The Great Ongoing Aviation Thread (general and other)

Miscellaneous and Personal Stuff I Must Share
8743

As with all this stuff …

The way you improve performance is to leave off payload. Or build longer runways. How much are we as a society willing to spend on every flight or on improving every airport to prevent these very rare compound occurrences? Good question.

Seems to me we can repair all the defective engine / pylon mounts (or ground these almost obsolete airplanes) for far less money. Which leaves more money in the safety improvement budget for other much more more likely accidents.


My bottom line on all these things. The mere fact we can imagine a remedy for a particular accident is not reason to implement that remedy, no matter how seemingly easy (or actually easy) it is. It’s only after considering the total costs and total likelihoods across all the accident types and all the proposed remedies that we know which ones to fund first.

8744

I agree but the old aviation adage remains true. You can’t have a runway that’s too long or a plane with too much power. Over time we’ve managed to landlock a lot of airports. And it’s often with homes. If the winds had been different and the plane departed 35L it would have been uglier..

So that leaves adding more power as a goal we can reach. I think that’s doable.

8745

That was always wives-tale BS. The real adage is and always has been

If you don’t care what it costs (read somebody else is paying for it), you can’t have a runway that’s too long or a plane with too much power. In the real world, you gotta pay for all that waste. And just enough is all you can afford.

8746

I don’t see a cost/benefit problem here if we do not need to retrofit old planes. Just build the cameras into new planes. Sure it costs more than a $100 GoPro but on a $121 million jet is $10,000 really an issue?

ISTM letting pilots see what is really happening to their plane is a good thing all around. More information lets the pilots make better choices.

Would a camera have helped these pilots? I dunno…maybe not…the whole thing happened very fast and those poor guys had seconds to make impossible decisions. That said, having more info available I doubt is ever a bad thing.

8747

Putting this in context though, this is an old plane. The MD-11 was type certified in 1990 as a derivative of aircraft developed over time going back to 1971. The certification basis goes to amendment 25-61 with rather significant portions using much older standards.

It’s quite likely that whatever technological root causes contributed to this crash are not “possible” in a new aircraft using modern standards (I say “possible” because this is an incredibly complex subject that I’ve posted about before).

No one is building 3 or 4 engine aircraft anymore. 2 engines (or one, after a failure) are more than enough to get pretty much anything off the ground. Modern reliability and system safety requirements far exceed anything the MD-11 was built with.

This problem may be unique to them and planes of their era. Which only strengthens the need to better implement oversight and the interpretation and application of regulations pertaining to incorporating newer standards, monitoring aging aircraft, etc.

ETA: the benefit would be higher in a retrofit than a new built aircraft, but much more costly. $10k is at least an order or magnitude off if not more to implement such a change per aircraft type.

8748

You keep on with costs.

So, how much? Not a retrofit, something built into a new plane. $10,000 you say is an order of magnitude off so is it $100,000? $1 million? $10 million for a camera anyone can buy for $100?

Once done the upfront costs are gone.

How much?

8749

The certification and parts development costs are kind of similar whether it’s a new plane or an old one.

Without much more of a defined scope of work I can’t be very specific and besides, I’m disinclined to work this hard on a day off.

But starting from scratch and taking into account all the stuff I listed in an earlier post, the development and approval costs could reasonably hit $500 000, possibly much more if existing avionics boxes need upgrades. Honestly, that might be underestimating it. You’ll want to recover that in sales, and the parts manufacturers and suppliers all want a profit too. Cost to customers per plane would be quite high.

I just looked up the cost of an airworthiness directive I was involved in years ago. No computers or wiring, just moving hydraulic lines around and mounting something differently. The cost to operators is estimated at over $50,000 USD once labour is considered (to be fair, it’s a bit of a pain in the ass installation to perform). And none of that has to do anything other than exist, in the sense that it’s not communicating or interfacing with anything.

I can’t give you a more solid answer, not without investing much more time than I care to put into the hypothetical.

A replacement coffee maker for one of my customers exceeds $20k. For the unit, just to replace it, already approved.

Whatever you think things cost in the real world, they have no relation to aerospace.

8750

Per plane or a one-time up-front cost? Big difference between those two. I can’t imagine a digital camera would cost $500,000 per plane. That’d be something like 0.5% of the cost of the whole plane. That can’t possibly be the case.

8751

See, I still don’t get what difference a camera would’ve even made in this case. These poor pilots are hurtling down a runway, past V1, with a cockpit full of alarms and task saturation to the extreme. In the maybe two seconds that they had, how could they possibly have benefited from taking the extra time to look at a different screen, try to parse with they’re seeing, and create an effective plan of action that differs from SOP that they’re already executing.

It would’ve made zero difference. They were committed to fly. Even if somehow through superhuman mental computation, they were able to understand the situation in full, they only had two options. Try to fly and maybe die, versus runoff the end of the runway into warehouses and certainly die.

Every piece of information they possibly could have benefited from was already there in the cockpit in front of them and then some.

I think, as you have a number of people in this thread working in this business and closely related to matters surrounding aircraft component certification, you need to start imagining bigger.

8752

This goes back to rejecting the takeoff. Rules are past V1 they MUST takeoff…no matter what.

My point is, if they could see how truly fucked the plane was, it is better to crash on the airport.

Likely all they knew was a warning that engine #1 failed. So they sailed off into the city.

I do not blame these pilots at all. They had an impossible situation to deal with in seconds. And maybe a camera in this case would have been of no use. But it might have and seems an easy thing to add to new planes for relatively little cost.

8753

Strong disagree. They could not have possibly ascertained the degree of fuckedness in the one or two seconds that they had at most to look at the screen and make the decision. In 99.9% of engine failures at that point of the takeoff, you’d be better trying to fly, even if the engine was completely off the plane because once again, crashing into a warehouse is certain death, but a plane ought to be able to fly missing an engine even when it’s because it’s on the wing.

Upthread it’s already been pointed out where they were on the runway, at the speed they were going they weren’t crashing on the airport one way or the other.

8754

Strong disagree.

I get the rules are there for a reason. I also get these pilots had seconds to make decisions.

If (if) they knew the plane was well and truly screwed and never going to be able to fly then best to reject the takeoff and crash on the airport. A camera showing them the extent of damage to the left wing would tell them that. And it is a fucking digital camera…you almost certainly have one on your phone right now…this is not expensive or super-tech.

Would it have made a difference here? I think not since it all happened so fast. What is amazing to me is that the mere notion of adding a camera so the pilots can see what is happening to their plane is such a point of contention to many here. Planes have had all sorts of bibs-and-bobs added over the years and we are all safer for it. This is a bridge too far though???

8755

Actual pilots here have told you that wasn’t possible

8756

I dunno. I’ve been in a modified twin that had an engine failure at rotation. The extra power got us over some trees. Plan B would have destroyed the plane. I have a Tim Taylor bias towards more power.

8757

Actual pilots said the plane would run off the end of the runway and maybe into a nearby warehouse at 100 knots (give or take).

Is that worse than falling out of the sky and crashing into the city at 180 knots? (give or take)

Pick your poison.

8758

You have your speeds backwards. They had reached rotational speeds and were slowed down by the impact into the building. They almost made it. every movement forward gave them more airspeed and lift.

There’s a decision point that allows them to abort. They apparently passed that point when it happened. They should have been able to climb out on 2 engines but were not able to. Maybe they had to add more rudder because of wing damage to keep the plane level and that induced more drag. We don’t know. There just wasn’t a way out in this situation and the training is to fly it out past V1.

8759

I think you underestimate the startle factor for an event like this and overestimate the cost benefit equation.

  1. This was an incredibly rare event. A straight engine failure is itself a rare event, let alone one that appears to take out another engine on a three engine jet.
  2. When an event like this happens you need to make a decision NOW, not … look down at the screens, select the camera feed, interpret what it’s showing you, come to an accurate interpretation, and THEN make a decision. By the time you’ve done that, it’s too late, the decision has to be IMMEDIATE from the pilot’ perspective. There’s a time factor built in already to recognise a straight engine failure, but that allows for a decision that seems to be immediate to the person making the decision. If you were to factor in the time taken to interpret a camera feed you would have to do overhaul the whole stop/go decision making process and calculation.
  3. There’s no reason to believe such a system would’ve been of any benefit here (did they time to use it?).
  4. A camera system certified for this kind of use would have to be used every time you had an engine failure. So now a straight forward engine failure that should prompt an immediate go/stop needs an extra step of selecting, looking, and interpreting, a camera feed. I wager that it would result in worse decision making in clear-cut situations.

I understand the temptation to engineer a solution to every unicorn :unicorn: event, but it must be tempered by careful consideration of what is practical, useful, and what the unintended consequences might be. Whenever you introduce a new safety layer you also introduce a new failure mode and that is not helpful. Creating a system like this would be like creating a system that can fly the aeroplane using thrust alone (Sioux City accident), it would likely never save another aircraft but would introduce a heap of cost and complexity when simplicity is the most desirable state.

It’s not the tech itself, it’s the procedure to use it. There’s nothing wrong with having a camera feed of various bits of the aircraft, that’s fine. On my little A320 we have a video feed from various angles of the forward galley so we can see who is attempting to enter the flight-deck before we let them in, all good. What is not good is having a camera feed that MUST be checked during a high speed, high stakes, stop/go decision making process. It’s just not practical and would cause more problems than it solves.

8760

Just a further note on the decision making process as it currently stands.

Pilots are human beings, we are simple creatures who can easily become overloaded with information. We need a simple clear-cut decision making process. At my company (Airbus aircraft) our take-off safety briefing goes something like this:

  • Below 100 knots* I will reject the take-off for any abnormality.
  • Below V1 I will only reject for an engine failure, a fire, or unsafe condition (to cover things that are obviously unsafe to take into the air like controllability issues).
  • At or after V1 I will continue the take-off.

Modern airliners will inhibit unnecessary warnings (“dings”) during the take off roll, so essentially they will only give a warning for something that you need to reject the take-off for. This means that line two can be reduced to “if it swings (engine failure) or dings I will stop”. Simplicity rules.

Note that the decisions making gets more and more simple as speed increases and time to make a decision decreases. Below 100 knots you’ve got time to look at what’s happening and bring the aircraft to a stop in a controlled manner, you can stop for anything. Between 100 knots and V1 a rejected take-off is a more serious matter and the decision is distilled down to just two scenarios with a bit of a lawyer clause to cover unicorn events. After V1 the decision is made, you are going. There may be very rare cases where you could decide to stop but they’re not worth thinking about because they are rainbow unicorn events :rainbow::unicorn:. Adding a complication to what should be the simplest of decisions is not helpful.

There is a magic number in aviation that relates to the chances of an incident of a certain type happening. I can’t remember it and my Googlefoo isn’t working adequately, @mnemosyne might know it or know of something similar. Essentially if the chances of something happening are below a certain amount, that is considered to be ok. There is no need to spend money or writing regulations to mitigate that particular risk. When we head across the ocean and spend periods of time where we are several hours from a suitable airport it’s been calculated that the risk of more than two significant failures is low enough that it’s not worth worrying about.

If we are three hours from an airport and have an uncontained engine failure that punctures the fuselage and causes a depressurisation, that’s ok, but if it also causes a fuel leak, then you’re fucked and ICAO, the FAA, EASA, and all the other regulatory authorities are basically ok with that.

You have to realise that having an engine fall off the wing and cause a fire that causes a failure of a second engine is getting into the we’re-ok-with-that part of the risk equation.

*100 knots is an arbitrary number used to delineate the low speed take-off phase from the high speed take-off phase. Boeing, as I understand it, uses 80 knots. The actual number doesn’t matter,

8761

There’s no single equation but it comes down to system safety assessments and compliance to certain regulations, in particular 14 CFR 25.1309 but also a few others when you get into specifics of software and interconnected systems.

The methodology to make those assessments have changed over the years, especially with the FAA’s latest overhaul which is actually the new rule (from last year) that I linked to above, the one that had 198 individual contributor comments (usually submitted as one document with dozens of specific comments each).

I noted when I looked at the TCDS for the MD-11 that compliance to 25.1309 was based on the requirement as it was written at Amendment 25-23, so the regulation from 1970. Similarly, the applicable rule for Flight Crew alerting (25.1322) was from 1977 and it originally only included rules on the colours of the lights for different alerts. It now includes criteria for detecting the alerts, interpreting them under all forseeable flight conditions, the timing of alerts, that they aren’t a nuisance, etc. Much more complex and it’s now integrated into the new vision of system safety. To my knowledge, no aircraft has been certified with this new standard (ETA: the new methodology. Compliance to the latest amendment of 25.1322 has been used on many aircraft) , but it would apply to any clean sheet designs currently in work.

The “magic number” you’re thinking about is, I think, the probability of a single failure, or interrelated common caused failure of two systems (if a damaged X causes both A and B critical system to fail, that’s one failure), to lead to the inability to operate the aircraft or continue safe flight and landing. In general, the threshold is that the probability of occurrence is less than 1x10(-9) per flight hour. In other words, it’s not likely to occur during the lifetime of an aircraft fleet.

The depth of assessment and understanding of that process has changed a lot since the MD-11 was built. Those issues with the process were direct contributors to the 737 MAX crashes. It’s only gotten more severe. Which is why I said earlier that the “problem” in the design may be limited to aircraft of that era.

8762

Here’s a link to today’s Aviation Week’s comments on the recent Emergency Airworthiness Directive. Which essentially recaps my post 8716 from yesterday. This link expires in about a week but should be accessible to all until then.


As a separate matter, here’s a link to the NTSB page that seems to be keeping track of their video briefings on this accident. NTSB is also being hammered by the childish government shutdown, so although the investigative process function is (supposedly) unaffected, the PR function, which includes their website, is very much reduced. NTSB Media Summary: UPS Flight 2976 Crash During Takeoff.

As always with me, I hate info on videos and would much rather read a transcript. So I’ve not watched the vids. There is a machine-generated transcript available (expand the lede panel just below the vid to find the [show transcript] button), but it’s displayed in a format you can’t easily read, save, or search. Gaah, so close.

Anyhow, supposedly somewhere in there they say that most of the pylon was still attached to the engine found laying near the runway. Which is not the way the wing-pylon-engine system is supposed to separate.