Asiana flight crash at San Francisco

[ross_M]

I respect the trying to keep up some camaraderie as to not convict these pilots too early, but there is a limit to that, please.

+1

For those who want to stop discussing this till the official report is out: Fine. But if any system changes are going to result, the sooner the better.

The media, policy-makers and general flying public have fairly short memories. Once this disaster fades into the background no one will want to do anything about it. Those who want status quo, and unexamined dirty details need exactly this haze of time.

So if there is got to be any lobbying for change it better be soon. Personally, I'll fully agree that a full conclusion about the causes of this disaster, we don't have yet. But I'm a Bayesian, and based on past data and whatever details have been confirmed so far it is fairly clear what the likely causative factors are here. Not with 100% certainty of course, but that doesn't mean we ought to ignore the obvious and pretend we have no clue as to what went wrong.

To sum, whatever flavor be your reccomendations for change, the sooner they are lobbied for, the greater the chance that they will stick. The more you wait, the greater the likelihood this will be yet another disaster only remaining in the memory of technical forums like this one.

[edmundronald]

Causative factors of Asiana loss of life

These are the holes of the cheese that lined up:

1. Disabled landing aids at SFO requiring visual approach from tired long haul crew.

2. PF unfamiliarity with type, and PIC unfamiliarity with training role, not a standard crew distribution, leading to loss of situational awareness of airspeed and CFIT.

3. Passenger ejection during hard landing as result of CFIT.

4. Ground crew driving over passengers.


I really think that (1.) deserves looking at, because otherwise it is just a question of waiting for the next incident, as aircrew training and fatigue will clearly not improve substantially in the next few years. As for (4) ...

[Indarra]

edmundronald

While early reports said passengers were ejected from rear, NTSB later stated this wasn't the case.

[glofish]

The typical 'slam-dunk' approach into SFO can be a challenge at the end of an ULH flight

It was not a slam-dunk approach, it was a very usual "visual" with the distance to height at a normal rate. If for the respective pilot this seems too high and close, say it and go around.
If either call is a challenge, then you belong to the children of the magenta, or there is definitely a problem with CRM/culture.

A visual in SFO should be a treat and not a threat.

You should not continue flying such equipment if you deem that a potential threat, as simple as that.
A T7 is ETOPS qualified and i don't know if you know, but most ULH sectors are ETOPS today and ETOPS is defined by "able to descend to resp. level, with resp. speed, proceed to the ETOPS Alt, 15 min hold at 1500ft, shoot an instrument appr. and GA, then land with a visual circuit". So there is a regulatory need to be able to do a visual at a most probably unknown airfield with a low ceiling!
I expect every ULH/heavy pilot to be able to do that, daily experience makes me doubt that, however. I have repeatedly pointed out the lack of skill of todays pilots. Now that some proof materialises, i hate the watering down of the obvious.

Disabled landing aids at SFO requiring visual approach from tired long haul crew.

This is not a hole in the Swiss cheese! They were augmented, thus fatigue is a small contributing factor. And spare us that "a visual is a hole lining up" bs, please.

[RHKAAF]

Asiana accident

I was based in Hongkong during the 80's and 90's when KaiTak was considered one of the more difficult airports to carry out a visual approach. Often , including during the Typhoon season, the ILS or IGS would be switched off for "maintenance" Then it could not be switched on again until it had been flightchecked--sometimes for weeks on end. Why cannot a backup ILS be on hand to cover outages?. The Military even have portable systems to call on. Are GPS approaches now permitted ?

When on visual approaches, the only need to look inside at the instruments is to check AIRSPEED. All other parameters are covered by eye ,especially at a field with VASI's, VGPI's or any other glideslope indicaters.

[Lonewolf_50]

oldbob @

It was a clear, a no-wind day and the airplane was doing everything that it
should have.
The target that was the end of the runway was in sight, if the target moved
up you add power. If it moves down, you retard power.
You concentrate on where you want to land and act accordingly...but then,
that is if you are a pilot.

That reflex/pattern/habit is apparently being trained out of long haul line pilots via neglect. That's the problem a lot of pilots have mentioned in the course of this discussion.

See, for example, the comments by the following gentleman:
HPSOV_L

The system is setting pilots up for failure. Does the industry care?
Fox3 suggests not, when it comes to cost benefit analysis.

How likely is it that the public will be receptive to the point Fox 3 makes?

I don't know.

Basic flying skills proficiency seems to have been a root cause of AF 447. A pilot could not fly straight and level in cruise while non flying pilot trouble shot a systems malfunction. Basic pilot skills, not in evidence.

Two captains flew an approach that got low and slow at the back end of the power curve. Flew an approach that was not stable. Dropped 30 kts below VREF without making a decisive correction. Basic piloting skills not in evidence. Go around not initiated in a timely fashion.

Lion Air: unstable approach, landed short, thank God nobody died on that crash . (And that one had Wx as an element. ) Go around not initiated in a timely fashion.

Are these isolated incidents, or are they symptoms of an industry with a disease?

"The pilots are to blame" is an easier sound byte to swallow. It also means less effort on getting systemic fixes put into place.

How costly has it been to implement and embed in the culture CRM? Bloody expensive, but it was done (thank God) in order to make flying better and safer. (I am

What is it going to take to resolve the issues like A330 flown by Air Afrique into the ground?

How many more body bags?

[BuzzBox]

It was not a slam-dunk approach, it was a very usual "visual" with the distance to height at a normal rate.

I don't know for sure, but previous posts on this thread would seem to refute your assertion that the distance to height was 'normal'. In my experience, long-haul flights coming in from the Pacific are typically held high (10,000-12,000ft over PYE), followed by a steep descent over the Bay area to join right base for 28L/R. It's certainly not impossible, but it can be a challenge to get down, configured and stabilised on the final approach.

A visual in SFO should be a treat and not a threat.

Yes, it should, but unfortunately that's not the reality for most long-haul pilots. The typical long-haul pilot does not get many handling sectors these days; captains might get four a month, while some FOs are lucky to get one. The simple fact is that most long haul-pilots are not as sharp as their short-haul colleagues, particularly when they are operating at the back of their body clock.

By the way, I was a 777 skipper, I have more than a few thousand hours on the aircraft and have operated into SFO on many occasions. I am well aware of the ETOPS requirements and I must say that your take on things is simplistic in the extreme.

[misd-agin]

All the drama about SFO being tricky. Please tell me what's so tricky about it?

At the end of a long haul segment? That happens hundreds or thousands of times a day across the world.

Transition pilot getting IOE? Happens every day, hundreds or thousands of times daily.

New TRI/LTCA/CKA? Happens monthly, and maybe weekly, at major airlines.

At the end of the descent is an approach. In this case a visual approach and an FMC could be used for basic(distance to touchdown...300' per mile calculation) or more advanced guidance (RNAV, VNAV, etc). And the approach ands at pavement.

Let's not make SFO into something it hasn't. The hardest part of flying into SFO is trying to take the entire view in, it's that beautiful.

[Fox3WheresMyBanana]

Especially if the experienced hands reckon an SFO approach is hard, shouldn't the decision to go around have been made a lot earlier from such an obviously unstable approach? Judging from the radar plot of the glideslope flown, they hit the g/s about twice as close as I ever did flying run-&-breaks in the RAF.

[BOAC]

Buzzbox - I do not agree with your analysis of Glofish's post - what seems to be filtering out here, and hinted at by you, is that given the understandable lack of 'handling' by long-haul crews an ETOPS diversion on near min fuel plus a subsequent g/a, all flown under a stressful situation, carries a high probability of an accident in the ensuing visual circuit and that it seems to be agreed that at the end of a long flight, a visual is not the desired approach as it is really not safe to expect a 'tired' crew to execute one. So be it, if true.

Surely then the regulations need to change to require

a) Extra ETOPS div fuel to allow a second coupled approach at the alternate

b) A mandatory coupled approach on arrival at dest (or div if required to achieve) on a long-haul sector as defined by over xxxx hours or by circadian factors

If that is what we need, that is what we need. Let's not just pussy-foot around talking about the bunnies being 'tired' and having body clock problems and 'slam-dunking' around the sky - adjust the ops to cope?

[flyingchanges]

They were 34 knots below Vref, with stick shaker active. They were definitely on the back side of the power curve and no amount of thrust would have been enough to get them out of this situation without increasing airspeed at this altitude.

At landing weight, TOGA power would have flown them cleanly out of this.

[ross_M]

These are the holes of the cheese that lined up:

1. Disabled landing aids at SFO requiring visual approach from tired long haul crew.

2. PF unfamiliarity with type, and PIC unfamiliarity with training role, not a standard crew distribution, leading to loss of situational awareness of airspeed and CFIT.

3. Passenger ejection during hard landing as result of CFIT.

4. Ground crew driving over passengers.

Edmund

Nice list!

I'll add some, perhaps less important, points:

5. General sense of unease in Flight Crew about visual ops. (Speculative, but should be tested. Perhaps by throwing visual approaches at this or other Asiana crew in the Sim or check flights. )

6. Failure of relief crew in drawing attention to the impending disaster forcibly enough or failure of flight crew to take notice of relief crew's warning calls. Bad CRM?

7. Failure of the other relief crew member to remain present on flight deck during landing. (Who can say if that one extra set of eyes might have caused an earlier go around)

8. PF mixing manual and automated landing modes without realizing the precise nature of control logic.

9. Boeing allowing combinations in control logic that are potentially dangerous or easily misunderstood or not widely understood.

[palm47]

one cannot say how many, simply do not fly commercially and so perhaps, understandably, shout "pilot error" as a reflex. It could turn out to be a far more complex situation than we realise. I've said this before but watch Captain Sullenberger's interview for an eloquent rebuttal of why it is wrong to start screamimg "pilot error". In fact he calls the very term outdated and talks about human factors.

We have arrived at a certain point of automation and "intelligent a/c" where I (as a passenger) will not longer accept "pilot error" as the cause of catastrophic events (and we can be lucky there were "only" 3 fatalities in this case).

Numerous pilots here have pointed out how fundamental errors where made in the simple observation of speed and altitude. And I as a passanger do not want to depend on captains, who may have an hangover, long haul, being from the wrong non-CRM "culture" or simply a bad day and being in the danger of making similar mistakes.

This event has tragically shown that a whole crew can make these kind of mistakes. The question is to the solution.
It could be a better pilot training/qualifications. But sorry, I as a passenger do not know who is flying the a/c. Is it the experienced one or are they basically training newbies in the cockpit and have additional divertion on top of all other problems of human factors?

The a/c had a vector into the ground (short of the runway) which could not be recovered when the stick-shacker went of. It was too late already (and it only warned of the approaching stall).
Yes, the pilots did let that happen. But the plane did, too. And I say the plane should have prevented that or at least warned the crew. The pilots here can say all they want about how simple it is to view airspeed and alt and act accordingly. The fact is: Some pilots didn't.

[PaperTiger]

Originally Posted by kaokao

However I could see that engine were dumb* for 20 secs.

Where's this coming from? If it was in one of the NTSB briefings then I missed it. If the Asiana pilots have spoken publicly then I missed that too.

Simple revisionism/PR methinks.

* in reponse to power inputs that is

[BuzzBox]

BOAC:

an ETOPS diversion on near min fuel plus a subsequent g/a, all flown under a stressful situation, carries a high probability of an accident in the ensuing visual circuit

I have no doubt that such a scenario would be extremely stressful and carry with it a high degree of risk. I understand that the ETOPS minimum fuel reserves are based on such a scenario, but I have to ask what are the chances of it happening in the real world? Would changing the ETOPS requirements actually do anything to improve safety outcomes and prevent the type of accident that occurred in SFO? I doubt it.

[SMT Member]

Lifted from aviation week dot com:

NTSB Chairman Deborah Hersman said July 10 that during the final 2.5 min. of flight, the flight data recorder shows transition to “multiple autopilot modes and multiple autothrottle modes.”

Uncertainty hangs over whether the mode changes were pilot induced, or commanded by the auto-flight system. In any case, does point to a crew busying themselves trying to make the autoflight system do as they want when, perhaps, they should have gone 'click-click', 'click-click' and simply hand flown the aircraft to a safe arrival. In this context a lack of currency induced by long-haul flights, restrictive SOPs and an institutional opposition to manual flight should play a major role in the investigation.

Perhaps this accident, along with the more unfortunate ones involving Continental and Air France, may finally motivate the regulators to impose more hands-on practice, both during initial and recurrent training, as well as an examination regime ensuring they boys and girls up front demonstrates their ability to pole the damn thing around the skies.

[BOAC]

Originally Posted by Buzzbox

Would changing the ETOPS requirements actually do anything to improve safety outcomes and prevent the type of accident that occurred in SFO? I doubt it.

- I am not following your question. Item b) would in theory HELP to "prevent the type of accident that occurred in SFO", not a)

a) would hopefully HELP to prevent an SFO Asiana event at an ETOPS div.

[Machaca]

Reuters reports:

Asiana Airlines said on Monday that it would enhance training for pilots seeking to convert to flying new aircraft, part of a series of measures it is undertaking after the fatal crash of an Boeing 777 plane in San Francisco.

"We will bolster our training program...when a pilot converts to a jet from a different manufacturer," Asiana said in a presentation to the government, which was shown to reporters.

The new measures will include enhancing training for visual approach and automated flight and conducting flight inspection on airports which are "vulnerable to safety."

The pilots aboard the Boeing 777 made a visual approach to San Francisco airport, and relied on automatic equipment to maintain airspeed.

Asiana said it will also seek to improve communications skills among crew members, and introduce a system to manage "fatigue risk."

Other measures include setting up separate maintenance teams for Boeing and Airbus planes, and enhancing safety management systems.

[ssites]

"With the greatest respect, I disagree. The typical 'slam-dunk' approach into SFO can be a challenge at the end of an ULH flight, even with a functioning ILS. Take away the ILS, put an 'inexperienced' crew in the seat and it's not hard to see how a break down in situational awareness could occur. Throw in some cultural factors on top of that and the rest is history."

And an inexperienced crew should not be flying 300 people around

[A Squared]

Originally Posted by BuzzBox

With the greatest respect, I disagree. The typical 'slam-dunk' approach into SFO can be a challenge at the end of an ULH flight,

And

Originally Posted by BuzzBox

I don't know for sure, but previous posts on this thread would seem to refute your assertion that the distance to height was 'normal'. In my experience, long-haul flights coming in from the Pacific are typically held high (10,000-12,000ft over PYE), followed by a steep descent over the Bay area to join right base for 28L/R. It's certainly not impossible, but it can be a challenge to get down, configured and stabilised on the final approach.

This is an interesting phenomenon in this thread. It was established very early on, within a day or so of the accident that there was nothing "slam dunk" about this approach. That is very clear from the flight track data available on Flight Aware, almost immediately and has been confirmed repeatedly by all actual sources of real data released.

Once again: Asiana flight 214 was aligned on final by 14 NM from the runway threshold at 4300 ft.

This has information been posted repeatedly in this thread, yet still people insist that ATC gave them a "slam dunk" and somehow, without a scrap of supporting information, this becomes "truth"