Yemenia A-310 accident 2009 report

[sitigeltfel]

As the original thread on this was closed, this may be of interest (French language)......

La Provence article

Link to simulation from black box info

[Mr Good Cat]

I can't speak French, but looking at the video clip am I right in assuming that ultimately the retraction of the flaps to UP with a simultaneous reduction of power by the autothrust resulted in a sink - which was then countered by increasing the pitch by pulling back the stick and putting the aircraft on the edge of the stall?

I'm not an Airbus guy but I noticed the selected speed was 210 knots on the FMA with autothrust engaged, yet the autothrust reduces the power when flaps UP is selected, despite the actual speed being 180-ish.

Is this the normal reaction of the Airbus autothrust when in a particular mode? What would cause this power reduction when selected speed is set at 210?

Of course the whole approach should have been thrown away long before this stage but I'm just confused by the automatics' reaction.

< Start the clip from 13:30 before the selection of flaps to UP >

Thanks.

EDIT: I think I get it - the autothrust had been disengaged earlier perhaps?

[CONF iture]

Now, where is the simulation ... for AF447 ?

[Clandestino]

I'm not an Airbus guy but I noticed the selected speed was 210 knots on the FMA with autothrust engaged, yet the autothrust reduces the power when flaps UP is selected, despite the actual speed being 180-ish.

There was no autothrust installed in this A310. As it's usual for early Airbi, it had autothrottle.

Reduction in thrust is not necessarily connected with flaps up selection.

EDIT: I think I get it - the autothrust had been disengaged earlier perhaps?

Nope, autothrottle was engaged until last 50 seconds. Reason why it went into RETARD was SPD F/D vertical mode and zero feet set in altitude selector.

Even without report, video is pretty interesting: night circling with two GPWS warning (sink rate -pull up and too low terrain) which proceeded anyway. Lethally wrong FMA selections make autothrottle cut the power, which goes unnoticed and when something is perceived to be wrong, grossly excessive pull up ensues with pitch reaching 30°. If only report were available two years ago, it would get mentioned in AF447 report along Colgan and Birgenair as another example of pull-up-no-matter-what.

[tubby linton]

I am an airbus pilot of a very similar type. Thrust was reduced because a pilot selected 0000 in the fcu and the athr tried to achieve this. There were numerous selections made on the fcu from when the aircraft became established on the ILS from GS LOC to V/S H/Sel and back to GS LOC. There appeared to be confusion as to how they were going to manage the automatics and fly the circle procedure evidenced by the amount of changes in altitude selected in the FCU. There also appears to be a lot of effort into trying to make an approach off the ILS and avoid the circle manoeuvre. I think that they became totally confused and descended very low on the downwind leg and then failed to manage the recovery to controlled safe flight.
What was the circling minima for this approach?

[M.Mouse]

Circling approaches are one of those approaches that are not normally made very often. They are not inherently difficult but a thorough briefing of how the approach would be flown with regard to which autoflight modes will be used when to take gear and flap, what speeds and timings for the downwind leg and so on is so vital in order that both pilots know what is going to happen and when.

It was all a mess but descending to circa 400' while not even halfway down the downwind leg shows clearly that neither pilot had a clear picture in their mind. The simulation was just staggering to watch.

[PLovett]

Yet another example of a crew that should have been nowhere near the sharp end of an aircraft. It was just a disaster from the moment the circling approach was commenced. No idea what they wanted to do, no idea what they were actually doing and no idea what to do when it turned pear-shaped.

[AtomKraft]

That was not pretty.

[b747heavy]

A total loss of awareness, that is apparent.

At no time during what appears to be a botched go around attempt, were the go levers selected. At no time did the FMA indicate THR, GO AROUND as the flaps were retracted to zero and the slats retracted, the S and Green Dot speeds looked far away on the speed tape. The gear came up with transitioning flaps causing the horn to sound, and that does sometimes happen. With the ALT selected to Zero and what was either V/S or SPD in the pitch mode, they were headed for the water!

Not nice to watch, but did they really have a clue? Not sure what their SOPs might have been, but I would fathom a guess that it all went wrong the second they selected V/S after LOC interception.

Another thing, whilst descending on the LOC, 1000' was selected for the circle. I do believe that on reaching 1000', the ALT was selected to the GA ALT (3000) perhaps a bit too early, at which point the vertical mode reverted to V/S with a descent, thereby fixing that in the window, which they seemed to follow until the AP disconnect?

Video was a bit hazy to see the FMA changes clearly on my small screen, but that's my guess?

Corrections welcome...

[jcjeant]

Now, where is the simulation ... for AF447 ?

I was thinking the same myself .. therefore, why the BEA has not produced such good video simulation ?

[tubby linton]

b747heavy. -The aircraft was in ALT hold at 1000ft and then 3000ft was selected and I think the alt hold pb may havethen been deselected. V/S would have engaged at the v/s at the time of deselection which was about 500ft/min.

[ahramin]

Slats retracted 15 knots before S speed.

[No Fly Zone]

I too read the extensive update/final report over at AVH. (Thanks to whoever made the rough translation.) While there a many contributing factors, it reads to me like two mature pilots with plenty of combined time-on-type, both lost situational awareness and flew themselves into the water.
I know that the several pilot's unions will scream like hell, but this event is one more example in strong favor of mandating CVR machines that record and retain the entire flight and that can be erased only after a normal/safe landing. In far too many cases, the critical events last more than 30". Investigators (and others) are simply missing too much data with the 30" limit. Pilots and their unions already enjoy protection from action based on CVR tapes (or digital recordings) and recordings from normal flights never see the light of day. Post accident investigators and the design engineers need this valuable information in their quest to prevent as many future events as possible.
No, I do not hate pilots. I respect them and I trust them. Still, in the interests of safety, we need far better records about what goes on in the cockpit and the '...last 30" is just not enough." My two cent's worth.

[CONF iture]

Most of cvr record the last 2 hours now.

Originally Posted by NFZ

Post accident investigators and the design engineers need this valuable information in their quest to prevent as many future events as possible.

And the pilot in question or his family needs that information as well to make sure that part of the story won't be conveniently forgotten ...
Where is the video animation in the AF447 case ?
Why the all data are not available to the families ?

[jcjeant]

On a side note ....
Since 2008, Airbus provides Yemenia technical experts and "controllers to train staff (pilots and engineers) and verify its performance in two specific areas: maintenance, on the one hand, engineering and operation of aircraft, to other. ". It was one of the conditions for the Yemenia is not on the European blacklist.
In April 2004, ICAO denounced the failure of Yemen to organize air transport.
In 2007, the DGAC banned flights over French territory by the 70 Airbus A310-ADJ following a SAFA inspection
In April 2008, Europe found that Yemenia did not meet certain safety standards established by the Chicago Convention
After the accident on 29 June 2009, EASA suspended the maintenance organization approval granted Yemenia and DGCA suspended the CDN 2 A310 registered in France and used by Yemenia
Comoros
No reliable administration of the Comoros to effectively monitor operators allowed to land on its territory. The ICAO audit performed in this country in March 2008 described its supervisory duties of "not implemented"
So .. some other elements than the pilots training were in the game for let this accident occur

[lomapaseo]

Trying not to be judgmental.

From earlier posts they seemed to have been confused and punched a lot of wrong buttons.


But once that aircraft started its video dance could they not have salvaged it by turning off the stuff and just flying the aircraft? I got the impression that they always were behind what the aircraft thought it was supposed to do.

[Tee Emm]

From earlier posts they seemed to have been confused and punched a lot of wrong buttons.

One of the things that is very rarely seen in a syllabus of simulator training is practice at seamless sudden transition from full automatic flying to manual flying without automatic features such as flight director and autothrottle.

Even simulator instructors themselves may be incapable of demonstrating this. Few simulator instructors demonstrate anything that requires skilful hand flying. They can talk the talk but few can walk the walk. It is natural for pilots to become apprehensive of their ability to instantly transition from automatics to manual operation if most of their career is on automatics. Then, when that rare occasion occurs requiring instant change of control to get out of trouble, particularly in IMC, they are in danger of losing situational awareness with the usual fatal results.

[worrab]

Some (most?) will know of this video which predates this accident by a dozen years or so and is horribly prescient.

http://www.pprune.org/tech-log/51492...ml#post7868343

[Herod]

Am I wrong here, or were they VMC the whole time? Circling minima was 1000' and the metar gives a cloudbase of 2000'. Watching the simulation it seems that right up until about -50s the whole thing could have been recovered simply by "flying the aeroplane".

[M.Mouse]

But once that aircraft started its video dance could they not have salvaged it by turning off the stuff and just flying the aircraft? I got the impression that they always were behind what the aircraft thought it was supposed to do.

Quite.

The automatics, when used correctly, when conducting a circling approach are a tremendous aid to flying an accurate and slightly more complicated than normal approach.

Planned, briefed and flown correctly using the autoflight system, whether Boeing or Airbus, would have enabled both pilots to have monitored a very accurate approach right up until late downwind where the aircraft would have been all but in its final approach configuration, at the correct height and speed when the autopilot would have been disconnected prior to a continuous turn onto final approach more or less on the correct approach profile.

The fatal flight brings into serious question the amount and quality of the training that two experienced pilots had received when they so completely mishandle a fully serviceable aeroplane to the point where they lose control and crash.