Braiding modelling

General discussion about TexGen.

Moderators: Martin, Developers

Post Reply
Jean-Vincent
Posts: 2
Joined: Mon Jul 28, 2008 8:46 am

Braiding modelling

Post by Jean-Vincent »

Dear Martin,

I am working with TexGen since few weeks and I is a really interesting software. So thank you for all : the software and also the documentation are usefull.

For my project, I need to create a braiding but not a plain braid. This braiding has to cover a mandrel (A tube, a conic or square prism).

I succed to create the python scrip to create it. You can have a look to the results :

Image

Image

Each interlacing creates a node for a yarn.

My problem is: As you can see it works but just for limited parameters (lenght, turns arround the mandrel, number of yarn etc ...)

For example if the number of turns arround the mandrel is too high, the interpolation doesn´t fallow the nodes. I tried to change the interpolation but without sucess.

To avoid it I would like to know if and how we can use the interference correction ?

Also if there is a special command to set the section to a node during the node creation. As i create the yarns and the nodes by loops in function of the parameters I would like to know if it can be done.

Thanks again for your software.

Regards

(Ps: Sorry if my English is not clear :? )
Jean-Vincent R.
joncrookston
Expert User
Posts: 27
Joined: Wed Apr 11, 2007 4:03 pm
Location: Nottingham, UK

Post by joncrookston »

Hi Jean-Vincent,

These braids look good. Can you show any examples of what you mean by this?
the interpolation doesn't follow the nodes
Hopefully that will help someone to answer.

Have a nice weekend,
Jon.
Last edited by joncrookston on Thu Aug 21, 2008 3:33 pm, edited 1 time in total.
wr
Contributor
Posts: 35
Joined: Tue Aug 28, 2007 8:33 am
Location: Ringkoebing, DK

Post by wr »

Hello Jean-Vincent,

I think what you are referring to is the fact that the order of the spline is limited which gives trouble when you make very long yarns with complex geometry, see here: http://mathworld.wolfram.com/BezierCurve.html
There does seem to be an implementation of cubic splines in there:
http://mathworld.wolfram.com/CubicSpline.html
this should always pass through all of the control points.
Since both implement the Interpolation interface you shouldn't have trouble switching.

A simpler way to solve your problem is making smaller unit cells (for the box-mandrel you might need a few of them) and linking them up.

This is also (in my view) more elegant and powerful in case you want to do simulation on the preform or composite later on (the lookups can be threaded and each subtextile can be queried fast for local properties).

Nice braid script though, let us know how you get on. It would look good in the applications section.

W
Jean-Vincent
Posts: 2
Joined: Mon Jul 28, 2008 8:46 am

Post by Jean-Vincent »

Thanks both of you for reply.

Your are right this is due to the lenght and the global dimension of the mandrel. But it is difficult for me to use a little part and reproduice it as you said because my scripts can cover irregular diameters for exemple.

But, I succed to solve a part of the problem by using the "SetResolution()". Now I am able to generate what I need but of course it takes time to do it (few minutes, max 5min).

My last problem is to set the yarn section for each node in the script, I try like that :

Code: Select all

section1 = CSectionEllipse(yarn_width, yarn_thickness)

for j in range(yarn_number):
    nodesection=CYarnSectionInterpNode(j,0)
    for i in range(2*yarn_number):
        section=section1
        yarnsection=nodesection.AddSection(section)
    Textile.GetYarn(j).AssignSection(yarnsection)
I am not sure that is clear but i have to ask my supervisor before share the all script. Actually he is in holliday, but for me i think it is normal to share with the communauty and especially in my case because I ask help :).

Once again thanks.
Jean-Vincent R.
Post Reply