I'm guessing, but I imagine the last fuel tanks to be emptied would be the fuselage tanks to keep the plane balanced. There is a tank in the centre wing box and on the 200LR versions 3 additional tanks ot the rear of the cargo compartment, under the passenger cabin.
Got this from Fortitude Foirums: Fuel is burned from the center tank first until the LOW FUEL CENTER EICAS illuminates at which time the center tank fuel pumps are shut off. On the B model, this happens when roughly 2000# of fuel remains. On the A model, approximately 3-500#. Then as fuel burns from the wing tanks down to a certain level (30,000#?) a scavenge pump brings some more fuel out of the center tank, but not to the point of being completely empty (remember TWA 800?). The idea here is to keep the center tank fuel pumps submerged
Pilot Peter Burkill has just given a press conference, here is what he had to say
AirSpace - more than just hot air
Have had a look at the video and latest pics, it is still not poss to say if ram air turbine was deployed or not. So unable to say if there was total power loss.
AAIB Initial report:
Following an uneventful flight from Beijing, China, the aircraft was
established on an ILS approach to Runway 27L at London Heathrow.
Initially the approach progressed normally, with the Autopilot and
Autothrottle engaged, until the aircraft was at a height of
approximately 600 ft and 2 miles from touch down. The aircraft then
descended rapidly and struck the ground, some 1,000 ft short of the
paved runway surface, just inside the airfield boundary fence. The
aircraft stopped on the very beginning of the paved surface of Runway
27L. During the short ground roll the right main landing gear separated
from the wing and the left main landing gear was pushed up through the
wing root. A significant amount of fuel leaked from the aircraft but
there was no fire. An emergency evacuation via the slides was
supervised by the cabin crew and all occupants left the aircraft, some
receiving minor injuries.
The AAIB was notified of the accident within a few minutes and a
team of Inspectors including engineers, pilots and a flight recorder
specialist deployed to Heathrow. In accordance with the established
international arrangements the National Transportation Safety Board
(NTSB) of the USA, representing the State of Design and Manufacture of
the aircraft, was informed of the event. The NTSB appointed an
Accredited Representative to lead a team from the USA made up of
investigators from the NTSB, the FAA and Boeing. A Boeing investigator
already in the UK joined the investigation on the evening of the event,
the remainder of the team arrived in the UK on Friday 18th January.
Rolls-Royce, the engine manufacturer is also supporting the
investigation, an investigator having joined the AAIB team.
Activity at the accident scene was coordinated with the Airport
Fire and Rescue Service, the Police, the British Airports Authority and
British Airways to ensure the recovery of all relevant evidence, to
facilitate the removal of the aircraft and the reinstatement of airport
The flight crew were interviewed on the evening of the event by an
AAIB Operations Inspector and the Flight Data Recorder (FDR), Cockpit
Voice Recorder (CVR) and Quick Access Recorder (QAR) were removed for
replay. The CVR and FDR have been successfully downloaded at the AAIB
laboratories at Farnborough and both records cover the critical final
stages of the flight. The QAR was downloaded with the assistance of
British Airways and the equipment manufacturer. All of the downloaded
information is now the subject of detailed analysis.
Examination of the aircraft systems and engines is ongoing.
Initial indications from the interviews and Flight Recorder analyses
show the flight and approach to have progressed normally until the
aircraft was established on late finals for Runway 27L. At
approximately 600 ft and 2 miles from touch down, the Autothrottle
demanded an increase in thrust from the two engines but the engines did
not respond. Following further demands for increased thrust from the
Autothrottle, and subsequently the flight crew moving the throttle
levers, the engines similarly failed to respond. The aircraft speed
reduced and the aircraft descended onto the grass short of the paved
The investigation is now focussed on more detailed analysis of the
Flight Recorder information, collecting further recorded information
from various system modules and examining the range of aircraft systems
that could influence engine operation.
Here's Captain Burkill's full statement courtesy of the BBC.
HHoffman:There were reports of very late deployment of landing gear - but why would that be if the fault happened so late?
Where did you here about the landing gear being deployed late?
Perhaps they hold it in, avoiding to slow down and keep the plane above the ground as long as possible? But this wouldn't make much sense, because it should have been deployed alot earlier than the power problem occured...
It's said they were troubled 2mins before touch down, so anyone can calculate if they should have deployed their landing gear long before...
Failure Pre Event
There are are only two options.
1. Technical systems failure.
2. Human error. (any human)
Prior to resolution and without factual analysis, the probability is always heavily weighted in favour of human error. This pre resolution probability is adjusted in this case by the following factors: -
Th nature of the failure is highly expected by human systems and technical systems designers. However the probability factors reduce each year at a greater rate for technical systems than for human systems. This is because technical systems can be made deterministic and human systems cannot. So overall a new Boeing aircraft with Trent engines, with UK ATC at a Heathrow runway final approach is most unlikely to have failed for unavoidable technical reasons. The probability of the failure (pre resolution, pre analysis) being for unavoidable technical reasons is undetermined but is unlikely to exceed one chance in ten.
Failure Post Event
The effect on failure probability of the facts as known. These facts have already been filtered so they may be
misleading or just wrong or even deliberately misleading (unlikely in this scenario).
The scenario seems straightforward final approach, non responding to auto throttle> insufficient thrust> insufficient lift > crash.
Now we are into a degree of speculation, not being a combined air accident investigator, 777 avionics, Trent systems designer, BA Pilot, Heathrow controller and failure mode risk analyser. The one probability that seems uncontravertable is that this is a very unusual (unique?) event with therefore rare causal factors.
The elements of this rare event are: -
The aircraft went through a change that took it from 100% reliability to 100% unreverable failure in under 30 seconds. It is possible this was random but almost certainly it was because of one of the above changes owing to landing. The possibility of random failure at the end of a history of unbroken reliability and in this case, many hours of a long flight, just 30 seconds before landing, is improbable.
Combining these changes to identify the cause is the next step. Easy with the flight recorder and engineers. Difficult without!
Complete Speculation only (Very Mickey Mouse so don't take seriously!)
It requires a major failure of systems design in a core scenario, hidden for 12 years for this failure to overturn the likelihood that this was human error. This is a one in a billion failure scenario. So what fails catastrophically one in a billion times? Something having failed (say 100), what are the subjective risk weightings of this 100.
Dual trent engines do not (themselves) both fail. = 1 (99 something else)
777's fuel system most unusual = 2 (97)
777's ILS most unusual = 4 (93)
777's other systems most unusual = 2 (91)
ATC (no effect known) = 1 (90)
Ground systems and avionics (little effect) = 3 (87)
Sabotage landing change related = 5 (82)
Cabin Crew or passengers inc deliberate = 5 (87)
Ground interference inc deliberate = 2 (85)
Pilots not noting warnings = 5 (80)
Pilots not following procedure = 5 (75)
Pilots overriding ILS = 25 (50)
Piliots engine error eg closing throttles = 30 (20)
Pilots fuel error eg not checking or shutting off wrong tanks = 5 (15)
Other = 15
The fuel systems do not allow you to run out or run the same fuel to both engines at the same instant from the same tank or bipassing the filters without warning in time to do something. Forget it. Only thing that can stop the fuel running those Trent's at the same instant is a person (or a catastrophic failure that would be obvious, like hitting a building).
The avionics could technically fail through each level of redundancy between the autothrottles and the engines but I assume the ILS has control of the throttles to balance the lift and airspeed, so why and how was the autothrottle under suspicion. Is the throttle control manual at this stage - I have now idea. This area seems the core of the problem but I defer to the experts. It seems the core of the interation between the computerized aircraft systems and the aircrew. I would always look at the interfaces for misunderstandings. The pilots used the manual throttles after, I assume this is not going to make any ILS computer systems happy if they are not disengaged or have hung but this is tested to oblivion so can't beleive it is a cause.
My rationale says that there was either a failure through all levels of redundancy of the avionics engine control or the pilots missed an avionics warning and took manual actions that misled themand had no effect. The humans had two levels of redundancy and two levels of failure correlation (ie two pilots) and no monitoring whereas the technical systems operate at many levels of redundancy with some correlation but much monitoring. No contest - an interface problem causing the pilots to interfere with the throttle systems is on the cards. But as I said - I know nothing.
This is so unusual that I only think a human could do it. But then I am biased
- by the probabilities.
More complete speculation.
It was not :-
Avionics prime and secondary control failure of each engine demand is possible but unlikely.
Either something unique happened that beat all the risk analysis at Rolls and Boeing or something unusual happened that gave warnings which were ignored/misinterpreted and resultant action was taken by mistake.
Interesting facts about the Malaysian 777 (1 August 2005) incident:- (ref. 1)1. one of several accelerometers had failed at the time of the [incident] occurrence2. another accelerometer had failed in June 2001.
This combined with the reports of a s/w fix (ref .2):-British
Airways said yesterday it had carried out an emergency directive to correct
that [Malaysian 777] “software” fault.lead me to ask the following questions:-
1. Why does a faulty accelerometer not get noticed immediately? 2. Could 1 faulty accelerometer explain the Heathrow incident?3. Could 2 faulty accelerometers explain the Heathrow incident?4. Was the software fix carried out on the Heathrow 777?5. If the software fix had been carried out on the Heathrow 777 then, as is not uncommon with software fixes:-a. did it not fix the problem and/orb. did it cause another problemQuestions, questions, questions... Let's see what the ADIRU data says...
ref. 1http://www.aaib.dft.gov.uk/cms_resources/AP-BGL%201-06.pdfref. 2 http://www.flightglobal.com/AirSpace/forums/incident-at-lhr-6435.aspx?PageIndex=2
My ref.2 was wrong. Cannot remember which source reported the software fix. But the TimesOnline mentions it and also the previous accelerometer issue.Here's the link<url>http://www.timesonline.co.uk/tol/news/uk/article3216746.ece</url>
The possible IEPR alarm failure was reported in the Gaurdian athttp://www.guardian.co.uk/transport/Story/0,,2243357,00.htmlSo one engine possibly failed (matches the state of the engines as pictured).What state was the other engine in? Pictures show it was running on impact. Why then did it not provide any apparent lift?
This one looks like we have a scenario where there was a major hardware failure and/or faulty sensor (or sensors) which was (or were) not handled correctly by the OPS.I can hear the s/w test team now "but that wasn't a requirement which we had to test for". I prey not.
Moe:Where did you here about the landing gear being deployed late?
Latest images courtesy of the Daily Mail's site.
Interesting to note that the landing lights are on - therefore there's obviously power, and also no sign of the ram air turbine extended.