Air India Flight AI171 Boeing 787 Dreamliner Crash

[EEVblog]

Keen to hear Petter Hörnfeldt's (AKA Mentour Pilot) take on this incident, which will inevitably come very soon.

I suspect he'll wait for the fog of war to clear.

Absolutely. He tends to rely on clear evidence in his videos over speculation. We might even have to wait until the preliminary investigation report comes out.

He did a live stream:

[jonovid]

pilot error?
the position of the landing gear lever and the position of the aircraft flaps? in the Boeing 787 Dreamliner cockpit.
can the two get mixed up? can one pilot accidentally retract the flaps? thinking it was the landing gear lever.
are there other aircraft that have different landing gear levers and flaps levers confusing the pilot?
If the flaps was accidentally retracted at V2 the jet aircraft would Not maintain safe climb or lift performance but sink or drop at this low speed.

aircraft maintenance error?
was there maintenance on the digital fly-by-wire system?
is it possible to accidentally swop landing gear lever for flaps lever? in either software or hardware?
is it possible to accidentally put a square peg in a round hole.

[NE666]

Everyone is saying flaps at zero.
Doesn't the system scream at you if you have flaps 0 at takeoff?

Yes.


(taken from a 747 simulator but they are all similar in sound and operation)

I don't think it would prevent takeoff but neither is it really plausible that two pilots could both manage to miss or ignore that during the takeoff roll. Hence the flaps must have been correctly set at the time takeoff started.

[tom66]

Ref the prior flight:  A/C not working on the aircraft, I thought they had multiple "packs" and could use the APU to provide that too?  Weird.

[Haenk]

That large cloud of smoke right at takeoff and the cockpit call of power loss seem to speak for a lost turbine. I'm no aviation guys, but a loss of a turbine at takeoff probably means no ascend. And in that densly populated area, there was simply no place to make a landing; due to low altitude there was not much time for other decisions anyway.
What an awful tragedy.

[BradC]

Twins are certified for a single engine climb out for just that eventuality, so I’d be surprised if it was a single failure.

[tom66]

This pilot suggests the possibility of flaps being accidentally retracted instead of the landing gear.

Explains why the landing gear was never retracted and the aircraft seemed to wallow.



But it doesn't explain the possible deployment of the RAT which is audible in one of the clips (but as of yet unconfirmed).

[NE666]

BBC News is now reporting that the 'Blackbox' has been recovered.

[EEVblog]

Twins are certified for a single engine climb out for just that eventuality, so I’d be surprised if it was a single failure.

Could very well be combination of an engine failure and a poor flap setting or mistake in flap vs gear as has been suggested.
Most airline accidents are a combination of factors like that.

[EEVblog]

BBC News is now reporting that the 'Blackbox' has been recovered.

And it should be completely intact, as it's in the tail section, and the entire tail section is intact and sticking out of the building.

[NE666]

Explains why the landing gear was never retracted and the aircraft seemed to wallow.

It might explain a couple of the observations made but would only lead on the question of 'how'?

The flap and gear controls are very distinctive, and located in very different areas. Never say never but again, it feels like a mistake that just wouldn't happen. Especially if the aircraft was being flown from the right hand seat, since then the Pilot Monitoring would have to virtually reach into the trouser pockets of the Pilot Flying in order to get to the flap control. With hundreds of flight and simulator hours, I just wouldn't expect muscle memory to fail a pilot like that.

[NE666]

Could very well be combination of an engine failure and a poor flap setting or mistake in flap vs gear as has been suggested.

Purely speculation on my part but, to Dave's point on incorrect flap settings, what if the aircraft loading data were incorrectly reported to the flight crew, or incorrectly entered into the FMC by that crew?

Could it lead to the Vref and flap setting data being sufficiently "off" to compromise the aircraft if there were another compounding factor i.e. single engine failure?

[MT]

How accurate are FR24, here they confirm the jet used all of the runway,

The positions are GPS derived and usually pretty accurate AFAIK.  To be clear, by "used all of the runway" they mean it started the takeoff from the far end of the runway not from a midpoint intersection.  That doesn't imply that the ground run lasted all the way to the other end.  9500 feet for the takeoff roll wouldn't be unusual.

Numerous pilots in videos e.g C.W.Lemoine even mentioning using the overrun and FR24 says they used "all of the runway", which should be interpreted as all of the runway" until else is proven, can bee sen in the take off video by the excess dust swirling up at the moment of lift off.  Whit such strong engines the 787 have why even using all of the runway unless something wrong with the planes system. Ahmedabad to London Gatwick at 26160kg payload 787-8 would require approximately 50 tonnes of fuel for this route, suggesting a takeoff weight in the approximate range of 205 tonnes. Given the Boeing 787-8's takeoff weight, it required a minimum runway length of 2,200 meters for a safe departure, assuming using flaps and in the right setting.

[tom66]

It appears that the standard behaviour for a flight like this is to take off about 3/4ths the way down the runway, from the turnaround point.  This appears to be the case in the accident flight, giving the aircraft roughly 2600m of runaway.  I'm not sure why you would not use the whole runway but it seems to be a consistent choice for this type of flight so perhaps the margin is deemed sufficient.  Either way I don't think it would have made any difference since the aircraft was airborne, if it was going to crash then it would always have done so outside of the airport boundary.

[MT]

It appears that the standard behaviour for a flight like this is to take off about 3/4ths the way down the runway, from the turnaround point.  This appears to be the case in the accident flight, giving the aircraft roughly 2600m of runaway.  I'm not sure why you would not use the whole runway but it seems to be a consistent choice for this type of flight so perhaps the margin is deemed sufficient.  Either way I don't think it would have made any difference since the aircraft was airborne, if it was going to crash then it would always have done so outside of the airport boundary.

Using the "overrun" suggest not normal take off.

[tom66]

It appears that the standard behaviour for a flight like this is to take off about 3/4ths the way down the runway, from the turnaround point.  This appears to be the case in the accident flight, giving the aircraft roughly 2600m of runaway.  I'm not sure why you would not use the whole runway but it seems to be a consistent choice for this type of flight so perhaps the margin is deemed sufficient.  Either way I don't think it would have made any difference since the aircraft was airborne, if it was going to crash then it would always have done so outside of the airport boundary.

Using the "overrun" suggest not normal take off.

I didn't say it was a normal take off...

The statement is if the aircraft needs 2200m according to the worst case calculations, then taking off with 2600m is probably sufficient, and it made no difference to the accident since if they had chosen to use the full runway then they would still have crashed into nearby Ahmedabad.  Except they may have hit something other than the hospital accommodation, but it's still a disaster.

[jpanhalt]

As others previously said, when an aircraft "uses the full length" that means starting at or near the runway threshold, not where they lift off.  Of course, the 787 taxied to the runway on an intersecting taxiway, but "back taxied" to near the threshold.  Every later report I have seen indicated Air India AI171 did exactly that.
   
Details of tower communications are always important.  Taxi to means "to" not on.  You get a second clearance to get on the runway.  A pilot at an intersection or when receiving ground instructions can request an intersection takeoff.  I have heard tower ask if a pilot will take an intersection, e.g., for spacing, but have never heard a tower to require that.  Intersection takeoffs are far more common for lighter, GA aircraft than for large commercial airliners.

[asmi]

As others previously said, when an aircraft "uses the full length" that means starting at or near the runway threshold, not where they lift off.

Correct. Most aircraft (we're talking about civilian passenger airliners here) lift off at or before 3/4 of the runway. The reason is to bring V1 (speed at which aircraft can no longer stop without running out of runway, often referred to as "committed to takeoff") as close to Vr (speed at which pilot is supposed to initiate a lift off) as possible - because between these speeds aircraft has no recourse for off-nominal situations - since it's above V1, it can no longer safely stop, but while it's below Vr, it can not lift off. V1 obviously depends on the starting point on the runway, while Vr does not, which is why CDR can make a decision to not use the full length of the runway if he is satisfied that given current aircraft configuration, loadout and weather conditions V1 is sufficiently close to Vr even if using only part of the runway.
   

Details of tower communications are always important.  Taxi to means "to" not on.  You get a second clearance to get on the runway.

Not just that, but clearance to cross runway needs to be requested from the tower controller, not from ground one, because runways are controlled by tower controllers. Which is why the tower is the one giving clearance to land, take off, "line up and wait" - enter runway, line up for takeoff and stop waiting for takeoff clearance, or cross runway.

[m k]

One end part video has enhanced sound.
If not fake it's clearly a propeller.

Some say also that there was a backtracking to the beginning.

[jpanhalt]

@asmi
I agree. In light twins (e.g.,Cessna, Beechcraft Barons & Dukes, etc.), one did not pay attention V1, V2... You looked at the "blueline" (MCA = minimum controllable airspeed), if below that, you  were better off just crashing straight ahead and take whatever happened.  If above MCA, you could try to make it around.  You are probably safer in most cases just staying on the ground.  One exception might be that recent accident where some genius put a concrete barrie at the end of the runway.

Modern transport jets are incredibly complex and have controls the pilots cannot override.  A non-woking radar altimeter becomes an emergency. 

[hans]

This pilot suggests the possibility of flaps being accidentally retracted instead of the landing gear.

Explains why the landing gear was never retracted and the aircraft seemed to wallow.

[..]

But it doesn't explain the possible deployment of the RAT which is audible in one of the clips (but as of yet unconfirmed).

I don't put so much weight on the landing gear still being down on these altitudes. On the video the jet clearly developed problems at or closely after V1 speeds. Raising the gear will improve drag, but if you need to bet on putting it down somewhere on the ground fast, then having the gear out can take a major part of the beating. Unfortunately, they didn't really have the luxury of some field to land.

Sadly, a similar confusing thing could be done for flaps too. The captain on British Airways Flight 38 made this intervention to glide a bit further to atleast crashland on the airfield perimeter. They could have changed it by accident after rotation, or thinking it would glide them just slightly further.

What I don't really fully understand from that video is the power on stall explanation. Typically stall recoveries start by pushing the nose down to gain speed and decrease the required AoA to fly safely. If this jet was still producing thrust from its engines, it would have been a very sticky thing to do so close to the ground.
Unfortunately, these big jets can be quite tricky to climb safely (in contrast to A320s/737s which climb like rockets). These jets can require high speed climbs when they are heavy and need to keep a higher speed for a safe climb profile. I'm not familiar with the 787 if it requires this, and subsequently what its rotation and initial acceleration speeds/altitudes are.

[tggzzz]

Those with long memories may remember this crash of an A320 when fly by wire airliners  were novel.
Compare and contrast, and see extensive discussion at  https://en.m.wikipedia.org/wiki/Air_France_Flight_296Q

[NE666]

The two recordings circulating aren't of great quality, to put it mildly. Whose knows what other local sounds are being heard in the mix.

I think I hear jet engines running but that's just me. I'm sure there's power of suggestion going on too, in terms of what people are being told to listen for.

My issue with the suggestion that RAT was deployed is: Why? In the takeoff and ascent, the APU must be running and can provide electrical and hydraulic power. Whilst the APU is doing that, the RAT is not required. By inference, the RAT being deployed would indicate not only the loss of both engines but the APU too, all within seconds of each other. Doesn't feel a very probable scenario, so I'm not sure how it has been arrived at so quickly by some.

[bdunham7]

In the takeoff and ascent, the APU must be running and can provide electrical and hydraulic power.

The APU is not required to be running (or even be functional) in this scenario and even if it were, it only directly supplies electrical power.  The hydraulic systems can be powered indirectly via electric pumps.

[Analog Kid]

Sigh.

I think I'm just going to ignore this whole thread. Y'all go right ahead and fill it with whatever speculation you feel like. I guess this brings out all the armchair aviation experts. The flaps were not set. They took off halfway down the runway. No,they did the whole turnback thing. The RAT was deployed. No, it wasn't, but it shoulda been. Dual engine failure. Pilot error. Blah blah blah blah blah ...

Me, I'll wait until at least the NTSB or Indian aviation authority comes out with an official report.