EC225 crash near Bergen, Norway April 2016

[n305fa]

Originally Posted by turboshafts

Thanks for your detailed answer and citations!

If there where no HUMS readouts, how can they conclude it is not a sufficient way of monitoring?

In G-REDL there was indeed seen several HUMS readouts,
some of them failed to cause the attention of the engineer because
the datacard was incorrectly loaded.

So no reading, or not an abnormal reading, is not definitely answered yet?

When it comes to no chip detection, if that is the case it is indeed a strange condition.
For the 300 flight hours done on LN-OJF the gearbox had, it can´t be argued against.
But what about the 1080 hours before?
Do we know already that the gearbox had 1300 hours without any flaws at all? No.

We also know that the individual gears are not due to visual inspection before 2000 hours of operation.

Also, if they claim that it have proven against the theory that spalling does not induce any chips, it could be because the gear had already induced spalling and left the surface in such a shape that a crack would be induced.
It was then shipped out to LN-OJF

We can clearly see that the gears on the pictures have spalling.
especially the bottom left.
from the gear starts to get light pitting, which may cause hairthin
cracks in the surface, it will also soon start to spall.
Can´t say how long, but with normal lubrication and temperature it should have at least been ok for 300 hours.

So again, I don´t feel that the report brings a clear view to several
relevant aspects that should have been cleared out until now.

For sure there maybe reasons to hold off the information, until
further investigations are done, but still if the conclusion is that the gear is without warning cracked causing the gearbox total seizure. I don`t believe it.
And person with experience in risk assessement and engineering should not believe it either.

Concentric: I agree with your theory, it could be the underlying reason it happened.
And that could also be the reason they changed the material on the bevel gear from 16NCD13 to 32CDV13 to allow nitrating.
Especially in relation to the crack-prone area near the welding.
However, that should have been a failure mode, that they are perfectly aware of. Seeing the process a gear goes through during manufacturing,
it is not very likely to be a problem with only 1 gear.
And even if it was, the way we can see the surface pitting on the other gears, and also cracks due to deformation, if not occured in aspects of the gear development and certification, that process is for sure useless

Turbo shafts

Just a couple of points

REDL s HUMS was analysed by both the AAIB, EC and the system manufacturer, all of who concluded that the CIs observed gave no prediction of catastrophic failure of the 2nd stage planet gear. There were also issues with the epicyclic chip detection system which called its correct function into question

We may not know the history of the MGBs 1300 hours since overhaul, but you can be absolutely sure that the AIBN and AH do, it probably hasn't been reported because there is nothing of significance.

You can't predict how long the gear should run as the complete gear is not available to fully quantify the damage to the raceway as (I assume) you do not know enough about the gears operational load cycle, effect esn of lubrication, rate of damage progression etc.

SPalling Is the progressive failure of a bearing surface due to rolling contact fatigue, surface damage continually increases, as does particle production. Carburisation of the surface introduces comressive stresses in the bearing surface, to a depth of about 2mm. Spalling in this area will result in particle release from the raceway but not crack progression into the body of the gear/outer race. THence the use of MCDs as the primary method of detecting epicyclic deterioration (HUMS doesn't work well due to the constant movement of the planets relative to the accelerometers).

The 225 and 332 have plenty of history of detecting spalling of planet gears without catastrophic failure, therefore it is possible that this failure is not due to normal spalling. If a crack forms at the interface of the carburised layer, or in the body of the gear it will only be detectable when it breaches a surface and results in particles being released. This was what the AAIB theorised with REDL and may be a factor in the accident.

Re the jamming of the gearbox, the epicyclic ring gear damage suggests that something was entrained between a planet gear and the ring gear which burst the case, If the epicyclic had jammed the ring gear and planets would have had significant stripping of gear teeth. There was no suggestion that the REDL MGB jammed