Pitot Tube Failure

[LH-B744]

Hello everyone,,
Involved in a discussion of the computerization of recently produced automobiles, trying to make an analogy to fly by wire. Attempting to reference pitot tube failure. Got hung up on how fly by wire systems compensate for loss of pitot tube sensing. I know that not all systems are dependent on working pitot tubes. But all I remember of the long detailed discussions of crashes involving pitot tubes is something about power and attitude. So maybe someone can remind me of where the pilots turn when they no longer have air speed data. It is my impression that air speed is very critical to flight management, but pilots can still aviate after the air speed is unreliable.

First of all, let me welcome you to jetphotos. In my eyes it's still somehow incredible. I'm here since, ok half an eternity. There are other seniors who are here since a full eternity.

So, you are completely new here on this platform. Ok, if you like to say that. But never on Earth you are new to aviation. Nevertheless, I was 30 (thirty) years old when I wrote my first jetphotos forum entry. So, I was a jetphotos junior at the age of 30!

Back on topic. You are not new to aviation, imho. So, what are we talking about? Come on. France is such a great country, and that's the opinion of a man who certainly knows more Italian melodies... The A320 was invented in France, you mentioned fly by wire. That's not new. LH ordered A320s beginning with the year 1985.
Two quite serious aviation accidents happened on board of an A320 or A330, in combination with 'pitot tube failure' and/or the really not new fly by wire in an A320.

1. 1988, June 26th. Which must have been an inauguration flight for the A320 in France. But somebody chose the completely wrong Flight Captain for such a procedure. 44 years old by then, so, born 1944, he must have thought that he is something like a God. Although in 1988 there definitely had been jet pilots who fly the A320 much more careful than this daredevil. Not necessarily the LH CEO, but Spohr spontaneously comes to my mind.
Now, I am 41 years old. So, old enough to be a thoughtless daredevil? People who are with me in a car when I drive do not confirm that. A pilot (passage airline) has to be aware that he does not fly dead pigs. But 130 passengers, back then, 1988, at Habsheim. You can be 44 years old with xx,xxx flight hours on whatever a/c types on this planet, I only count the flight hours on type A320,
especially for a so called flight captain who is
a) on a completely new a/c type and
b) on a tiny airport where neither he nor his F/O has ever been before and
c) who does not obey the 11th commandment in aviation,

You shouldst never fly a passenger jet in low-level flight.

Definition: not below 1970 ft AGL !

If I know this 11th commandment, then he also knew it, I'm sure. And he ignored it, without the exact knowledge of the 'alpha protection' rule, which is a killer combination together with low-level flight in unknown terrain.

Who am I to teach a man who was born 1944 how to fly a jet full with passengers?! There is a difference between Cargo and Passage. Passengers do write down what they experienced after the flight, in contrast to dead pigs.

The second case, AF 447, is the best example for 'how to NOT react to a temporarily pitot tube failure'. Again, passengers died. But not due an egomaniacal Flight Captain. During AF 447 the Captain, Monsieur M.D., obviously was the only one who could have rescued that flight.
No pilot names. But that Captain deserves my respect. Within seconds he saw what went wrong, with his 6300 flight hours as a Flight Captain (responsible Flight Commander).

I have heard of flights with more than 1 pilot in the rank of a Captain (four stripes), so, rather 2 Captains and 1 'apprentice' (F/O). So that always 1 instructor is present in the cockpit, to avoid the 'sleep gap', to avoid misunderstandings, especially on the long haul. That's not cheap. But in case of AF 447 it obviously had saved 228 souls..

Two newbies in an A330-200, and the Captain needed his sleep, so left the cockpit. That's the first thing which I've learned when I was a young aviation enthusiast, 30 years ago. You pull the yoke, that's one of the best ways to slow the bird down. In a Cessna 172, in a Fairchild Swearingen Metro, in a LH-B744, and also
in an Airbus A330-200.

So, what should you do in case of pitot tube failure. Please do me a favor and do not pull the yoke. You can try to descent or accelerate. But the best way of reaction is, and this is also what the BEA (Bureau d’Enquêtes et d’Analyses pour la sécurité de l’aviation civile) published back then,
do nothing.

And I have to second that. All jets which I know since I am here on this platform have one very cute quality. A320, A330, A340, A380, B763ER, B777, B757, B747, all of them try to gain speed after a descent which was (apparently) begun with a loss of speed.

That means, all these jets are able to regain a stable flight situation , if not,
yes, if not a F/O who knows nothing
tries to pull the yoke out of that jet.

Please don't pull the yoke if you try to avoid a stall.

Leave the elevator alone, and your aircraft will recover! That's what the experienced AF 447 Flight Captain tried to say, but the newbies called him too late..

[Gabriel]

Such as...

Such as if you don't remember the memory items, were not trained for high-altitude UAS, are startled, or a a betting man that like cowboy improvisation and willing to not go around when you just lost visual contact with the runway after passing minimums and receiving 13 GPWS alerts.

[Evan]

Do follow the procedure (memorized items included) but, if for whatever reason you are not going to do that...

Such as...

[Black Ram]

This system was optional in the planes types that even offered it as an option.
And then I heard critiques because AoA is very non-sensitive to airspeed at high speeds, meaning that you need big variations in airspeed to make a meaningful change in AoA (always keeping 1G) and that, with that considered, it was even less accurate than P+P=P (although I don't understand why, since pitch -one of the P- is as sensitive as AoA, in delta-pitch = delta-AoA if flight path remains constant).
But anyway, I've seen this system being dis-recommended by some experts.

Came in handy during the 2018 MH134 incident where they lost all pitots. Also I recall reading the BUSS system is standard equipment on the A350.

[3WE]

as I recall it.

If that's your position, then I maintain that my position was absolutely do NOT pull the stick 100%. Instead remember that P+P=P is going to keep you right where you are while you figure out what the hell is wrong with all of the warnings going off/sensory overload (and hell yeah, check the book to be sure what the hell is really going on).

[Gabriel]

Awesome selective memory bro'

My selective memory is that just about every single time I brought up pausing to use a familiar pitch and power while you deal with sensory overload, you said we could only go immediately to the Book of Acronyms.

My selective memory also includes some dude named ATLcrew thinking he might initiate a robustly-gentle descent to get away from coffin corner (assuming he might be pretty high, of course).

Gabriel- what's your recollection?

That the memory items came first. That prescribes a P+P that is supposed to be memorized so you don't need t go and look into the Book of Acronyms where they come from.
THEN... you go to the Book of Acronyms to fine tune the P+P.

That was always Evan's position as I recall it.

Your position was that using "typical" P+P in lieu of the memorized values was about equally ok (what actually seems to match the procedure at the time that called for the memorized P+P "if needed to stabilize the flight").

And my position was somewhat in the middle: Do follow the procedure (memorized items included) but, if for whatever reason you are not going to do that, use a typical (or even slightly atypical) P+P for the condition. Whatever except pulling up so hard that you make the stall warning sound immediately and then make a 1.5G 7000FPM climb to 2500 higher than the max altitude you were willing to climb and when the stall warning sounds again at the top, pull up hard again and keep pulling up hard all the way to FL 0.

As far as I recall.

[3WE]

For (guessing here) the 1,568,669 time: first stabilize with memorized P + P, then, when stabilized, reference the QRH to get more accurate P + P. But forget that, you're never going to get it (or acknowledge that you get it).

Awesome selective memory bro'

My selective memory is that just about every single time I brought up pausing to use a familiar pitch and power while you deal with sensory overload, you said we could only go immediately to the Book of Acronyms.

My selective memory also includes some dude named ATLcrew thinking he might initiate a robustly-gentle descent to get away from coffin corner (assuming he might be pretty high, of course).

Gabriel- what's your recollection?

[Evan]

I said, "I agree"...But please, argue on and again, thanks for re-confirming your black and white mindset that you can only look up P+P=P values from the Book of Acronyms.

For (guessing here) the 1,568,669 time: first stabilize with memorized P + P, then, when stabilized, reference the QRH to get more accurate P + P. But forget that, you're never going to get it (or acknowledge that you get it).

(By the way- I'd ALSO say that P+P=P IS IN FACT ALREADY incorporated into MOST flight automation logic...I (yes, literally 3BS) can tell most airliners to go to such and such altitude and fly at such and such speed...All I have to do is watch a youtube and briefly practice working the autopilot on MSFS. I know you can explain it in 5000 words or less, but hey, THE LOGIC IS THERE. (Extra significance to the period))

The key is to modify the logic to allow the autopilot to remain engaged without airspeed data, by replacing it with 'calculated airspeed' derived from the remaining data sources and the P + P that autopilot commands. The PFD would have to replace the speed tape with something like the red-green scale. It could also drive the FD's if the pilots opt to take manual control. I don't even see why the pitch envelope (AoA) protections can't remain active, since everything needed to establish AoA is still functioning, however low energy and overspeed protections might have to go.

Now run the AF447 scenario:

- Airspeeds are lost.
- Nothing happens, specifically, there is no sudden startling event, no autopilot disconnect, no suddenly having to take the controls in a turbulent situation.
- Therefore there is no upset and no opportunity for pilot error due to startle factor and stress.

PM: Uh... we're in UAS mode, we've lost the speeds.
PF: Ok, lets keep on top of it. It should reset in a minute or two. QRH please.

[Gabriel]

Shall we agree to disagree that unintended consequences and temporal accuracy (while the speed settles) are valid answers to this question:

(quote=Gabriel, verbatim): A good question could be why, to this day and to my knowledge, P+P=P has not been incorporated in the flight automation logic.

No, I can't. DOING A has accuracy challenges and unintended consequences. NOT DOING A has accuracy challenges and unintended consequences.

You cannot take a decision NOT to do A based on the consequences of YES doing A and disregarding the consequences of NOT doing A (i.e. the consequences of the decision taken). I mean, you can, but don't ask me to agree.

If you see my attempt of an answer you see both sides of the coin:
(quote=Gabriel, verbatim): "manufacturers preferring a bigger risk of a deadly crash where the airline is liable rather than a very much smaller chance of a deadly crash where the manufacturer is liable."
Blue is NOT DO A and purple is DO A.

(By the way- I'd ALSO say that P+P=P IS IN FACT ALREADY incorporated into MOST flight automation logic...I (yes, literally 3BS) can tell most airliners to go to such and such altitude and fly at such and such speed...All I have to do is watch a youtube and briefly practice working the autopilot on MSFS. I know you can explain it in 5000 words or less, but hey, THE LOGIC IS THERE. (Extra significance to the period))

Correct, a lot of the logic needed to manage a UAS without sudden unexpected disconnection of the autopilot is there, and the part that is missing is just code. No need for extra hardware, sensors, wiring or annunciation lamps. Very much like AoA indication and TOPMS. (in less than 5000 words).

[3WE]

I said, with an automated UAS mode engaged. "you do nothing" about flight control (unless you opt to take over).

I said, "I agree"...But please, argue on and again, thanks for re-confirming your black and white mindset that you can only look up P+P=P values from the Book of Acronyms.

Blah Blah Blah

Ok, I appreciate that there was acknowledgement on many points.

Shall we agree to disagree that unintended consequences and temporal accuracy (while the speed settles) are valid answers to this question:

(quote=Gabriel, verbatim): A good question could be why, to this day and to my knowledge, P+P=P has not been incorporated in the flight automation logic.





(By the way- I'd ALSO say that P+P=P IS IN FACT ALREADY incorporated into MOST flight automation logic...I (yes, literally 3BS) can tell most airliners to go to such and such altitude and fly at such and such speed...All I have to do is watch a youtube and briefly practice working the autopilot on MSFS. I know you can explain it in 5000 words or less, but hey, THE LOGIC IS THERE. (Extra significance to the period))

[Gabriel]

Essentially NONE of your aeroengineer colleagues ever decided this was a proper thing to do...

And then you ask why.

"Because computers are good" doesn't answer that.

I didn't offer that for an answer. I did offer this as a plausible one: "manufacturer preferring a bigger risk of a deadly crash where the airline is liable rather than a very much smaller chance of a deadly crash where the manufacturer is liable."

And in any event, 20 stupid answers that I might have given wouldn't make your answer any better than it was.

Acknowledged: My answers were the purest of parlour-talk ass-hattery.

I am not questioning your answer for being parlour talk (this forum is mostly parlour talk after all) but for not being an answer to the question. Not a meaningful one at least.

OR you can do it they way you have done every last departure from cruise altitude for the past 6 months.

You mean let the FMS tell the AP to do it while the human pilots closely monitor the instruments and make the relevant ATC calls and PA's??

Final thought- put your P4FA system in the CONTEXT that a plane is near coffin corner...Is +/- 10 knots good enough? +/- 5 knots?

Unlikely to be that narrow, but even if so, it is irrelevant (as an answer to my question) for the reasons already explained above in this post and previous ones..

And again, what caused the pitot failure? Are there some modes where P4FA is going to do the wrong thing?

Oh, I'm sure there are. Everything has ways to go wrong. Again, what's your point?

Sitting here today, I can't imagine what that is, but a lot of smart people approved a stall-preventing-nose-over system for not having bad unintended consequences.

That's an example of how NOT to do things. Among other stuff, they did NOT have triple readings. Or double.

I am guessing that aeroplanie designers have put lots of thought into how to handle Pitot Failures and designed what might be the most robust system to address it. Aggie ass-hat guess, but...

Aeroplaine designers have other stuff in their agendas other that safety. As the previous example you mentioned shows. And as my previous answer to my own question also speculates.

[Evan]

Essentially NONE of your aeroengineer colleagues ever decided this was a proper thing to do...

I doubt that. It may be true that none of the high-level decision-makers have thought it was necessary, in a climate where every unnecessary penny is shaved in the name of profits.
If you've been paying attention, the current philosophy, even at Airbus, is to place trust mainly in correct pilot actions. Pilots are generally trusted over systems, which is why so may pilot actions are still needed when things go wrong, despite the fact that these actions could be automated. I think this may be due in part to...

"manufacturers preferring a bigger risk of a deadly crash where the airline is liable rather than a very much smaller chance of a deadly crash where the manufacturer is liable."

I find it plausible that airframers do not want to take on such liability.

As I told Evan, is there anything specific you want me to tell my neighbors at Boeing to see if they can get the suggestion over to the development department? I repeat my concerns that they seem a bit disinterested regarding the subject of software modifications.

This is the hostile, dismissive and belittling argument I would expect from insiders like ATL, but you of all people know that this is a discussion forum for the purpose of discussion, for the sake of discussion.

As to another post above- where Evan yelled, "You do nothing" and assumed I'd disagree...I actually agree...Hell no, you absolutely do not do a full pull up...and the plane WAS flying fat dumb and happy [due to P+P=P no less] so don't screw with it.

Weren't you just complaining about context? I said, with an automated UAS mode engaged. "you do nothing" about flight control (unless you opt to take over). In the present say, you absolutely DO NOT do nothing. You follow procedure and CRM. You run the very short QRH checklist. You assure that there are no stealth factors (such as inadequate thrust-locked power that differs from the apparent thrust lever position or misleading flight directors) that can potentially work against you in a complex, modern aircraft. Once stabilized, you establish engineering-derived pitch and power settings to ensure a safeguard against your own vulnerable human judgment.

Why?

Because it's part of your job.

Who do you work for?

The people who own the airplane and the people riding in it. They prefer safeguards in addition to airmanship.

Why?

(see: Aviation Accident FIndings Throughout History)

[3WE]

***context***

Context? Indeed, context matters.

Did you note (from your own words) that there is essentially ZERO use of P+P=P for automated compensation for pitot failure.

How's that for context?

Essentially NONE of your aeroengineer colleagues ever decided this was a proper thing to do...

And then you ask why.

"Because computers are good" doesn't answer that.

Acknowledged: My answers were the purest of parlour-talk ass-hattery. But "computers being good" is not a reason why no one in the aerospace industry has seen the need for some sort of P+P=P Pitot-Failure Algorithm.

Context? The clearest context is that you have developed some sort of fixation on computers and that is your answer to everything.

Hey, I have a basic model new pickup truck...Chevy 1500 crew cab with a 6 cyl engine...I recently saw 30 MPG (60 MPH back roads with a tail wind). It drives nicely, pulls trailers OK (no not big ones), has ABS, and who knows what all engine controls- that run by computers...Yeah, computers are used in aeroplanies and they will be used MORE. But I got a beer that the Gabrielian P4FA routine ain't coming. I know you don't like my guessed explanation, but you asked the question. What's your answer as to why? All the other aeroengineers are stupid, unlike you?

As I told Evan, is there anything specific you want me to tell my neighbors at Boeing to see if they can get the suggestion over to the development department? I repeat my concerns that they seem a bit disinterested regarding the subject of software modifications.




As to another post above- where Evan yelled, "You do nothing" and assumed I'd disagree...I actually agree...Hell no, you absolutely do not do a full pull up...and the plane WAS flying fat dumb and happy [due to P+P=P no less] so don't screw with it.

The key question there is if you are high up near coffin corner (or more blandly where your maximum speed and stall speed begin to get closer to each other)...maybe you want to do something about that since you don't have airspeed indications....well in that case, it might be wise to descend a bit (ATL crew actually suggested this). You can do it Evan's way and look up THE settings in THE book of acronymns, OR you can do it they way you have done every last departure from cruise altitude for the past 6 months. Whether you look it up in the book or chose something with which you are familiar- it is indeed based on the fundamental concept that P+P=P, but Evan doesn't like it when we refer to things in fundamental terms.


Final thought- put your P4FA system in the CONTEXT that a plane is near coffin corner...Is +/- 10 knots good enough? +/- 5 knots? And again, what caused the pitot failure? Are there some modes where P4FA is going to do the wrong thing? Sitting here today, I can't imagine what that is, but a lot of smart people approved a stall-preventing-nose-over system for not having bad unintended consequences. I am guessing that aeroplanie designers have put lots of thought into how to handle Pitot Failures and designed what might be the most robust system to address it. Aggie ass-hat guess, but...

[Gabriel]

As I told Gabriel, if you want to propose a new system go for it. The best I can do to help you is talk to a few neighbors who work at Boeing...unfortunately, most of them deal with F-15/F-18 parts and stuff- not so much systems development. I also suspect they may be a little less interested in designing new automated stuff at this particular point and time.

Bobby, ATL, could you forward this suggestion to your former and current colleagues and maybe it would work back to designers? It sounds cool.

I'm the wrong person, and this forum is the wrong place.

Do you mean that the industry didn't change their stall procedures because of the thread I started here on the subject?

[Gabriel]

What are you talking about???

You asked this:

A good question could be why, to this day and to my knowledge, P+P=P has not been incorporated in the flight automation logic. It doesn't seem to be so complicated.

I answered 'because maybe it takes a few minutes for the plane's speed to stabilize'.

To which you asked aren't computers better than humans?

That's a totally different question.

How is that a totally different question? When you have an UAS event you have basically 3 choices: Automation relinquishes flight control to the human pilot and let them fly accordingly, automation keeps flight control and adapts to fly the plane accordingly, or nobody does anything and we are all along for the ride.

If your answer "because maybe it takes a few minutes for the plane's speed to stabilize" is not done in that context, then how is that an answer to my original question? Say that it takes a few minutes for the plane's speed to stabilize. So what? Why has P+P=P not been incorporated in the flight automation logic? Yes, it takes some time for the speed to stabilize, and the sky is blue, and the water is wet. So what? I don't see a valid "because" there.

'Tweak' is key word. If the pitot is working things can be adjusted (over controlled if you will) to promptly establish a new speed. If it's not, you 'wait' through the asymptotic approach.

Of course, "pilot" above (the agent doing the "tweaking" or "overcontrol" can include Flesh or Otto. So we are back to square zero with the question.