G'day ywdmf (got a name?),
Since we have refused to buy AMDL on principle (it should be a part of Synergy, especially at today's higher price!), I have been quite successful analyzing insert/core deflection analyses in 3D with a careful but tedious modelling process. If you start by modelling the insert at your desired mesh density, and save the DD mesh, and then make the 3D mesh from a copy of this. With the insert DD mesh, delete any elements on surfaces that are not common to the plastic part, i.e. ones that protrude into the mold and are restrained by the mold.
Then mesh the plastic part in its desired density of mesh in a DD mesh and copy it. Delete all of the elements on the common surfaces between the insert and part from the plastic part DD mesh on one of the copies, and then merge with it the partial DD mesh of the insert. Connect up the interfaces between the two meshes, being certain to retain the integrity of the insert mesh so that it will match-up to the insert 3D mesh once you run the analysis. Now you can make another copy of this DD mesh into the 3D plastic part mesh.
I recommend that you keep the DD mesh copies along the way just in case you forget about an insert area, then you don't have to start all over. Believe me, you only make that mistake once! You end up with a lot of interim models if you have more than one insert, but if you do it correctly and the 3D mesh and mesh repair does not alter the surface mesh (with a bunch of what I call "spider webs"), you can then conduct a successful insert and/or core deflection analysis with good results as long as you constrain the part correctly.
Hope this helps,
Bob Sherman, RTP Company