Not a problem, and thank you for the comprehensive feedback.nLooking at your images more closely I noticed it is always a very short beam that you are referring to. Are these small beams the only locations you would like to apply the end releases to?nIf this is the case, you may want to consider using link elements. This can be done by assigning a duplicate cross section in SpaceClaim, giving it a name such as LINKS, then selecting those bodies in the mechanical tree (they will now be separate bodies upon import due to their different cross section) and setting the Model type the Link/Truss. This may be applicable in this instance as you are releasing all rotational degrees of freedom. If you are only releasing one or two rotations, or translations, then this will obviously not work.nAlso note for link elements, that their cross section definition is actually different in the solver, and the result output is slightly different. The are essentially uniaxial, so if you want the correct transverse stiffnessm I would not utilise this method.nA disdvantage of this method is that you will have many more bodies in the mechanical session, which slows down the application, and requires the user to have to filter through more stuff. An advantage is that this model will solve faster (probably not observable) as the endrelease command in APDL actually creates duplicate nodes and couples which will add to the matrix.n