I trying to model crack propagation using XFEM in Abaqus. I applied variable load until my crack initiate and propagate to failure in my model. I would like to define how much load is implied to initiate (LI) and failure propagate (LF)?
your question contains the whole of fracture mechanics. I recommend to start careful studying of the fracture mechanics. the successful reading will hopefully lead to the answer for your question.
I know that You can define critical energy or critical strain to failure, nothing about critical load to failure. Theoriticaly, the concept of critical load is not in accordence with the failure.
Given the load in your system and having computed your FE solution, you ckeck if the corresponding value of the J-integral (or the energy release rate G_c) exeeds the prescribed/theoretical one => crack propgation
Generally the problem can be approached using to methods, one using the bifurcation analysis and one, using a failure criterion.
Let's deal with the tensile crack now:
To have crack initiation, the stress must surpass the tensile strength of material. Using basic mechanics, you should be able to find a load that results to a principle stress that is equal to materials limit, give you the limit load for crack initiation.
For crack propagation, J-integral gives you a measure of stability of crack. Also tensile strength itself can be used.
For compression crack (localization), Mohr-Coulomb is a good criterion.