Dead stick approach
Well...with the greatest respect, and after having read the entire thread this very morning, I feel that 'most pilots' were not so sure that a great job had been achieved from 750' there is more than a 'very limited amount of time' available.
Although the aircraft 'cannot stall' the landing was so hard that an undercarriage leg went up through the wing. Stalled or not that aircraft virtualy fell the last few feet.
Of course, I wasn't there and it is easy to speculate from the lhs of my sofa these days but allow me to offer a sample of my mornings 'cherry picking' and again simply express the hope that the eventual enquiry will publish the results of a pilot perfomance evaluation.
Happy reading...
I would expect that the natural reaction would be to take manual control, knowing that the autopilot would attempt to maintain the glideslope to the detriment of the airspeed. The way I read this report (and I may well have got the wrong end of the stick of course) is that the aircraft must have stalled (read 'descended rapidly') after the autopilot 'flew' the aircraft to 175' and 108kts and disconnected itself.
--
As far as the kudos for the "wonderful" job the Autopilot did...
If I were handflying this approach, and lost thrust, would I sacrifice altitude for airspeed ( to prevent a stall ), damm fckng right I would.
And YES, with a loss of thrust, I KNOW I could FLY it to the ground a WHOLE lot better than the A/P. This is airmanship 101 guys. The only question I have is at what point do I realize I had no thrust and then act.
--
It seems apparant that the A/P remained engaged down to 175 feet. Although the A/P is a pilot`s best friend in most emergency situations, perhaps it was not the case in this instance.
The speed at 750 feet must have been about 140 Kts.
The speed at 200 feet was 108 Kts.
As the the A/P was engaged, the AFDS was trying to maintain the glideslope, with a probable linear speed decay. The speed loss was approximately 32 kts in 500 feet.
Every kt of speed below VRef 30 would result in a worse Lift/Drag ratio, with the result that the aeroplane could not eek out as much distance as it potentially could if it was flying at it`s best L/D ratio speed (approx VRef). At 108 Kts, the L/D ratio would be significantly reduced.
---
The AAIB is explicit in saying that flap 30 was selected. No mention is made of decreasing the flap selection.
The usual jet airliner L/D of 18 is for the clean configuration and I suspect that flap 30 yields substantially less. A simulator run would present the raw F30 L/D, but in any case we did have some thrust and the achieved slope and L/D is derivable from the FDR.
There are a number of flapped gliders that use flaps to add drag. The big caution with flapped gliders is not to add too much flap as you can lose considerable altitude reducing flap. Many glider approach accidents have happened when flap was reduced with insufficient altitude.
I would not want to explore in the air how the 777 with thrust restricted to the accident setting reacts to a flap reduction attempt at 600', but it's possible the AAIB will decide to investigate that in a simulator.
In a high drag situation, increasing airspeed increases drag substantially. With flap 30, the best L/D speed would be lower than with lesser flap selections.
With gliders that have powerful spoilers and/or flaps, you can select full spoilers and/or flaps full on and if you are still not coming down steeply enough, add airspeed to steepen the slope.
In this accident, we see a trade of airspeed for glideslope, mostly at the behest of the autopilot. Given the touchdown point and the 108 kt. cited by the AAIB, they were amazingly lucky.
It was an interesting decision by the crew to leave the autopilot engaged as the speed decayed.Perhaps there was a good reason for this. Perhaps they were understandably so preoccupied with trying to work out why there was no response from the thrust levers, that the speed decay went unnoticed.However, the reason why the autopilot allowed the speed to decay was not because it was attempting to fly at the best speed for the situation presented to the crew. It was because it was attempting to maintain a glideslope that it was commanded to follow. Unfortunately, a decaying speed from about 140 knots at 750 feet to 108 knots at 200 feet, resulted in a severely degraded flight path angle. If the autopilot/authorottle is not performing what it is commanded to do, (in this case maintaining the commanded speed), then it is best to disconnect, and correct the situation manually.In this instance, as the speed started to decay, an autopilot disconnect followed by flying at a speed of between VREF and VREF minus 10 would, quite probably, have resulted in a different outcome, and a more controlled landing.
This is true...to an extent. However, in my company the following is beaten into us with a metaphorical big stick (and rightly so) at every recurrent: Aviate, Navigate, Communicate.
Someone has to be flying the aircraft at all times, either manually or through the automatics. With a loss of thrust at that height, it would not do to have both crew trying to troubleshoot the problem. One of them HAD to be flying the aircraft. With that in mind, personally I find it suprising that the PF would choose, seeing that the AP was trying to maintain the glideslope thereby rising the nose causing the speed to bleed off, to leave the autopilot in to the point that it disconnected itself at 108kts.