Convergence issues with 'Stress Intensity Factor' (K results) for Composites

Hello, 

I am doing a contour integral simulation for Composites. Even with very finer mesh, my K- factors do not converge- an example of my results  given below:

My SIF result -  220.6 (1st contour)           197.7 (2nd contour)             214.1 (3rd contour)              218.0 (4th contour)             219.1 (5th contour)   

I observed good convergence for the isotropic materials. Please, any comments on what the reason could be that causes this issue specifically for composites?

Best Regards,

Wong.