Composite Torsional Buckling

[njc204]

I am attempting to find the buckling eigenvalues for a shaft under torsion that has sustained through damage. When I run the analysis all the mode shapes have just a few nodes near the damage moving while the rest of the model does not deform at all. I am using using shell elements.

I am now just using a single ply with the equivalent properties as the original layup and but I was wondering if there is a better way to fix this problem.

Thanks in advance.