Not 100% sure what you are trying to achieve, but a kinematic coupling between the surface node and the RP of the display body should do the trick.

Thank you for your quick answer David Sziroczak!

I reformulate for clarity then: is there a possibility that the coupling with the surface is done before meshing it? Or do I always need a node to tie it to? It would be interesting if I wouldn't need the node so that the RP would automatically be linked with the element that forms later in that area, but the more I read about it, the less feasible it seems.

Another solution that occurs to me is that perhaps I could generate mesh nodes at specific coordinates (I've read there's the possibility to edit them, not sure if they can be created from scratch though) and then create the sets or RPs linked to those nodes, for which I would probably need some Abaqus programming skills (that I'm a bit short of atm). If there's this possibility though, I'd be willing to try!

I hope this clarifies a bit more my intentions.

Cheers,

Sara

A very interesting question! A thought I share.I hope I have interpreted the Q correctly.This looks akin to finite element patch. So if the point (circle) is on the

patch , a deformation of the matrix would still contain the point in its element. Another possibility is perhaps to use block matrices to transform points from

original patch to deformed patch.

Good Luck

Cheers

Sara Medina Lombardero

Thank you for explaining. If I had to do it, I would definitely go down the route of sectioning, that is by far the easiest way to get a node where you wanted to.

Alternatively you could use a python script to find the closest node and automatically apply the coupling. It obviously involves scripting if you haven't done it before, it's probably not the easiest task to get started with.

There is another possibility to use a tie constraint with a large distance, and then the software will project your remote node on your surface to determine which nodes to connect it to with an MPC. Keep in mind that using ties over large distances is not the intended functionality, you will get a warning message, and you should check if the results you get are sensible. I would try this if you don't want to do sectioning.

Also a very ugly option could be to model a small part of the geometry "sticking out" of the surface, and define the remote node on the end of it. The small part could be defined with negligible stiffness so it doesn't affect the rest of the calculation, but still carries the node. I wouldn't do this, as any other solutions are probably quicker and definitely better.

Just for the full picture, you can define an element with the user element subroutines which has an extra remote node, this is by far the worst solution in terms of work and usefulness, but it can be a solution.

Also you could check the integrated surface outputs in case you can get rid of the node altogether and get what you need in a different way.

I hope these thoughts will help you to solve your issue!

Narasim Ramesh thank you for your input as well! I take it you suggest there is a mathematical possibility, sort of like applying an equation to the point so that it would have its own deformation path? I wouldn't know how to create these block matrices you mention, but perhaps it's a possibility worth investigating too, thanks for the contribution!

Yes That is my thought. Another idea comes to mind.Say the patch co=ordinates

of the polygon you are interested in is available before and after deformation, then

there must be atoplogical transformation matrix which can be evaluated as follows (like shape analysis)

1. A = mx2 matrix of x/ y coords before deformation.

2. B = matrix as above with coords of polygon after deformation.

3 We seek a transformation( deformation) matrix say C such that

A* C = B .. equate elements to get cij approx use Penrose perhaps , C is 2x2 matrix

4 Apply C to circle point circle cords

May help

Cheers

Narasim Ramesh it is indeed an interesting proposition. The problem would then become a problem of locating the coordinates in each state (deformed/ undeformed) for particular elements of the mesh and calculating the mentioned transformation matrix. This reminds me of an image correlation kind of solution, which could be relatively easy to implement.

I am however not too familiar with Abaqus when it comes to this sort of mathematical implementations (I'm still learning), but I hadn't found before any solutions related to this problem, and these suggestions are giving me lots of ideas, so thank you again for your input!

Cheers,

Sara

My pleasure..I am not familiar with Abaqus too.I use Scilab (FOSS) .Since you mention image processing , it might be worth checking out Imagej (FUJI) also FOSS . Basically this is to me a Linear algebra problem involving active transformation. If the marked sensor point is presented in both pre and post deformation process then Imagej would be useful to track the movement.

Good Luck

Cheers

I think above would be a solution for an image of sensor movement on a physical

patch which is deforming.

Cheers