partitioning and meshing

[CSAPL]

As ABAQUS is not equipped with Triangle element for Coupled soil-pore elements, I have used Rectangualr element. However, the mesh is now awaful. This is because the several part-partitions I made while creating my model to accomodate the several materials I have.
Anybody has a suggestion.
-Should I creat a mesh first and then partition it and then assign materials to it partitions.

Please advise

 

[brep]

Try using a different mesh algorithm - the advancing front quad mesh might give you something better than the (default) medial axis.

 

[hamadadf]

i'd recommend using finer mesh around the partitioned parts.

 

[corus]

Some times you can get a structured mesh but the mesh produced looks awful. This occurs with such regions as L shaped corners. brep's advice does produce a better mesh quality but some times it doesn't look so 'neat'. Try partitioning regions so you get rectangular regions, or better still square regions, as near as possible. Failing that you just have to mess about with biasing the mesh until it all looks reasonable. The mesh verification tool will help you decide on the quality you've achieved.


corus

 

[CSAPL]

I have a better mesh now but still I get warning at several elements. That is because (as i beleive)they are trapezoidal with sever acute angles. I hope this warning is not critical to convergence of the problem

 

[brep]

Corus,
The mesh may not "look" as nice, but beauty of a FE mesh is not skin deep There are many good examples of better, ugly meshes as generated by automatic mesh adaptivity (v6.6). Perhpas you could try this too?

 

[CSAPL]

advancing front has realy improved the mesh. I still get some warning (yello color)at several element (I think because of their Sever Acute angles)

By the way does any one know how can I color the mesh elements (based on the materials) in the mesh Module (I know one can do that in the visualization Module but I need to color before runing the job. It looks that ABAQUS can only color the model (Color Code) after runing the job !!!

 

[corus]

brep, I've noticed that with an L shaped structured region that the mesh will go outside the bounds of the region. That isn't ugly, just wrong. In addition half my audience look in awe at my work for two reasons, firstly because it's come from a computer, and secondly because it looks nice. The niceness factor has a certain appeal, particularly if you can add in plenty of colours. Failing that I supply crayons when issuing reports to management so they can colour the pictures in themselves.
I sympathise with CSAPL's dilemma regarding colouring the mesh in CAE. Of course you don't have to run the job completely to see the coloured mesh in Viewer,just run the job with a data check. Failing that have you considered crayons?
Only kidding.

corus

 

[brep]

Corus,
Have you tried Photoshop?

 

[brep]

CSAPL,

In V6.6 (which is the current release as of last April), you *can* color code by material. Just use the combo box in the top right of the GUI. You can also use the more flexible (and rigorous) GUI but clicking on the color palette icon. Color code by XXX (where XXX = sections, materials, parts, sets, surfaces, loads, BCs, interactions, element types....)

 

[CSAPL]

Unfortunately I am working with the version 6.5.1.

Coming Back to Partioning issue,so according to Cours,if one wants to apply Intitail Conditions on certain nodes, then one (in advance, while doing the partioning) has to marke these nodes as Vertices. Is that right ?

 

[brep]

If you have a geometry-based model (i.e. NOT and orphan mesh), then creating a set containing vertices will imply that an nset is created. Try it and look at the inp for verification.

Similarly, for a solid model, if you create a set containing and edge or a face, then an nset is created with the nodes atached to the edge/face. If you create a set containing a cell then an elset will be created.

Also, if you have a shell model, a set containing faces will result in an elset of associated elements...

 

[CSAPL]

BREG
Regarding your answer on partitioning, i do not agree with you please see my comment to Corus in the thread " On Creating Nodes and Elements Sets" it is very common thing in ABAQUS modelling but it is puzlling us!!!!
Look forward to hear your answer on this issue

 

[johnhors]

Csapl,

Why are you still using v6.5.1 ?

 

[CSAPL]

Johnhors Yes I am still using 6.5.1. DO you have any suggestion regarding this Set and partitioning issue

 

[CSAPL]

Johnhors Yes I am still using 6.5.1 because I have no access to a newer version. do you have any suggestion regarding this Set and partitioning issue