[ander]

A number of shell elements have only 5 degrees of freedom (3 translations and 2 rotations). In explicit FEM, it is common to use the central difference method, and for instance in the paper of Belytschko et. al. https://doi.org/10.1016/0045-7825(84)90026-4 it is stated that this method is used.

However, to me it doesn't seem trivial to extend the central difference method for translational DOFs to a setting with rotational DOFs, ESPECIALLY when there are only two rotational degrees of freedom. I have some idea of how this can be done with three rotational DOFs, then you can use an exponential map to update a rotation matrix or quaternion that describes the nodal rotation.

If someone has any pointers or sources on how this can be done in the 5 DOF setting, I would be happy to hear it.