I have been a little reluctant to comment on these failures, not knowing for sure what the cause might be. It does appear, however, Hydrogen Embrittlement (HE) may indeed be the cause. In recent newer GE engines (CF6-80C/E or newer), the LPT shaft connects into the Fan Mid Shaft as described by BadgerGrowler. I believe the FMS is made of Maraging steel and the LPT shaft of a different high strength steel. Maraging steel is somewhat sensitive to HE like normal high strength steel and it is only used because of its high shear strength and grain size does play a role in HE. The HE may come from factors in the manufacturing process or raw material (alloy) production, not as a result of operating service. Higher strength is developed by heat treatments but comes at the expense of ductility. Corrosion resistance is also reduced. Typically, high strength steels becomes HE sensitive when heat treated to give tensile strengths above 130,000 psi, perhaps higher for Maraging steel. The part must be subjected to tensile stress as would be the case of a forged shaft. Shafts do not need to be assembled or in service to be under tensile stress. The heat treating process provides most of the residual stresses. Manufacturing processes that utilize acids or electro-plating introduces hydrogen into the part material. I think shafts are electroless plated or vapor deposition coated to prevent corrosion in service.
No matter what, there is a process as barit1 described as "baking". And as he pointed out, parts must be baked within 4 hours of hydrogen exposure, less is better. Parts must be baked at 400℉, and parts must be held at 400℉ for a minimum of 4 hours, longer depending on thickness or size. If the baking process is not done properly, parts can be sitting on the shelf and can begin to show signs of HE. 70% of aerospace components that fail because of HE are related to improper "baking", one or more of the steps outlined above. Since "baking is a batch process, that is more than one component per batch, if one is bad, you can bet the farm that all the components in that batch will be bad. And this takes us to the GE situation.
Boeing essentially put out an original schedule for aircraft build for both the 787 and the 747-800 series. GE, in turn ordered parts to that build schedule, placing orders for long lead time parts first. Then technical issues at Boeing resulted in long delays/postponements for the need of engines. So the engine parts were on hand or partially processed when new delayed schedules for aircraft builds were announced. I am sure they sat on the shelves until needed. Meanwhile, the engine testing and certification programs went on and engines were delivered for flight testing of both aircraft models. IMHO, this is why the problem wasn't discovered during the normal certification process, just an opinion/speculation. Nevertheless, the delay in the program may have hurt rather than helped.
So if the problem is indeed HE, the NTSB AD to GE is to find the rest of the shafts from that "baking" batch or alloy batch ASAP assuming the "baking" process or alloy batch might be the cause. Either way, it would be traceable.
Lyman,
The joint of the LPT shaft to the FMS is not in the hot section of the engine.
TD