Machaca

TopBunk -- Curious if the removal or discharge (or both) check valves are the subject of deeper scrutiny.

Green-dot

If the boostpump discharge check valve or the inlet valve are suspect:

- That would suggest that all 4 boost pumps in the LH/RH main tanks failed to deliver due to above suspect reason (as explained by TopBunk) at almost the same time?

- What happened to suction feed in that case? That should have provided engine fuel feed in such a scenario. Certainly sufficiently enough with engines normally required to accelerate to somewhere above approach idle during adjustments on final approach in landing configuration.

By the way, the inlet valve (removal check valve) is a valve that closes when you remove the pump. This permits you to remove the pump when the tanks contain fuel. Therefore, i don't see how this particular valve could be a factor unless clogged with ice.


Green-dot

MUNT

Not sure if this has been answered. The -400 is only EEC controlled, which, though a componenet of FADEC, is a little different on its own.

Chris Scott

Green Dot, you are right. It's a pity Top Bunk's information is not more specific: perhaps it does in fact refer to the HP Pump (independently on each engine, with an interval of a mere 8 seconds). Hmmm.

By the way, I think you meant "gravity feed" rather than "suction feed"? Wasn't it you and Swedish Steve who proved a long time back in this thread that we are talking about wing tanks? And the aircraft was at sea-level, with plenty of atmospheric pressure available.

Quote from pacplayer [yesterday, 1212z]:
You keep talking about what fadec is designed to do from the factory. That's not the context of the conversation I was attempting to delve into. A rollback is a well-documented software created anomaly, as I understand it, in which, for whatever reason the fadec has either miscalculated the thrust solution (thinks it's at high power and stays at idle) or has had a dual failure of both channels and is presumably trying to reboot itself at idle (since both channels are susceptible to the same software bug.)
[Unquote]

I would be very surprised if both channels are susceptible to the same bug. Although Boeing decided not to follow Airbus's lead, I believe, in trying to create and segregate two independent software design teams, they would have done their utmost to avoid this obvious trap.

Chris

Mechta

Some years ago, a particular aircraft type with which I was involved, was experiencing false fuel indications. I observed research into this, which involved a transparent vessel of fuel being vibrated at various frequencies to represent vibration levels and frequencies seen in-flight.

When the vibration level was varied, it was possible to create what can only be described as a 'black hole', into which bubbles dissolved air from the surrounding fuel would head. Slight variation of the vibration level could make the 'black hole' move up or down in the fuel.

The supposition for our investigation was that if such a black hole occurred in a fuel probe stillwell, this could lead to a loss of signal and hence the erroneous level reading.

As I recall, the vibration levels in this case pretty well matched those seen when the aircraft throttled back from climb to cruise RPM, which tied in with the in-flight occurrences.

What I'm wondering is if a particular vibration level and rate of change could have caused a similar 'black hole' to occur near the fuel pick up point.

The fact that this event occurred on the final part of the flight, after the fuel had been 'degassed' of air in the cruise, would suggest not, and it seems highly improbable that it would occur in both wings at the same time, but if the system is symmetrical, who knows?

Green-dot

To Chris Scott:

The OEM AMM officially refers to it as the Engine Fuel Suction Feed.

Swedish Steve, CONF iture and I had analysed that there was sufficient fuel on board in the main tanks (wing tanks) and that the center tank was empty. Determining approx. volume by the area of frost on the wing lower surface and comparing it with the fuel quantity reading on the flight deck of another B777-200ER. See posts #844 thru #878 for details.


Green-dot

pax2908

I apologize for yet another ignorant question (please delete if inappropriate). To get the fuel out of the tanks, air has to go in... What happens if that vent is obstructed? (Can it?)

Mechta

In response to Pax2908; if the tank vent is blocked, the air pressure in the tank will drop, and probably result in at least partial deformation of the structure, as wings are not designed to be a vacuum chambers!

This is what happened to an HS125 with a blocked vent (type 'blocked' in the find box): www.flightsafety.org/fsd/fsd_jul02.pdf

ChristiaanJ

pacplyer,
Your reply to Chris Scott makes me doubt your competence, if not your actual provenance and education....

Where I come from (admittedly some time ago), channel A and channel B software were maybe not actually written by two different companies, but definitely by two different teams, using different compilers, etc. etc.

"I submit that both channels use the same code."
If so, things have changed a long way from "my days", and maybe... "Houston, we have a problem"?
Duplicating a mistaek doesn't make it any less of a mistaek.

Duplex hardware monitoring works fine. The chances of any two op-amps, RAMs, PROMs, transistors, or whatever, failing identically, in the identical location in channel A and B are infinitesimal.
But if both channels run exactly the same crudware... when both of them execute exactly the same faulty instruction at the same time, no kind of monitoring is going to catch it.

I'm not suggesting this is really relevant to BA038. I DO think it's relevant to pacflyer's misinterpretation of software reliabilty.

CJ

DC2 slf

AAIB says the computers ordered the valves to open and the valves did open and were found open but the engines acted as if they didn't get fuel.

How is software relevant?

barit1

The "blocked vent" theory might be viable; there could be enough air entering the tanks to sustain the engines through TOD down to approach altitude, albeit at less-than-spec inlet pressure to the HP pumps.

But once the controls demanded accel in late approach, the blocked vents placed a real limitation on the amount of fuel delivered to the HP pumps. Ergo cavitation and failure to accel.

It might not be enough "vacuum" (i.e. tank pressure below ambient) to permanently deform the wing skins, but enough to starve the donks.

Chris Scott

Quote from Green Dot:
The OEM AMM officially refers to it as the Engine Fuel Suction Feed.
[Unquote]

I stand corrected: thanks. The point I was trying to make was that, in the event of failure of all the tank pumps we are not actually relying – as you know better than I do – on the ability of the HP Pump to suck fuel from the wing tanks to the engine: gravity will do the job. On all the aircraft I flew with pylon-mounted engines, this mode of operation was known as gravity feed, not suction feed. I wonder why the B777 uses a different term.

Thanks also for the references to the posts in which you established beyond reasonable doubt that the wing tanks had plenty of fuel remaining: I remember the discussion only too well. It's worth mentioning that, if the remaining fuel had all been in the centre tank and the centre tank pumps had failed, the only possible way of getting fuel to the engines would be by suction. This was, however, plainly not the case.


pacplyer,

Suggest you read my post again more carefully?

pacplyer

Chris,

I misread you post, my apologies. I have deleted it.

pacplyer

ChristiaanJ,

Good post, that's what I wanted to know.

Sorry all, I was not quite myself last night.

Cheers

gas path

It doesn't! the same term is used for the 747 'Engine fuel suction feed' AMM 28-22
Better make that two blocked vents then because there is a NACA duct and flame arrestor on each wing. Plus of course the relief valves!

gas path

What's that a drawing of Machaca It doesn't look much like a triple seven BP housing...Try this
Incidently the outlet check valves have a habit of shedding the their 'rubber' face seals and are being modded for valves without the seal. Awkward but can be changed without going into the tank... After emptying of course

Oilandgasman

SLF but 40 years operating pumps. Interesting comment on the discharge NRV.
When a Booster pump discharge NRV fails in service and the pump is stopped for operational reasons, what stops the flow of fuel from another booster pump in service from discharging its fuel back through the defective pump NRV and into the fuel tank? (Are there motorised discharge valves on each booster pump which close automatically when a booster pump stops?) In the oil busines this is know as pump cross circulation. It has happened many times and action similar to what you have mentioned taken to minimise the risk. In systems where a failure of this nature would be a life or process threatening event, two NRVs of different manufacture are mounted in series on the critical pump discharge.This cross circulation would also result in a very significant fall in the booster pump header discharge pressure.
Obviously if all booster pumps are running whilst A/C in flight, there would be no backflow and no prospect of fuel starvation to the HP pump from this NRV failure mode.

Green-dot

To be clear, if a scoop or flame arrestor becomes blocked, a pressure relief valve in the surge tank opens to make a vent.

There is a pressure relief valve in the inboard access door of each purge tank.

The pressure relief valve is normally closed. When it is closed, the valve is in line with the bottom of the wing. If a pressure difference opens the valve, it moves up as it opens. A spring holds the valve open until you close it. You have to pull a reset handle to move the valve back to the closed position.

When you do an inspection, you must look at the pressure relief valves to make sure they are closed. An open pressure relief valve is an indication of a blocked vent scoop or flame arrestor. The pressure relief valve can also open to relieve air or fuel pressure if there is too much pressure during refueling.

Sounds to me both the vent scoops/flame arrestors and pressure relief valve position are pre-flight/walk around items and should be snagged/corrected if any deviation from the normal was noticed.

Personally, while most issues focus on possible causes at subsystem level, i remain wondering if the cause to this accident should be viewed against the total scheme of things. Observing the aircraft as a whole with all systems interacting with each other. Why is no effort made to rebuild the original frame?

It is a unique opportunity with the aircraft relatively intact. Attach 2 serviceable engines (as far as the AAIB reports sofar engines are no suspect), re-attach the original gear, LRUs and wiring as much as feasible, put it on jacks and in the freezer, and do an air mode simulation with all systems operating and selected between TOD and moment of failure? Oh, and put the tail back on, the only reason i can figure why they took it off is either because it makes the frame sensitive to strong gusts or to remove the reference to the BA logo on it.

Who knows what such simulations may reveal?

Like i mentioned, just a personal thought.

Green-dot

gas path

Not BA's property anymore! As soon as the a/c was recovered the insurance co. payed out and the title changed hands. The fin was removed (by BA) at the request of the insurance co. so they could sell it. and also to prevent the hulk from 'weathercocking'.
Only the trouble is the fuel manifolding has been removed.

Green-dot

When they have tested it and find absolutely nothing out of the ordinary with it, they can put it all back in.


Green-dot