I would like to try and take a step back and explore why current sumping procedures, even when correctly applied, might be inadequate and allow significant water (whether free or entrained/dissolved) to accumulate in the 200-ER model centre tank in certain narrowly defined cold, but within certificate, conditions. By significant I mean capable of causing a restriction in a fuel line, not that it would prevent the fuel burning in the engine.
It seems important since the accumulation of significant water in the centre tank may be a contributory factor in creating the unusual restrictions in each fuel line of BA038 which triggered a double rollback.
1.
Centre tank accumulation of free water
There seems to be a recognition unconnected with BA038 that significant water can build up in the centre tank and restrict fuel lines:
- the Boeing-issued 777-FTD-28-07002 (as revised 20 Feb 2008) at
JetPhotos.Net Forums - The Friendly Way to Fly - View Single Post - BA 777 landing accident at LHR (and for a more summary version not mentioning 200-ER see
http://www.pprune.org/3958000-post608.html) mentions that the 200-ER centre tank is reported to have accumulated water which has frozen and restricted the flow of fuel along at least the jet pump driven fuel scavenge lines.
- In the absence of any information on the water content of the fuel in the BA038 centre tank due to the damage on landing, this FTD seems to offer at least some strong circumstantial evidence of the accumulation in similar centre tanks of water and ice on long, cold flights and that this ice is capable of causing restrictions in fuel lines - this seems worth mentioning in the context of the BA038 analysis. In particular, I note from the FTD that:
while the freezing water is causing restrictions in part due to cold temperatures in the centre tank, the freezing appears also at times to be due to the centre tank water being subjected to the even colder main tank fuel temperature.
the centre tank is prone to the build-up of water sufficient, if not properly dispersed within the fuel, to cause enough icing in the fuel scavenge lines to restrict them in part or completely.
free water built up is only dispersed by the operation of the water scavenge and when this is not operating the free water sinks to the bottom and is drawn into the fuel scavenge lines.
also without mixing by the water scavenge pumps, entrained and dissolved water in the centre tank fuel is likely to separate out and end up in the OJ or fuel scavenge pick-ups.
the pipe which takes the centre tank fuel to the outlet in the main tank on the outboard side of the wing is extended a little further outboard in the 300ER and that this extra distance in itself, due it seems to the cooling effect of the cold-soaked main tank fuel, may contribute to the accretion of ice in the fuel scavenge line sufficient to block it.
corrective action in the centre tank on the 300-ER is being taken to move the water scavenge inlet closer to the fuel scavenge inlet to minimise ingestion of free water in the fuel scavenge line.
200-ER centre tanks (ie like in G-YYYM) are not covered by the FTD but ... the FTD notes that One operator advised of this problem [of restrictions due to ice in fuel scavenge line blocking scavenge to main tanks] occurring to a lesser extent on the 777-200ER model as well.
the underlying cause of free water in the centre tank is not addressed, nor is the fact that the water scavenge does not operate to disperse water after the OJ pumps are turned off.
there is a contributory factor here common to BA038 in that the FTD states that: A review of operator QAR data to date shows that this problem primarily occurs on long distance flights from Europe, suggesting a potential operations component to the problem, but long distance flights from JFK have not been affected. Further review shows the problem to primarily occur between the months of October - April, suggesting an ambient ground temperature component to the problem.. Is this a coincidence?
- There is no indication of the highest levels of free water which operators have found in affected 777 centre tanks (eg: 1, 10, 100 litres?) does anyone know?
- Water will accumulate in the centre tank:
where (a) there is free water or dissolved/entrained water in uplifted fuel; or (b) water condenses in centre tank during flight or ground turnaround, AND in each case, if it is not removed from the centre tank because it has separated out of the fuel and frozen or hidden itself in pockets inaccessible to the OJ pump or fuel scavenge or sump drain valves, then in the next sector one starts off with a base amount of water in the centre tank derived from (a) and (b).
The next uplift of fuel may melt and/or disperse it (by entraining or dissolving it) but not necessarily or permanently.
The process will repeat itself until the environmental conditions cease to apply or some other event intervenes as it normally would to cause the water to be consumed by the engines or drained.
- the volume of water contained in an uplift of properly specified fuel for a long sector is no more than a few litres so, if significant volumes of water are to be explained in the centre tank, it would seem condensation would need to be responsible. However, condensation from air drawn into the centre tank to replace fuel drawn out would, particularly at high altitude, have minimal water content (much less than a litre!). To explain higher levels of water from condensation there would seem to need to be a stream of outside air coming into the centre tank, preferably at low altitude. Question whether sufficient circulation might be provided by either:
the continued operation of the two jet pump fuel scavenge lines after the fuel is scavenged in the last hours of flight at the end of each sector; and/or
the descent at the end of each sector with an empty centre tank into humid air at ground level?
2.
200-ER centre tank - layout
- There are various baffles across centre tank and other dead spots in which free water may accumulate. The floor of the centre tank becomes more curved in flight and this may also hinder ensuring free water's dispersal in the fuel at certain times of flight. The layout of baffles in the central section of the centre tank can be seen in figures 6, 11 and 12 in the AAIB report produced in respect of the omission of G-YMMEs purge door at
http://www.aaib.dft.gov.uk/cms_resou...ection%201.pdf but there are also significant ribs in the colder inner wing section of the centre tank.
- There are only two sump drain valves in centre tank would anyone be able to point out (on picture of centre tank mentioned above) exactly where the two sump drain valves are located on lhs and rhs, relative to low points, baffles, OJ pump inlets and fuel scavenge inlets?
3.
Weaknesses in detection of water
It seems difficult for operators to monitor easily whether free water or dissolved/entrained water is accumulating in the centre tank and thus they may allow a contributory factor to a double rollback to develop. Issues include:
- Sumping would only have a chance to remove free water if: the free water has had a chance to settle or (if the water has frozen) the ice had had a chance to melt and then settle and in each case such free water is not in an inaccessible blackspot as a result of the baffles or compromised location of the sump drain valves. Anecdotal evidence suggests that one or more of turnaround times, OATs and centre tank design (among other things) may effectively hinder these conditions occurring in long cold-soaked back-to-back winter sectors.
- Sumping will not work properly in cold conditions and is therefore potentially an unreliable procedure if required to ensure safe operation.
- Are Karl Fischer (or similar) tests to be routinely conducted several hours after uplift of warm fuel into cold-soaked centre tank at cold stopover as part of the routine Boeing/operator recommended sumping procedures on daily/transit basis I wonder? If not, then melted and dispersed ice in warm fuel uplift is arguably unlikely to be detected anyway and even then may require pumps to be run for some period before testing (though perhaps running pumps would not help sumping of free water)?
- If the water scavenge pumps operate successfully after a warm uplift to dissolve or entrain melted ice into the fuel in manner, then water sensors, which can only indicate if free water has melted and settled to the bottom of the tank, also seem imperfect as an early warning of significant levels of entrained/dissolved water. It is not even clear that excess free water MMs which are triggered are seen as significant in any event, or have immediate pre-flight or in-flight remedial actions/procedures associated with them? In particular:
Excess water maintenance messages are not drawn to the attention of the pilot (unlike say an EICAS advisory or caution say about low pump pressure). The excess water maintenance message seems to self-clear even if correctly triggered.
It is only designed to be triggered in centre tank if more than 627 litres (138 gallons) or so of free water is settled at bottom of particular part (any idea which?) of centre tank. This is insignificant amount in context of burning 80 tons of fuel but even a part of such amount might represent more than enough water, suitably frozen, to create a significant restriction in each of the fuel manifolds.
Perhaps, if free water in the centre tank is seen to be a contributory factor of the double rollback of BA038, more prominence needs to be considered for this message?
- perhaps in winter after each couple of long very cold sectors out of LHR test all of the centre tank contents by draining the centre tank in a warm hangar and then examining and measuring whatever else was inaccessible in the centre tank by opening it up. Also perhaps to repeat process but on final sector to disable fuel scavenge completely (though this might reduce centre tank condensation (if any) too) in order to enable examination of contents of 800kg of remaining fuel for any build-up of free, dissolved or entrained water. The key seems to be to test the correct categories of flight legs and not simply to rely upon a sump test.
4.
Centre tank scavenge
It may be worth considering, for more risky types of long, cold flights:
- scavenging the centre tank fuel into one main tank only in order to break the centre tank fuel being a common cause of a failure condition and to preserve the independence of the main tank fuel supplies to each engine? Downside would be an imbalance between the main tanks.
- not running the fuel scavenge pump when the centre tank is empty, by removing the motive power to the fuel scavenge after 30 or 40 minutes of operation (or no later than centre tank indicating empty)? Otherwise is there a risk that continued operation of the fuel scavenge while the centre tank is empty:
could draw into the centre tank towards the latter stages of the flight a significant stream of (potentially humid) cold outside air from the vents (causing unnecessary condensation build-up)?
could result in a significant volume of air being entrained into the main tank fuel (causing issues by its likely release at high altitude elsewhere in the fuel supply system)?
- requiring a water scavenge at all times while the centre tank has fuel in it, rather than just when the OJ pumps are working, in order to minimise the settlement and accumulation of free water in 800kg of fuel remaining after the OJ pumps are turned off?
5.
Centre tank OJ/jettison check valves
Is there any possibility of undiluted centre tank water being injected directly into the fuel manifolds without passing through the main tanks first?
- The AAIBs comment in a different context about the possibility of the OJ/jettison valves opening to admit air from the centre tank into each fuel manifold if pressure difference was sufficient presents an intriguing prospect if it might also allow the possibility of dregs or slugs of free water (maybe a bucket or two) being swept into the cold-soaked main tank fuel in each fuel manifold towards the end of the flight.
- Pressure differences as a plane rapidly descends are to be expected the fuel manifold might just be at least on a localised basis be at lower pressure temporarily, perhaps as a result of having pockets of undissolved air in it (if the scavenge pumps operation while dry could provide a mechanism temporarily to introduce air into otherwise unsaturated fuel) which might allow the centre tank pressure to overcome each OJ/jettison check valve.
- However even then would water be able to pass through each switched-off OJ pump with the same ease as air say?
- Clearly more than one contributory factor is required for undiluted centre tank water to end up in each fuel manifold at the end of the flight, but it might offer an alternative explanation as to how significant water might unusually move into each fuel manifold, near simultaneously, towards end of flight, sufficient to cause a restriction to each engine and without increasing the concentration of water in the main tanks.
- Should there be a cut-off valve (in addition to a pressure dependent check valve) which could be closed after OJ pumps turned off to ensure that this could not happen? Is there a safety reason why the operation of the OJ/jettison check valve must be allowed after OJ pumps turned off?