Printed parts would be difficult to model accurately, because the strength is highly dependent upon layer direction. So you'd have to use both an orthotropic material model (to take into account layer fusion or lack of it) and a non linear stress strain curve to model plastic behavior. You could ignore infill and just consider the shell strength, but that might be overly conservative especially for all loading except bending and torsion.
In my experience with FEA (extensive I've been using it since the mid 80's professionally) it's not really practical. But you might be able to get good enough results to compare designs. But as far as predicting actual failure loads, that would be a crap shoot.