fcFEM - FEA from start to finish

[-alex-]

Thanks for your reply, very promising and exciting stuffs. Hope you'll overcome this challenge.
Alex

[HarryvL]

In the two examples I present above it works well, but there are plenty cases where the mesh generation does not deliver what I need. So this is still very much work in progress ...

I now have a reasonably robust way to generate interface elements and handle jumps in boundary conditions (see below). Time to focus on FEA again. Next I will focus on implementing an elasto-plastic interface with cohesion, friction and tension cut off.


Block with insert pushed diagonally down. The interface normal stiffness is high, but shear stiffness zero:

block with insert.png (128.49 KiB) Viewed 1537 times

And the square axle example shows lateral slip between axle and block. In front the block laterally expands and at the back it contracts (as would be expected from the Poisson ratio effect).

square axle compsolid slip.jpg (55.33 KiB) Viewed 1537 times

[HarryvL]

A horizontal cut of the last example nicely shows the differential movement of axle and block:

square axle compound shift (2).png (157.48 KiB) Viewed 1515 times

What appears penetration at the corners isn't This is small deformation element and the differential movement of the two components is relative to the original configuration, i.e. in y direction.

[HarryvL]

I am working on displaying interface results. This is the idea:

square axle compsolid interface display.jpg (62.2 KiB) Viewed 1458 times

Got the interface mesh to work with displacement results. Next display of stresses and figuring out how to have multiple meshes in the result object. This is done in natural frequency analysis, so should be possible?

[HarryvL]

oh .... and of course with the great flexibility of FreeCAD it automatically works for the deformed mesh. I am really enjoying this.

square axle compsolid interface display (2).jpg (37.95 KiB) Viewed 1457 times

[bernd]

cool

in frequency there are multiple result objects, but they all use the same result mesh. I do not 100% understand what you would like to archive. As I understood it, you would need multiple result objts with an own result mesh each.

[HarryvL]

Nope I want to show two meshes at the same time. One for interface elements and one for volume elements. Just like in the above picture. Now I just show the GMSH mesh and the result object has the interface mesh. I would like to show both deformed (for example).

[bernd]

Just use two result objects. Each result obj has one result mesh. I do not see the problem which probably means I still have not understood the problem ...

[HarryvL]

The problem was that I thought I needed two meshes in one object, but 2 result objects should work Thanks.

[HarryvL]

This time with stress results:


A simply supported beam with uniform load. Two blocks in a Compsolid, with automatically generated interface element between them. The nodal values of contact stress are in very good agreement with bending theory. For some reason the graphical interpolation (out of the box functionality in both FC and ParaView) struggles with a simple linear distribution. Anyway, the numerical values are spot on:

Beam Contract Stresses Mid Section.jpg (66.46 KiB) Viewed 1375 times

Stresses for the square axle (in hindsight I should have called it a square pin ... this axle won't rotate very well ) are high at the bottom trailing edge and fairly uniform on the leading face. Please keep in mind that contact stiffness in both compression and tension is taken high. I need to implement the Coulomb plasticity element with tension cut off next to show what happens in reality:

square axle contact stresses.jpg (72.54 KiB) Viewed 1375 times

and with VTK export to ParaView the results are even clearer:

Square Axle Contact Stress.png (110.14 KiB) Viewed 1375 times

Finally the contact stresses for the semi-circular punch. As expected, the stress is concentrated at the bottom:

Circular Punch Contact Stresses.jpg (91.5 KiB) Viewed 1375 times

Next I will play with 2 result objects, as suggested by Bernd, followed by plasticity