Problem with Hexablock (T-junction)...
Problem with Hexablock (T-junction)...Posted by Yohann Eude at February 19. 2013
I read the doc and I thought that the associations between vertex edge and quadrangle of the model and the geometry are implicit for simple geometric as cylinders and T-junction. When I computed the mesh of my T-junction, I can note that all edges are Ok but not for the T-junction close to the main cylinder (see the attachment). In the picture we can remark that some edges don't follow the surface of the main cylinder... This problem exist only when the second cylinder don't go through the main pipe.
Do you have an idea to solve this problem?
I tried to do a quadrangle association and it seems to be ok for the meshing of surfaces but not for the edges... I don't know how I can do a edge association in this area because there is no lines in this area in the geometry...
Thank you again for your help
Re: Problem with Hexablock (T-junction)...Posted by Yohann Eude at February 19. 2013
So, I solved the problem manually. I created the necessary edges in the geom module and I did the edges and quadrangle associations in the HexaBlock module in the area of the T-Junction.
Now, the mesh follows the surface but there is a strange deformation close to edges created by Hexablock.
I hope that this problem will be solved in the future version of salome.
Re: Problem with Hexablock (T-junction)...Posted by Kloss Francis at March 06. 2013
I have reviewed your case and you are right there is a problem.
This is a bug in the case of T-junction.
This bug will be fixed for the next
SALOME version 6.7 provided that day in June 2013.
Re: Problem with Hexablock (T-junction)...Posted by Yohann Eude at July 01. 2013
I just tested the module Hexablock from the new version of salome 7.2 and I observed that the module is more fluid and robust on this new version. However, I observe again a problem for the volume meshing of a T-Junction with the "make cylinders" tool. The mesh obtained is different from the version 6.6 but still bad... There is a very high concentration of nodes on the wall in the T area and the surface doesn't follow the boundaries surface.
You will find in attachment the Hexablock xml of the case. Is there a solution to solve this problem?
Thx again for your help.