In radial RMA, the deviatoric parts of trail stress and iteration stress have the same direction, so the calculated shear stress after convergence is just a scale down to the trail shear stress calculated from the elastic assumption.
This constraint does not make sense. Anyone has an explanation?