Hacker News new | comments | ask | show | jobs | submit login
Air India Express 737 Hits ILS, Damages Wall on Departure, Flies for 4 Hours (flightradar24.com)
224 points by sassyboy 4 months ago | hide | past | web | favorite | 238 comments

Just saw the pics of the damage on the twitter thread linked-to in the article. See below.


Looks quite severe.

I'm impressed the plane flew for 3 hours without getting destablized and crashing....

Air India has the worst safety reputation in India, and several accidents / incidents. I flew Spice Jet a few times over the last few weeks when I was there visiting. Spice Jet is like India's Virgin America. Service was excellent and I did not experience any issues while flying.

The most surprising part is, trichy to mumbai takes hardly 1hr flight time. What caused them to take 4hrs?

In 4hrs, they could have landed in Dubai itself easily

Being Indian, I agree with statement about AirIndia. I would never fly in AirIndia even if the ticket price is too low, laughably they usually cost way above than any other competitors.

Looks like they went halfway across the Arabian Sea before turning around, strange... https://www.flightradar24.com/data/flights/ix611#1e2da319

What caused them to retun in half way? Report says Trichy airport informed pilot about posssible damage to plane. Pilot ignored it saying everything is green.

What happened in 4hrs to divert back to Mumbai is still mystery. Will be interesting to see full report

Is it possible that Dubai refused to let them land with a flyable but potentially pretty broken plane?

Do you think it's even theoretically possible that an air traffic controller will say "looks like you might be having trouble with your plane... we're not going to allow you to land here. Go elsewhere"

No. If the pilot declares an emergency the pilot is in command, effectively.

It is much more likely that Air India dispatch wanted the airplane repaired at a base where they have maintenance.

I also suspect it was better for the crew to end their last airline flight as crew in their home country...

A few years ago, was on a Delta flight out of Detroit to Boston. Over halfway to Boston, we turned back as the airplane had a maintenance issue that Delta wanted to deal with at DTW rather than BOS. (They do have a permanent presence at BOS, but a larger one at DTW.)

It would be the opposite, every major airport will be open for a plane in trouble.

And even non-major ones! I hear you can even land at an air force base in the US as a civilian private pilot without any negative consequences, providing you have declared emergency.

Yes, if you have nowhere else to go, but it's a last resort and there's definitely a high chance of fines or more serious consequences, especially if it's an active base with sensitive material or operations.

No, that almost certainly wouldn't happen. If you declare, you're basically given cart blanche provided you can reasonably justify your actions, and if an AFB is closer than a civilian airport, for example, its absolutely okay to land there (though you should probably give them a call first saying you're landing). Your first priority is not dying; legal consequences don't matter when you're in that situation, and the FAA knows that. If breaking a law will save you they'd rather you do that than die, and they almost certainly won't punish you for it.

Nobody is arguing that it's not ok, but consequences always exist regardless of conditions.

In this case, it's not about the FAA. An military base can choose not to prosecute once they investigate, but don't assume that they are just fine with strangers landing. Emergencies can be staged and, as stated, if there are sensitive combat and intel operations ongoing then it will be a serious incident.

This is no different than calling search and rescue to save your life, and then getting a bill afterwards. Still an emergency, and still has consequences.

I'm not aware of any more serious consequence having ever happened for a legit emergency, and I am aware of fatal accidents where private pilots were e.g. out of fuel and didn't realize the AFB was an option even though they could've made it.

You know, it's funny, you save the plane, then you get second guessed about whether you used air Force resources wisely, which can have serious repercussions. Or as a professional, have consequences to your career.

It is a crap shoot, based on optics, and partially based on real data. You have to hope your investigator, or boss, sees the data the same way you did....

You made the opposite choice, lose your airframe and your life, and then you should 'clearly' have made the other decision, because the data's in...

Sure. I just thought that they weren't in as much trouble that they needed to land immediately, but could have caused some trouble after landing.


No. That's really not how anything works.

If they went to burn the fuel, going somewhere and returning is easier than circling for 4 hours.


1. You would not want to ascend to flight level 360 (particularly with a damaged plane).

2. You would not want to leave the vicinity of an airport (particularly with a damaged plane).

Easier, but more dangerous; surely not protocol.

You would prefer to stay near an airport, rather than far away over the sea.

And you would want to stay at low altitude, where you can burn the fuel much, much faster as well as divert more quickly if anything went south.

Nothing gets better at FL360 in this situation if your intention all along was to divert.

"Being Indian, I agree with statement about AirIndia. I would never fly in AirIndia even if the ticket price is too low, laughably they usually cost way above than any other competitors."

Since I have flown with Air India, would you mind being more specific? Are the statistics? My flights were okay. Okay, zero status miles credited at Star Alliance and a bit stingy with alcohol on board. But food and service was good, plane looked well maintained.

Please tell me more.

I used to be a government employee and was required to fly Air India so I flew it a fair bit. Personally, the flying bit was not that different from others. Their website was a nightmare to book on, usually I found them generous with alcohol on international flights, but sample size is small here. Their miles were hard to use, too many conditions. Staff was a bit low on professionalism. Saw a couple of fights between flight attendants.

> Saw a couple of fights between flight attendants.

In flight entertainment.

"Their website was a nightmare to book on"

I book 90% of my flights on Vayama, so no problem here. I once decided to book direct only on airline websites to make a possible ticket change easier and cheaper. The one time I tried it I had serious trouble. VPN blocked by airline, based in China, google captcha required which requires a VPN etc. I tried the booking process 4 times until I bought a ticket on Vayama. Later got an email from the airline ("your ticket got issued") so I ended up with two tickets for the same flight. I emailed them and they reimbursed the money. It was Sri Lanka airlines.

I've flown on them thrice. All three flights were delayed, even when it was a 35 minute flight from Jaipur to Delhi (supposed to land at 9:00pm, landed at 11pm).

When I last checked, their website literally did not work (clicking the booking button after entering your details took you to a 404 not found page)

It's a disaster of an airline, especially when cheaper options such as Indigo are way more punctual

I had a flight around 2 AM on night, and man, the crew were so sleepy. The crew who made announcement during landing was close to yawning. I am highly suspicious they were taking nap on the back of the plane.

Fun fact: Air India is a state owned airline that has been bailed out with tax payer money multiple times. Attempts to sell/privatize the airline has failed too so far.

They had to expend the fuel as dumping is not always an option, when you have a tail strike you don't take risks at landing, consuming the fuel to reduce weight and minimize the risk of fire is a basic option.

Going to 36000 feet was taking a massive risk.

I don’t think we have enough information yet to establish whether excessive weight was a factor.

Also, in a plane that cannot dump such as the 737, to burn off fuel one would prefer to stay in the thicker lower part of the troposphere. Going to altitude not only risks decompression, but it actually preserves the remaining fuel due to lower drag.

That is not a maneuver a pilot with a known-damaged aircraft would make to burn off fuel.

On the other hand being higher buys you more time to correct for problems when they occur.

It buys you more time when the engines fail and you turn into a glider. If structural failure occurs, it just takes you longer to fall to the ground.

Not really. If there was a decompression the very first thing that the pilot would do is immediately descend to 10,000 feet. The aircraft is one giant pressure vessel with wings strapped to the sides. Taking that up into thin atmosphere when it might be compromised is a bad idea.

The flight is from Trichy to Dubai, so they were on course.

Is this a hit and run with an airliner?

I see some wires coming out, they can't possibly have not noticed, they must have lost some instruments, surely some warning lighted up in the cockpit.

I hate articles introducing three-letter-mysteries and leaving the readers in confusion. To add insult to injury, none of the comments seem to acknowledge that the topic is impentrable.

What is the mysterious ILS that was hit?

Even after reading the Wikipedia article, I only have a vague idea - seems to be a radar system, both on ground and in the air. Did the plane hit a radar tower?

Others have noted that ILS stands for Instrument Landing System. It’s a system used for landing airplanes when you can’t see the runway, typically getting you down to 200 feet, but with special training and equipment you can get down to 100 or even zero (autoland). ILS is the world standard for getting airplanes onto runways in bad weather.

The ground portion of an ILS consists of two antenna arrays, one at the end of the runway called the localizer, and one just off to the side called the glideslope. These arrays produce two fan-shaped signals that vary left-to-right and up-and-down in such a way that an aircraft can determine its location relative to the runway with remarkable precision. The localizer signal provides lateral guidance to the runway centerline, while the glideslope provides vertical guidance down to the touchdown zone, usually on a 3° glidepath. One or both of these arrays is what this airplane ran into.

The system is passive, in that the signals are simply broadcast from the ground continuously, aren’t unique to a particular aircraft, and there is no return or response from aircraft. The ground antennas simply “shout into the void” as it were, and aircraft receivers determine their location based purely on the shape of the signals at the aircraft’s location in space.

For illustration here is a localizer antenna (though note that the airplane in this picture is “backwards,” i.e. the localizer for a given runway is at the far end of the runway, though it is possible to “fly the backcourse” and land with one at the near end): https://image.slidesharecdn.com/instrumentlandingsystemils-1...

And here is a glideslope antenna: https://i.pinimg.com/originals/8f/56/b3/8f56b3b52fb9b81c84f7...

Another one of these comments that make this forum always a joy to read. Thanks for the detailed walkthrough.

There are lots of great YouTube videos explaining how these work with some visualizations.

Here's a good one: https://youtu.be/FeELh0kMSIA

Thanks for that, I was wondering if the grounds portion of the ils was hit, or the plane component.

Superb comment. Thank you for taking the time to post it.

The article is very clear for pilots and people with knowledge in aircraft industry, it is meant for them. ILS, ADS-B, stuff that is basic domain knowledge. There are strange things in the article, like how a plane with severe damage was allowed to raise to 36,000 where catastrophic decompression is possible (like https://en.wikipedia.org/wiki/Japan_Airlines_Flight_123), but the basic information is there and clear.

ILS stands for instrument landing system. It's a set of antennae that produce radio signals to guide the aircraft to the runway. The plane must have struck one of those antennae.

I think that the audience most interested in such stories especially from this source is familiar with the terminology so they don't waste time detailing it.

How many articles posted here on HN tell you what a CPU is? Or a pointer? Or a container? And that's just to pick a few common terms. News about a certain topic will build on the assumption that the basic concepts related to it are already known to the reader.

You raise a good point, but in the end the situation is not similar.

The situation in the article can be very well explained in layman's terms. Most people know enough of the ground-world and air-world to understand how things hit and what's important about hits.

However, the article uses an overly specific term, and on top of that, it overloads it with a meaning which is not the one found in literature. Notice that Wikipedia's diagram of "ILS" shows it as an airport-sized system made of several components, none of which is called "ILS".


This doesn't make understanding possible, unless you know what "ILS" means in the lingo.

In IT terms, this is akin to saying "A sysadmin tripped over a CISC and fell". To someone familiar to the terms, this translates to "someone tripped over a silicon chip with a CPU, most likely an Intel".

Meanwhile, a layman would look up "CISC" on Wikipedia and be confused about whether someone trippped over a computer box, an instruction set, or a processor. All of this is confusing and irrelevant to the point, which can be understood by anyone: someone tripped over a few cm wide piece of silicon.

Back to the aviation article: "A plane hit a few meter high tower housing an ILS antenna." conveys all the information, but doesn't leave anyone in the cold.

I stand by my initial complaint: any speech that obscures the actual topic behind lingo available only to a small group of people should stay in that small group or improve.

> I stand by my initial complaint: any speech that obscures the actual topic behind lingo available only to a small group of people should stay in that small group or improve.

These days, with the internet at everyone's fingertips, I find it very hard to understand the "I couldn't find what it means" explanation when a simple search for ILS returns some pretty clear explanations on Wikipedia or Quora on the very first results page. No need to even click a link. And if you do click you get pictures, explanations, everything you could want including how they can be hit [0]:

>> A localiser antenna. Point of interest: In case of an aircraft over-running the runway, it is the localiser antenna which gets smashed!

And keep in mind that FlightRadar24 is not a site for the laymen. It's for people with an explicit interest in aviation and there's a definite assumption that you understand these concepts. Perhaps other sources were clearer. Just like it's assumed you know what ADS-B, kts, or FL(350/360) are. They aim for a different audience.

[0] https://www.quora.com/How-does-the-ILS-Instrument-Landing-Sy...

That's an aviation website. Would you expect HN posts to explain all computer related terms?

I think it depends on the audience and in this case the site is flightradar24.com. I also had to google that BTW.

As other response mentioned what it is, look at this image for get an idea how it looks: https://www.airteamimages.com/---__-_94189_large.html. It is present on the entry side of most runway.

ELI5: ILS tells the plane how high/low, left/right it is in relation to the runway when landing in foggy conditions

I have been trying to go over how this could happen. The damage to the wall seems so severe that I would image even over the sound of the engine the pilot would hear and notice that they had hit a wall. Flying the plane at this point would be risking their life. The first thing that came to my mind was the only reason a pilot would risk his life is if his life was already at risk and if he had been drinking when he hit the wall he would face many year in jail I assume for putting lives at risk. In a panic says to himself the plane is damaged but I know it can still fly and takes off without a plan really but perhaps just to let the alcohol he had leave his system. I hate to speculate here on HN. This is just crazy to me and I really look forward to hearing more details as to how his happened.

Over at http://pprune.org the suggestion is that the pilot's seat was not locked into position correctly and slid back during takeoff. This caused the pilot to pull back on the stick too early and his instinct was the push forward to keep the plane on the correct trajectory. He likely over-corrected and ended up flying 6ft off the ground at the end of the runway where the landing gear caught an antenna and the chainlink fence on top of the brick wall.

It's all speculation at this point.

Ugh, there have been Cessna deaths caused the same way. It's my least favorite failure mode to think about, since there's no obvious way to correct. In a Cessna, best you can do is release the yoke and tiptoe on to the rudders, but professional instructors have still died when this happens to them, since low altitude errors are often not correctable by the time you've lost too much altitude.

Usually there ought to be more than one locking mechanism involved, though..?



In contrast, I can't say I've ever heard of seats in cars having this problem. Is it because plane seats are highly optimised for weight?

Suppose your car seat hasn’t locked. You drive, then you suddenly go up a hill, and the seat slides back. Well, fine, you hold on to the steering wheel (or not), stop, take a breath, adjust the seat, and carry on.

Suppose your plane seat hasn’t locked. You taxi around to the runway (slow and always level), then you begin the takeoff run. Already, you have some acceleration pulling back. Then you gently pull back the yoke to rotate. Now, suddenly your seat slides back. You hold onto the yoke, stall, and die. Or you let go off the yoke, and now the plane is unpiloted in the crucial take-off phase. And, note, stopping is not an option at this point.

That’s why vigorously thrashing around in your seat is part of the before-takeoff checklist (“Seat - LOCKED”).

My wife knows of two fatal car accidents caused by this. Don't adjust your seat whilst driving peeps, and make sure it clicked into position before you drive if you adjust it.

Mind sharing what causes the accident, does a sliding seat cause the driver to yank the steering wheel and lose control?

Also, what about adjusting the steering wheel while driving? Sounds like a similar threat.

Wow, thanks for the info. If you happen to have links for them, I'd be curious to read more.

Inability to apply enough force to the brake pedal at the right moment.

One of them was adjusting the seat whilst driving, it flew back and she swerved. The other one I understand was a seat which wasn't clicked in.

Pulling back on a car’s steering wheel doesn’t do anything. Also, cars have a very low stall speed.

A car's stall speed depends heavily on the car's altitude.

Could be because people don't generally drive 40 year old cars! The aircraft in the first fatal link above was manufactured in 1979 and the seat failed in 2016; the second was manufactured in 1969 and crashed in 2014.

And because it would be less likely fatal in the car, so you'd be less likely to hear of it happening.

(Less likely fatal because car collisions are less fatal due to e.g. airbags, and less likely to impact terrain since it only takes a second to undo your car seatbelt, and because there's no pitch control to have messed up.)

My front driver's seat adjustment is all motorized - but not the passenger seat. Originally it annoyed me (the motors are so slow to move!) but now I'm wondering if it might be an intentional safety feature - it's never "unlocked", so you won't suddenly slide away from the pedals and wheel because a part failed to latch right.

If your car seat slides back unexpectedly, the most likely problem will be a decreased ability to press down in the accelerator. The seat would generally slide back only under forwards acceleration. Under braking the seat would slide forward giving positive pressure onto the brake pedal.

It also diminishes your ability to brake. And if you do slide back, and the seat then locks in place, that could be serious.

You would keep steering the car, while also attempting to unlock the seatbelt. Depending on the car, you could also activate the parking brake. (In my car, it's a button on the center console, and it can be activated at high speed and will provide emergency braking.)

"Serious" is still not the same as "fatal", even if your car hits something. If your plane stalls and falls hundreds of feet, my understanding is that the fatality rate is going to approximate 100%.

If your car seat is electric is probably based on a worm gear which is more or less self locking.

I've flown A380 and B747 on a commercial (Lufthansa) simulator. Both had seats which could be only moved with motors (locking in any position). Don't know how it is with B737 but I would assume the same (for this very reason).

From that thread:


Those passengers just won the lottery.

And quite probably many people living just past the end of the runway.

How will the alcohol ‘leave his system’ in the 4 hours it takes to fly to Dubai from India?

Pilots would have to blow 0.0 but that doesn't mean zero evidence of alcohol. In 4 hours you can metabolise around 4 standard drinks out of your bloodstream. It's not like drugs where any trace is bad because it's illegal.

That said, I'm not sure I am on board with the explanation. It could just be that with everything going on in a cockpit it wasn't noticed.

Does reduced air pressure speed up the removal of alcohol via breathing?

If a fever can, wouldn’t alcohol do the same?

A fever is, at it’s core, an increase in metabolic rate. Your hypothalamus is telling your body that it’s temperature set point is higher than usual, and it takes energy to raise your temperature. I don’t see how a difference in air pressure would have a similar effect.

other way around - reduced air pressure is reduced oxygen, which is lower metabolism. Higher metabolism causes EtOH degradation to go more quickly. Fever is higher metabolism.

Your blood oxygenation will stay at 95-99% even while in a cabin.

Metabolism in aircraft is pretty low: you’re generally sitting idle, but I think the alcohol metabolism is still constant except the terminal metabolism.

Fever is partly metabolism, but largely reduced cooling.

I've taken an oximeter with me on flights before, and can give a data point of n=1 to say this isn't the case.

My oxygen % decreased initially with altitude until it stabilised around 90-92% (at ~10k feet, IIRC). It sat at this point for the whole flight (8 hours), with a low of 87%.

Through forced rapid deep breaths I could get it back up into the mid-high 90s, but also got light headed and funny looks.

Moving around the cabin would also raise it.

Interesting, the studies i’ve Read pegged the average reduction at 4%.

After V1, the accelerate-stop distance, the pilot is committed to continue flying.

The cargo area of most airliners is not pressurized, so the gash is not an immediate problem.

In a case like this, the pilots would want to climb to several thousand feet and evaluate the situation before landing.

Looks like the pilot did a great job once it was realized the airplane was damaged of remaining calm and flying the airplane.

the cargo hold of a 737 is indeed pressurized https://aviation.stackexchange.com/questions/8252/are-cargo-...

Cargo is definitely pressurized. Not heated/aced, but pressurezed the same as the rest of the cabin. Everying from the skin inwards is under 5000ish feet of pressure.

It's heated above freezing too as far as I know, I've never had anything in my luggage freeze.

Heated in that the air becomes warmer as it is compressed. But not the additionally-heated air provided the main cabin. Even so, at low air density even -30 would take a while to cool the large mass of bags.

The front cargo hold is heated with engine bleed air.

Not a pilot, and not trying to be an armchair pilot, but why would he have ascended all the way to 36000 feet? Isn't there a real risk of depressurization with hull damage?

No. If there was a leak then it wouldnt pressurize properly on the way up. If it was holding pressure, it would probably keep doing so.

It totally depends. A weakened structure can hold pressurization for a while, and then suddenly (and possibly catastrophically) fail.

A famous example of this is American Airlines flight 96 (1972): https://en.wikipedia.org/wiki/American_Airlines_Flight_96

Also Japan Airlines Flight 123

One minor UX gotcha: the alarm for a pressurization problem on a 737 is not all that intuitive.


I'd rather not trust my life to "probably". If there was some damage to the hull or a cargo door that could lead to failure, it's under more stress and is more of a hazard as the altitude increases.

Less stress imho. At altitude the forces on wings/structure are far less than during takeoff.

The takeoff has already happened, and a landing (of some sort) will have to happen (which also means flying near ground level).

The thing that doesn't need to happen is flying at full flight altitude when the pressure differential between the inside and outside is greatest.

Incorrect. In level flight the wings have - by definition - 1g of upward load. Regardless of speed or altitude.

That’s what sand worm was saying - level flight puts less stress on the wings than takeoff (where you accelerate vertically, thus have more than 1g).

Still, you don’t want to take a structurally damaged plane high up.

Consider cabin pressure differential versus the outside ambient air.

1g in up/down force. But thicker air means more forwards/back forces. Higher engine power = more force, regardless of G.

Engine power is a fraction of weight. Even at full thrust (which is rarely used in commercial service, even for takeoff), the engines are only capable of 0.25g or so.

To take a structurally damaged plane to altitude with that argument (if it holds pressure, it probably won’t explode) is reckless folly.


> After V1, the accelerate-stop distance, the pilot is committed to continue flying.

I'm not a pilot, but I'm 100% sure V1 means you can't safely abort the takeoff, _not_ that you're committed to the remainder of the flight.

Also, the ILS is at the end of the runway, they were surely airborne by this point (i.e. past V2), so I'm not sure why you even brought up V1 in this discussion.

In fact, I can't imagine being so low at the end of the runway. At somewhere like Midway (MDW), you'd almost certainly hit a house, at the very very least the perimeter fence.

> In a case like this, the pilots would want to climb to several thousand feet and evaluate the situation before landing.

I'm fairly certain that in an emergency situation you don't "evaluate the situation". Short of actually being unable to fly, you have to take off after V1. If there is any issue whatsoever, you alert the tower and begin to come around to do an emergency landing using "normal" emergency procedures.

V1 is the takeoff decision speed, not the accelerate-stop distance.

The accelerate stop distance is the distance it takes to accelerate to V1, reject the takeoff, and stop the aircraft. V1 is in KIAS (knots of indicated airspeed). ASD is in feet/meters.

You are correct that passing V1, the aircraft is committed to fly unless an emergency relating to controllability surfaces.

@ codeisawesome If he only had a small drink he might have though a couple hours flight time is all his body needed to metabolise the alcohol to where it wouldn't be detected. Again I hate to speculate here and hope that there is a more rational explanation as to why things went the way they did. For example like captain_perl said at a certain speed the pilot is commited to take off. Ok, but why continue flying and not get enough height to turn around and land immediately? Was he following some protocol that states continue climbing then assess damage? Did air traffic tell him to proceed as planned? To the casual observer here, things don't add up. I am hoping there is some explanation but at this point something weird seems to have happened or we are missing a part of the picture. Very interesting.

You replied to the wrong comment.

Who wants to wager that the pilots did an aviation version of the five stages of grief, with some help from passenger / flight attendant reports?

FO: Tower said we hit something, may have damage. Capt: Just close, we'll be okay. FA: We heard a thump, and so did the passengers. CA: Just a noise, we'll be okay. Dispatch on ACARS: You hit the localizer and wall past the far end, you morons. Descend, lower your cabin diff pressure, and GTF to Mumbai. CA: How do we explain this? FO: Ask the passengers for newspapes, we must find new jobs..

Not inconceivable that they decided to keep flying to overwrite the CVR (assuming it was the old 120 minute type).

Given the apparent structural damage, that was a risky decision.

Unless I'm misinterpreting the satellite image its ~150m from the end of the runway to that wall[0] and it also looks a bit downhill from the runway to the wall[1].

I don't see how this could happen without the pilots noticing; they would have been skimming the ground well past the end of the runway.

[0] https://www.google.com/maps/place/Tiruchirapalli+Internation...

[1] https://twitter.com/shukla_tarun/status/1050581391828836352

Perhaps they were slightly airborne, like 0.5m above the ground, but ground-based instruments recorded the plane as still being "on the ground" (i.e. same altitude as the runway) because of the downhill slope? If they really were on the ground, the pilots would have continued to feel the vibration of the wheels against the ground, a very clear indication that they aren't flying.

I'm even more surprised that the plane could land on its own wheels despite the huge gash in the fuselage. Do landing gears retract so quickly, or was the wall simply not wide enough?

I have no first-hand information, but I think this is going to turn out to be a reduced thrust takeoff with the crew being unaware of that (meaning they screwed up).

Airliners have enough power to lose an engine at a decision speed (V1) while still on the runway and on the one remaining engine continue to accelerate on the runway to rotation and eventually liftoff speed and still clear obstacles. There is no mention of an engine loss causing this issue (and surely the airplane wouldn't launch for FL360 with one engine INOP), so this almost has to be a mis-set takeoff thrust accident as, with both engines set for proper takeoff thrust, it's rare to use more than 2/3 of the runway and the initial climb is brisk due to the large excess of power a turbojet engine has at sea level.

There appear to be two holes knocked in the wall, spaced about the right distance to be made by the two sets of wheels. I'm not surprised that the wheels are strong enough to knock a few bricks over & still work fine.

I also think the wall is lower than the ILS gear, so probably that is what made the gash in the belly of the plane. Amazed that nothing serious was cut.

I cannot believe the pilots were completely unaware, I mean just reaching the end of the runway without having pulled up must be a never-in-your-career kind of scary moment.

Measurements from those instruments would be from the aircraft itself as these are ADS-B transmissions from the aircraft.

Question for professional pilots here: does this feel like normal behaviour for a pilot?

From my limited experience of pilots, this seems entirely unlike the kind of behaviour you'd expect - they're usually conservative when it comes to safety or is that just the projected image to reassure the public?!

Unfortunately don’t have time for a long post at the moment, but as an airline pilot and instructor/check pilot, I can say that extreme incompetence exists in the airline world, especially in Africa, India, SE Asia. There could be some very good reasons that they chose to continue, and often I learn those reasons when the facts are released, discovering that a good decision was made. I’ll be following the investigation, but this one seems somewhat alarming to me. Here’s another of mind-blowing incompetence, also from India - https://www.ajc.com/news/national/airliner-forced-land-after....

Also, more automated, single-pilot/ground operated or pilotless aircraft can’t come soon enough.

More like Air India in specific. Other Indian airlines are pretty good.

--I can say that extreme incompetence exists in the airline world, especially in Africa, India, SE Asia

including Hong Kong?

Cathay has a pretty good reputation, I believe. The HK civil aviation authorities, so-so, but all right. Struggling to get the expensive new ATC center to work.

(Note that HK is considered part of East Asia.)

Self-reply to add:

See here for some news articles on the ATC centre:


Hong Kong is East Asia.

I stand corrected.

No, plus that's not what he said.

My, what a broad brush you have. With of course no ‘time’ to back it up with facts.

Some facts:

Hull loss rates by region of operator per million departures (Jet / Turboprop) 2012-2016: [0]

- 2.21 / 7.38 Africa

- 1.17 / 20.59 CIS

- 0.74 / 3.42 Middle East / North Africa

- 0.53 / 1.55 Latin America / Caribbean

- 0.48 / 1.45 Asia Pacific

- 0.22 / 0.98 North America

- 0.14 / 0.73 Europe

- 0.00 / 8.73 North Asia

India falls below Myanmar, Nepal, Pakistan in air safety audit[1]

>The audit — ICAO Universal Safety Oversight Audit Programme — seeks to identify if countries have effectively and consistently implemented the critical elements of a safety-oversight system.

>India is one of the 15 countries that are below the minimum target rates.

[0] https://www.iata.org/pressroom/pr/Pages/2018-02-22-01.aspx

[1] https://economictimes.indiatimes.com/industry/transportation...

I appreciate the hull-loss facts, but do they directly support "pilot incompetence"? I've read enough NTSB reports to know it's usually multiple congruent factors (aircraft maintenance, organisational culture, pilot work-load).

“extreme incompetence exists in the airline world” does not equal “pilot incompetence” but would rather include aircraft maintenance, organizational culture, ... So the facts support the statement perfectly well.

This is highly misleading and deceptive. You are mixing up the headline and data from link 2 which is mainly focused on which government agency licenses ATCs as the reason for lower scores, to the IATA hull loss data from link 1 which does not single out any country or region for below 'minimum safety target rates'.

This is the IATA data for hull loss rates for 2017 per million departures:

Asia Pacific - 0.18

CIS - 0.92

Europe - 0.13

Latin America and the Caribbean - 0.41

Middle East and North Africa - ​​0.00

North America - 0.00

The second link does not mention or link to anything about hull loss, safety issues, aircraft, training but talks about lower ranking due to ATC licensing by government agencies.

These are two independent data points.

The first is the bulk loss stats.

The second is an article discussing India's poor audit results in the ICAO safety audit, per the quote above the link and sourced to it.

I'm not sure how you can claim that the second link doesn't "link to anything about hull loss, safety issues, aircraft, training". It is the result of an independent safety audit by the ICAO, which encompasses licensing, operations, airworthiness, and a number of other categories[0].

[0] https://www.icao.int/safety/pages/usoap-results.aspx

For uncommon events like aircraft hull losses, I very much prefer the larger sample size of a few years, even if it does put the data slightly out of date. Though note - the IATA there notes that 2017 indicates a recent (last-few-years) trend of increased safety in sub-Saharan Africa.

GP poster was quite clear that the two pieces of data were from different sources, and the five-year data shows clear trends (though with

There are different failure modes

Asiana missing a landing in SFO on a perfectly clear day is one example https://en.wikipedia.org/wiki/Asiana_Airlines_Flight_214

Cultures where hierarchy prevails over CRM and recent graduates can recall page 243 of the manual but can't do a visual landing in a perfect day are also dangerous.

About aviation in India, two recent incidents:

Jet Airways 9W-697 where crew apparently "forgot" to pressurize the plane. http://avherald.com/h?article=4bded8e6&opt=0

Air India AI-676 had "inexplicable loss of performance". Crew "forgot" to retract the landing gear: http://avherald.com/h?article=4ac18ec7&opt=0

Or I was going to sleep and didn’t want to stay up to write a report on my phone. Use your data but consider my anecdotal comment above. The aviation world is more opaque than one might think, and the governing bodies in certain areas are extremely protective of what is a great source of pride for their countries. I’ve worked with many people from these regions, talked with many people who work with people from these regions, and know many people who have worked in these regions, in addition to the exposure that comes from simply being in this industry. The system is designed to protect you and does its job, even when certain operators aren’t up to par, including the pilots.

The accidents rates by country paints a completely different picture. [1] This wikipedia list of glaring pilot incompetence leading to fatalities doesn't seem to show any bias for specific regions. [2]

When using anecdotal data one just can't make sweeping judgements and conclusions data does not support, that takes discussion into the region of prejudice and bigotry.

This anecdotal one person usually 'an insider' finding one thing wrong or incident in some country and casually conflating it to the whole is seen too often and creates grossly inaccurate stereotypes, when the exact same issue or incident exists everywhere else.

This is not harmless and derails informed discussion. It creates prejudice and in this case stress among travelers when they travel to other countries thinking about 'training, hull loss, safety' based on uninformed discussion not supported by data on the ground. If you are going to travel in a Cessna the standards including pilot training is not comparable to jets whichever country you are in. But no stable functioning country in the world takes passenger and running jets casually as the accident and safety data shows.

1. https://www.independent.co.uk/news/world/america-russia-and-...

2. https://en.wikipedia.org/wiki/Pilot_error#Notable_examples

That first link is pretty inapplicable - the numbers it shows are total fatalities and total accidents since 1945, not scaled for population of the country, not scaled for the quantity of air travel (in a time range where, especially early on where air travel was least safe, there were vast regional differences in wealth and frequency of flying), and over a long enough period that it doesn't really reflect the current reality.

Your second link is exactly the same kind of anecdotal BS you're complaining about - a list compiled by English-speaking editors on a Western-centric website of pilot errors they've personally heard about.

Discussion with actual numbers is here: https://news.ycombinator.com/item?id=18206505

These numbers seem correct, but I don't see how they add up to "extreme [pilot] incompetence" mentioned upthread. The countries that do worse are poorer, and therefore consistently fly older airframes with worse maintenance schedules, and regulations aren't as strict.

For all we know Africa has the best pilots in the world but the crappiest airframes, or fantastic pilots and awesome airframes but their maintenance is atrocious.

I'm not saying I find that especially plausible, but until these numbers are at the very least adjusted for hull loss rates by region of operator per million departures per age of airframe we're comparing apples to oranges.

This is the IATA data for hull loss rates for 2017 per million departures from your link:

Asia Pacific - 0.18

CIS - 0.92

Europe - 0.13

Latin America and the Caribbean - 0.41

Middle East and North Africa - ​​0.00

North America - 0.00

And it ends there. No mention of any problems or concern for 'safety limits' for any region. Infact the IATA data is overall happy with the rates across regions. If you can find anything on that page that points out a problem for a specific region please link to it.

From same link: "2017 was a very good year for aviation safety."

The link includes a comparison with the previous year (2016) - onboard fatalities were ~90% lower and jet hull losses were ~75% lower. You're cherry-picking a very atypical year.

I don’t think your links show that.

That first link tells you little without knowing the denominator for each country. How many flights did they have over that period of time?

And it covers more than 70 years, which is way too long to tell us whether a country or region has a bad safety record today...maybe America had 99% of its crashes in 1945, whereas the other countries on the list had all of theirs last year.

I used to work as a flight instructor and flew corporate. I’ve flown small airplanes with a bunch of airline pilots, whatever that is worth.

This does not seem like normal behavior from an airline pilot at all. I can see about 10% of private pilots doing something like this. It would probably be someone who has been flying a long time and gotten way too comfortable. Things always work out, so why not this time? The difference with an airline is its incredibly public flying. Even if the pilots disregard everyone’s safety (including their own) they know that they’ll be caught. You do something dumb and it’s going to be on the news, even if it all works out. How many of us can say that about or jobs?

The fact that the crew thought that this was a risk worth taking reflects really poorly on the safety culture of that airline. They thought that flying an airplane that was compromised some unknown amount was worth keeping the schedule going (or maybe people just wouldn’t find out?). That idea got informed somehow.

Here's the statement from the Airline. Not sure if 3600 hours is good experience in the industry? May be it is.. https://pbs.twimg.com/media/DpSNZO4V4AAPjKa.jpg:large

3600 hours is a good measure for routine, not for skills. I saw stupid accidents done by pilots with over 20 years and 10,000 hours of flying.

Note: 3600 hours on type, not total flight hours.

No, it isn't normal behavior. And it is not unknown that some airlines are more lax than others.

There is nothing about the ‚why’. Maybe they had an engine problem past V1 and decided to go. The long go around can be explained with them having to focus on stabilizing the situation, maybe involving problems with the landing gear. Even with the severe damage, the inner cabin and pressurization must have been intact. Assuming human failure should come last, when we have no details about the cause.

> The long go around can be explained

... by their desire to overwrite the CVR??

Well, I hope the pilots were not so dumb. I mean, there is still the FDR and recorded radio comms. If they took that time to overwrite the CVR, then they will probably lose their job anyway due to the suspicion of hiding, even when they would be acquitted of wrongdoing in a legal proceeding.

Not a pilot but familiar with the aviation world and procedures.

Obviously the plane had issues getting off the ground. Probably because it was too heavy. Typically, pilots calculate a speed beyond which they can no longer safely abort the takeoff. Beyond that speed, the plane is basically going to fly or shoot off the runway and crash. That point is decided based on weight, wind, runway length, etc. It's entirely possible mistakes were made with that or with loading the plane. Maybe there was some wind shear as well, which could explain a sudden drop or unexpected challenges getting off the ground.

Usually the co-pilot's job is to call these speeds out and if either of them calls to abort, there is supposed to be no discussion or debate on this and pilots are trained to act right away because every second counts. Obviously that did not happen

So, the pilots were not aware of issues before the abort speed (or they should have aborted) and committed to getting off the ground. I'm sure the in cockpit recordings will be part of the investigation. They sort of succeeded in the sense that they hit some objects but ultimately did not crash and got off the ground. Their climb rate must have been terrible. Usually the gear comes up as soon as you have a positive rate of climb (reduces drag). I imagine they called gear up seconds after leaving the ground before they hit anything even and the gear was likely transitioning.

Wind shear could have caused enough change in airspeed to cause the plane to not climb or even descend a bit. A heavy plane would have used up most of the runway in any case.

They then proceeded with what looks like normal procedure to get to altitude. Presumably they would have almost climbed out and changed frequencies from tower to local traffic controllers after 20-30 seconds or so. Recordings of that are going to be interesting. Presumably the damage to the ILS equipment would not have been noticed right away on the ground and people would have needed some time to figure out what the hell happened and what caused it and what plane hit it. Likewise the damage to the wall would not have been reported right away. By then the plane would have transitioned out of their area.

The question is whether the pilots noticed that they hit any objects and what the communication was with these controllers on this. The plane would have been pitched up (restricts visibility to the ground in front of you) and the damage occurred pretty far behind the pilots. So they may not have seen the obstacles immediately in front of them or heard/felt hitting them. These planes are big and heavy and there's lots of noise and vibrations when a plane takes off.

The damage looks dramatic but you wouldn't be able to see it until after the landing. Obviously the plane was flying and climbing and cruising normally. And they also landed safely. So flight operations seem to not have been directly affected. So, I can see them concluding that they had a stressful takeoff but had gotten in the air successfully. Then some time later they got the news that they hit some obstacles and that there is probably damage to the plane.

The big question is why that took 3 hours. I imagine it involved a lot of communication on the ground.

No. Go-around and assess the damage on the deck.

A go-around wouldn't give much chance to assess the damage beyond superficial. Would be much better to fully land and stop.

Not before consuming fuel. If overload was a factor, immediate landing is not a good idea.

The AVherald article: http://avherald.com/h?article=4bedd321&opt=0

As some of the others have mentioned, the reason the plane survived is likely because that was not a pressurised area. If it had breached the pressure vessel they would've noticed it very quickly (lack of pressurisation).

Not really - If there had a been a gash in pressurization, they would have caught it on the way up. The 737 is a _tank_. It has more or less the same wing load as a F-16. It takes a lot to shake around - it's every bit the 1950s build 707 chassis, and like the old 50s lathe - it can take a hit and keep going.

Plenty of 737's have had explosive depressurization with injuries or fatalities, most recently this year


and one of the most famous incidents (in the U.S.)


Both were from fatigue, but fatigue doesn't have to be a gradual thing; it could come from a single hit that causes the pressure vessel to fail at altitude.

See the whole list here


The difference is that happens at altitude. When the plane has just taken off there is no pressure differential. There’s no energy TO explode.

Damage can be less severe than an actual puncture of the pressure vessel. Damage can weaken the pressure vessel such that it only fails (possibly explosively) once under pressure.

And yet, it landed!

Indeed, but the two previous posters were discussing counter-factuals.

Don't the NGs have a nasty habit of splitting in two on hard landings (KI172, 4C8250, BW523, JT904, AA331)?

Luckily for this flight, that they didn't damage the pressure vessel and then blow it out at high altitude. Better to stay low, not pressurize, and land "at the nearest suitable airport."

Yeah, the decision of the pilots to ascend to higher altitude, ie lower pressures (ie pressurising the plane) is very questionable.

Edit: clarify, typo

The airplane does not require pressurization to fly. If the cabin fails to pressurize, there'd be an automatic warning and the pilots would put on oxygen masks and lower the altitude.

Impressive resilience by the airframe. Good job Boeing.

Yes, I'd say this incident makes for a good Boeing ad - "Even a brick wall can't stop a 737 from flying a route!"

That seems mildly irresponsible. "Eh, I don't need to pull up, it's a Boeing."

Lets hope they scrap that aircraft..

i could not find a reference but i remember an incident where an aircraft only scraped its backend on take off and repairs were made, but years later the aircraft was lost with everyone on board after cracks propagated up and around and the aircraft lost its whole tale section

Those incidents (CAL611, JAL123) only happened because the repair was completed improperly to begin with and subsequently missed by every inspection.

There's no reason the aircraft can't be properly repaired, and be as strong as it was when it was delivered. The Boeing Structural Repair Manual is very precise. It was just ignored.

The problem here is that the plane climbed risking catastrophic depressurization - you cannot repair in flight, you need to land first for that.

GP was proposing scrapping the aircraft, which is not necessary for ground repairs.


Worst single-aircraft disaster in history.

Yet another reason not to fly Air India, just in case you needed another beyond price, on-time performance, service, and cleanliness.

It wouldn't be the first time a crew errs in calculating their takeoff performance and uses too low a thrust or flap setting. Everyone on that plane is lucky to be alive, as others have died after similar takeoffs.

Too low a thrust?

I thought that on airliners pilots input runway length, altitude and wind strength and direction, and the computer optimized the thrust accordingly?

Yes, the autothrottle commands the thrust as calculated for a Flex take-off. But it's only correct if the input parameters are correct. There was a Sunwing 737 which bent the runway lights at Belfast in 2017 because the pilots had entered the wrong temperature variable:


Can we take a moment and reflect on the fact that all of this data is freely available to anyone? I can sit in front of my computer, or anywhere with my phone, and follow planes around the world with an accuracy of a few meters, and replay the data at a later date.

The various flight trackers all use thousands of receivers around the world, and most of them probably use the same kind of SDR chip that was originally designed for USB TV and radio receivers.


If anything this is great free advertisement for Boeing's quality.

As a pilot once told me: "If it ain't Boeing, I ain't going"

Brand loyalty is an amazingly powerful thing.

I wonder if anyone says, “I won’t be there Gus if it isn’t an Airbus”.

Time’s a plenty in an a320

and the quality of Air India pilots too :) I mean... seriously how bad can you be to put so many people at risk.

The obvious mystery is why you would keep flying for so long after an incident like this...

A good reason would be fuel weight on possibly damaged landing gear.

Also, the airport wasn't good anymore. They had just destroyed the Instrument Landing System at the end of the runway. Nobody should land there.

A lot of pilots don't use ILS and use a visual approach. That was one of the issues with the SF airport, as the left runway was closed, the lights were off and the pilot lined up with the taxi way.

> the lights were off

While it is true that the normal lighting was off (because it wasn't a valid runway at the time), there was a lighted, flashing X to indicate that fact. Along with the fact that the taxiway was not illuminated as a runway, and the pilots were notified about the runway closure…


> A lot of pilots don't use ILS and use a visual approach.

I feel like if you are so visually impaired, you can't see an object you are facing, you really should't land visually.

The ILS is not necessary for landing in VFR conditions, although non-USA pilots tend to rely on it as a crutch.

This is being downvoted, though is true to my understanding -- not uncommon for even the largest airports to have both ILS and non-ILS runways. (e.g. Runway 1L at SFO is non-ILS.)

In this particular case, the departure airport only had one runway, and immediately after the botched takeoff it was at least in dire need of a debris check, so I wouldn't be surprised if the airport was closed while that happened. Wouldn't have taken four hours, though, and I suspect that they would do whatever they could to accomodate an emergency aircraft had the pilot declared.

Maybe it didn't feel that bad in the cockpit, and the pilots were hoping they would wing it. They may have thought it would have endangered their jobs (which it certainly will...)

Usually in times of partial emergencies, there is a threshold point where you full own the emergency and declare it.

Maybe burning fuel to reduce the chance of explosion during the emergency landing? That's quite common.

No, from the description, they clearly ignored reports that they "may have hit something", they said "nah, we good." It'll be interesting to see, when the passenger reports start coming in, whether it's plausible the pilots just didn't feel it.

Boeing planes can dump fuel, iirc.

The 737 (subject of this article) cannot.



The rules are based on the fact that landing is harder on the airframe than taking off, so as long as max takeoff weight is only a fraction higher than max landing weight you get a pass.

And at any rate you get the greater of 10 minutes or 1% of max takeoff weight per minute to dump enough fuel to land safely. I wonder how long it takes a 737 to circle the runway before it's safe?

20 percent of fuel is used from take off to cruise altitude. On a plane like a 777, that still leaves a very substantial amount of fuel onboard. However on a smaller plane like a 737, you don't have to circle for too long before you're below MLW.

That's quite a design-flaw then.

Care to provide your credentials to make such a claim? Especially around one of the most widely used and longest produced airliners in human history.

The Airbus A320 can't either. It's not required as long as the landing system is designed to handle the fully-loaded weight.

Correct. Even on the much bigger A330 wide body, fuel dump is only an option: https://www.airliners.net/forum/viewtopic.php?t=1351703

Usually they can dump it

I thought they track things like small vibrations in plane body due to such incidents. How would they detect Bird strike(https://en.m.wikipedia.org/wiki/Bird_strike) if hitting wall go unnoticed? Don't airports have sensors and cameras for such situations to alert Air traffic control immediately?

It wasn't clear from the article. What kind of wall did they hit?

It's in the linked Twitter thread. It was a brick wall.


Well a single layer, unreinforced brick wall isn't really that strong, especially near the top edge. It's perhaps plausible they didn't feel anything or at least nothing very alarming. I mean, the gear presumably still retracted properly and deployed later for landing, so it wasn't damaged too badly.

Yea I'm interested in seeing what the passengers say. If they just heard a bump or didn't suspect a thing, the plane may have just graised it and the pilots didn't notice. If the passengers heard a noticeable sound or reported concerns and were dismissed, that's grounds for negligence.

I don't know about you but I hear plenty of clunks and wirring on planes during take off and landing, sometimes uncomfortably loud and even though I know a bit about planes, I don't know if it's normal or not. (I'm alive so I assume it is normal!)

One takeoff I was on, the plane was vibrating like crazy, like I've never heard before but after a few minutes, it stopped and nothing happened!

The other photo posted here shows that it was a rather thin brick wall too:


I guess the majority of the damage was done by the large steel antennae it hit before the wall.

Good grief that's a lot of damage, how come they didn't immediately land?

They hit a wall and the Instrument Landing System at the end of the runway

That doesn't really answer irrational's question...

There's a photo in the article -- a brick wall.

Honest question: why aren't pilots forced to take alcohol tests before flights? There was some JetBlue incident about 10 years ago, IIRC.

They are, but not for every flight. I've been told by pilots flying in Europe they get randomly selected for a breathalyzer test once every 20 flights or so.

Are forced alcohol tests given anywhere?

Not that I know of. Mandatory drug testing in sports is now common. When operating heavy machinery, and when a lot of lives are at stake, I feel some safety procedures are called for.

In some places bus drivers have mandatory breathalyzer testing at the start of their shift.

Its mystery how plane survived such a huge impact

Airport perimeter damage: https://mobile.twitter.com/ANI/status/1050582092688629760

The landing gear is built to withstand hitting a concrete runway with a 737, whereas a thin brick wall can be brought down with just a hammer. It's like rock-paper-scissors: wheels beat wall, ILS antenna beats fuselage skin.

A similar incident that happened with a bigger plane that completed a 13h flight: http://avherald.com/h?article=48c78b3a

It looks like the plane couldn't properly take off(likely overloaded), and it's not just an overrun. See the speed graph in the post, and the fact that the gears were retracted that low.

I don’t think we have enough information to make that statement yet. Could be they forgot to set flaps, could be they miscalculated TO speeds, could be they misconfigured auto throttle... As usual, need to wait for investigation and report.

Why is there a brick wall so close to runaway?

Because most pilots can fly professionally enough not to hit it.

There are failures modes a pilot has no chance of resolving.

There are worse places to take off/land. Midway (MDW) has very little distance from the runway to houses and highway. Princess Juliana International Airport on Sint Maarten (SXM) has very little distance from the end of the runway to the ocean. Gibraltar International Airport (GIB) is also basically on the water.

Overrunning the runway is often considered avoidable from my limited understanding. Planes are flight worthy under _a lot_ of issues. Even single engine failure or engine fire after V1 (the speed at which you can't abort/reject takeoff without overrunning the runway) isn't cause to abort/reject takeoff.

With landing, the plane needs to touch down early enough and at the correct speed to stay on the runway.

Overrunning the runway is never considered a "good" thing. Rejected takeoffs that shouldn't have been and bad landings are often causes of death. They're expected to not happen.

Usually? Because there's an urban area with people and highways and houses past the runway.

As far as I can tell this wasn't unusually close to the runway, the plane just stayed on the ground far past the end of the runway.

Air India, an old antiquated government subsidized airline that should die.

The government has actually been trying to kill it for years but no one wants it: https://money.cnn.com/2018/05/31/investing/air-india-privati...

This might provide the government the opportunity to shutter it for good.

Right, there have been no incidents in privately run airlines over the past few years.

Sounds a lot like Alitalia.

Trust me, if procedures are found at fault, it will die if the EASA-FAA-CASA bans them.


I think it's not appropriate to generalize. How many such incidents happen?

Applications are open for YC Summer 2019

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact