Indonesia Air Asia flight QZ 8501 loses contact with ATC

[Awesom Andy]

Apparently report finally released but details still a little sketchy at the moment. From reports I have read it looks like some sort of aircraft specific recurrent rudder problem with the pulling of circuit breakers and reversion to some sort of alternate flight law. Apparently conflicting side stick control inputs from both pilots resulting in a flat stalled spin into the ocean. No mention of loss of flight instruments, airspeed indicators or contributions of weather.

Rough translation as I understand it: There was a technical issue with the rudder, and the pilots pulled the circuit breaker for one of the computers. This also disabled the autopilot, and they lost control of the plane. I think I would put this down to pilot error.

 

[Quickstatus]

Sounds like a repeat of AF477.
The common denominator is inexperienced co-pilots operating highly complex aircraft but unable to be aviators when the computer fails. Current aircraft design permits the employ of low hour "flight computer operators".

I wonder if regulators can fix this by demanding pilots have a certain number of hours flying in non automated aircraft.

 

[jb747]

Sounds like a repeat of AF477.
The common denominator is inexperienced co-pilots operating highly complex aircraft but unable to be aviators when the computer fails. Current aircraft design permits the employ of low hour "flight computer operators".

Aircraft design has allowed that for at least the last 30 years. It's just that airlines have decided that it's cheaper than employing pilots.

I wonder if regulators can fix this by demanding pilots have a certain number of hours flying in non automated aircraft.

You need to work out what you're really after here. Automated? FBW? Neither are bad technologies. FBW is how all the new airliners fly, and it isn't going away.

A loss of automation, or reversion to a degraded FBW law shouldn't even be an item of news.

If some of the reports are correct, the aircraft dropped to either alternate law II or direct law, and when it did the rudder went to a slight (2º) offset. Inputs (to the controls) when the system reverts to a lower law are the norm, though I don't think the sims necessarily reflect them. The secondary effect of rudder is roll, so the aircraft then started rolling, apparently getting to 50º or so. I don't know where the pitch input came from, it certainly wasn't necessary. In any event, a perfectly flyable aircraft.....

 

[Awesom Andy]

AV Herald has a detailed description on what happened. Appears to be a loss of situation awareness, with no pilot input till 9 seconds after the auto pilot disengaged while the plane rolled at 6' per second, and continuous back stick input despite stall warnings. Might be an interesting conversation if these pilots ever run into the AF447 crew in their after life.

 

[BAM1748]

AV Herald has a detailed description on what happened. Appears to be a loss of situation awareness, with no pilot input till 9 seconds after the auto pilot disengaged while the plane rolled at 6' per second, and continuous back stick input despite stall warnings. Might be an interesting conversation if these pilots ever run into the AF447 crew in their after life.

Isn't the root cause multiple rudder failures on this aircraft and the fault not fixed.

 

[Awesom Andy]

Isn't the root cause multiple rudder failures on this aircraft and the fault not fixed.

It certainly contributed, and they wouldn't have to fly the aircraft manually if it wasn't for this issue. However, it's still a perfectly flyable aircraft with the issue and the autopilot off. Any pilot (from ATP down to students in a Cessna) should be able to keep the wings level and maintain altitude without an autopilot.

 

[BAM1748]

It certainly contributed, and they wouldn't have to fly the aircraft manually if it wasn't for this issue. However, it's still a perfectly flyable aircraft with the issue and the autopilot off. Any pilot (from ATP down to students in a Cessna) should be able to keep the wings level and maintain altitude without an autopilot.

I'm not saying it isn't, but this aircraft has a history of it. I wouldn't know if this is a regular thing on this type or any type and I would hope not, but suggest if it is and designers have a problem.

 

[jb747]

AV Herald has a detailed description on what happened. Appears to be a loss of situation awareness, with no pilot input till 9 seconds after the auto pilot disengaged while the plane rolled at 6' per second, and continuous back stick input despite stall warnings. Might be an interesting conversation if these pilots ever run into the AF447 crew in their after life.

"Loss of situational awareness"? He was in a window seat of an airliner....I wonder what he thought he was there for?

This is so close to AF447 that it almost makes no difference. The aircraft dropped out of normal law, and became, for want of a better term, a 737 (or 310, anything not FBW). And, non FBW aircraft don't make a particular habit of dropping out of the sky.

9 seconds before a control input, with the autopilot alarm merrily yelling its head off, is an eternity. This should have been a nothing, not an accident.

The rudder issue was trivia. It did not mean the rudder was doing something it shouldn't, but only that a limiting system wasn't working. The 2º of input after the law change is a normal effect.

The aircraft are not hard to fly in either alternate or direct law. In some ways, especially for ex Boeing pilots, they are easier.

The one issue that this does bring up yet again, is the lack of interconnection between the joysticks. Dual input should not be possible, and the lack of feedback across the coughpit is a weakness in every FBW Airbus. That's one reason why Boeing are still using a yoke in the 777 and 787.

 

[RooFlyer]

SMH report, which goes into the issue of training/FBW.

Perhaps the most disturbing aspect of the report involved the pilot training - or rather, the lack of it - by AirAsia Indonesia. After the autopilot was disengaged and the flight controls switched from the highly protective "normal law" to "alternate law", the aircraft rolled to the left by 54 degrees - beyond the 45 degree angle considered to exceed normal parameters, deeming it an "upset".

and

In the wake of the Air France crash, the Civil Aviation Safety Authority required that pilots receive upset recovery training, although it was long a standard practice at Australian airlines. The AirAsia Indonesia crash report said Indonesia's Director General of Civil Aviation has no such requirement and neither of the pilots were trained in upset recovery on an A320.

"The effectiveness of fly-by-wire architecture and the existence of control laws eliminate the need for upset recovery manoeuvres to be trained on protected Airbus aircraft," the Airbus A320 flight crew training manual said, in a statement that AirAsia Indonesia relied on but one that should raise questions within the aviation industry.

JB, would it really say it in such basic terms (given the words are in quotes) and does the A380 'manual' say anything similar?

 

[jb747]

SMH report, which goes into the issue of training/FBW.

When you convert to an Airbus, there is actually a need to convince pilots that they can pull right up to the stall, without things going awry. That's because you need to pull all the back stick you can get in some events, notably GPWS (ground proximity) warnings. Previously you might have pulled to stick shaker onset, but that's still a little way off maximum angle of attack. For those of us with a past (i.e. other brands), we store that away, but need to remember that it only applies in some cases. As soon as you end up in a degraded law, all of these fancy tricks are gone...but, it's still not substantially different to any non FBW aircraft. So, when working properly they are very safe, and even when degraded, they are no worse than the tens of thousands of airliners that came before them.

JB, would it really say it in such basic terms (given the words are in quotes) and does the A380 'manual' say anything similar?

I think hubris is a good word for this. Back in 1985 (or so) Airbus developed the A320 family. At the time, it was probably the most advanced aircraft in the world. But, it also had a number of flaws. In particular, the coughpit design. Instead of building upon all that had gone before, they moved off in a radical direction. The two most obvious variations are the adoption of sidesticks, and thrust levers that don't move for most of the time. AB engineers decided that they knew better than the accumulated wisdom, and that pilots would simply learn to love these changes. Fast forward 30 years, and both ideas remain as poorly thought out as they always were, but AB is locked into them. Like a religion, pilots who don't like their ideas are accused of being locked in the past (and all of the other phrases applied by believers of any cause).

As for the phrase you've mentioned...yes, their manuals did say that, but it was removed a number of years ago. Presumably after some pilots showed them the error of their ways. But, I always saw it as an advertising slogan, not really meant for the pilots, but rather for the airline management that would be able to seize upon it to lower standards (and so save money). The Airbus chief designer basically stated that his aircraft could be flown by a milkman...though he didn't say how well or for how long.

Whilst I'm critical of some aspects of Airbus design, it has to be remembered that it has functioned pretty well for the past 30 years. Yes, I'd change it, but only to the extent that I'd remove the traps and black holes. The biggest change needed is one of mindset, and I don't see that coming any time...soon or otherwise.

 

[RooFlyer]

Thanks again. My aversion to fast growing airlines is re-inforced by all of this.

 

[PaulST]

Thanks again. My aversion to fast growing airlines is re-inforced by all of this.

Exactly.
Please excuse the tangent, but I flew with EK last month (SYD-AKL in an A380) and the captain, who was an Aussie, sounded like a teenager.
I'm planning a few trips for next year and whilst I'm keen to try some non-QF airlines (which are considerably cheaper), events like these make me a bit nervous!


Thanks for the expert commentary JB!

 

[RooFlyer]

Exactly.
Please excuse the tangent, but I flew with EK last month (SYD-AKL in an A380) and the captain, who was an Aussie, sounded like a teenager.
I'm planning a few trips for next year and whilst I'm keen to try some non-QF airlines (which are considerably cheaper), events like these make me a bit nervous!


Thanks for the expert commentary JB!

Personally, I didn't have EK in that category but maybe I should! But your post made me think of the 'old saying' that: You Know You’re Getting Older When Policemen Look Young.

To me the coppers definitely look like teenagers (except for my local guy ... a grizzled ~60yo sergeant who I'd put my money on in any fight, armed or otherwise :mrgreen: )

 

[PaulST]

Personally, I didn't have EK in that category but maybe I should! But your post made me think of the 'old saying' that: You Know You’re Getting Older When Policemen Look Young.

And "you know you're getting older when your doctor looks like your children*". Fortunately for me, I tend to look older than I am (compliments of my fair complexion and premature balding) so I've only been told I look too young to be a doc once.

I'd like to think EK are fine but they're expanding incredibly aggressively. I'm flying again with them on Jan 1st so we'll see how we go.


Edit: * or your grandchildren!

 

[RooFlyer]

My GP looked after my mother before she died 10 yrs ago, so he's safe!

 

[jb747]

The common denominator is inexperienced co-pilots operating highly complex aircraft but unable to be aviators when the computer fails. Current aircraft design permits the employ of low hour "flight computer operators".

Let's do some maths on just how experienced, or otherwise, a 2,000 hour Airbus LCC FO is. Assuming he starts with 200 hours, with 100 hours in real aircraft, and then 100 in the sims.

If he's very lucky, he'll go one for one with the Captains (i.e. getting every second sector). There's no guarantee of that at all though, he he could get substantially less. If the aircraft is manually flown to 1,000' on departure, and from 3,000' on arrival, then he'd be flying for about 90 seconds at the start of the flight, and about 250 on arrival. But those numbers could easily be as low as 10 seconds, and 30 seconds respectively. Whilst they could be greater, that's not all that likely. Anyway, assuming 90/250, he gets 340 seconds of flying per sector.

If there's an average sector length of 1.5 hours, then that 2,000 hours is 1333 sectors, of which, in the best case, he's flown 666 (!). At 340 seconds each, that's 226,666 seconds. Which converts to the dizzying total of 63 hours. That's what I would see as best case...it could easily be a fraction of that, but it's unlikely to be much more. Of that, zero would have been in anything other than normal law.

That's one reason why I have been so anti cadets. Whilst cadet schemes will produce many decent pilots, they're actually the blokes who would be decent no matter how you trained them. There will also be many who simply never develop any skills. You certainly don't develop flying skills in an airliner...they have to be there already.

In a few years time, in the normal course of events, this FO would most likely have been promoted....so now you can most likely understand how Asiana happened.

 

[Quickstatus]

Let's do some maths on just how experienced, or otherwise, a 2,000 hour Airbus LCC FO is. Assuming he starts with 200 hours, with 100 hours in real aircraft, and then 100 in the sims.

If he's very lucky, he'll go one for one with the Captains (i.e. getting every second sector). There's no guarantee of that at all though, he he could get substantially less. If the aircraft is manually flown to 1,000' on departure, and from 3,000' on arrival, then he'd be flying for about 90 seconds at the start of the flight, and about 250 on arrival. But those numbers could easily be as low as 10 seconds, and 30 seconds respectively. Whilst they could be greater, that's not all that likely. Anyway, assuming 90/250, he gets 340 seconds of flying per sector.

If there's an average sector length of 1.5 hours, then that 2,000 hours is 1333 sectors, of which, in the best case, he's flown 666 (!). At 340 seconds each, that's 226,666 seconds. Which converts to the dizzying total of 63 hours. That's what I would see as best case...it could easily be a fraction of that, but it's unlikely to be much more. Of that, zero would have been in anything other than normal law.

That's one reason why I have been so anti cadets. Whilst cadet schemes will produce many decent pilots, they're actually the blokes who would be decent no matter how you trained them. There will also be many who simply never develop any skills. You certainly don't develop flying skills in an airliner...they have to be there already.

In a few years time, in the normal course of events, this FO would most likely have been promoted....so now you can most likely understand how Asiana happened.

So a pilots professional "flying hours" only equates to manual flying?. Are flying hours logged to the minute? second?

I decided not to fly LCC if possible when a colleague's son did a B Aviation. Had 200 hours in private instruction then did the BAv then got employed by LCC as a cadet. Rapidly became FO..... Still alive.....

 

[jb747]

So a pilots professional "flying hours" only equates to manual flying?. Are flying hours logged to the minute? second?

I decided not to fly LCC if possible when a colleague's son did a B Aviation. Had 200 hours in private instruction then did the BAv then got employed by LCC as a cadet. Rapidly became FO..... Still alive.....

Flying hours is recorded from the time the aircraft brakes are released, until the first door is opened. My point is that the vast majority of the time, the autopilot is doing the actual flying, and that's normally a good thing, in that it lets you get on with managing the operation. But, it isn't a way of improving actual flying skills....which is what is needed when the automatics decide to fail.

As an aside, the military only recorded wheels up to touchdown, but that's a way of maximising your flying budget, given that governments liked to cut hours.

 

[Quickstatus]

Flying hours is recorded from the time the aircraft brakes are released, until the first door is opened. My point is that the vast majority of the time, the autopilot is doing the actual flying, and that's normally a good thing, in that it lets you get on with managing the operation. But, it isn't a way of improving actual flying skills....which is what is needed when the automatics decide to fail.

As an aside, the military only recorded wheels up to touchdown, but that's a way of maximising your flying budget, given that governments liked to cut hours.

Ok I get it. You are saying this: the flying hours for pilots who operate highly automated aircraft in a commercial environment with single bottom line imperatives do not necessarily equate to actual flying. Some if not most of the flying hours are as a computer operator.

I wonder what the rate of manual flying hours accumulation needs to be to maintain a steady state of flying skills in case the flight computers gives a "blue screen of death"

I can see the rate of hours accumulation for some airline economic models will never increase flying skills.

 

[RailFlyer]

In particular, the coughpit design. Instead of building upon all that had gone before, they moved off in a radical direction. The two most obvious variations are the adoption of sidesticks, and ....

Out of interest, with two seater military jets which use joysticks, are they mechanically linked in all/most/few cases? What about the thrust levers? I could imagine this being much more confusing as with tandem seating you have no chance of seeing what the other pilot is doing to the controls?

Or does the problem more come down to the Airbus ability to "average the inputs" rather than just have 1 person in charge with a big toggle switch in centre of the dashboard to select which person is in control?