@Lonewolf_50:
My question was purely hypothetical as you know, and proposed on the observation that hundreds of these planet gears have been manufactured and not failed catastrophically in the same application. But this one did fail, so why? Transit damage might be one answer and is being further investigated but other answers may also exist. Fatigue obviously has propagated a crack but it would need a source, which could be quite small, probably outside the residual compressive stress zone from the carburising, and if so might not cause sufficient spalling at an early enough stage to be detected. That seems to be what AIBN are reporting.
I would not rush to the conclusion that there exists a ‘bad batch’ of components and with the affected type(s) currently grounded any worrying would be commercial and not safety based. The failed bevel gear shaft of G-REDW was initially put down to a bad batch with roughly drilled holes and we know the rest of that story.
The manufacture of these safety critical gears would be subject to many controls and inspections. Baking would most likely be among them and that process would be tightly controlled (possibly empirically), but I don’t know how you could actually determine the removal of hydrogen had been homogeneous and irreversible. Is baking done only after carburising or also after etching for final polishing?
There are so many factors that need to be achieved in a complex manufacturing system to make this gearbox safe. The holes in the Swiss cheese do not necessarily need to be concentric, just the slightest overlap may do it, like a build up of tolerances. That the AS332 up to L1 variant has never suffered this fate once in 4m hours says a lot for the system, but does the L2/EC225 just occasionally nudge one cheese a little too far?