Hacker News new | past | comments | ask | show | jobs | submit login
PIA Airbus 320 Crashes near Karachi (bbc.com)
109 points by billfruit 13 days ago | hide | past | web | favorite | 87 comments





This incident occurred in my neighborhood. The plane crashed in a residential area and the overall death toll is still unknown. There are a few confirmed survivors so far: President of Bank of Punjab and a 3-year-old kid.

I wonder if it was a fuel issue. The "we lost two engines" near the destination airport makes me think of that.

Based on pictures circulating on Twitter, it looks like the plane's engines may have scraped on the runway (presumably in a landing attempt without landing gear down) before taking off again.

Very unlikely, if the landing gear has problems to the extent that you scrape the runway with the engines, then taking off is almost impossible and useless, you just stop to a halt. If your landing gear does not extend and lock properly, you abort the landing way before touching it.

A gear-up touch-and-go has been done before though:

https://en.wikipedia.org/wiki/Mal%C3%A9v_Flight_262


The engines were mounted above the wings on that airplane, so they wouldn't have scraped the ground during the touch-and-go.

khuey's interpretation was my assumption too having seen the Twitter pictures of the damaged nacelles, and lack of landing gear, during the first go-around. The RAT being deployed suggested to me some sort of hydraulics fault too; or maybe engines had failed at that point?

So, what would you say caused that damage?

https://mobile.twitter.com/PlaneSpottersPK/status/1263810587...


To further elaborate (since it's the first time I'd heard about this system), the RAT ('Ram air turbine') as used on the A320 is a mechanism for generating power for flight-critical systems during emergencies.

A good default stance to take in situations like this (especially so soon after a tragedy) is that the pilot and crew were likely fully capable and well-aware of what they were doing.

[1] - https://en.wikipedia.org/wiki/Ram_air_turbine


That depends on when the pilots became aware that the gear was not down. If thier first indication was touchdown, then ya, taking off again isn't going to happen. But remember that these aircraft don't always respond instantly. Perhaps the pilots were aware some time before touchdown. They could have touched down half way through an attempted an abort, only just grazing the ground.

You are completely right, based on the pictures I saw this is most probably what happened.

This assumes experienced pilots who were paying attention and following the checklist. If they weren't and forgot the gear on the first pass they might have panicked and tried a go-around.

The A320 has a very loud gear up alarm, and it is one of 3 that cannot be silenced or disabled by any means.

Initial speculation (because that's all anything will be for weeks at the very least) and comments from other A320 pilots indicate what likely happened is they were attempting an intentional gear up landing, were too fast and burned up too much runway, and attempting a go-ahead. Raising the nose can cause the rear to hit the ground if you're close enough, and this would explain the charring on the bottom of the engine nacelles.


I'm not an expert (y'know, like everyone in any given HN comments section), but surely there must be automated systems in place that warn the pilots if the landing gear is not deployed when within x meters of the ground?

Yes there will be an automated warning. On the airbus it calls out "retard" below 20 ft. Edit: This wrong, see reply below.

"Retard" is a normal callout during touchdown. The gear up warning is a separate "too low gear" callout from TCAS.

"Too low, gear" (and the other terrain alarms) are from GPWS, not TCAS.

Thanks.

The RAT was deployed too, so clearly they were having power issues. Just haven't seen if they had power plant problems before or after the engines appear to have scraped the ground.

Not much more than an armchair flight engineer here, but even if there's a hydraulic failure, does the A320 have a failsafe mode where the landing gear can be deployed via gravity?

Yes, you can use the alternate gear extension mechanism to manually unlock the wheel bay doors and rely on gravity to pull the gear down and lock it into place. This sometimes doesn't lock the nose wheel (you'd have to be nose down a lot for it to work) but that's usually okay.

There were also multiple landing attempts, which supports that theory as well.

The FlightRadar24 plot shows only one go-around attempt, with a climb to 2,000.

I agree there's no indication of more than one go-around in FlightRadar24's data [1].

But during takeoff, the data shows the flight jumping from 0 to 1000 ft in a single measurement, so it's probably not as robust and granular as we might hope.

[1] https://www.flightradar24.com/blog/pakistan-international-ai...


Yea, ADSB rarely shows up on airplanes below 1000ft. On many of my flying lessons the Flightaware flight path goes haywire because it doesn't always pick up pattern traffic.

Ran out of fuel after 3 aborted landings, pilots perhaps distracted by an engine failure?

Based on the photos of the damage to the engines the aborted landings didn't go well.

https://avherald.com/h?article=4d7a6e9a&opt=0


The FAA and ICAO regulation for flights above 18,000 feet is that you need to have enough fuel to reach your destination, an appropriate alternate airport, and cruise for an additional 45 minutes. I doubt they ran out of fuel after 3 attempts to land at the original destination airport.

I'd love for a heavy pilot to comment on the math involved. Fuel burn on takeoff is a lot more than cruise. I'd assume ~3x as much. Throw in the amount of time to fly around a pattern and that 45 minutes doesn't give a lot of flexibility if you start botching landings.

This may surprise you, but the fuel burn for holding is essentially constant vs altitude. What suffers is the range, flying at low altitude is much worse for fuel range because at high altitude, true air speed increases for a given indicated air speed.

Page 17 shows ~5000 lb/hr at mid weight: https://www.air-septimanie.com/pdf/technical/A320/en/A320PER...


For a large jet, a single go-around can take about 12 minutes.

Longer if they're having problems turning the plane due to hydraulic issues.


3 aborted landings is a lot.

Isn’t it considered good policy to redirect after a few aborted landings?

(Not a Pilot).


I don't think there where three aborted landings. My understanding is there was only one go around. This is backed up by flightradar24's altitude graph: https://twitter.com/flightradar24/status/1263800350060003328...

Generally there are alternate airports chosen for if a plane can't land at its desired destination + fuel for multiple aborted landings


> Isn’t it considered good policy to redirect after a few aborted landings?

If weather is below minimums.

But an airliner doing more than 1 go-around in good weather is something that I've never heard of - it means the pilots don't know how to control airspeed.


Seems they couldn't get their nose gear down.

In US training, generally it's recommended to configure the airplane for landing, and if the nosewheel collapses, just ride it down.

Because airplanes land on their main (back wheels) first, the nosewheel is not considered very important, and the incident is considered to be very survivable.

However, in foreign countries, there can be a greater fear of career repercussions for damaging equipment or delaying flights. This was also seen in the SFO Asiana accident that a foreign crew just fell apart in an atypical scenario:

https://en.wikipedia.org/wiki/Asiana_Airlines_Flight_214

Once in a while in the US you hear of a pilot trying to land firmly to snap the nosewheel into position, but it would be unusual for them to take off and try it again, esp. with passengers.

Some factors are: the pilot doesn't trust the gear indicators for some reason, or they decide burning off fuel before landing is safer, or the airport doesn't have adequate firefighting equipment or a nearby hospital.

Source: commercially-rated airplane pilot, small airplane maintenance test pilot.


Just curious: is it possible that they knew they had a gear malfunction and didn't know whether it was nose or main, when they decided to go around?

There is a single go-around; the plane burst in flames, there was enough fuel for that so there was enough fuel to run the engines, the minimum reserve is 30 minutes, the go-around takes a lot less.

Planes have a lot more combustible material on them than just jet fuel. Engine oil and oxygen canisters for a start.

A transmission of the pilot’s final exchange with air traffic control, posted on the website LiveATC.net, indicated he had failed to land and was circling around to make another attempt, reported AP.

“We are proceeding direct, sir — we have lost engine,” a pilot can be heard saying.

“Confirm your attempt on belly,” the air traffic controller said, offering a runway.

“Sir - mayday, mayday, mayday, mayday Pakistan 8303,” the pilot said before the transmission ended.

source: https://www.dawn.com/news/1558944/at-least-80-killed-as-plan...


How big are the chances that the most prominent passenger is the sole survivor of a plane crash (yet, but I don't expect many more)? The Bank of Panjab president, whow! Lucky guy.

Decent, given that the current fatality count is 11 out of 91 passengers and an unspecified number of crew.

Certainly the fatality count will rise, but with modern planes survival tends to be more about escaping the post crash fire. If >50% of the passengers and crew live, I wouldn’t be surprised at all.


I'd expect a lower survival rate as it appears to have crashed into buildings, and not into a field or similar.

Crashing in a city means that you crash into buildings that are both flammable and hinder escape, as you’ve said. But it also means that you’re crashing near a means of rescue, rather than having to wait precious minutes for rescue services to locate you. Only time will tell which of these factors was more important.

Good news is that crashes on landing tend to involve a lot less fuel, which should improve survival chances significantly.


I read this earlier today and was hopeful that fatality count would be low. Just shocked to see a final count of 2 survivors total.

There is no information that he is sole survivor. There is no information about how many survivors there are exactly.

Is he really lucky? Any person taking part in plane crash is unlucky in my view. Also "survived" does not yet mean he is unhurt.


The subset of people who are on an airplane at any point in time generally skews heavily prominent. Especially more so when you subtract almost 100% of tourists from the equation.

That might be sampling bias - how many people do you know that fly around these days? (Corona, lockdown and all)

There are 40 survivors reported in local media.

Even if it was the case I wouldn't draw any "conspiracy" conclusions from it. As part of a demanding CEO role he may of be quite physically fit, or maybe he just had the whole of business class to himself.

If so, it would be a reverse conspiracy. The conspiracy folks certainly would come to weird conclusions if he would have died. And prominent folks usually sit in the front rows, the most insecure places in a plane.

But I wrote that idea very early after the crash. I didn't expect that so many survived a plane crash, even into 4 houses. Now it became clear that many survived. Almost a miracle to me. Very lucky.

My initial idea was some dust in the engines, because that was the very first flight after the lockdown. Bad maintenance.


My understanding is that the front of a plane is usually more survivable in a crash. Usually crashes are tail-first, and the back is closer to the engines / gas to star with. (pilots are generally trying to pull up at the time of a crash, so the nose is up and tail is down).

So yeah, if he was in business class, I don't think it'd be crazy for that to boost his survival odds (assuming anyone does survive).


That is not correct. The back of the plane is slightly safer based on historical data, largely due to the tendency of the front of the plane to break off during a crash landing. [0]

If the pilot is still pulling up when the plane contacts the ground, then you’re probably talking about an extremely high speed crash, where no seat is safe. The types of crashes you need to analyze for seat specific survivability are things like runway excursions and relatively level landings onto fields or bodies of water.

O: https://youtu.be/kJZ1eHU_JZg


Lots more details here, as usual:

http://avherald.com/h?article=4d7a6e9a&opt=0


The mention of Pakistan's "chequered aviation safety record" got me wondering how it compares to other airlines.

I came across this site [0], and found the data to be really interesting. Pakistan International Airlines ranks decently well, while Southwest Airlines ranks among the worst.

[0] https://www.airlineratings.com/safety-rating-tool/


This finding is an artifact of their methodology. Southwest is not EU admitted because they only recently expanded to routes outside of the US and have not so far expanded outside of the Americas (they may not want to as it would introduce real complications to unified-fleet their business model). They are not fatality-free in the last ten years, but their fatality in 2018 was the only passenger fatality they have ever experienced. This incident also seemed fairly fault-free on the part of Southwest as it resulted in guidance changes requiring more frequent inspection of the engine model involved than previously.

Perhaps they should undergo IOSA audit, but it is entirely optional and is viewed as less necessary in the US because the FAA's mandatory safety oversight is similarly strict. It is mostly a certification obtained by airlines to demonstrate that they are "as safe" as their peers, and as the largest carrier in the US with an excellent safety record, Southwest is probably just not very inclined to go to the extra expense for a certificate to hang on the wall.

I would say this is a basic problem with evaluating airline safety based on external accreditation rather than on their actual safety record. However, I'm sure the latter would be much more difficult to do globally.


This is like ranking software engineers solely on the number of Microsoft certifications they have and the total number of open issues in their GitHub repos.

Given that Southwest Airlines has killed one passenger in its entire history I'm going to say that a methodology that ranks them among the worst in safety is likely suspect.

I was going to say, that doesn’t pass the sniff test.

Excluding this current incident, PIA has killed 102 passengers and crew since 2000.

The most expansive view of Southwest’s record would include 4 fatalities: one passenger, two people on the ground, and another passenger who died of a medical issue while being restrained.

One of those ground deaths was because someone walked onto the runway and was hit by a landing 737, so I’m not inclined to blame Southwest for that.


https://en.wikipedia.org/wiki/Southwest_Airlines_Flight_1248

One of the crazier things I've seen. Really the only fatality a Southwest pilot is responsible for.


Incidents and other things that normal passengers don't hear about a large role in such statistics.

While I agree, one would expect incidents to scale roughly with the rate of fatalities. With PIA having 2 orders of magnitude more fatalities in this century, it would take some pretty extraordinary non-fatality incident statistics to declare that PIA is much safer than Southwest.

Also worth considering is fatalities per flight hour. While I couldn’t find accurate flight hours for all the airlines, Southwest has a much larger fleet (752 vs. 31). Combined with a much lower fatality count (arguably 2 vs. 102), this implies that Southwest is much, much safer than PIA. It’s possible that non-fatal incidents can cover that gap, but at a high level analysis this seems implausible.


Checking further, it seems that at least some of the ratings depend on having passed IATA audit, which is voluntary. So a Low Cost immediately loses on that.

They are dinging Southwest mostly because it has never asked for IOSA certification. There's no requirement for them to do so, because they aren't an IATA member.

I do not trust that web site. Their methodology is suspect (see Southwest Airlines, as you mention) and their data is suspect (Alaska Airlines has no fatalities? No, they have had 9 accidents that resulted in at least one death, including several that resulted in all lives on the plane lost).

Flight 261’s crash killing 88 because of “poorly conceived and woefully executed“ maintenance procedures sure feels like something that should show up on that list.

The first thing I thought of here were some of the concerns around return-to-flight after planes have been sitting around due to the pandemic.

Hope there are more survivors both from aircraft as well as on the ground -- Karachi, Pakistan is a densely populated City.

I am wondering why editing the original title was allowed in this case?

[flagged]


Not to dismiss the gravity of the situation, but engineering and process failures are interesting to this audience. This is one example, likely of the latter.

"On-Topic: Anything that good hackers would find interesting. That includes more than hacking and startups. If you had to reduce it to a sentence, the answer might be: anything that gratifies one's intellectual curiosity."

https://news.ycombinator.com/newsguidelines.html


"Off-Topic: [..] If they'd cover it on TV news, it's probably off-topic." (and it appears previous submissions of this indeed did get flagged. Wheareas an after-the-fact analysis of the causes IMHO would be on-topic)

(But also: "Please don't complain that a submission is inappropriate. If a story is spam or off-topic, flag it.")


Is there an option to hide tech unrelated news ?

Not generally, but if there's a thread you just don't want to see, there's a "hide" link among the links underneath the thread title. Just search for hide on the page and it'll be the first result.

No.

:(

I think I will be the first of many to say this: When they will finally get PIA privatised for good?


When Air India gets privatized for good ;)

Jokes aside, airlines around the world are pretty much nationalized one way or the other -- and we have to re-evaluate things on the other side of pandemic 2 years from now.


Can you provide more context? Why would this help?

For example in the national flag carrier I know (I used to work 5 years in that sector and I am still a licensed pilot) many people were hired and appointed in positions based on political influence, not on competence. It happens less in private companies.

But do we have evidence that privately owned airlines have a smaller record of accidents and incidents than those not?

Hard to compare, because not many countries have at the same time government owned airliners and private. In the case I know, the record is flawless for private companies (zero accidents) and 2 crashes for the flag carrier (a relative was part of the crew in one crash). I would not make a general statement based on a single country, but looking for airplane crashes in the past 20 years most involved government owned companies.

I wrote a thesis on place name ambiguity [1], so not surprised. There are tons of people flying to the wrong place with the same name all the time - they only told me when I mentioned my thesis topic.

There's even a book about a guy who visited half of the places called 'Aberdeen' on earth, which took 10 years.

[1] J. Leidner (2007) Toponym resolution in Text


You think the plane flew to the wrong place?

probably wanted to comment on https://news.ycombinator.com/item?id=23270232 instead

How can you accidentally comment on the wrong article?

Well to be fair, commenting on the wrong article with a comment about going to the wrong place kind of proves the point.

open several articles up in multiple tabs.. accidentally click the wrong tab when they go back to comment?

This is interesting. It makes me think of the way we cluster thematically-related placenames in cities. It's well intentioned, but I think people can mentally store place names based on other categorical groupings (eg, we might remember a street had something to do with the sun- only to find all the area streets have some relation).

I'd love to read more on your thesis!




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

Search: