What are you trying to do, Cryo1? Is there some need to model the entire cylinder? It sounds as though you could model a small piece and get the same results without all the effort of modeling the entire cylinder. My first thought is a 2-D plain strain cross section. Fix the inner edge by changing the surface designation and apply a surface boundary conditions to fix the nodes. You could then tie the outer surface to a single node at the center and apply a torque (moment) at that node.
This may still be the "easy" way to model it. If you need the entire cylinder, you could then "copy: join all copies" to extrude the cylinder. Is the rubber compound homogeneous? Non-linear (I'm assuing this one)?
These are my quick thoughts...
Garland
Garland E. Borowski, PE