhattarai

Re: Meshing adjacent geometries with huge dimension differences

Posted by Saint Michael at August 19. 2019

Hi

Since the two geometries have equal edges and faces, it seems that you will get equal meshes on equal geom entities using most of algorithms and hypotheses.

On the other hand, you can join the two geometries in Geom module using Partition and release from care about hypotheses.

St.michael

Re: Meshing adjacent geometries with huge dimension differences

Posted by A. Bhattarai at August 20. 2019
Dear  Saint Michael,
 
Thank you for the reply. I think my question might have mislead you. I have used the cut/partition tools between two given structures. But my problem is the cylindrical structure has diameter of around 0.3 mm and the other structure has the dimension in the order of cm. Using any and same algorithm for both of them will undoubtedly connect the neighboring nodes. 
 
Problem:
Selecting (Create mesh->1D--->Wire Discretisation--->Max Length) small sized elements will only define the geometry of the cylinder properly and using same algorithm for other larger structure will unnecessarily create a bulky mesh. I just want to get rid of this bulkiness by defining same algorithm on the surfaces which are connected and different where they are not connected. I tried to use sub-mesh tool but gave me nothing. I was wondering if there is any other trick to make a compatible meshes with relatively lesser elements other than by using same and small size elements.
 
Thank you again.
 
Best regards,
Bhattarai
 

Re: Meshing adjacent geometries with huge dimension differences

Posted by Saint Michael at August 20. 2019

Dear Bhattarai

Defining a sub-mesh is a right way to get meshes of different density on different parts.

Also you can try NETGEN 1D2D[3D] algorithm which creates more refined mesh on curved parts and less refined on more planar parts. Its "Nb Segs per Radius" parameter defines how element size depends on surface curvature. Also you can try its "Chordal Error" parameter for that. "Local size" parameter can be used as an alternative to defining sub-meshes.

St.Michael

Re: Meshing adjacent geometries with huge dimension differences

Posted by A. Bhattarai at August 22. 2019

Dear  Saint Michael,

 
Thank you very much for the reply. I tried firstly with the 2D meshing with NETGEN 1D2D algorithm (2D parameters--->local size in sub-shape--->size = 0.1) to construct the larger geometry which is little strange in some parts. For example: localized element concentration occurs in some regions. This simply increases the number of elements in the mesh. Nevertheless, the curved common faces are meshed properly with local size = 0.1.
 
Is there any way to refine this algorithm or use other hypothesis where there will be no localized elements?
 
The meshed result can be found here:
 
 
Thank you again.
 
Best regards,
Bhattarai

Re: Meshing adjacent geometries with huge dimension differences

Posted by Saint Michael at August 23. 2019

To prevent too small elements on some faces, put them in a group and create a sub-mesh on it where set a larger Min Size. Or alternatively, set the larger Min Size in the global hypothesis and create a sub-mesh on faces where smaller elements are allowed.

St.Michael

Powered by Ploneboard
Document Actions