Podcast: AI 171 Preliminary Crash Report Adds Facts, Raises Questions | Aviation Week Network

Listen in as Aviation Week's Guy Norris and Sean Broderick join Jens Flottau to share their thoughts on the preliminary report into the first crash of the Boeing 787.

Subscribe Now

Don't miss a single episode of the award-winning Check 6. Follow us in Apple Podcasts, Spotify or wherever you listen to podcasts.

Discover all of our podcasts at aviationweek.com/podcasts

AI-Generated Transcript

Jens Flottau (00:22): Welcome to this episode of Check 6. On June 12, India's worst aircraft accident in decades took place. A Boeing 787-8 operating Air India flight 171 to London Gatwick took off from Ahmedabad. It was 1:38 p.m. and everything looked normal, but only for a few seconds. Literally while the aircraft was still inside the airport perimeter, both GEnx engines lost power and the aircraft crashed into several buildings. 260 people on board and on the ground died. Miraculously one passenger survived. Just before the deadline defined by ICAO Annex 13, India's Air Accident Investigation Bureau published its preliminary report. 

I have read numerous stories about the report over the last few days and new opinions range from "it raises more questions than it answers" to "it prematurely blames the pilots" to "it's very objective and fair." My name is Jens Flottau. I'm the executive editor for Commercial Aviation at Aviation Week Network. And luckily I'm here with two experts in Boeing engines and safety: Sean Broderick, our senior editor for Transport and Safety in Washington, and Guy Norris, our senior editor for propulsion and technology based in Colorado Springs. Welcome to the program guys. Let's start with Sean. So what is in the report?

Sean Broderick (01:53): Thanks Jens. Well, we certainly know more than we did before the report was released. Unfortunately there's a lot that it seems that's not in the report that would help explain what is in the report. The most important piece of the 15 pages of the report to me is the passage that confirms that the fuel to the 787's GEnx-1Bs was in fact cut off shortly after the aircraft lifted off within about four seconds, three or four seconds. And the report is pretty specific. It says that the engine one and engine two fuel cutoff switches transitioned from run position to cutoff position. That to me suggests that the investigators determine that those switches, they were physically moved. The functional flight data recorder on the 787 records these positions. And so the switches were moved and then the engine N1 and N2 began to decrease. So the engines then began to lose power one after the other. So the moving of the switches does not seem to be in response to some other engine problem, which wouldn't be by the book. That's not what the pilots would do three seconds into a rotation. But I think that's one thing to underscore the report seems to make that clear.

(03:33): But beyond that, we don't know. There's no indication as to why these switches would have been moved. And the report does not expressly say who moved the switches, but it does say that there were two pilots on the flight deck, at least two pilots on the flight deck. The report does not indicate that there were any more pilots. It says there were two pilots on I believe 10 crew. And if there was a relief pilot on this airplane, I would think that they'd be listed as a pilot. There is no reference of anybody else being on the flight deck. Doesn't mean they weren't there, but it's not in the report and there's no indication as to who moved these switches right after that important exchange. In the report, the cockpit voice recording report says one of the pilots, it does not say which one. The pilot flying or the pilot monitoring. The first officer was acting as pilot flying the captain, more experienced captain, but both had pretty decent amount of experience on the 787.

(04:40): One of the pilots asked the other pilot, "Why did he cut off, why did he cut off the fuel switches?" And the other pilot responded that he didn't. That's the only reference to any conversation during the entire taxi and liftoff and 32-second flight. So we'd have no idea what was happening before that or after, at least in terms of what the pilots were saying. And beyond that, we know what we knew before the airplane. The flight lasted about 30 seconds. The airplane went up. The airplane came down. Seems to be that fuel starvation and the inability to get the engine restarted, which was an automatic function guy, I'll talk about this a little bit, probably just simply did not have enough time to get the engines re-lit and to fly. They did again, somebody did flip the switches back to the run position. That's where they were found in the wreckage.

(05:36): Everything around that is basically a question and there are very few answers in the report. There's no indication of any warnings, not even a discussion of what came up on the EICAS when this happened. I can't imagine that there wasn't an alert, there wasn't something that popped up during this. And so if there was anything, it's not mentioned. If there was something that might've prompted pilot action, it's not mentioned. The report also does say, and I'll leave it at this, the report says that at this point there is no recommended actions on the 787 or GEnx-1Bs. And the FAA underscored part of that by issuing an international notice on July 11th had tip to Reuters for first reporting this. But the notice references a part of the report that talks about 737 fuel switches being installed without their locking mechanism. And there was concern that on this, it's similar to what the 787 has and there was concern that if the locking mechanisms aren't working, then possibly you might have a situation where the switches could move uncommanded, it potentially could be bumped or something and move.

(06:54): But FAA on July 11th put out a note that says it acknowledges the reference in the AAIB's preliminary report regarding this issue that was flagged seven years ago. And it says the FAA does not consider this issue to be an unsafe condition that would warrant an AD or on any Boeing airplane models including the 787. Now if this issue was somehow believed to be linked to this accident, I don't think the FAA would be putting out a notice saying it doesn't consider it to be an unsafe condition. To me, something like this contributes to an occurrence, let alone a fatal accident that qualifies it as an unsafe condition. It's a pretty loud statement to make I think from the agency because AAIB is the only one based on ICAO Annex 13 that can speak freely. They can say reveal whatever they believe needs to be revealed in the report and publicly, everybody else participating has to get sign off from AAIB to release any specifics about the accident investigation. That's why this note, if you read the entire thing, it's simply the only reference to the accident is the report, the preliminary report. And it says that this issue is referenced in the report. It doesn't say that this issue didn't contribute,

(08:20): But by saying it's not an unsafe condition. That's basically what the FAA is saying.

Jens Flottau (08:26): I want to ask guys something here. Is it possible that the aircraft system could have thought that the fuel switches were moved to cut off when they actually had not been moved? So is there a scenario where there's a mechanical failure that could have happened here?

Guy Norris (08:49): Yeah, hi. I think the short answer to that is no, as far as we are aware anyway, there is no mechanism by which you would have an engine shutoff that was not commanded essentially from the flight deck. And I think to Sean's point, it's interesting that the report itself does indicate that there was actually a transition of the fuel cutoff switches. And interestingly there was a difference in the time at which those transitions occurred. Engine one and two cutoffs occurred separately. They went from run to cutoff with the time gap of a second says the report. So that would mean that there was a physical movement of toggling of those fuel switches and secondly that they were done separately so that this was something that was an interaction with a control, a physical interaction with the physical control on the flight deck. So yeah, the short answer is I don't think there is one other than physically how it is architected on the flight deck in the systems.

Jens Flottau (10:10): On a normal flight, you would only touch these switches before takeoff and after landing, right?

Guy Norris (10:16): Correct.

Jens Flottau (10:16): Or in an emergency when you have to relight an engine.

Guy Norris (10:21): Yes, exactly. Yeah. And the fact that you got both the N1 and N2 spooling down, that's the on the GE GEnx engine, that's the fan. And the core speeds essentially both beginning to immediately spool down, indicate that you were talking about an immediate fuel starvation situation. And the system began to automatically, as soon as the ram air turbine deployed, which we know that it did, because it was following its emergency procedure, it will deploy if there's a failure, multiple failure of both engines simultaneous. It did actually deploy and immediately the APU inlet door opened, I mean everything happened that was supposed to happen in an emergency. Essentially the aircraft was trying to look after itself and recover. And the report actually does indicate that the exhaust gas temperature, the EGT seemed to be actually rising again because both engines were in the process of relighting.

Jens Flottau (11:35): But they just ran out of time. Right? But it was so low. How long does that process usually take for the engine to restart?

Guy Norris (11:43): I think literally another 30 seconds would've given them enough time. But obviously that was time they didn't have. I think one of the things that we perhaps should talk about is what could be a scenario for the inadvertent movement of those switches. And there are very few that I know of, maybe Sean could chime in here as well. But the only one that if you look at where the cutoff switches are located on the flight deck, they're immediately behind the throttle quadrant. They're guarded by these two sort of small brackets as it were to prevent inadvertent contact with them. They also have this sort of locking mechanism by which you have to physically pull them out towards you before you can shift the actual toggle that switch from the cutoff to run position or vice versa. And they're also very central in the pedestal. So if you imagine if you are a jump seat rider and Sean mentioned, we have no idea whether there was anybody on the jump seat, you would have to very much have to lean physically forward to interact with those fuel controls.

(13:02): Not an easy thing to do when you're at takeoff power accelerating. And the only other sort of switches that are remotely close are the stabilizer cutout switches, which are just off to the left-hand seat position. So one of the things that we don't know yet, of course, is whether there was perhaps another reason that the left-hand seat pilot would potentially reach across to potentially activate those cutout switches on the stabilizer. And that's a scenario. I mean there's no doubt that the investigation will be covering everything like that. So why on earth would that happen? I don't know, we've been thrashing around on this at all. But one of the things that is mentioned in the report is that on the previous flight, the engine indicating and crew alerting system did flag up the stab POS XDR, which means stabilizer position transducer, exceedance or error.

(14:13): And this was flagged up. It suggests an issue not really with the horizontal stabilizer itself, which has been obviously key to controlling the aircraft pitch, but it probably means more to do with the transducers, the sensors that are indicating what its position is. So the cutoff switches for that stabilizer I mentioned are right beside the fuel control switches. Is it feasible that perhaps a recurrence of that error either flagged up or the crew thought they saw it as the aircraft was rotating and instead of cutting off those stab norm in the stabilizer position, did some kind of muscle memory kick in and they accidentally went to the fuel control switches?

Jens Flottau (15:07): But they went to the wrong switches and not intentionally, but by mistake in that scenario?

Guy Norris (15:13): In that scenario. Yeah. So it's a very strange situation because if you look at the pedestal, there's very few options for an inadvertent activation of those switches.

Jens Flottau (15:28): If we take a step back and look at how difficult it is to find out what's happening here, is this reopening the debate about cockpit cameras?

Sean Broderick (15:43): The camera thing? To me it's all separate podcast. I'm going to say this, I would like to see what else the investigators have that is not in this report because...

Jens Flottau (15:59): There's a lot we don't know that they know and we don't yet.

Sean Broderick (16:04): Yes. And so while I'm not going to advocate for or against recorders, would it make it easier if the information was shared transparently to know what is going on? Of course it would. But the question is, is the issue here that we don't know? We don't have enough data or information to really, we don't as sitting here today we don't, but are investigators starved for information? I'm not sure that's the case. For example, Guy just walked through that scenario where maybe the pilots were reacting to something, but we have no indication that they had anything to react to either an alert or a caution on a flight deck or conversation between them that says, "Hey, well problem, I'm going to take care of this, stab something." So they say something as this airplane is rotating that would cause them to take an unusual, a non-normal action.

(17:02): We don't have any indication to that. And so before we start going down these roads of, well, it could have been this or it could have been that. I mean we know we know very few things based on the report if we assume the report is accurate. And getting back to a question you asked Jens, if there's a failure that would cause the fuel system to stop working without the switches being toggled, I don't see what that has to do with this because the report says the fuel switches were cut and then N1 and N2 began to decrease. To me, that's pretty clear. The switches were cut and that cut the fuel and the engines then began to immediately lose power. That's it. There's nothing else to indicate that something else happened. Now if something else happened, then we don't know about it. It's not in this report. There are lots of other scenarios, but based on that particular line right there, it's pretty clear that that's what happened in the first four seconds after rotation. So...

Jens Flottau (17:59): I want to talk about the investigation itself for a little bit. You touched on it on this Sean and Guy. And we all have covered numerous accidents over the years. So India has really waited until the last minute with this report. And I guess unlike in other cases, has given no factual updates in between that could have yielded crucial information. And there's other points that I want to point out. The process of getting the flight data and cockpit voice recorder from Ahmedabad to the AAIB and then to start extracting the data that took I think a week or at least very long. So overall, how do you view the quality of the investigation and then following up on that of the report itself, but first the investigation.

Sean Broderick (18:53): Well, I wrote about that on the same day that this report came out and I wrote about it before it came out. And my opinion really hasn't changed too much. I think Annex 13 is clear. While it says you are recommended to issue a preliminary report like this within 30 days, it also says that in situations where you have a high profile accident that's getting a lot of attention, you release information. As soon as... I don't have the wording in front of you, it basically says, look, update the public to help with facts to help allay any concerns and to prevent the exact kind of speculation we're seeing here. Any fatal accident today to me meets that threshold. First 787 fatal accident certainly meets that threshold. I think that more information could have been communicated to at least say, this is what we know, this is what we found from the accident site.

(19:51): Now interesting in this case it could have... if they release as soon as they find those switches in the run position and then all of a sudden the report comes out and says they were moved to the cutoff and then to the run, there would've been a whole different line of speculation based on that line, that pattern of factual releasing. But sometimes that happens to release facts from the accident site. I don't think is out of the ordinary. It's not out of the ordinary even. It's not that I don't think it's out of the ordinary. It's not these days. And it also, I think it helps allay public concerns. So we don't know about the FDR. We read our auto reports. I have not confirmed that chain of custody of the FDR. We've read a lot of things. To me it does not sound like that it was handled expediently. And then the information, the information certainly in the report, just judging by the report, it seems to be there was a lot of picking and choosing done for the information in this report for reasons that are not clear, that's only causing more speculation and that's not good.

Jens Flottau (20:55): Okay. You agree?

Guy Norris (20:57): Well, yeah, absolutely. I mean it's interesting when you think about the aircraft's got two flight data recorders, essentially. The first one they found on top of that, the roof building where the impact site was pretty quickly within a day essentially. But it took another, what, four days for them to find the forward flight data recorder where obviously there was a lot of debris and that's not surprising. But what surprises me is that it took until the 24th, which took 12 whole days before both of them were even transported to New Delhi, to the AAIB facility. I don't get why there was such a delay there. Of course, it's maybe on site investigation beginning and that sort of thing. But to me that's curious and a bit disappointing.

Jens Flottau (21:49): Do we know anything about how AAIB cooperated with the NTSB and others in this investigation?

Sean Broderick (21:57): Only what's been released publicly and that's that everybody's happy with everybody else and...

Jens Flottau (22:03): Do we believe that's the case?

Sean Broderick (22:03): I don't think that's ever the case. I think getting to the point where you're releasing information and agreeing on what the facts are telling you, I think that that's always a rough path. I think we're not getting less parochial and political I think today. So the problem that creates is that the Annex 13 puts the responsibility of investigating the accident into where the accident happened, right? And in this case, often the accident happened wherever the accident happens, there's a local airline involved or maybe a local manufacturer involved. So automatically you have bias built in. I mean you do. And the U.S. system is pretty good where the NTSB is independent, they're not under the FAA. Not every country is like that. I think that it's becoming more challenging to have complete trust in the process, in the way it is set up now. I think that's my conclusion, not just from this report, but from other things we have seen in the past handful of years on how information's released or not released.

Guy Norris (23:22): Again, just going back a little bit to your question on flight deck cameras as well as being part of the airborne flight recorder system, one of the things, I know it's a very sensitive subject, particularly amongst a certain generation of aircrew, but obviously there's a couple of things really to think about. One is the technology is obviously now available and ready. And I think there's no question that it's doable from a technology perspective and that you would maintain and retain that data as part of a strengthened, enhanced airborne flight recorder system. You'd be able to see what was going on. And I think that the other thing is that the new generations, current generations and new generations of aircrew, it's a bit like adapting to sidestick controllers from yokes. And it's one of those generational things. I feel that modern generations would be more open to accepting it.

(24:26): It's a bit like the use of Auto GCAS, for example, the ground collision avoidance system that was really pushed back that pilots did not want anything that was going to take over from their F-16 in the event of a near ground collision. But lo and behold, the new generation of air crew actually says, well, why not? It's a good idea. It's gone on to save lives and aircraft and is now being introduced throughout different combat fleets in the U.S. Air Force. And a version of it will eventually be introduced with the Navy too. So what I'm saying is that generational changes do actually begin to look differently at safety. And they don't see it as somebody taking, looking over their shoulder necessarily. And of course the third aspect is that a visual, a recording of what was going on then at that moment would almost immediately show you what happened. Wouldn't necessarily say why it happened or what the motivation was or why somebody did something that they did, but at least you would clearly see who turned that switch. If anybody turned it on at all, what was the circumstances? It would be an additional aid to our investigation. So anyway, that was just my point. I know it's a thorny subject.

Jens Flottau (26:00): So before we wrap up, let's look ahead. The preliminary report is out. What needs to happen next? What do you hope will happen next? More information released before the final report?

Sean Broderick (26:16): From my point of view, the most important thing is to make sure that you don't have a problem that affects more than just that particular airframe. And then beyond that, that doesn't affect more than just the operator. And then that regulator, you got to identify the potential risks or the potential hazards, sort 'em out and take any appropriate action. And the only action we've seen so far after this accident was in the first few days, the India DGCA did a whole bunch of varied inspections on, or had Air India or Indian carriers inspect 787s for disparate things. All of them potentially related to what might have caused this based on two days into the accident, including pilots double checking their flap settings because that was the first, everybody looked at the video, well not everybody, but people looked at the video and decided that they knew the airplane's configuration based on 15 seconds of grainy video.

(27:24): It turns out that was wrong. But no, back to your question again, it's take appropriate action to prevent this from happening again. And if they don't find anything, then they, as far as I'm concerned, they don't have to put out anything additional, but if they find something, they better put it out quick. And I would hope that from a larger lessons learned standpoint, if they do find out it was some sort of confusion on the flight deck, if it was some sort of whatever the issue was up to and including a deliberate act, which has to be on the table still, then you find ways to help prevent that from happening again in whatever systems and processes that we have in place in this industry. I mean, that's what accident investigation is all about. Figure out what happened and make sure, prevent it from happening again. So as long as they're doing that, they're doing their job, it would've been nice in the first 30 days to have a little more reassurance that this is where the investigation is going and this is where it is not. And right now it does not seem to be that there are any technical issues to worry about based on what we know and the FAA knows at least today.

Guy Norris (28:34): Yeah, and just following on from Sean there, I mean one of the things that really strikes me is that if you look at the overall trend of air safety, really certainly over the past two decades at least, what strikes me now is the fact that more and more of these accidents do turn out to be deliberate acts because the level of safety has got so good. These are now standout occasions. I know it's a strange thing to say, but when you look at, alright, just go back. We don't know about Malaysia. That's never been proven. The MH370, who knows? But even regardless of that, you've got China Eastern 2022, the Germanwings A320 accident, 2015, Mozambique, the LAM Mozambique accident in 2013, not an accident really. And then going back beyond that, EgyptAir from 1999, the SilkAir 737 accident in '97, Royal Air Maroc in '94. I mean, you can go all the way back to the Japanese Airlines accident in '82. So what I'm saying is that I think there's got to be a huge focus regardless of how this eventually comes out on somehow looking at the deliberate action aspect of these accidents, does this imply even more of a automated robotic control assistant in the flight deck, a third pilot or even a human pilot? And that's not going to let a human pilot do these sort of things, but you're getting into such dangerous territory there. I know...

Sean Broderick (30:20): Guy has found a more controversial topic than flight deck video recorders.

Jens Flottau (30:27): Okay. And before we find another even more controversial topic, I think we're going to wrap it up for now. Thanks, Sean. Thanks Guy. I'm sure we'll have to talk about this again, the consequences when we know more. We'll definitely do another Check 6 on this one. Also, thank you to our podcast producer Guy Ferneyhough in London, and of course to all of you for listening. If you haven't already, be sure to follow Check 6 on Apple Podcasts, Spotify, or wherever you listen so you never miss an episode. If you found today's discussion helpful, please consider leaving us a star rating or a review. Better yet, share this episode with a friend or a colleague.

Announcer (31:17): College students qualify for free digital subscriptions to Aviation Week and Space Technology. That includes access to our archive, a valuable resource that contains every issue back to 1916. To sign up, go to aviationweek.com/student.