## You are here

# Voronoi tesselation

Mon, 2011-03-14 07:38 - Panos Efthymiadis

Dear all,

I would like to ask you how is it possible to import into abaqus a 2-D or a 3-D Voronoi structure. so far I ve been able to create my grains into Matlab or fortran but I dont know how to import them into abaqus.

All the best,

Panos

»

- Panos Efthymiadis's blog
- Log in or register to post comments
- 8153 reads

## Comments

## Geometry import in ABAQUS

Dear Panos,

You can define any geometry in the .inp file. It has text format and geometry is defined there in anatural way; specific section of ABQUS Manual covers all posiible commands for .inp file

For example..

*Node

1, 0.0, 0.0

2, -0.5, 1.0

Obviously you can code a procedure in MATLAB (or use other language...) which generates such file for you.

However in this case you lose all the advantages of ABAQUS CAE graphical environment and have to work with text file to define entire model.

Theoretically other ways are posiible, but i didn`t try them.

1. You can export your geometry to some of the supportable formats (they are covered in ABAQUS manual).

2. You can code a tool in Python to perform a tesselation directly in ABAQUS. This way realtes to so called parametric modelling and you can also find detailed information in the manual. But actually i am not sure if Python API gives enough possibilities to do this.

## Re: Geometry import in ABAQUS

Hi Alexandre,

thank you for your answer. As I see it, it s better if I only import the geometry and then work with abaqus cae for further processing. The porblems I encounter when I need to import a Voronoi plot as an input geometry in Abaqus are the following:

1. Voronoi tesselation creates convex polygons; but you also get some polygons at the edges of the plot that do extend to infinity. So if I want to import only the square/cube plot I need to find the crossing between the borders of the plot and the respective polygons. But I cannot do that since some extend to infinity.

2. So far

I am usingMatlab , I have no experience with python; and the documentation of python with respect to geometry creation is rather poor; it only has 2 examples, simple ones. I think that people generally create the geometry into Matlab / fortran etc. and then import the points into python and then abaqus..All the best,

Panos

## I have no expirience with

I have no expirience with Voronoi tesselations but i see no reaons, why you can`t find crossing between the border and polygons, as they probably both allow simple analytical representation. Also if it is really difficult, i think any ABAQUS tricks are useless untill you can represent your mesh in a standart way (as a domain of triangular, or rectangular elements with known vertex coordinates).

Also i have a new idea. As i already mentioned, in the *.inp file you have full freedom of defining nodes and elements. However it is incompatible with high-level feature-based modelling scheme, of ABAQUS CAE. To combine both approaches, you can try a trick. When you create a model in CAE environemnt and launch job, .inp file is created, and it is the file, which is actually transfered to computation engine. So you can create a fully working model with geometry in ABAQUS CAE and only then inject your mesh into the .inp file. Then you lauch computation from command file, but you still can use CAE to open and visualize *.odb file with results.