go around flaps15

Here here!

TIMA9X

Yep could not agree more, everyone walked away including a couple of (modest) hero pilots who did a damn good job in a matter of a few seconds!

S.F.L.Y

Very interesting remark about the AP and pilot workload. For sure the AP is here to assist us, provided we use it correctly. Good remark from Suitecaseman regarding using the AP to hold best glide speed. In this particular case the AP was not set to assist the crew in maintaining this best glide speed. The AP basically reduced and maintained the speed well below this best glide speed, thus shortening the gliding distance and causing high Vz impact. The AP which was supposed to assist the crew basically badly deteriorated the situation and shouldn't have been kept for so long on the ILS. The aircraft was trimmed for Vref, simply disconnecting the AP would keep it flying at a more appropriate speed and extend the gliding distance. Instead of that the aircraft was maintained on a wrong glide path at cost of precious speed. The AP not only lead to reduce the gliding distance, but also Vz control.

Captain Airclues

S.F.L.Y

You say that the actions that you propose would increase the glide distance. But what if you had done that, but not quite made it to the runway. The gear would have been damaged by the soft earth, possibly causing a fuel leak. However, because of your longer glide, the aircraft would be travelling much faster when it reached the concrete, with the possibility of sparks igniting the fuel.
The actions that the BA038 crew took resulted in everybody walking away. Can you be certain that your proposed actions would have achieved that?

Dave

Cloud Bunny

First off, Rainboe, couldn't agree more with the post.
Just to add something to this and I'm surprised none of my colleagues have got in on this (unless they have and I missed it) but we have had the benefit of some Boeing guys flying around with us for the last few months including some guys who test the 777s. We're also looking into this and other incidents in our SIM checks this 6 monthly period. What is being preached by Boeing to us is that in this scenario there are 2 main points:
1) Leave the flaps where they are.
2) Maintain VRef, even though this means you have to drop the nose and "accelerate" towards the ground. It gives you enough energy to flare the aeroplane enough to reduce the 'hit' on the ground.

Even applying this you aint gonna make the runway and as the guys from Seattle say you have to do the best with what you're left with. Which essentially sums up to me what the guys on board did that day and in doing so managed to get the aeroplane on the ground in a state so that everyone could get off, aided in no small part by the fact that Boeings are built like tanks!!

40&80

Good idea to run the whole approach through in the sim. and see if the aircraft would have done a better job doing an auto land with land flap and no pilot input at all.
If it would have crashed...the pilots did good...if it would have auto landed on the runway they did bad.
My money is firmly on the pilots and against the folk who think an aircraft does not need pilots and can be flown by an Indian call centre.

WojtekSz

the more i think of it the clear it is that:
(1) the crew made incredible job! chapeau bas! this situation must have been tried in all sims by whoever had a chance to try it - nobody so far has claimed successful landing. Two years after the accident! And the crew has done it in some 40 seconds
(2) from what i read and analyze the flying crew actions (a) at first tried to reach the airport and miss the obstacles on the way and only then (b) concentrated on just landing the plane. For me this explain why they first went for distance - aviate - and only later for speed to flair and land.

After reading comment from Cloud Bunny i did some very rough estimations and it appeared that (the numbers are for single wing of B777 with area of 215m2, excluding the fuselage drag):

• 3deg AoA the geometrical wing drag area would be equivalent to about 17m2 with Cd=0,05 => drag 0,85
• 8deg AoA the geometrical wing drag area would be equivalent to about 36m2 with Cd=0,07 => drag 2,54
• 12deg AoA the geometrical wing drag area would be equivalent to about 44m2 with Cd=0,08 => drag 3,44
• 14deg AoA the geometrical wing drag area would be equivalent to about 58m2 with Cd=0,09 => drag 5,22

estimated difference in geometrical drag area for 30deg flaps (over 25deg ) would be some 5m2 with Cd=0,3 => drag 1,5.
The overall wing drag estimation would be:
- flaps 30 + AoA 8deg : 1,5 + 2,54 = 4,05
- flaps 25 + AoA 12deg : 0,0 + 3,44 = 3,44
Result: total drag was reduced but this gain was not significant (about 10-15%) and was quickly lost due to constantly growing AoA.

so

SFLY had a point - confirmed by experience from Boeing sims - reducing flaps had much smaller effect than i have expected.

(the calculations are very rough and are for comparison only - do not build your next a/c with these numbers )

Mmmayday38

You are all mostly talking about theories, but I was dealing with gut feelings and I now know what 'flying by the seat of your pants' means. I had limited time during the 35 seconds that we knew we had a problem and there were a couple of theories that I recalled. However, although theories have 'their place', gut instinct and the view out of the window on the day also had it's place. I am not going to discuss or discredit the various theories that are being bandied around as they all are relevant in various situations.

I am struggling as to what I can write here. The final report will be released soon and I think you will all be relieved when it is out, I know I certainly will be. All of the speculation and assumptions will be laid to rest as the facts will be presented.

I know that S.F.L.Y is wanting answers and explanations and at this stage I am sorry that I feel I am unable to clear up his questions.

I can tell you that on the day a lot of things happened within a short period of time. At 480 ft it became apparent to me that we had double engine failure (in all intense and purpose, that's what it was to us), I chose to leave my First Officer handling the aircraft to allow me to attempt to diagnose and fix the problem. As a pilot you are taught that 'flying' the aircraft is only about 5% of commanding the aircraft! I checked the systems for obvious reasons that would cause the severe power loss and could not find a solution. I followed the acronym PPP - Plane, Path, People.

After I had checked all I could for the 'Plane' I moved onto 'Path'. In my opinion there was no way anyone would survive if we continued on the Path we were on. At the speed of 115kts I clearly felt from my visuals that the aircraft was not holding the glideslope and we were dipping under it and that I wanted to G/A but I knew we couldn't. The visual location of the impact from my seat at this point, with a vert speed of approx 1800fpm and a gusty headwind was well before the perimeter road - more akin to WojtekSz' post 2743.

After 12 years on the 777, I know how she feels and she felt very draggy, I had to reduce drag. I couldn't raise the MLG as it was obvious we were going to crash and we would need it, plus the drag would have increased as the main doors are lowered first, before the decrease. At this stage I felt certain there would be 100% fatalities. At 240ft I selected flap 25 and I immediately noticed the difference, I felt that fatalities had reduced to 50% from making this action.

Boeing have been testing this in their sim and I have seen their findings, for this reason I am bewildered by Cloud Bunny's post 2757 and what is being taught in the sim.

fotoguzzi

(SLF, here). Thank you for your posts! (I had been hesitant to respond until I gathered from the professionals that you were definitely who your nickname and bio indicate.)

I just wanted to say that I appreciate any pilot who continues to make the most of the situation right now, no matter what caused the problem in the first place.

Even the late Captain McBroom in the fuel-exhausted United Flight 173 DC-8 here in PDX continued to fly long past the time when it would have seemed to make a difference. As a result, he and almost all the passengers walked away.

So, thank you and your crewmates for keeping at it until the eventual happy resolution.

WojtekSz

Mmmayday38: thanks for this insight!

Wannabe Flyer

Mmmayday38

First of from an SLF thank you!

Secondly as I browse various scenarios and analogies of what is being said in various posts I would as a non technical person (humble SLF in awe)... ask this question.

Given your real life unique experience from which all walked away and god forbid it were to play out exactly the same, some day again in the future is there anything you would do different (as learned from this experience) which would have resulted in the outcome being even better than it already was?

korrol

We are very fortunate indeed to have Mmmayday38 contributing here - and it's very gutsy of him to actually spell out what happened and how he felt so frankly and in such detail. It's a real privilege to read his comments. Hats off to him.

We know that when the flap settings were changed by Mmmayday38 (at 240 feet) the autopilot was still engaged - so perhaps the move to Flaps 25 prompted the system to just tweak something somewhere to give the 777 just that tiny bit less drag to stretch the glide and gain those crucial extra yards.

Once the AP was disconnected though it was all down to human skill. ...But phew - it was right on the margin!

S.F.L.Y

Mmmayday38 thanks for your comments and I'm glad that you understand my point is not to blame anyone but to understand the mechanisms that lead the sequence of actions, wrong or not. We all are subjects to these mechanisms and understanding how and why we do things is very important.

I found this very interesting point in your comment:

This is definitely how tasks should be split, my next question is why the manual aircraft handling happened so late (or what processed contributed to delay the AP disconnection).

WojtekSz's analysis is very interesting, but it is not totally correct to compare flaps 25 and 30 with constant AoA. However, the interesting point is to see how much of drag is added by deviating from the best glide AoA.

Provided that his numbers are correct, we see that between 3 and 14 degrees of AoA the drag difference is around 4.37. Maintaining the best glide AoA would not only save 3 times more of drag than the simple flap retraction, but also produce more lift (thus significantly increase the gliding distance and Vz control).

Quality Time

Having done a few double engine failure on final scenarios in the sim,I think the reality is you are doing well if you get out a Mayday call and Brace Brace to the cabin.

Don't stall the aeroplane and keep energy for a flare.

I'm surprised at the Boeing recommendation to keep leave the flap alone too although I know that's the case.

If you are flying at Vref +5 the speed drops below Vref before the failure is even recognised in most cases. (B737-800 F30)

Better lucky than good and so much the better if you are both!

TIMA9X

Mmmayday38,
Just to say thank you for your post, I don't want to enter into the what ifs; flap settings, AP etc etc etc full stop.

On the day that this happened which I remember so well (living on the flightpath for ten years) knowing what 480 ft is when the problem occurred, and it was a 777 my thoughts recall the same words until this day, Those guys in the cockpit did a great job!
Sir, next time I have to SLF I hope you are in the left seat!

To quote again a PILOTS PILOT
Yep they sure did!

Mmmayday38

Quality Time; I am interested to know what Boeing are now recommending pilots/airlines to do if this scenario happened again please? If, for instance, on finals, with gear down and F30, final approach speed at 1500 ft and a double rollback occurred, or similar double engine failure; what are they saying to do? This would be twice the height of the BA38 incident. Are they saying the same for all wind conditions/runway clearways?

Wannabe Flyer; In answer to your question... We had the best outcome on the day, and I wouldn't change the decisions. I would, however, have appreciated more time; both to deal with communications, options and the fact that more time would have meant more height! I have seen the figures, and wouldn't want to end up anywhere other than the BA38s' impact point... no further forward, no further back! I'm not saying we briefed to put it down where we did, but, at the time it appeared to be the closest 'safe and unobstructed area' to aim for. With hindsight, my impact point appears to be the location that would result in the least number of fatalities.

bearfoil

Excellent idea. It would save fuel as well. Had 038's GS been even slightly higher, with a 'below' figured in, no need for thrust to make the field. It seems to me a low GS is a margin that doesn't allow for minor drift below it. It is a hard deck. It also led the AP to seek a higher AoA. A long 3 degree is asking for fuel/waste.

misd-agin

Exactly the point, at some point allowing the aircraft to slow to 108 kts probably resulted in worse performance than maintaining Vref Flaps 30.

Quoting Boeing's recommendations probably has to be clarified or expanded. I doubt they recommend maintaining Flaps 30 if you're at 2000' AGL. At some point there's a tradeoff. Again, that's experience, experimenting in the sim, and a smattering of luck if you'd ever have to make that decision quickly.

Everyone's familiar with the drag increase at greater than Flaps 20. Getting to mid slats is key.

You can verify stall information for your a/c's stall speeds. First notch of leading edge slats typically is 50-60% of the total stall speed reduction. Additional flaps have a lot of drag vs. amount of stall speed reduction.

RatherBeFlying

As you increase flaps from zero, the L/D ratio decreases; i.e. you fly a steeper approach or you need more power to maintain the glideslope.

When there's no more power available and you're coming up short, you need to increase L/D.

Decreasing flaps does that, but you are also increasing stall speed; so, in a low energy situation with very few seconds to weigh alternatives, not decreasing flaps further once a survivable impact point has been achieved sure looks like the best way to minimise the risk of something else going wrong.

Mmmayday38 had to negotiate a tight squeeze between landing too short and increasing the stall speed too much; he looked out the window and did exactly what was needed

Dont Hang Up

These type of incidents are self evidently extremely rare. Therefore any change in aircraft or airfield operations to mitigate the risk must have pretty much zero downside in terms of safety.