I am trying to run a centrifugal loading (628rad/s=6000rpm) on a rotor; I have simplified this as much as I could, the contacts are no penetration with friction and where possible, bonded and cut the assembly to the 1/32nd portion (cyclic symmetry).
The problem is that sometimes I get very distorted results with high localized stresses where they shouldn't be. Note that I am constantly changing the mesh on certain parts in order to find the real values of the stresses at concentration sites (i.e. fillets, holes).The simplfied bolt (EN24T steel) is covered with an insulation of Glass fibre. These then go through a series of rings made of different materials (steel, GF, aluminium). In between I have lamination steel 21 mm thick which i have modelled as 7 bits of 3mm each with no penetration contacts between them.
As you can see, there is no way that that stress is real, especially that a few elements away the stress drops to almost nothing. Does anyone have any ideas why am I getting such results for some meshes? I ve also tried to see if the max aspect ratio of the mesh is responsible for this, but even for a low max AR like 7 I get this kind of results.