Your finite element results have a good agreement with the experimental data. in my opinion this is acceptable (but you have to put both the graphs in one chart above each other so we can see better, I think it will look like the attached which is very good). back to your problem in case you want to have better results>>> in your simulation did you add post debonding frictional sliding??, if you consider this additional type of contact in your simulation then you can delete it or (suppress it) and tun the simulation again in order to have zero stiffness after debonding . Also depending on the slope comparison of your (Experimental vs FEM results) you can play with the tangential stiffness and/or the CZM parameters of your contact region in order to have a better fit between your experimental and FEM results. 

you can refer to the posts Here and Also Here for your reference