I am trying to do a post-buckling analysis of a cylinder under bending load usig riks method in abaqus. As imperfection i used only the critical buckling mode from linear buckling analysis.The Riks algorithm was able to follow the primary equilibrium path until the limit point was almost reached, but encountered severe numerical difficulties when attempting to go into the post-buckling range. It cannot trace post-buckling behaviour and only update the load proportionality factor with a very very small number ( for example -1e-11). How can resolve this problem?