I placed a constraint across the face of the bracket that will be in contact with fork. Is this sufficient, or should I model a face to represent the fork, and then bolt the bracket to that?
I think I'd do the latter, because the bolt head contact is where the load actually gets transmitted. Constraining the interface with the fork crown is more like a glue joint. You don't actually have to model the bolt, just create a spot face on the bracket the size of the bolt head and apply a load consistent with the bolt tension.
Now you've made me check that you can simulate the friction between the fork crown and the bracket, because without it the bracket will just fall off as soon as you apply a load, or it will rest on the bolt shank which is not what actually happens in real life.
I think I'd do the latter, because the bolt head contact is where the load actually gets transmitted. Constraining the interface with the fork crown is more like a glue joint. You don't actually have to model the bolt, just create a spot face on the bracket the size of the bolt head and apply a load consistent with the bolt tension.