As an alternative to a kinematic coupling constraint, you could create a rigid body and connect it to the end face with a tie constraint. You can then control the end rotation by applying it to a reference point on the rigid body.

Thomas Thienpont I tried but still have the same problem. I really don't know how to solve it. If it could be useful I am attaching my model

Virginia Lo Gatto, I took a quick look. I think there are some issues with you assembly. You have two instances of the part "PROVINO", while I assume you only need one. I recommend removing the duplicate part, and reassigning all boundary conditions. I also recommend working with a courser mesh when you are trying out new models. That way you can debug the model faster, as it takes only a few seconds to run it.

Hi Virginia,

I'm currently having the same issue with torsion of a hollow cylinder. Abaqus obviously converts the rotations or moments applied on the reference point to tangential displacements or forces on the coupled nodes, respectively. But these displacements or forces are not updated in their direction as the part is deformed, so the part blows up radially. I've tried almost everything (NLGEOM, activate "follow nodal rotation", more substeps, even "rigid" elements between reference point and coupled nodes show an elongation ...) but nothing worked.

I formerly used ANSYS and as far as I can remember I applied the displacements for the final configuration directly on the nodes via a small APDL-Macro which reads out the coordinates of the nodes and calculates the individual nodal displacements for a given rotational angle.

But I haven't figured out yet how to implement this workaround in Abaqus.

Kind regards

Markus