I made a series of tests:
I reduced the displacement of the roller than 5 mm.
I deleted the "contact sizing" and the result was:
http://www.youtube.com/watch?v=M-PYBV-PQ8A&feature=youtu.be
Do not understand why the ends do not rotate. I tried using contact type "frictionless" and get the same result.
Also if I want to impose a greater displacement fails.

Sebastian, I work with CATIA not ANSYS but the methodology looks similar judging by your screen shots. I think the problem is with restraint "D". Try removing it and applying restraints to both edges of the plate but only in the direction I illustrated. Allow free movement (including rotation) in all other directions. Let me know if that works. I'm curious. Nice job by the way. I'm guessing this is homework but it's cool!

Hello Terraman! Thank you for your comments. I am inclined to think this is because the boundary conditions or contacts.

I have restricted movement in the direction perpendicular (Z) to the plate in two faces. Free movement in the other directions, do not look bad. However I tested what propose to me and now does not converge!!

About this work, I tell him that it is the design of a rolling machine three-roll pyramid type.

I am interested in verifying the force required to plastically deform the plate and reaction forces in the supports.

I will continue proving. I will keep you abreast of developments. If you have other suggestions let me know.
Thank you very much, greetings

Sebastian