6 ms·

And here is the kind of catastrophe, that was prevented by Walter's calculations: https://en.wikipedia.org/wiki/Aeroflot_Flight_8641 https://en.wikipedia.org/w

by monort 7y ago

And here is the kind of catastrophe, that was prevented by Walter's calculations: https://en.wikipedia.org/wiki/Aeroflot_Flight_8641 https://en.wikipedia.org/wiki/Aeroflot_Flight_8641

WalterBright 7y ago
The 757 jackscrew gearbox is a scaled down version of the 747 one. There was one failure of the 747 jackscrew: https://en.wikipedia.org/wiki/National_Airlines_Flight_102 https://en.wikipedia.org/wiki/National_Airlines_Flight_102 A loose armored vehicle crashed into the jackscrew snapping it. There is no possible recovery from that. I cringe every time I see the video of that one. Horrible.
- rtkwe 7y ago
  I thought the main cause of the crash was the cargo weight, didn't know it had also damaged the control hydraulics. Wonder if it could have been recovered if they'd had those controls.
vijayr02 7y ago
Going down the wikipedia rabbit hole, the Aeroflot page refers to https://en.wikipedia.org/wiki/Alaska_Airlines_Flight_261 https://en.wikipedia.org/wiki/Alaska_Airlines_Flight_261 A few days ago I'd asked how the NTSB gets good co-operation from the airlines in these cases (thanks for the answers at the time!) The interesting point to me was the contrast between the 2 cases. In the Aeroflot case: > Three engineers who signed the jackscrew drawings were convicted. In the Alaska Airlines case: > The investigation then proceeded to examine why scheduled maintenance had failed to adequately lubricate the jackscrew assembly. In interviews with the Alaska Airlines mechanic at San Francisco International Airport (SFO) who last performed the lubrication it was revealed that the task took about one hour, whereas the aircraft manufacturer estimated the task should take four hours. This and other evidence suggested to the NTSB that "the SFO mechanic who was responsible for lubricating the jackscrew assembly in September 1999 did not adequately perform the task". Laboratory tests indicated that the excessive wear of jackscrew assembly could not have accumulated in just the four-month period between the September 1999 maintenance and the accident flight. Therefore, the NTSB concluded that "more than just the last lubrication was missed or inadequately performed". I couldn't find any mention of personal sanction. Just surprised by the difference in treatment between the 2 cases...
- starfallg 7y ago
  If you read further, there is an even more interesting bit - > In 1998, an Alaska Airlines mechanic named John Liotine, who worked in the Alaska Airlines maintenance center in Oakland, California, told the Federal Aviation Administration that supervisors were approving records of maintenance that they were not allowed to approve or that indicated work had been completed when, in fact, it had not. Liotine began working with federal investigators by secretly audio recording his supervisors. On December 22, 1998, federal authorities raided an Alaska Airlines property and seized maintenance records. In August 1999 Alaska Airlines put Liotine on paid leave, and in 2000 Liotine filed a libel suit against the airline. The crash of AS261 became a part of the federal investigation against Alaska Airlines because in 1997, Liotine had recommended that the jackscrew and gimbal nut of the accident aircraft be replaced, but had been overruled by another supervisor.In December 2001 federal prosecutors stated that they were not going to file criminal charges against Alaska Airlines. Around that time Alaska Airlines agreed to settle the libel suit by paying about $500,000; as part of the settlement, Liotine resigned. The deficiencies of the FAA regulator regime should be obvious from this case, but it seems that no-one was held accountable for Alaska Airline's lack of safety culture in its maintenance operation.
bumby 7y ago
Not to nitpick, but the failure mechanism in the crash and what Walter was testing are slightly different. Fatigue failures like the one in the crash are from multiple sub-maximal loads (look up "S-N" curves for more info). In other words, fatigue comes from cycling lower loads many times on a part. Buckling calculations look for maximal load failures on a column (i.e. a single larger load event that can make it fail). The link is a great example of a real world case study (and thank you for finding it!), and no doubt somebody was doing fatigue calculations on the jackscrew at Boeing as well. I just wanted to add a bit of nuance
- beerandt 7y ago
  >no doubt somebody was doing fatigue calculations I'm always surprised by the general publics' lack of awareness and understanding of fatigue. And then seeing the figurative lightbulb illuminate after asking them how they would break a metal clotheshanger, without any tools. Anyway, it may not have been his personal responsibility, but certainly fatigue must have been a significant consideration during design, since (afaik) it's most often the controlling factor in lifetime estimates for structural components in airframes. I would strongly assume the same would apply to a component who's expected failure mode might be buckling.
  - digikata 7y ago
    True, the design is one important part, but the design can be 100% on, but the manufacturing and the operational inspections are important to present failures too. Variance from controlled manufacturing conditions can result in parts that basically look and perform ok at first, but are at risk of failing early, that's why aircraft have inspection schedules - that are sometimes increased due to increased cycles on some routes/airports. (and of course there is the case of the part being damaged somehow, where some of the other comments have linked cases where this happened catasropically).
    - beerandt 7y ago
      Hmmm... I'm not real familiar with that mass production qaqc. Is there a term that describes a minimum design spec (or similar aspect), that takes into account the actual manufacturing (and other) variances? I'm thinking that in construction, a rough equivalent would be a designs constructability / bidability review.

by monort 7y ago

And here is the kind of catastrophe, that was prevented by Walter's calculations: https://en.wikipedia.org/wiki/Aeroflot_Flight_8641 https://en.wikipedia.org/wiki/Aeroflot_Flight_8641

WalterBright 7y ago

The 757 jackscrew gearbox is a scaled down version of the 747 one. There was one failure of the 747 jackscrew: https://en.wikipedia.org/wiki/National_Airlines_Flight_102 https://en.wikipedia.org/wiki/National_Airlines_Flight_102 A loose armored vehicle crashed into the jackscrew snapping it. There is no possible recovery from that. I cringe every time I see the video of that one. Horrible.

rtkwe 7y ago
I thought the main cause of the crash was the cargo weight, didn't know it had also damaged the control hydraulics. Wonder if it could have been recovered if they'd had those controls.

vijayr02 7y ago

Going down the wikipedia rabbit hole, the Aeroflot page refers to https://en.wikipedia.org/wiki/Alaska_Airlines_Flight_261 https://en.wikipedia.org/wiki/Alaska_Airlines_Flight_261 A few days ago I'd asked how the NTSB gets good co-operation from the airlines in these cases (thanks for the answers at the time!) The interesting point to me was the contrast between the 2 cases. In the Aeroflot case: > Three engineers who signed the jackscrew drawings were convicted. In the Alaska Airlines case: > The investigation then proceeded to examine why scheduled maintenance had failed to adequately lubricate the jackscrew assembly. In interviews with the Alaska Airlines mechanic at San Francisco International Airport (SFO) who last performed the lubrication it was revealed that the task took about one hour, whereas the aircraft manufacturer estimated the task should take four hours. This and other evidence suggested to the NTSB that "the SFO mechanic who was responsible for lubricating the jackscrew assembly in September 1999 did not adequately perform the task". Laboratory tests indicated that the excessive wear of jackscrew assembly could not have accumulated in just the four-month period between the September 1999 maintenance and the accident flight. Therefore, the NTSB concluded that "more than just the last lubrication was missed or inadequately performed". I couldn't find any mention of personal sanction. Just surprised by the difference in treatment between the 2 cases...

starfallg 7y ago
If you read further, there is an even more interesting bit - > In 1998, an Alaska Airlines mechanic named John Liotine, who worked in the Alaska Airlines maintenance center in Oakland, California, told the Federal Aviation Administration that supervisors were approving records of maintenance that they were not allowed to approve or that indicated work had been completed when, in fact, it had not. Liotine began working with federal investigators by secretly audio recording his supervisors. On December 22, 1998, federal authorities raided an Alaska Airlines property and seized maintenance records. In August 1999 Alaska Airlines put Liotine on paid leave, and in 2000 Liotine filed a libel suit against the airline. The crash of AS261 became a part of the federal investigation against Alaska Airlines because in 1997, Liotine had recommended that the jackscrew and gimbal nut of the accident aircraft be replaced, but had been overruled by another supervisor.In December 2001 federal prosecutors stated that they were not going to file criminal charges against Alaska Airlines. Around that time Alaska Airlines agreed to settle the libel suit by paying about $500,000; as part of the settlement, Liotine resigned. The deficiencies of the FAA regulator regime should be obvious from this case, but it seems that no-one was held accountable for Alaska Airline's lack of safety culture in its maintenance operation.

bumby 7y ago

Not to nitpick, but the failure mechanism in the crash and what Walter was testing are slightly different. Fatigue failures like the one in the crash are from multiple sub-maximal loads (look up "S-N" curves for more info). In other words, fatigue comes from cycling lower loads many times on a part. Buckling calculations look for maximal load failures on a column (i.e. a single larger load event that can make it fail). The link is a great example of a real world case study (and thank you for finding it!), and no doubt somebody was doing fatigue calculations on the jackscrew at Boeing as well. I just wanted to add a bit of nuance

beerandt 7y ago
>no doubt somebody was doing fatigue calculations I'm always surprised by the general publics' lack of awareness and understanding of fatigue. And then seeing the figurative lightbulb illuminate after asking them how they would break a metal clotheshanger, without any tools. Anyway, it may not have been his personal responsibility, but certainly fatigue must have been a significant consideration during design, since (afaik) it's most often the controlling factor in lifetime estimates for structural components in airframes. I would strongly assume the same would apply to a component who's expected failure mode might be buckling.
- digikata 7y ago
  True, the design is one important part, but the design can be 100% on, but the manufacturing and the operational inspections are important to present failures too. Variance from controlled manufacturing conditions can result in parts that basically look and perform ok at first, but are at risk of failing early, that's why aircraft have inspection schedules - that are sometimes increased due to increased cycles on some routes/airports. (and of course there is the case of the part being damaged somehow, where some of the other comments have linked cases where this happened catasropically).
  - beerandt 7y ago
    Hmmm... I'm not real familiar with that mass production qaqc. Is there a term that describes a minimum design spec (or similar aspect), that takes into account the actual manufacturing (and other) variances? I'm thinking that in construction, a rough equivalent would be a designs constructability / bidability review.