Beam Element to Solid Element in CCX

VMH

Hi, I haven't looked much into this yet. But, has anyone tried transferring all DOF from a beam end to a solid face using CCX solver and also Mecway solver and obtained reasonable results? See attachments. I know we can do this in Abaqus. Thanks

VMH

Attachments

Sergio

I though to sugest you to use TIE, but I just read in the documentation that is only for 3d elements. Then I though in MPC, but true is that I don't understant how to use according to the CCX docs. On the other side, in the Mecway docs, MPC looks usefull for the matter.

I will follow this post to see how can be solved.

VMH

This will be an interesting topic for many who are using beam elements for the global modeling connecting to shell/solid elements for the connections/parts they are interested in. This give a better boundary conditions for more complicated structures (civil/structural engineering).

Victor

CCX doesn't seem to use the rotational DOFs in its constraint equations (*EQUATION) so this type of connection doesn't work correctly though it's OK with Mecway's built in solver.

I've constrained it so only translational DOFs are relevant and it produces a sensible solution, liml and png attached.

VMH

Thanks Victor. I tried axial, bending, and also combined axial and bending. They all had localized stress concentration from the beam to solid element which wouldn't exist if they are all beam elements or all solid elements. Thanks

VMH

Victor, this is what I was looking into:

http://www.femnews.pl/en/2016/04/18/benefits-of-using-beam-models/

Victor

Bonded contact won't do that because it only connects the beam's node to a point on the master surface, rather than the whole surface.

The poor-man's way would be to use multiple beams. Or you could probably do it using CCX's *RIGID BODY but that makes the surface rigid, which might not be ideal.

I hope to add a kind of "soft" rigid body connection in future, like Nastran's RBE3 which would serve this purpose with Mecway's solver but I'm not sure how/if it can be made to work with CCX because of the lack of rotational DOFs.

VMH

Thanks Victor, I'm always interested in deformable rigid body/link. It's like dummy member I used in structural analysis software like SACS by Bentley that transfer loads but removed from the stiffness calculations (save and delete joints/members from the stiffness calculations).

davefo

How did you analyse this when i try i get an error, see attached. Appears to not be able to analyse it as a 3 point beam member.

VMH

davefo, which version ccx were you using. I tried the rerun my original model attached in above message and didnt see your error message.

Victor

That screenshot shows the internal solver, not CCX.

davefo

yes i was using internal solver produces the error. I thought the above was this worked with the internal solver, "though it's OK with Mecway's built in solver"
when using the CCX solver i get the attached error. CCX version is 2.10

Victor

Oh I see. 3-node beams aren't supported by the internal solver. You should also clear any other red errors in the outline tree that can be ignored with the CCX solver.

If it says "Solver did not produce an output file", click "CCX output" to see the detailed error message.

davefo

found the reason for the calculix error.
I had pointed the solver path to the actual ccx.exe
when i changed it to point to the .bat instead it now works.