Since rigid body and shell edge no longer supported by CalculiX solver, it seems required to have feature of distributed nodal loads on shell edge. This known as elastic methods similar approach can be found in bolt group analysis or pile group reaction in foundation. Center of node groups first need to know, then distance to center related to inertia in purpose of converting concentrated moment or torque to individual nodes of perimeter shell edge. Eccentricity n both direction is allowed as input, but some problem may exist for non ortogonal plane.
p.s currently Surface Traction feature also capable to distributed loads on shell edge even limited to concentrate force only not for moment/torque.
The newest one - 2.23. But I asked Guido if he could relax this limitation to allow such combinations in linear analyses (the problem is only with nonlinearities incouding Nlgeom) and he said he would check it.
right, since newest version of 2.3 no longer supported clearly shown in analysis logs. I’m not sure if shell element and rigid body will work properly in general case. This is long discussion; shell edge rotation is problematic even restraint has been assigned manually. Several cases shown working properly, but overal is not even at condition of single point constraint. Multi point constraint such as rigid body usually leading to be more problematic.
I haven’t noticed any issues with rigid body constraints and shells in linear analyses (namely without Nlgeom), but they could occur since:
Internally, a rigid body is enforced by using nonlinear multiple point constraints (MPC).
and shell expansion/knot generation is also nonlinear - see Guido’s explanation here.
However, I am not a fan of hardcoding such limitations, I would rather leave it up to the user to take the risk. It’s up to Guido’s (and his team’s) decision, though.
converting concentrated force and moment to individual nodal loads on shell edge can have advantages of this limitation as explained by Guido. The case becomes less problematic since it uses single point constraint not multi point.
Sometimes making the shell faces/edges rigid while applying load or BCs (that’s another common application) is actually desired. But the thing is also that fixing rotational DOFs of shells in CalculiX often leads to issues.
since it uses single point constraint thus rigid in plane can be easily assigned i.e cylinder did not ovalize. Also, rotational restraint at edge shell of individual nodes in local coordinate system (SPC) usually works even at case large deformation and plasticity (nonlinear buckling with imperfections).