k-epsilon not converging after y+ adaptation
 

I am trying to solver a vehicle ZX symmetry model of a streamlined motorcycle (basically a streamlined body with 2 wheels protruding at the base)

The solver converged on the first run. Afterwards I decided to adapt the mesh to meet the y+ max constraint of 500 for k-e realizable with standard wall function. The model will not converge. What are my options?

-2.5M cells
-20 meters x 5 meters(symm) x 5 meters
-k-e realizable with standard wall functions
-moving ground (93.56m/s)
-air speed (93.56m/s)
-rotating wheels

Thank you.

Ok it seemed to be a problem of mesh quality. Once I got my mesh quality to 0.25 (ICEM) by allowing 0.1 tolerance on the mesh smoother it quickly converged initially and during several Y+ cell refinement adaptations.

Thanks anyway and sorry for the beginner questions.