Home' Asian Aviation : AAV May 2010 Contents Safety
post-accident tests showed that ice-restricted fuel ow
may limit achieved engine thrust, there is no requirement
to record the position of fuel-metering valves (FMVs)
that control fuel delivery.
Such information (obtained during the investigation)
was "invaluable" in showing that engine controls had
functioned correctly and that fuel restrictions led to
engine rollback. e AAIB recommends that EASA
and the FAA require, where practicable, that operational
positions of engine fuel-metering devices be recorded
on digital flight-data recorders (FDFRs). It also
recommends that regulators research ice accumulation
and subsequent release mechanisms in fuel systems
and review requirements to ensure systems tolerate
formation and sudden release of ice.
As the 777 descended through 720 feet above ground
level on approach to London Heathrow, the right engine
stopped responding to autothrottle commands as fuel
ow became restricted at the engine's fuel/oil heat
exchanger (FOHE). Power fell to 1.03 engine pressure
ratio (EPR) and, seven seconds later, le -engine power
reduced to 1.02 EPR. Airspeed dropped and the aircra
touched down on grass 330m short of the paved runway,
fuel restrictions having continued until ground impact.
Investigators concluded that the le engine su ered
similar restricted fuel ow, but limited available data
made it impossible to eliminate a possible restriction
elsewhere. Testing and data "mining" suggested this was
very unlikely, as was a separate restriction within seven
AAIB says that ice formed (from water occurring
naturally in the fuel) during the long ight, which had
operated with low fuel ows and in low temperatures.
Accreted ice was released from the fuel system upstream
of the FOHE, restricting engine fuel ow at the face of
Testing confirmed the FOHEs' susceptibility to
restriction by highly concentrated so ice at a fuel
temperature below --10 degrees Celsius and with fuel
ow above ight idle. Some 10,500kg of fuel remained
when the engines rolled back, and investigators con rm
it had not cooled enough to su er "wa xing".
AAIB says the crew knew that they would y through
unusually cold air and had monitored fuel temperature,
which was always within limits. e pilots conducted
cruise "step climbs" in the autopilot vertical-speed (VS)
mode which involves lower than full-climb power
settings and signi cantly lower engine fuel ows than
if the vertical navigation (VNAV) mode had been used.
The minimum fuel temperature of --34 degrees
Celsius was never cool enough to require crew action,
but did contribute to the --22 degrees Celsius low fuel-
temperature during approach. When both engines
rolled back less than a minute before touchdown, the
crew faced a "highly unusual" situation, for which no
speci c training existed, say investigators.
When the aircra descended too low, the autopilot
tried to maintain the glide slope by increasing aircra
pitch, which further reduced airspeed, generated
an initial "airspeed low" warning , and (eventually)
triggered the stall-warning stick-shaker. At this stage,
the co -pilot pushed the control-column forward, which
disconnected the autopilot and made a short landing
inevitable. To lower drag, the commander reduced ap
by ve degrees, which had little e ect but moved the
2.9g-impact point nearer the runway.
e AAIB says an evacuation checklist dividing pilots'
independent tasks could result in " re" handles being
operated before fuel-control switches, while damage to
wiring between the handles and engine-fuel spar shut-o
valves could leave the valves open -- as happened in this
accident, leading to release of 6,750kg of fuel.
There was no evidence of excessive or unusual
amounts of water in the fuel or fuel tanks. Investigators
say that if large quantities of free water had been in the
main fuel tanks at the start of the accident ight, then
the fuel temperature would have caused it to freeze
before those tanks started to supply the engines.
Post-accident contamination of the centre tank with
re- ghting medium made it impossible to establish
how much water was in the centre tank at the end of
the ight, but there is no evidence that either engine
consumed large quantities of water during the ight.
Extensive tests showed that water injected into the boost
pump inlet -- leading to a high water concentration in
the fuel -- could form su cient ice to partially block the
FOHE and restrict the fuel ow to the engine HP fuel
Tests also demonstrated that, with normal
concentrations of dissolved and entrained water in
jet fuel, ice can form inside fuel-feed pipes. In certain
environmental conditions, ice can accumulate when
the fuel temperature is between +5 and --20 degrees
At approximately --5 to --20 degrees Celsius, ice will
adhere to its surroundings. It was shown that increasing
fuel ow can release su cient ice to clog up the FOHE.
As the aircra descended, an increased temperature in
the engine strut also could cause ice to be released from
local fuel pipes.
Tests by engine-manufacturer Rolls-Royce showed
that to achieve the accident- ight engine-control system
response to the reduced fuel ow, the restriction had to
be at the face of the FOHE or LP pump just before,
or during, the nal acceleration of the engines during
Testing also showed that an FOHE face restriction
could not have been from a gradual accretion of ice,
but more likely from an 'avalanche' or 'snowball' of
ice. AAIB says that a combination of other factors
-- such as turbulence, aircra -pitch changes, and
an increase in the strut temperature -- could have
contributed to a sudden release of so ice in the fuel-
e accident ight had cruised for more than eight
hours at an average fuel ow of about 7,000lb per
hour, during which time, fuel temperatures remained
below --20 degrees Celsius. Use of VS mode for the
step climbs meant fuel ows did not exceed about
8,900lb per hour. Investigators say tests showed that
ice may be released from fuel-feed pipes at higher ow
rates, similar to those achieved during nal approach,
which reached almost 12,300lb per hour.
e AAIB says the accident ight was unique
among 35,000 Rolls-Royce-powered ights in its
combination of the lowest cruise fuel ow combined
with the highest approach fuel ow at the lowest
approach temperature. Just two out of 142,000
Pratt & Whitney-powered flights had similar
Ice inside the 777's fuel pipes "probably" began
to form on the feed lines passing through the main
tank, while the centre tank supplied engine fuel, say
investigators. Although centre-tank fuel would have
been above the freezing temperature of water, the feed
pipes passed through cold main-tank fuel as the fuel
temperature fell from --2 to --21 degrees Celsius.
Tests revealed no other mechanism that would
have restricted ow elsewhere in the fuel system and
triggered the subsequent engine-control response on
the accident ight, concludes the AAIB. ●
Testing showed that the formation of ice in jet fuel depends on a complex combination of circumstances.
46 AsianAviation | MAY 2010
Links Archive AAV April 2010 AAV June 2010 Navigation Previous Page Next Page