Dry Fibre export - periodic boundary conditions

[jessica90]

Hi!

I fear that I have encountered another problem with the periodic boundary conditions, this time with the dry fibre export.

The set Bound0A and Bound0B contain the expected nodes, set Bound1A and Bound1B however not (see picture). I also tried shifting the domain to not split the yarn length wise, with no success. Do you have any idea how that comes?

Thank you in advance!

Jessica

[louisepb]

Hi Jessica,

I'm afraid I can't really see what's going on here. Would it be possible to send the model and/or the inp file so I can see more clearly what the problem is?

Best regards,
Louise

[jessica90]

Hi Louise,

Yes, of course, sorry. I attached the tg3 file and the inp file (as generated by TexGen). In the inp file the surface definition of the elliptical yarns is also missing, I already took care of that manually though.

Thanks,
Jessica

[louisepb]

Hi Jessica,

Sorry to have taken so long to reply to this. I've taken a look and it seems to be an issue with the repeats, firstly not corresponding to the size of the domain and secondly having both an x and y component. Basically the code makes assumptions about how the repeats match up to the domain so I need to work out what to do when these assumptions aren't true.

You may be able to get around the problem by specifying separate yarns with repeats that correspond to the domain size rather than having the same yarn repeating within the domain. Not the ideal situation but I'll take a look at the code and see if I can sort it out.

Best regards,
Louise

[jessica90]

Hi Louise,

Thank you for your reply. I tried using separate yarns with the repeat vector only assigned alongside the respective yarn direction, unfortunately without much success. It think a problem with this method is that since the yarns run across the domain diagonally, the yarn parts that have to "connect" over the domain edges via the boundary conditions are not from the same yarn.

I also tried using a volume mesh and I ran into similar problems. I meshed and exported each of the three yarns in a separate tg3 file, deleted the matrix and renumbered the nodes and elements as to not have doubled ids in the different yarns with a script and then copied everything together. Defining the yarns in the same way as in the file that I sent you a couple weeks ago led to extra yarn elements connecting the yarns to each other (see picture). Defining the yarns separately with repeat vectors only in yarn direction resulted in non compatible meshes alongside the edge domain. I'm starting to run out of ideas...

Best regards,
Jessica

[louisepb]

Hi Jessica,

Sorry to take a while to reply.

The algorithm for generating tetrahedral meshes, as you've seen, doesn't work well for yarns which repeat within the domain. (Something else to add to my todo list!)

I think I have come up with a solution that may work for you. I have attached a script which creates a second yarn with vertical positions reversed, as in the middle yarn in the textile. I have changed the repeat for yarn 3 so that it matches the size of the domain and then added the new yarn so that it sits inbetween. I have also duplicated yarn 0 so that there are two yarns, each with the same repeat size as the domain.
The resulting tg3 file is also attached

I did a dry fibre export and these yarns now seem to have the correct boundary node pairs. You will need to do the same for the other set of interwoven yarns.

I hope that makes sense. Get back to me if you need clarification

Best regards,
Louise

[jessica90]

Hi Louise,

thank you for your reply. I exported a dry fiber mesh from the tg3 file you uploaded, unfortunately, the node sets containing the nodes for the boundary conditions appear to still be incomplete. On the face normal to the y-axis the nodes for the new (blue and yellow) yarns are missing:

nodeset.png (79.93 KiB) Viewed 10370 times

Sorry my model makes things so complicated for you

Best regards,
Jessica

[louisepb]

Hi Jessica,

I think I've solved it! I have changed the code so that it uses the domain size rather than the repeats to find matching nodes. I think that the original code was written with the assumption that the repeats and domain size would always match. As you have found out this isn't always the case. This will go into the next release.

Best regards,
Louise

[jessica90]

Hi Louise!

Thank you! Do you know yet when the next release will be available?

Best regards,
Jessica