Braid model

[Bromissimo]

Dear All,

I am trying to model a braid having 4 interlacing threads.
After defining relations, I can model each thread individually, i.e. in four different sketches. It works.
When I try to draw and define all threads in a single sketch (and then having the possibility to easily modify all of them just varying the right parameter), however , a message appears in the message area (when I try to close the sketch):

Multiple loops must all be closed in this section.

This message does not appear when just 2 threads are defined, but just when adding the third and the forth one.
Where do I do wrong? Can anyone help me, please?
Can also anyone suggest me if there are other ways to model a braid (I use variable section sweep)?

Thanks!

 

[kenppy]

Are you saying that the loops are closed and it's still not working?

 

[Bromissimo]

Hi Kenppy,
thank you for your reply!

In my design, it should not be any loop. I do not understand why ProEng let me define and generate two yarns in one sketchboard but shows the above mentioned problem when I try to draw 4 in the same sketch.

This is the procedure I use:
-Sketch a line (which is my overall length of the braid)
-Variable sweep section
-Define a line in order to control by relations the rotation and radius of the sweep section along the length.

Again, this procedure only works for 1 or 2 threads. In order to model a braid having 36 or more intertwining yarns, I should sketch them one by one, and this procedure is very impractical.


Thank you!

 

[kenppy]

There appears to be no limit to how many closed loops can be swept along a common path, although this is not a braid, how many curves are you following?
View attachment 5152

 

[Bromissimo]

Hi Kenppy,

in order to make my sweep, I have used this video (first 3 minutes) as reference:
http://www.youtube.com/watch?v=V4R8kQh3bos

If relations are defined for one sweep, the procedure should be valid for all the sweeps, though. But I get the error message when closing sketchboard.

Can you tell me the procedure you have used to create your model (above picture)? This will be very useful for my next step (where I need to model a thread made of many filaments).

Thank you!



Edited by: Bromissimo

 

[kenppy]

I used the same method to create the path and then VSS to follow it. Then yo have any number of sections following the same path, but this is not braiding as I understand it. Have you a picture of what you want?

 

[Bromissimo]

View attachment 5153

This is what I have done.

 

[kdem]

Based on what you have shown in your picture it looks like your geometry is repeating so an axial pattern might work. In other words you 2 sets of four identical VSS as far as I can tell. So you would have two sketches on for the two sets of VSS and axis pattern the VSS.

 

[Bromissimo]

kdem, that is what I have tried to do, but when I add the third and/or the fouth element in the sketchboard, the error message appears, as previously said.

I did not define a relation for the axial pattern, but for the radius and the angle (there are 8 hidden surfaces) and the cilindrical sweeps are made using a protrusion along the edge of those surfaces.
How can I define a relation for the axis if no one angle has been defined? I have tried with trajpar, 360*trajpar and sin(360*trajpar).... It doesn't work. Problably I did something wrong.
Any suggestion?

Thanks!

Edited by: Bromissimo

 

[kdem]

kdem, that is what I have tried to do, but when I add the third and/or the fouth element in the sketchboard, the error message appears, as previously said.

We may be talking about different modeling methods and I may just have a bad understanding of what you have modeled. Can't really tell without looking at the model. I'm not sure I'm understanding what you mean by 3rd and 4th sketches in the sketchboard. Based on the other posts it sounds as if you are trying to define four of the elements in one sketch. What I'm seeing just from what you have shown is two sets of four features with the four features in each set having identical geometry. If that's correct what I'm suggesting is working on getting the VSS trajectory for only one feature of each group, you would end up with two trajectory curves which will define two VSS features. You would create a geometry copy of both of the curves and create axial patterns using the pattern feature. For each pattern you specify four instances for each pattern to get the eight trajectories that you have shown in your picture. You would then reference pattern the VSS to get the eight VSS features. Assuming I'm understanding the geomerty correctly.

 

[kenppy]

This one intrigued me enough to see what I could do with it.

Let me know if this helps.

resume all features




2011-06-12_080133_braid.prt.zip

 

[Bromissimo]

Hi kenppy.

I just cannot see your model because I have Pro/Eng WF 5.0 University Edition. It does not let me open it!!!
I will try to open it tomorrow from another workstation having a different version installed.

Thanks!!


Edited by: Bromissimo

 

[kdem]

Is your model something you can post? I do havethe student edition so I can open it if you're allowed to post it. It would be helpful to get an understanding of how you have modeled it so far.

 

[Bromissimo]

kdem,

The model is going good. I start drawing all filaments for each sweep.
However, the model will be just an approximation of the real shape. This because the cross section of the wire changes along the length of a real braid, and this overcomplicates the model and its definition.

Thank for your help!

 

[Bromissimo]

kennpy,

your model looks great! (I can open it even with PE WF 5.0 University edition..I did not resume the features!)
It is very similar to what I did. However, there is an essential difference between the models.
In my model I have used these relations:

sd5= 360*trajpar*n
sd6=A+B*sin(m*sd5)

I have used these relations to create a sweep section along a straigth line.

On the other hand, you have used these relations:

sd5=360*trajpar*n
sd6=A+(B*sin(trajpar*360*braid))

You have used these relations to create a sweep line along a curved line.

Intent design is very similar and both applicable!

Can you please tell me more about the "braid" command parameter? It is the first time I see it.

Thanks!


Edited by: Bromissimo

 

[kenppy]

'braid' is just a name I gave to an axial pitch for the right and left handed surfs so that the change could be made in one go.

 

[Bromissimo]

kenppy,

you are very kind!

I have another little problem to solve (and I actually struggle on it): what if one filament has to pass above and below 2 other? The picture clarifies the statement.


2011-06-13_101428_2x2_architecture.zip

I have problems in defining the relation.

Thanks

 

[kenppy]

I assume you are doing this as an exercise, can't imagine this has a practical use as in being used for production.

This weave has a longer pitch so it's a matter of playing around with that but that will be approximate and not look like the picture.

It's complicated by the fact that it's not now a pure sine wave but a harmonic with flats (or stretched peaks). I suggest you define this shape with a datum graph. Experiment with this on a flat weave until you get it right before doing the sock version.

 

[Bromissimo]

How about a datum graph? Drawing "by hand" this function? How to make it peridodic?

It is about a week I am trying to define harmonic equations to describe that function! Fourier series (even 10k order) will always leave two "humps" at the edges of the flattened maximum. I have tried in excel, but I think ProE would not either accept such a terrible polynomial equation or incur in some trouble of any kind.
A mathematician would be very helpful!

I have to define this equation...I will post some suggestions as soon as I got something "linear".

Thanks

 

[kenppy]

Using an equation would be be most flexible.

I was thinking that if you have a sensible number of pitches then you could draw a graph with a repeat of something like this shape.

View attachment 5156