PIA A320 Crash Karachi
Join Date: Oct 2010
Location: 5° above the Equator, 75° left of Greenwich
Posts: 411
Likes: 0
Received 0 Likes
on
0 Posts
It’s pretty ingrained in almost all of us, apparently, to fire the APU at the hint of a dual (or even a single!) engine hiccup. However, I’m with vilas on this one: it wouldn’t have made any significant difference in the final outcome.
MikeSnow: Notice the report says “normal law”. As vilas pointed out, this means the aircraft always had at least one normal generator working, meaning the APU generator came online before the second generator got offline (if it ever did, as one of the engines had only partial thrust loss). As sufficient hydraulic pressure is generated with very low N2 values, losing only one hydraulic system (I.e. the one supported by the completely dead engine) doesn’t get the aircraft out of normal law, and most likely if the generator wasn’t kicked offline, then the partially failed engine was still supporting its hydraulic system.
It seems the PIA case is quite different, as the photos of the airplane with the gear up show the RAT deployed, meaning it did lose power to both AC busses (read: generators). If I venture a little analysis of the altitude and speed charts in the preliminary report, we see the airspeed after the go around never got above 250-ish, then it goes down to about 200 and stays there until the end of the plot. The point where the speed starts to decrease from 250 could be where the first engine failed, then they seem to have traded speed for altitude (since that one kept going up), speed stabilized in regions consistent with normal green dot speeds, and it’s just one mile (starting from the runway) since the speed starts to go down to the end of the plot, which is when the second generator got offline (I.e.: second engine failure). You need a whole minute to get the APU running and online in the A320. There was never going to be enough time to get it running. And even then, the second engine still failed, so they would’ve gone into alternate law anyway. They only would’ve had double instruments, the critical bits of the plane would be the same as without APU. And obviously that didn’t help them before...
lomapaseo: I think yours is a very difficult question to answer with the information we have available, at least to guesstimate where this particular aircraft would have (crash)landed. They weren’t at typical values of speed, altitude and configuration to try and estimate the flight path. And I guess it all depends on when did they lower the gear
MikeSnow: Notice the report says “normal law”. As vilas pointed out, this means the aircraft always had at least one normal generator working, meaning the APU generator came online before the second generator got offline (if it ever did, as one of the engines had only partial thrust loss). As sufficient hydraulic pressure is generated with very low N2 values, losing only one hydraulic system (I.e. the one supported by the completely dead engine) doesn’t get the aircraft out of normal law, and most likely if the generator wasn’t kicked offline, then the partially failed engine was still supporting its hydraulic system.
It seems the PIA case is quite different, as the photos of the airplane with the gear up show the RAT deployed, meaning it did lose power to both AC busses (read: generators). If I venture a little analysis of the altitude and speed charts in the preliminary report, we see the airspeed after the go around never got above 250-ish, then it goes down to about 200 and stays there until the end of the plot. The point where the speed starts to decrease from 250 could be where the first engine failed, then they seem to have traded speed for altitude (since that one kept going up), speed stabilized in regions consistent with normal green dot speeds, and it’s just one mile (starting from the runway) since the speed starts to go down to the end of the plot, which is when the second generator got offline (I.e.: second engine failure). You need a whole minute to get the APU running and online in the A320. There was never going to be enough time to get it running. And even then, the second engine still failed, so they would’ve gone into alternate law anyway. They only would’ve had double instruments, the critical bits of the plane would be the same as without APU. And obviously that didn’t help them before...
lomapaseo: I think yours is a very difficult question to answer with the information we have available, at least to guesstimate where this particular aircraft would have (crash)landed. They weren’t at typical values of speed, altitude and configuration to try and estimate the flight path. And I guess it all depends on when did they lower the gear
Join Date: Jan 2020
Location: Bucharest
Posts: 93
Likes: 0
Received 0 Likes
on
0 Posts
Last edited by MikeSnow; 10th Jul 2020 at 10:33.
Join Date: Jun 2007
Location: Wanderlust
Posts: 3,405
Likes: 0
Received 0 Likes
on
0 Posts
I repeat again APU in 1549 case didn't do anything else. In dual Engine flame out checklist the APU start is way down because in ELEV EMER. you cannot start APU above 25000ft. and because of high speed windmill relight is available at higher altitude. He started it from memory and it would have been life saver had a flamed out Engine restarted. But Engines had not flamed out they remained crippled. Normal Law or alternate law makes no difference to ditching. In normal law due to dropping speed he got into alpha prot and couldn't flare sufficiently and in alternate law he would have got stall warning and he would have pushed the nose down but the result would have been same, hard impact.
Join Date: Jan 2020
Location: Bucharest
Posts: 93
Likes: 0
Received 0 Likes
on
0 Posts
I stand corrected, reading an interview with Sullenberger confirms that being in normal law didn't help them, because they didn't reach the limits that would have triggered the envelope protections. But he felt that having them available was a good thing, he said about the protections "I had that as an ace in the hole".
https://www.historynet.com/sully-speaks-out.htm
So, not sure why the NTSB felt that starting the APU was so critical for the outcome.
https://www.historynet.com/sully-speaks-out.htm
So we stayed in normal law—all the flight envelope protections were intact. That helped provide essentially guardrails to prevent us from exceeding certain extremes in terms of speed or attitude, for example. So I had that as an ace in the hole. It turns out we didn’t need those protections; we never got to the limits at which they would have protected us from ourselves.
Join Date: Jun 2007
Location: Wanderlust
Posts: 3,405
Likes: 0
Received 0 Likes
on
0 Posts
The main issue with flight 1549 was that the crew was faced with a serious emergency not practiced nor was there a precise check list suitable for the situation. Sullenburger's greatest virtue was that in one minute he decided he will land in Hudson, going back as he rightly realized would have deprived him of any change later. That ensured nobody was going to die, full credit for that. However as we know now all other things like starting the APU or asking the copilot to restart Engines turned out superfluous and distracting. The Engines had not failed they were not going to restart to new life. Had he not done that the copilot would have guided him to maintain correct speed and proper flare would have prevented serious injuries. Even selection of Flap3 may have released from alpha prot. Anyway that is for others to understand for the future.
Last edited by vilas; 11th Jul 2020 at 11:42.
Join Date: Jul 2013
Location: Within AM radio broadcast range of downtown Chicago
Age: 71
Posts: 842
Received 0 Likes
on
0 Posts
vilas, does your description of the events of 1549 and the flight crew's decisions lead to any conclusion about whether or not algorithms alone could have dealt with the emergency as well? I get that this might be thread drift but . . . . I haven't yet been able to understand how a batch of code could cope with an "omitted case" - like no checklist for this situation specifically.
Only half a speed-brake
Never.
There will be more and more automation, replacing human handling skills. Yes, perhaps a ditching flare mode one day.
But never a reliable system that can take decision and prioritize, adapt and reevaluate in the real time.
Automated yes, autonomous never.
Liveware will stay for two reasons:
- backup and takeover from the automation, beyond its programmed scope
- public perception, marketing
The second will be the more important, especially as various dirty games about autonomous vehicles will come to light.
There will be more and more automation, replacing human handling skills. Yes, perhaps a ditching flare mode one day.
But never a reliable system that can take decision and prioritize, adapt and reevaluate in the real time.
Automated yes, autonomous never.
Liveware will stay for two reasons:
- backup and takeover from the automation, beyond its programmed scope
- public perception, marketing
The second will be the more important, especially as various dirty games about autonomous vehicles will come to light.
Join Date: Jun 2007
Location: Wanderlust
Posts: 3,405
Likes: 0
Received 0 Likes
on
0 Posts
WillowRun 6-3
It's not impossible. NTSB in 1549 inquiry gives some references about research on systems that were to give more precise state of Engine damage to Pilots. They were given up due to lack of funding. If it was available then crew wouldn't have wasted time on relight. Now in A350 Autopilot is available with dual Engine flame out. Automaic decisions and executing a landing or ditching should be possible.
It's not impossible. NTSB in 1549 inquiry gives some references about research on systems that were to give more precise state of Engine damage to Pilots. They were given up due to lack of funding. If it was available then crew wouldn't have wasted time on relight. Now in A350 Autopilot is available with dual Engine flame out. Automaic decisions and executing a landing or ditching should be possible.
It's mostly a pinch point on time vs prioritization of tasks flying the airplane. If EGT% is pegged out at max, you don't have a chance unless you shut it down as the turbine will melt, not just overheat. Sometimes you don't have time to even care (as heard in CVRs). If the egt is simply overlimits you might be able to recover some thrust by reducing the throttles or restart if you have the time, but otherwise just leave it alone and work higher priorities (damaged engines rarely cause significant hazards after the first seconds)
If you have time to even think about the engines, then consider the restart itself needs time (airspeed, altitude or APU). The only problem I can see in past history has been the prioritization of tasks or not reducing the throttles for simple overheats
If you have time to even think about the engines, then consider the restart itself needs time (airspeed, altitude or APU). The only problem I can see in past history has been the prioritization of tasks or not reducing the throttles for simple overheats
Join Date: Feb 2008
Location: victoria bc
Age: 82
Posts: 77
Likes: 0
Received 0 Likes
on
0 Posts
The backup and takeover could be done from the ground, public perception and marketing with the same smoke and mirrors applied to everything else..
There will always be a shortage of experienced pilots in remote places but maybe not so likely in the major airlines of the future.
There will always be a shortage of experienced pilots in remote places but maybe not so likely in the major airlines of the future.
Psychophysiological entity
But never a reliable system that can take decision and prioritize, adapt and reevaluate in the real time.
Systems will have an encyclopaedic knowledge every air accident ever recorded and know how to calculate around a reoccurrence. It will know the world's weather with fine detail on the route. It will know the the location of hostile armies and the moment planning permission is granted for the simplest tall object. It will create pages of ACARS while acting on the information - scrolling lists of horrors will not faze it one jot. It will never lose situational awareness, knowing where it is relative to the planet and what used to be called the fixed stars. It will not suffer stress, or any other kind of work overload - in fact it will rarely use a significant proportion of its computing power.
Decades away? Probably, but the incredible thing is we are within touching distance of such technology now. The public's acceptance? Quite a different matter.
Last edited by Loose rivets; 13th Jul 2020 at 01:24. Reason: proof reading requires really reading stuff.
Only half a speed-brake
Loose Rivet Today the decision making element does not need to be certified.
If the umpteenth next generation was to be what I called autonomous, the software "AI" would need to be formally reviewed and approved.
Through that the development cost will be enormous and unpredictable as well. Not to make a code that works, but a code that you can prove won't do anything else apart from what it says on the tin.
My take is that the investors would not approve a launch of a project where the worst case cost are impossible to estimate.
You can pretty y much always fix and aeroplane but code sometimes just cannot be. And a replacement one is prone to another full set of unknown bugs, just like the original one.
If the umpteenth next generation was to be what I called autonomous, the software "AI" would need to be formally reviewed and approved.
Through that the development cost will be enormous and unpredictable as well. Not to make a code that works, but a code that you can prove won't do anything else apart from what it says on the tin.
My take is that the investors would not approve a launch of a project where the worst case cost are impossible to estimate.
You can pretty y much always fix and aeroplane but code sometimes just cannot be. And a replacement one is prone to another full set of unknown bugs, just like the original one.
Only half a speed-brake
For PAX ops, I do not think so. Whilst I agree about your smoke and mirrors observation, it still feels like there are sufficient number of sub-critical masses to keep a human element inside a steering chamber on board forever.
Yes, words flight deck or pilot will become obsolete.
Pegase Driver
Join Date: May 1997
Location: Europe
Age: 74
Posts: 3,684
Likes: 0
Received 0 Likes
on
0 Posts
Yes the automated technology is there and already on trial ( check Airbus ATTOL on google) Autonomous , not yet and not for a long time I would say .
As to the critics about automated in modes of transport, we should remember we are in the very early beginning of automated cars.sea and air transportation .
The first one that tried was the railroad : 2 dimensional . There are numerous trains and metro systems that are automated today . Public resistance and of "it will never work" cries by professionals were similar in the 1970-80's when they were introduced.
To take an example today 2 main metro lines in Paris are fully automated since years , run every 2 min at peak times ( compared to 4 before) and were the only lines operating without interruption during the long strikes last winter..
That last part also is an interesting one regarding automation and public acceptance..
As to the critics about automated in modes of transport, we should remember we are in the very early beginning of automated cars.sea and air transportation .
The first one that tried was the railroad : 2 dimensional . There are numerous trains and metro systems that are automated today . Public resistance and of "it will never work" cries by professionals were similar in the 1970-80's when they were introduced.
To take an example today 2 main metro lines in Paris are fully automated since years , run every 2 min at peak times ( compared to 4 before) and were the only lines operating without interruption during the long strikes last winter..
That last part also is an interesting one regarding automation and public acceptance..
Join Date: Dec 2001
Location: what U.S. calls ´old Europe´
Posts: 941
Likes: 0
Received 0 Likes
on
0 Posts
Certain trains between airport terminals run autonomous since 20 years... Not at all the same technology you would need for autonomous flight.
But maybe we do not need it anyway, as our conventional aviaton today is also not fully autonomous but mainly ground controlled. We only talk about a different level of ATC. Which, as we can see in this example, may have saved the day.
But maybe we do not need it anyway, as our conventional aviaton today is also not fully autonomous but mainly ground controlled. We only talk about a different level of ATC. Which, as we can see in this example, may have saved the day.
Right now we have people complaining about the cost of actions to prevent the spread of a disease that kills or injures a few people per 100. But imagine the uproar if an aircraft manufacturer said, "it's too expensive to introduce a technology that would prevent one crash per 10,000 takeoffs" !
WillowRun 6-3
It's not impossible. NTSB in 1549 inquiry gives some references about research on systems that were to give more precise state of Engine damage to Pilots. They were given up due to lack of funding. If it was available then crew wouldn't have wasted time on relight. Now in A350 Autopilot is available with dual Engine flame out. Automaic decisions and executing a landing or ditching should be possible.
It's not impossible. NTSB in 1549 inquiry gives some references about research on systems that were to give more precise state of Engine damage to Pilots. They were given up due to lack of funding. If it was available then crew wouldn't have wasted time on relight. Now in A350 Autopilot is available with dual Engine flame out. Automaic decisions and executing a landing or ditching should be possible.
are we supposed to do that?' It was literally a case of not knowing where or how to start.There are plenty of engine monitoring algorithms that measure the health of the engine and send out reports (e.g. ACARS). But they all depend on the same thing - the engine has to be running before they have viable inputs. I think some of the R&D types went off and started thinking about the problem (particularly at the engine companies), but I doubt there is much that can be done with current technology.
All true
to add
Fancy thinking has a way of biting your a$$ sometimes when it interacts with human performance under emergency conditions. Pilots with negative buoyancy may react in ways that negate the automatics. Then there is the difficulty of how to handle things like ash ingestion, where the engines likely were not restartable initially until repeated attempts finally wore away enough ash-buildup to get one or two going again.Like I said earlier, I can find fault with the pilots only in the first minute of an event whilst realizing minutes later no amount of engine restart software is likely to help
to add
Fancy thinking has a way of biting your a$$ sometimes when it interacts with human performance under emergency conditions. Pilots with negative buoyancy may react in ways that negate the automatics. Then there is the difficulty of how to handle things like ash ingestion, where the engines likely were not restartable initially until repeated attempts finally wore away enough ash-buildup to get one or two going again.Like I said earlier, I can find fault with the pilots only in the first minute of an event whilst realizing minutes later no amount of engine restart software is likely to help
Join Date: Jul 2013
Location: Within AM radio broadcast range of downtown Chicago
Age: 71
Posts: 842
Received 0 Likes
on
0 Posts
In the wake of PIA 8303 (Karachi)??
IFALPA has issued a Safety Bulletin: Unstable Approaches During Reduced Operations; it does not reference the Karachi accident in any manner yet, an unstable approach was a factor in the accident (was it not?) and thus, the bulletin would seem pertinent here.
https://www.ifalpa.org/media/3550/20...operations.pdf
https://www.ifalpa.org/media/3550/20...operations.pdf
from the last paragraph.
While this is hardly controversial, it might be worth considering the formal and informal norms in a given company.
The formal norms are a written policy, SOP, rules, etc. The informal norms are the unwritten rules, incentives, and 'reputation points added and subtracted' based on various decisions made.
How many of you have been in companies where the informal norms and the formal norms (policy) didn't quite match? (*There is a current thread on R&N in re Easy Jet that seems to cover this mismatch a bit regarding sickness...)
Can anyone comment from their own experience in this airline? (PIA)
It is important to highlight that the decision to execute a go-around is not, in any way, an indication of poor flight crew performance but rather prudent decision-making. There should be a clear non-punitive go-around policy.
The formal norms are a written policy, SOP, rules, etc. The informal norms are the unwritten rules, incentives, and 'reputation points added and subtracted' based on various decisions made.
How many of you have been in companies where the informal norms and the formal norms (policy) didn't quite match? (*There is a current thread on R&N in re Easy Jet that seems to cover this mismatch a bit regarding sickness...)
Can anyone comment from their own experience in this airline? (PIA)