787 "unrealistic airspeed" AD

[peekay4]

Boeing, FAA warn 787 pilots of bad airspeed data


Boeing 787 pilots are being warned not to make sudden control inputs in response to a “sudden, unrealistic” drop in airspeed shown on cockpit displays.

The US Federal Aviation Administration (FAA) will adopt an airworthiness directive on 1 April requiring 787-8 and 787-9 operators to update the flight manual with the warning message.

The FAA accelerated the release of the airworthiness directive, bypassing the normal rulemaking process to make operators adopt the change as quickly as possible.

Boeing made an identical recommendation to 787 operators on 4 March, which the FAA directive will make mandatory.

The fleet has made three reports of displayed airspeed plunging significantly below actual airspeed, the FAA says. In each case, the 787 was flying in conditions involving significant water ingestion and possibly icing of two of the three pitot tubes feeding speed and altitude information to the air data system. ...

In one case, the pilot reacted to the inaccurate data by commanding a “significant” nose-down dive, over-riding the auto-pilot in the process. ...


More from:

https://www.flightglobal.com/news/articles/boeing-faa-warn-787-pilots-of-bad-airspeed-data-423735/

[stilton]

Shades of AF447.


I thought the 787 had automatic switching to an AOA display in event of unreliable airspeed ?


Designed in exactly to prevent this sort of thing.

[ironbutt57]

pitch attitude and thrust...

[Pontius]

In one case, the pilot reacted to the inaccurate data by commanding a “significant” nose-down dive, over-riding the auto-pilot in the process. ...

Now there's a great idea......NOT!

No other indications of any wrong-doing, normal pitch attitude, normal thrust but the airspeed suddenly drops so let's just shove the nose down. Probably the same sort of person who sits there overriding the auto throttles and wanking the thrust levers back and forth for every knot of speed change in turbulence :rolleyes:

[M.Mouse]

I thought the 787 had automatic switching to an AOA display in event of unreliable airspeed ?

If the AIR DATA/ATT selector is in AUTO and either the PITOT or STATIC sources disagree such that the output from the ADRS is not trusted then the EICAS message NAV AIRSPEED DATA (in the case of pitot problems) or NAV AIR DATA SYS (in the case of static problems) will display and the PFD/HUD will display AOA airspeed (NAV AIRSPEED DATA) or display AOA SPD airspeed AND GPS ALT altitude (NAV AIR DATA SYS).

If the ADRS and IRS calculated speeds disagree you get an 'UNRELIABLE AIRSPEED' EICAS and you DO NOT get automatic AOA airspeed display.

[Sailvi767]

Now there's a great idea......NOT!

No other indications of any wrong-doing, normal pitch attitude, normal thrust but the airspeed suddenly drops so let's just shove the nose down. Probably the same sort of person who sits there overriding the auto throttles and wanking the thrust levers back and forth for every knot of speed change in turbulence :rolleyes:

At least on the 787 the pilot had to shove the nose over. The Airbus will do it automatically!

[glad rag]

Can't argue with that. Does it also call retard as well?

[Downwind Lander]

It sounds to me that some pilots need a bit more dual experience at the weekend on a single engine piston.

[IcePack]

Interesting in that I have seen a sudden loss of airspeed on a couple of occasions. If I hadn't pitched down immediately & got the airspeed increasing. Well you would be reading about it on here. Jet streams & surprise CAT are not uncommon so IMHO get the nose down so you know your airspeed will be increasing & then asses, confirm the situation & act appropriately. To ignor seems to be an odd thing to do.
Well I guess Boeing suspect pilots can no longer fly so will end up crashing anyway.

[Denti]

It sounds to me that some pilots need a bit more dual experience at the weekend on a single engine piston.

Damn, that would be nice. I believe i have had one weekend free in the last two years though...

[dash6]

Single engine piston available mid week too.

[ATC Watcher]

Well I guess Boeing suspect pilots can no longer fly
Reminds me of the discussions (pre-Internet) on the introduction of the A320 in the 80's. Funny in a way to see similar remarks now on the 787.
1980 Ziegler : " this aircraft cannot stall, even if the pilots wants to .." or " it can be flown by my concierge ( janitor) "
Welcome back in time. :rolleyes:

[Aluminium shuffler]

Why have manufacturers stopped having proper redundancy on the latest aircraft? Look at the 737, from Jurassic to NG, and the information on the primary flight displays or main panel all comes from completely different sources, be it independent gyros in the Jurrassic's FDs and independent pitot static systems, or with parallel but independent ADIRUs in the NG, with independent sources to each. The standby instruments are also completely independent. Look at the 777 and you have a single ADIRU supplying all main flight instruments, even the standby attitude not being truly independent of the main instruments, fed by the SAARU, which is the back up for the main instruments. How is it that the bigger, dearer and supposedly more advanced aircraft have less redundancy and less indication to pilots of which system is malfunctioning?

[sleeve of wizard]

For those of you that don't have access to the FCOM Bulletin.

There have been several events where the indicated airspeed shown to the pilots
dropped rapidly (within a few seconds) to below 50kts. These events are
suspected to have been caused by temporary blockage of multiple pitot tubes
when near weather systems with high moisture content. In all events, the
blockage cleared within several seconds. During some events, the FLIGHT
CONTROL MODE, AUTOPILOT DISC, and other messages were shown
(along with a flight control reversion to secondary mode). In one event, the PFD
automatically switched to AOA speed; and in another event, the AIRSPEED
UNRELIABLE message was shown. Depending on the duration and nature of
the pitot blockage, FLIGHT CONTROL MODE, NAV AIRDATA SYS, NAV
AIRSPEED DATA, AIRSPEED UNRELIABLE, AUTOPILOT DISC,
AIRSPEED LOW EICAS messages can show during similar events.While, in all these events, the air data system recovered and the fault detection
system provided alert level EICAS messages where expected, a possible flight
crew response to an initial sudden drop in indicated airspeed could be to apply
pitch inputs which may result in ‘G’ excursions. If the crew applies strong nose
down force to the yoke without disconnecting the autopilot, when the autopilot
disconnects (as a result of the manual force applied), a sudden nose down
command would be applied to the airplane. This nose down input could be
amplified by the low column forces being fed back to the pilot as a result of the
low airspeed being indicated.
As a result of the above events, Boeing is making improvements to the flight
control system and force feedback to mitigate the negative effect of large control
inputs in these scenarios, and is also making improvements to the EICAS Alert
logic related to air data faults. In addition, Boeing is continuing to investigate the
pitot blockage event rate to determine if any design improvements are warranted.
In response to concerns about possible negative crew reaction to an initial sudden
drop in airspeed, the FAA intends to issue an Airworthiness Directive to require
operators to incorporate a procedure into their AFM which instructs crews to
avoid abrupt control inputs in response to sudden unrealistic drops in indicated
airspeed. This procedure also reminds pilots that if manual flight is desired, the
autopilot should be disconnected prior to any manual flight control inputs.

[semmern]

This procedure also reminds pilots that if manual flight is desired, the
autopilot should be disconnected prior to any manual flight control inputs.

.....oh, really?

[olasek]

... Really. It is sad but pilots have to reminded sometimes of absolute basics.

[stilton]

What does 'Boeing is making improvements to the flight control system and force feedback really mean ?


Sounds like a change that would have to be recertified.

[NSEU]

Look at the 777 and you have a single ADIRU supplying all main flight instruments

I wouldn't underestimate the abilities of one 777 ADIRU. It has no less than 6 ring laser gyros, 6 accelerometers, 3 sources of pitot, 6 sources of static, 3 power supplies, multiple processors, etc.... The SAARU has four fibre optic gyros, 4 accelerometers, 3 sources of pitot and 6 sources of static. The ISFD uses the centre pitot/static system only and has an independent source of attitude.

If there are certain failures, the selection of good data is being done by the boxes automatically through fault monitoring and voting (i.e. mid value selection). Mid-value selection of certain data is what an older generation aircraft (such as the 747-400 or 767) would be doing during the critical phase of Autoland.

The boxes are automatically supplying the best available data that the pilots would use under normal and some abnormal situations. And if things get really bad, you still have the option of going back to the basics by selecting SAARU "single channel" data if you, say, want to know the data that a particular pitot/static system is producing.

[Aluminium shuffler]

While I understand that the 777 ADIRU is more advanced then the 737 NG's, I'd still rather have completely independent systems for left and right sides. I just don't trust any single system, regardless of its internal redundancies, to automatically ignore bad inputs and always display the right data. This seems to be borne out by the 787 cases here.

Right from the start of our careers we are taught to trust our instruments, but more and more often, the modern instruments are having simultaneous widespread failures. AF447 was a good example, but that was well after issues with pitot heads had been identified and the accident could have been avoided if the faulty units had been replaced. Now the 787 is getting similar issues. So, do we now have to change mindsets and start second guessing the instruments?

With single source systems, any failure can lead to all displays showing the same error. I don't know if that was a factor in the AF447 accident as I don't know their architecture, but I think that despite its flaws, you have to go a long way to beat the 737's simplicity and redundancy. More modern types seem to have too much networking and too many vulnerabilities to a single fault.

[safetypee]

AS, to answer to your concerns and increase confidence in aircraft systems, it would be necessary to study CS 25 certification requirements together with the manufacturers design submission and system implementation.
Simply, aircraft instrument displays must not present hazardously misleading information (definition and risk level in CS 25). This requirement is often achieved by using independent sensors and some form of cross monitoring or confidence voting; where ‘bad’ data is eliminated or the display shut down to prevent display.

A simple analogy is like the glass half full, half empty dilemma. Arbitration between the two opposing views requires a third party and only then in avionics terms is ‘amalgamated’ data used.
The Boeing problem appears to be in the inability to prevent the display of erroneous data, either because arbitration has failed or the inability to determine ‘bad’ sensed data and thus shutdown the display. In terms of beer glasses, there is little value of agreeing half full etc, if the beer is already flat.

The Boeing problem may be similar to that of ice crystals and pitots, where either the atmosphere is not as expected or that the sensor is susceptible to unusual conditions (or both); in this instance water.
This condition is very serious and requires (technical) rectification in the sensing system; however as a side problem, human reaction to an erroneous display appears to have identified a further issue in the control system.

Meanwhile in our ill-experienced litigious society, manufacturers have to explain the basic responses to pilots.
Something like warning people who have been brought up and now dependant on automated alcho-pops and fizzy mixes, not to drink flat beer, which of course requires an understanding of what ‘good’ beer is in the first place.

[macdo]

Flying a modern FBW aircraft? Then stay well away from TS. I have experienced the ice crystal thing flying behind a line of CB's over the Indian Ocean @ 410. Lost PFD speeds, odd noise of something hitting the screens, then 10 secs later all gone. Very disconcerting. Will happen again at some point I'm sure.

[Aluminium shuffler]

Thanks for the expanded information SafetyPee. Good stuff to know. The trouble with being a paranoid cynic is that I don't trust the certifiers to do their job correctly, and given all the 787 problems, many of which were predicted, I think that mentality is justified. If the 777 system is near bullet proof, then the 787 system should be near flawless, and yet here we are. I fear many of these changes are cost driven, not safety or performance driven.

[safetypee]

:ok: AS.
Forget the paranoia, join the optimistic cynics club.
Consider humans as an asset, why is the design and certification as it is; we might assume that this did not plan for failure, thus what was the cause; the system, certification (beware assumptions), atmospheric situation, etc.
Consider what can be learnt for these events and similar preceding ones. Do we really have to tell pilots what is a sudden and significant change in airspeed, not to manoeuvre the aircraft with autopilot engaged (beware negative experiences with CWS).

Recovery actions are necessary, but the precursor is avoidance – macdo, but how are the conditions identified, then to be avoided by how much; what is ‘safe’?
If the conditions involve high water content then then WXR might detect them and thus avoid … but what level of intensity, deviate by how much.
If not detectable by radar … ice, or associated with CBs, then deviate, but by how much etc.
Deviation might be easier for some 787 pilots because certain engine variants have had restrictions or deviation re CB avoidance - ice crystals. So if engines had suffered in these conditions, why not airspeed.

There is much to learn from this issue, not just for the 787, but for all aircraft, operators, and individuals; how do we think about these things, learn from them and apply lessons learnt.

[Ian W]

One of the problems is that some atmospheric processes are still not understood. One is the way super-cooled water behaves just in simple freezing on nucleation particles. ( http://www.pnas.org/content/99/25/15873.full.pdf ). So there are bound to be the occasional surprises with high velocity flight through this kind of weather. I would expect work is going on researching what weather conditions caused the problem, such as size of droplets, density of droplets, temperature etc.

[GlobalNav]

It just seems that for A/S to truly drop so low so quickly, a lot of other things would need to be happening at the same time. The 787 avionics architecture is so integrated and redundant, I don't see why such drastic changes in A/S can't be checked against inertial attitude and acceleration. Likewise, I'd think a pilot would know that such a change in airspeed, if valid, would be accompanied by some very noticeable vestibular cues. Lack of such cues has be a cue that the airspeed has indeed not changed so drastically.

[Nemrytter]

One of the problems is that some atmospheric processes are still not understood. One is the way super-cooled water behaves just in simple freezing on nucleation particles. ( http://www.pnas.org/content/99/25/15873.full.pdf ). So there are bound to be the occasional surprises with high velocity flight through this kind of weather. I would expect work is going on researching what weather conditions caused the problem, such as size of droplets, density of droplets, temperature etc.One of the problems is that airlines and aircraft makers are very reluctant to share their experiences with researchers - they try to do as much as possible in-house. Good for corporate privacy, bad for actually getting results*.
Unfortunately, whilst good, most in-house researchers are not as knowledgeable as a lot of others in academia and research institutes (as they're expected to cover a widfer range of topics rather than specialise). It's a real shame that there's such an atmosphere of distrust and secrecy, something that prevents everyone from working together to understand this phenomenon.

* I say that as someone whose salary is actually paid by an aerospace corporation to research ICI, yet still can't access their information and databases.:ugh:

[Aluminium shuffler]

Nem, I agree. This industry always bangs on about how safety is its number one priority. I have yet to see any evidence after 20odd years in it that it is even in the top five.