Air India 171 Accident: RAT Got Deployed | Hacker News Reader

Air India 171 Accident: RAT Got Deployed

youtube.com

6 pointsbelter11 comments

I think people are making too much of some very grainy videos. There are some indications that the RAT deployed, but I think these have to be weighed against the inherent improbability of a completely simultaneous dual engine failure. There's no indication of yaw at all in the available videos, which suggests that both engines would have had to fail immediately at almost exactly the same time. That's an imaginable scenario, but one that you have to hope is extraordinarily unlikely. People will point to BA 38, but I don't think the engines in that case shut down so completely that electrical power was lost. Also, the loss of thrust in each engine was separated by 7 seconds, which is a long time in the context of Air India 171:

>Shortly after the co-pilot had assumed control, the autothrottles commanded an increase in thrust from both engines. The engines initially responded but, at a height of about 720 ft, 57 seconds before touchdown, the thrust of the right engine reduced. Some seven seconds later, the thrust reduced on the left engine to a similar level. (Accident report, p. 5)

Having done a bit of internet research I'm convinced the RAT deployed. There is a very good quality video of a Thai Airways 787 coming in to land with the RAT deployed which sounds identical to that of the rooftop iPhone video, if you get the better quality version. It sounds like a Jet aircraft being pursued by a P51.

I was never very convinced that a 787 which managed to get airborne (suffering a tail-strike or a runway overrun would certainly be a possibility), would not be able to climb away on two good engines, regardless of the flap setting. There is a good demonstration of the takeoff configuration alarm from a simulator, it would be very difficult to ignore, and unlike some older airframes, the exact reason is displayed on the MFD.

Incidentally in that video the aircraft rotates and climbs out without flaps set, I don't place a lot of emphasis on that though because, firstly the simulator is not real life, and secondly I don't know if the conditions the simulator were trying to recreate were at all representative of the accident flight.

I agree that the evidence of the RAT being deployed is somewhat convincing, but I've looked at the images and listened to the sounds, and I don't think it's as clearly established as the online 'consensus' would suggest. It's quite common for entirely erroneous theories to do the rounds in the aftermath of accidents. Some of these theories are backed up by expert opinion and an apparently convincing analysis of the available evidence. For example, here [1] is an completely incorrect early theory of Air France 447 supported by William Waldock, "who teaches air crash investigation at Embry-Riddle Aeronautical University in Prescott, Arizona".

>[Waldock] examined the photos and video of the stabilizer and rudder and said the damage he saw looks like a lateral fracture. "That would reinforce the idea that the plane broke up in flight," he said. "If it hits intact, everything shatters in tiny pieces."

It's always worth bearing prior probabilities in mind. Dual engine failure and in-flight structural failure are both possible failure modes, but they are very rarely what causes modern airliners to crash.

[1] https://eu.seacoastonline.com/story/news/2009/06/09/vertical...

Yeah, I've seen examples of that in the past. The 737 Max incidents were fertile ground...

However, to come at it from the other side. If we assume the RAT didn't deploy, and by implication there was no loss of engine thrust. What could have brought the airplane down? Certainly it wasn't a loss of control incident.

This AAIB conclusion makes interesting reading:

> The operator realised that the flap mis-selection event to G-EZEW was not an isolated event and carried out a study into similar incidents with assistance from the manufacturer. The operator was concerned about the risk associated with aircraft being in a low energy state near to the ground, including performance risks and the possibility that crew members would become confused leading to a loss of situational awareness.

> Information from the manufacturer indicated that properly-computed takeoff performance calculations, combined with the aircraft’s Alpha/Speed Lock and Alpha Floor protection functions, would allow the aircraft to climb safely following a flap mis-selection event, even when combined with other, adverse, factors. Aircraft climb performance following early flap retraction would exceed one engine inoperative climb performance.

To be clear, that is a conclusion reached on the smaller A319/A320 Airbus aircraft but I'd be surprised if the conclusion was not similar on the 787.

My wild guess (which, to be clear, is no better than all the uninformed speculation that I’m complaining about) is that (i) some kind of software fault caused both engines to throttle back to idle and that (ii) the RAT didn’t actually deploy and all the RAT commentary is a red herring. Otherwise I don’t see how to explain the fact that thrust (or lack thereof) appears to have been symmetrical throughout the flight. Both engines failing is already unlikely, but both failing at exactly the same moment? It seems incredible that any kind of failure internal to each engine could be the cause. Surely in that case we’d observe at least a momentary rudder deflection, if not a visible yawing motion of the aircraft itself.

Yeah, I initially wondered about a botched single engine failure. It's not uncommon in twin piston and turboprop aircraft for someone to shutdown the good engine. It is less common in turbofan aircraft because their performance is typically pretty good, even one engine out, and there is no immediate need to manage the dead engine.

As you say, the complete lack of any unintended yaw motion and lack of any evidence of failure of either engine in the videos make it seem less likely.

An auto-thottle cock up does seem like a likely, if scary, scenario. Two competent pilots on a CAVOK day losing a plane under complete control just by failing to advance the throttles manually. I'd prefer a mechanical fault.

I see that the WSJ now claim to have a source inside the accident investigation for deployment of the RAT: https://www.wsj.com/business/airlines/boeing-787s-emergency-... Looks like dual engine failure really is the most likely scenario.

Glad I don't own any Boeing shares at the moment...

>both engines would have had to fail immediately at almost exactly the same time. That's an imaginable scenario...

Not only imaginable but similar has happened on a 787:

>An ANA 787 had an unexpected dual engine shutdown after landing on January 17, 2019, activated by a system called the TCMA (Thrust Control Malfunction Accommodation), intended to shutdown runaway engines on the ground. Its logic should only activate it on the ground with weight on wheels if it senses the thrust lever is at idle but the engine is not (from https://www.reddit.com/r/aviation/comments/1ld9yb8/air_india...)

I'm guessing something like that - a system screw up of some sort.

That’s interesting. It’s a little unclear from the reporting whether the engines rolled back to idle (in which case electrical power would still have been available, I assume) or shut down completely. A software bug rolling both engines back to idle seems kind of conceivable. But shutting them both down completely, simultaneously, while in flight? It seems so unlikely that I’m inclined to be skeptical of the RAT claims. But obviously, no-one really knows anything at this point.

Edit: I see that the WSJ now claim to have a source inside the accident investigation for deployment of the RAT: https://www.wsj.com/business/airlines/boeing-787s-emergency-...

There seems clear evidence the RAT was deployed from the sound on videos - you can analyse the frequencies and the like.

For whatever reason in the ANA flight the engines both shut down completely

https://www.aerosociety.com/news/ana-787-engine-shutdown/

not sure why you'd want to do that rather than idle them. It's reminding be a bit of the 737MAXs crashing due to a faulty sensor triggering some badly designed system.

If you watch the takeoff video the plane climbs normally for about 10 seconds and then suddenly starts descending without any change of direction or rudder suggesting both engines went simultaneously which probably wouldn't be the case with fuel or birds - they found no birds when they looked.

Some info on that system design https://www.pprune.org/rumours-news/617426-ana-787-engines-s...

It's designed to trigger if an engine is at high thrust while the throttle is in the idle position and the plane is on the ground. Maybe there was a low probability combination of sensor failures which could explain why it only happened after much successful flying and maybe why it was when they raised the gear if that effected the on the ground sensor?

Reply on news.ycombinator.com