Originally explained on the blancolirio channel on YouTube -
The timing and manner of the break make a lot more intuitive sense when you consider that the engine is essentially a massive gyroscope. As the plane starts to rotate, the spinning engine resists changes to the direction of its spin axis, putting load on the cowling. When the cowling and mount fail, that angular momentum helps fling the engine toward the fuselage.
From a failure analysis perspective that is much less relevant though. The first failure was the rear engine mount if it had been a secondary failure it would have been deformed first and then broken, and it clearly is not. It just tore in half on one of the four connections and then the rest deformed slightly due to overstress.
It was however relevant to the survivability of the accident: if the left engine wouldn't have detached, or would have detached in a more "manageable" way, the other engine (probably the tail engine from how it looks) wouldn't have been affected too, and the pilots would have had a better chance to take off. Plus the whole "when an engine detaches, it shouldn't start a fire in the wing it was attached to" part of course...
Yes, agreed, the secondary safety wasn't there either. There was a Boeing accident near Amsterdam with the plane crashing into an inhabited area, it had dropped two engines but kept flying, at least for a while...
A 747 with 2 engines out that's already got some altitude and speed is much more "survivable" (still extraordinarily difficult) than a trijet trying to take off on 1 engine, which is impossible.
Planes can safely land with 0 engines, though this is obviously "not ideal." See: Gimli Glider.
Yup. That's exactly what experts said of American Airlines flight 191 which was basically the same engine mount, same failure. Engine flipping over the wing.
The failure of the pylon appears to be different. On AA 191, the pylon rear bulkhead cracked and came apart. In the case of UPS flight 2976, the pylon rear bulkhead looks to be in one piece, but the mounting lugs at the top of the rear bulkhead cracked.
American 191's engine mount failed because of improper maintenance. It remains to be seen whether this failure had the same cause or if it was something else, such as metal fatigue.
A failure due to metal fatigue would still be a failure to properly maintain the aircraft, right? I know by "improper maintenance," you're referring to actual improper things being done during maintenance, and not simply a lack of maintenance. But I'm reading things like "the next check would've occurred at X miles," and, well... it seems like the schedule for that might need to be adjusted, since this happened.
Yes, when I said "improper" I meant the American 191 maintenance crew took shortcuts. The manual basically said "When removing the engine, first remove the engine from the pylon, then remove the pylon from the wing. When reattaching, do those things in reverse order." But the crew (more likely their management) wanted to save time so they just removed the pylon while the engine was still attached to it. They used a forklift to reattach the engine/pylon assembly and its lack of precision damaged the wing.[0]
Fatigue cracking would be a maintenance issue too but that's more like passive negligence while the 191 situation was actively disregarding the manual to cut corners. The crew chief of the 191 maintenance incident died by suicide before he could testify.
> The crew chief of the 191 maintenance incident died by suicide before he could testify.
To be clear, a crew chief (Earl Russell Marshall) did. But he wasn't directly involved in maintenance of the specific DC-10 that crashed. Or at least, I haven't found a source saying he was, and some sources say he wasn't. https://www.upi.com/Archives/1981/03/26/The-wife-of-an-airli...
If the (FAA-approved) maintenance schedule says "the next check should occur at X miles" and X miles hasn't happened yet, then it's not going to be classified as improper maintenance -- it's going to be classified as an incomplete/faulty manual.
Now, of course, if that maintenance schedule was not FAA-approved or if the check was not performed at X miles, that's going to be classified as improper maintenance.
According to various comments the plane was nowhere near the cycling for a special detailed inspection of the aft pylon mount lugs: SDI is at 29200 cycles and the plane had 21043.
There was a lubrication task in October, but according to tech comments that would just in greasing the zero fittings, no taking apart anything.
Those pictures of that torn up part are pretty hefty, that's a clean break, no stretching as far as I can see it just tore the material in half, you can see the grain. There does not seem to be any torsion either so most likely that was the first part to go, if the problem had been in the engine then I would expect this part to be mangled, not pulled apart. What stress damage there is occurred shortly after that first break. A valid question would be whether or not that crack was there before take-off or not.
I'm very curious what the metallurgic analysis of the mirror part on the other wing will come up with, especially whether there are any signs of stress fractures in there. If there are that will have substantial consequences for the rest of the still flying MD11's, about 50 or so are still in service.
The preliminary report mentions fatigue cracks on both sides of the aft lug, and one side of the forward lug, with the other showing no trace of fatigue, only overstress.
From this it seems like the aft lug was way fucked, and the forward lug was hanging on for dear life, until it could not.
Yes, that's exactly how I read it. The aft lug was the first to go, the forward lug shows signs of stress so it held on longer.
I don't think they're going to be skimpy on the metallurgy report so looking forward to the analysis of the mirror parts on the other wing. Those will tell without a doubt whether it was maintenance related or age related fatigue. Right now I would bet on the latter but the former could also still be a factor, for instance, that bearing might not have had enough lubricant.
It depends. This aircraft was made near the beginning of the MD-11 production and if the original analysis for the fatigue life of this location was wrong, then you would expect to see that appear in older aircraft first. If that ends up being the case then it's not an inspection or maintenance issue, it's an engineering failure. Given aerospace accident history I would say that is less likely than some maintenance issue but we won't know for sure for a bit.
Even if it was an inspection or maintenance issue (which given the kind of failure and available data looks increasingly doubtful, though it can not yet be ruled out) this part failed in a catastrophic way when it should have had ample engineering reserve over and beyond the load to which it was subjected. It just snapped clear in half those breaks are indicative of a material that has become brittle rather than that the part deformed first and then broke due to excess stress.
In other words, a slow motion video of a camera aimed at that part during the accident would have shown one of the four connections giving way due to fatigue cracks and then the other three got overstressed and let go as well, in the process damaging the housing of the spherical bearing.
The part at the bottom of page 9 is the key bit. Now I very much want to see the state of the mirror part on the other wing, that will show beyond a doubt whether it was maintenance or an over-estimation of the design life of that part.
It would also be interesting to have a couple of these pulled from the fleet and tested to destruction to determine how much reserve they still have compared to the originally engineered reserve.
According to the preliminary report, 3 of the 4 showed fatigue cracks, and the 4th overstressed. So yes, agree a random sample of these parts should be pulled from the fleet and tested - but something pretty crazy was happening here re: fatigue.
That it was so far from the maintenance schedule to be inspected AND that the fatigue cracks seem to have formed in areas that would be hard to visually inspect anyway points to either a engineering problem (especially bad, since the DC10 problem of a similar nature happened in roughly the same parts, albeit due to different abuse - you’d think the engineers would overdo it there, if nothing else), or some specific type of repeated abuse that particular pylon received, which is pointing more to a design problem.
Re-reading the 1979 report might be helpful here. This isn't my field, but it seems that the engine is attached "hard" to the pylon, then the pylon is attached via a bearing mount system to the wing frame. The bearings wear out, and hence have to be replaced (not sure how often, but they were doing it on the entire fleet prior to the 1979 crash). The 1979 investigators thought that the fatigue cracks were caused by removal of the entire pylon/engine assembly as one unit (because that put excess stress on the aft bearing, they suspected due to support being provided from below by a fork lift). After the 1979 accident engines had to be removed first, then pylon, supposedly removing that cause for mount cracking. Perhaps there was another cause.
Before the 1979 accident, engines also had to be removed first.
Airlines have to follow the approved maintenance manual procedures; that manual called for engine removal and installation from a pylon that was on the wing. American was improvising a maintenance procedure without the legal authority to do so, resulting in 191.
There is nothing here to say it being a maintenance issue is doubtful. It could quite literally be a similar issue to Flight 191, we don't know yet.
It did have ample engineering reserve beyond the load it was subject to... before fatigue damage initiated a crack which then grew until there was no reserve left. The question is why the fatigue crack initiated prematurely? Maintenance damage? Analysis mistake? We don't know yet.
If you read the original 1979 report in full, I think you'll begin to realize that this "improper maintenance" thing was a cover-up. Actually quite similar to the 737MAX -- find someone or something to blame other than the design of the aircraft, then move on.
The picture of that part that is torn into two pieces certainly seems to suggest so, that's a clean break, not an overstressed part deforming and then breaking.
Ironically, AA flight 191 could have been salvageable, because the engine detaching didn't start a fire. However, it led to loss of hydraulic pressure on that wing, which led to the flaps/slats retracting on just the left wing, which led to the plane becoming uncontrollable. After that accident, the DC-10 was retrofitted with hydraulic fuses to prevent something like this happening again. Unfortunately, that didn't help the UPS crew, because in their case, the detachment caused more damage to the wing...
Flipping backwards is what caused the engine to fly to the right and land to the right of the takeoff runway. The stills in the NTSB preliminary report clearly show the engine flying over the aircraft, to the right, and then heading straight down.
Because making every jet engine in two different models would make them a lot more expensive. It would also cause maintenance issues because of non-interchangeable parts.
Not an aviation expert at all, so take this with a grain of salt, but I think "the spinning engine resists changes to the direction of its spin axis" offers two important insights:
* why it failed at rotation (the first/only sudden change of direction under full throttle) rather than as soon as it was mounted onto the plane, while taxiing, as soon as they throttled up, mid-flight, or on landing. This is important because at rotation is the worst possible time for this failure: no ability to abort take-off, no ability to land safety under no or severely limited power, little time to react at all, full fuel. Knowing these failures are likely to manifest then stresses the importance of avoiding them.
* why it failed in such a way that it damaged the rest of the plane.
Not so much what was wrong with the mounting in the first place, if that's what you're asking. Presumably it was designed to withstand the forces of this moment and clearly has done so many times before.
Well, some force flung it inboard and above the fuselage (gods, that CCTV stills sequence.) Knowing that the engine rotates CCW, there are not many candidates.
> There are lots of candidates for a failing engine yeeting itself in any direction.
For the precise trajectory, certainly; for the general direction, not so much. Could you describe a combination of forces that would have thrown that engine to the left of the direction of travel? (We're talking about this accident, not any engine anywhere.)
My broad comment is that gyroscopic precession having any critical role in this is incredibly far fetched. That said, I've never flown or worked on a turbofan so ¯\_(ツ)_/¯.
Gyroscopic forces might have changed the direction of travel a few degrees, but the motive force comes from the engine's thrust, the power of its spinning blades pushing air. An engine cut loose at full power moves forward like a missile.
Yes, obviously; MD-11s aren't flinging engines off the wing every single takeoff. A 34 year old airframe may or may not actually match design strength, though.
Yes, but the point is that this moment of the takeoff is when a failure that's been waiting to happen is most likely both because of the thrust and the gyroscopic resistance.
Did I understand the report correctly that the part was scheduled to be replaced in the future after a certain number of hours, it just hadn't hit the threshold yet ?
If you're referring to this quote (excerpted from the AVHerald article linked elsewhere in the thread), I don't think so:
> At the time of the accident, N259UP had accumulated a total time of about 92,992 hours and 21,043 cycles [..] A special detailed inspection (SDI) of the left pylon aft mount lugs would have been due at 29,200 cycles and of the left wing clevis support would have been due at 28,000 cycles
This isn't talking about replacement, only inspection; and it wasn't going to happen in the near future: 7k cycles at four flights/day means inspection is due in 5 years.
It wasn't doing four flights per day. As a long-distance cargo aircraft it was doing two flights per day, and I doubt it was flying every single day of the week.
So we are talking about at least 10 years before that inspection was due.
I'd be very surprised to read that the aft lug that cracked (and the bearing it contained) were made of aluminum. They were almost certainly steel or Inconel.
I know Veritasium gets posted here a lot, but a few days ago he posted a deep-dive into the the engineering of jet engine turbine blades. Turns out they're made from a single crystal of a superalloy that entangles itself at a molecular level such that it actually gains strength as it's heated, only losing strength above 1200 degrees C / 2200 degrees F. Below that temperature, as long as the strain on the part is below the plastic deformation threshold, it's not really losing any strength at all over time.
No; roughly, yes. Based on the crystal structure of the metal, fatigue works differently.
> The fatigue limit or endurance limit is the stress level below which an infinite number of loading cycles can be applied to a material without causing fatigue failure.[1] Some metals such as ferrous alloys and titanium alloys have a distinct limit,[2] whereas others such as aluminium and copper do not and will eventually fail even from small stress amplitudes.
The timing and manner of the break make a lot more intuitive sense when you consider that the engine is essentially a massive gyroscope. As the plane starts to rotate, the spinning engine resists changes to the direction of its spin axis, putting load on the cowling. When the cowling and mount fail, that angular momentum helps fling the engine toward the fuselage.