It seems that node displacements are not computed correctly for shells.
Also I suspect the corotational framework should be adopted to write the displacements to consider properly the rotations