Sunfish
 

At the point when the right
engine began to lose thrust the data indicated that the
right engine EEC responded correctly to a reduction
in fuel flow to the right engine, followed by a similar
response from the left EEC when fuel flow to the left
engine diminished.


it's in the report.

Throttle-up. Perceived discrepancy in FOB
 

Quote from autoflight:
Is anyone claiming that the crew opened the thrust levers to max and kept them there?
[Unquote]

Quote from AAIB Bulletin S1-2008 (Feb18):
The engines failed to respond to further demands for increased thrust from both auto-throttles, and subsequent movement of the thrust levers fully forward by the flight crew.
[Unquote]

Does that go any way towards answering your question?


CONF iture, Thanks for reminding me about the "250 kts 12D before LAM". I also think they are unlikely to have entered the hold at much above 230 kts, decelerating to their ideal holding IAS. I'm guessing this might have been about 210 - 220 kts, as they were light.

So I think it is likely that a significant amount of thrust would have been used in the hold (but see my previous post, above) to average only 400 ft/min ROD. By the way, a typical time taken from exit of the LAM hold to touchdown is 10 - 15 mins, depending on traffic. Even taking the quickest case, they are still likely to have used thrust (above idle) again while flying at a steady 160 kts, normally required by ATC for the segment between (roughly) 7 miles and 4 miles from the runway.

But, in the absence of the FDR/QAR data, this is all speculation...

Quote from CONF iture (discussing/quoting AAIB Bulletin):
“the total fuel on board was indicated 10500kg” and even if “both of the eng spar valves were found to be open, allowing the fuel leak evident at the accident site” they avoided to mention how much fuel they drained from the main tanks.
[Unquote]

You seem to be suggesting that the AAIB's economy in words may have been designed to cover some doubt they had about the indicated fuel quantity (FOB) on landing. I think it is likely that the indicated FOB the AAIB stated would have represented the lower of the following 2 figures:
1) total FOB in real time as measured by the FQI sysyem;
2) departure FOB (as entered into the FMS by the crew at Beijing) minus fuel used (as measured by the engine flowmeters); known as the CALCULATED FOB.

Method (2) does not, of course, take into account any leaks; nor (probably) any APU fuel consumption. The latter is not normally used after engine start; the former self-evident post-flight.

The stated FOB on landing being 3600 kgs higher than the F/P destination fuel could be accounted for by a number of (presently unknown) factors in combination. A ZFW lower than planned (as you are thinking?) would have saved 300-400 kgs of fuel per tonne on a 10-hour flight. Again, as you imply, we do not know how much flight time was saved (if any, due to better wind-components and/or direct routings). Contingency fuel is - for the purpose of calculating F/P FOB at destination - assumed NOT to be burned, so is probably irrelevant.

Finally, you will have noticed that TOD was at FL400, which is probably higher than planned, saving fuel for the last part of the cruise with their light payload. Also, the F/P included a temporary descent from FL341 to FL318 early on, which would have been costly in fuel; they avoided this at the request of ATC. The latter is the only known source of fuel saving for us at this stage.

In summary, there is already much known scope for fuel saving, and the unknowns may have even more to offer. I see no reason to infer (as I assume you are) that the AAIB may harbour doubts about the validity of the indicated FOB they published.

B777fly

The aircraft is refuelled through separate refuelling valves. There are 2 in each tank and they are not powered in flight.

Jetdoc,

Thank you for the further information. How do the refuel valves function physically and where are they in the main (wing) tank? I recall that they are powered open during refuelling. If they are closed and depowered in flight, presumably they are locked in position? They must also have a NRV in the out-of-tank direction only?

With such an inexplicable accident it is quite likely that two rare and non-related system failures have combined to produce the engine power loss, so bear with me if I try to eliminate all possible ways in which air might reach the tank pump inlets ( given that there must have been 2 or more system failures for that to happen)

Sit back and relax 777 - there is a quite professional body, far better resourced than you, trying to achieve just that.

has that been definitely established?

The refuel valves are solenoid controlled from the refuel panel and they require fuel pressure to push them open. That fuel pressure is supplied by the pumper truck or when the aircraft is on the ground and you want to transfer fuel around, you can use boost pumps in the other tanks to move the fuel. Of course then, you need to open the defuel valve.

The question is not well formulated.

The transmitting power would be dependent on how far away the transmitter is (inverse-square law and all that) and whether the transmitter is broadcast or directed-energy.

One needs to know how much energy is required to move the relay. This is not dependent merely on the voltage rating of the part, but on its specific design and manufacture and that of the system of which it is part. Then you have to guess what kind of transmitter you want to know about and where it might have been. Then you can figure out how much power it would have needed.

And then you have to figure out how it might actually achieve the action proposed. It is obviously not sufficient just to put the right amount of energy there - one needs to suggest a causal mechanism.

PBL

The AAIB report states there was fuel, There wer pumps capable of supplying pressure the metering valves were open but the fuel flow did not increase and infact fell from an initial appropriate rate to a rate that was too low.


To state it was restricted is probably not the right term but it was inadequate.

Quote from PBL:
The transmitting power would be dependent on how far away the transmitter is (inverse-square law and all that) and whether the transmitter is broadcast or directed-energy.
[Unquote]

We also need to know the frequencies most likely to affect: (a) valve solenoids; (b) servos/motors; and (c) computers.

Chris ScottAs PBL says, that depends ... on its specific design and manufacture ... each solenoid or motor has it's own windings and it's susceptibility to interference will be directly related to where it is in the a/c, and the other components around it.

I dont think that an EMI scenario where the RF field is sufficient to engergise or de-energise a 28 volt relay is likely. I have never seen any direct effect to an electromagnetic relay from significant RF fields from 2 - 2000 MHz.

Problems only start when the relay has a semiconductor drive circuit (either integrated in the relay or remote from it). With insufficient or defective decoupling and filtering unwanted operation can be caused down to a few Volts/metre at VHF/UHF frequencies. ie 3 or 4 orders of magnitude less than my experience above.

Correct response - wrong circumstance
 

Warning: I'm non-professional; not crew, not engineer - just guest here, thanks.

In an earlier post I asked the following but no one has answered. Can anyone assist?

"An article on smartcockpit, written by an Airbus employee about Low Fuel Temperatures, has a table listing the Minimum Inlet Temperatures for engines from various manufacturers. RR has temp of 3C (This figure seems generalised for RR).

What happens next if the fuel heat exchange systems cannot raise the fuel temperature to, or above, 3C?"

I'm interested in the answer because:
1) as far as I have read, no one on the thread has explicitly discussed the systems designed to protect the engines in the event that fuel quality/quantity is below operating standard.

2) I also wonder if there has been a failure in applied logic, meaning, the aircraft systems did exactly what they were designed to do but under the wrong, or unforeseen, circumstances, namely in this case, about to land.

3) It's intriguing to me that some form of self-limiting system operated to reduce the engine speed to just above flight idle. Presumably (?) all of the ways in which an engine can be throttled back are fully known, that the actions are 'programmed' and follow some logic pathways developed during the design stage of the engines/aircraft. If so, the examination of the cause of incident would start by examining these logic pathways and associated systems. Of course, I'm assuming that an engine is never, under any circumstances, allowed to be un-controlled unless there has been some catastrophic failure.

And so, combining 1 to 3 above, and while reading about the 3C minimum fuel temperature, I wondered what the engines are designed to do under such circumstances. Leaving aside the provisions made for starting engines on the ground in freezing conditions, under what other conditions would designers have reasoned an aircraft’s fuel would be about to enter the engines below 3C and what should the engines do to protect themselves and the aircraft?

Could it be that the designers thought that those conditions are ‘only’ to be met at high altitude in very cold air, in which case, the best option is to command the engines to throttle back, to protect themselves, and force the aircraft to descend to warmer air?

But, suppose fuel below 3C was about to enter the engines on flight BA038 at very low altitude; what would the engines do?

You see my point – planned/designed response but wrong circumstances!

By the way, I’ve discussed the 3C temp. issue because that is what triggered my thoughts but you engineer types can probably think of other issues. But the issue must result in some system commanding a throttle back.

Okay, heading for bunker ….

Thanks in advance.
Regards, Tanimbar

Vacuum in tank + rising outside pressure = fuel backflow?
 

When I read about the cavitation evidence in the fuel pumps, I immediately wondered if maybe a vacuum had built up in the fuel tanks, for example because the vent valve (used to let air enter the tanks as needed) might have been stuck.

A vacuum in the fuel tanks, combined with rising outside ambient pressure (plane approaching sea level), might ultimately lead to air bubbling backwards from the engine through the pumps and into the tanks - or not?

Two factors may have precipitated this condition exactly during the landing approach: The first factor being the ever increasing outside air pressure; and the second factor being that the engines were temporarily put into idle, presumably also reducing the pressure in the fuel supply pumps, which then would be more vulnerable to air flowing backwards than they were during cruise.

Once air starts flowing backwards through the pumps, they probably would not recover for quite a while.

To follow up on Chris Scott's query and ionagh's reply, here are a few links from twenty years ago to problems the US Army may have had with EMI affecting its Black Hawk helicopters (I emphasise the words "may have had", in contrast to "did have"!). The Navy apparently did not have similar concerns with its Sea Hawks: they were better shielded because of the extremely EM-loud shipboard environment.

All from the Risks Digest:
From Risks 5.56
From Risks 5.58
Also from Risks 5.58
From Risks 5.59
From Risks 5.60
and a cautionary note from me:
in Risks 5.59

PBL

Quote:
Originally Posted by Green-dot
how much transmitting power (radiated field emission) would be required to make a 28 Vdc relay move from run to cutoff position?

"The question is not well formulated."

PBL, i agree but you clearly filled in the blanks. Thanks for your clear explanation regarding a complex matter in such few words. :ok:


Regards,
Green-dot

...I need more popcorn...
 

Firstly, congrats to everyone (including me :)) that read through more than 2000 posts in this and previous threads...:D

Secondly, thanks to all the posters (and/or mods) for keeping this latest thread very much to the point as well as exceptionally "civilised" compared to previous ones.

I am starting with my 777 conversion in May this year and I will probably not have to study anything on the fuel system :p - thanks for all those enlightening us with technical specs (be it 777 or not).

I do not want to speculate on this incident - not because I have anything against speculation, but because I just cannot think what could possibly cause this!

However, to all those with theories involving multiple failures / unlikely events; I think you might be on the right track.
This incident was completely out of the ordinary and very unlikely. Therefore one could expect something unthinkable to go wrong.
If it was a simple answer, we would have had it by now.

Regards
PP

...now where did I stash that popcorn again?

Gustep and all of the other "Vacuum/Blocked vent" theorists, there is a "low fuel pressure" warning that would no doubt have been triggered (and recorded) should this have occurred.

tanimbar

Would a system commanding a throttle back send the command to the EECs? As I understand it, the EECs reacted to a physical reduction in fuel flow, not a throttle back signal.

It is highly unlikely that any amount of RF energy (within reason) would actually move a mechanical device like a relay. What might happen is that the rf field might effect the electronics that drive the relay but any such circuit that was safety concious would normally be shielded.

Quote

To follow up on Chris Scott's query and ionagh's reply, here are a few links from twenty years ago to problems the US Army may have had with EMI affecting its Black Hawk helicopters (I emphasise the words "may have had", in contrast to "did have"!). The Navy apparently did not have similar concerns with its Sea Hawks: they were better shielded because of the extremely EM-loud shipboard environment. endquote

Many years ago flying Canberras we had a situation where hitting the push to transmit button energised the explosive bolts in the canopy blowing it off. The VHF antennas were moulded into the perspex canopy and the feed to the antennas ran parallel to the explosive bolt wiring. If the shielding on the antenna feed broke the entire output of the vhf transmitters was coupled straight into the explosive bolt circuit and generated sufficient current to ignite the bolts.
This was considerable amount of rf energy in very close contact with very sensitive circuits and I doubt that in modern aircraft rf energy sources of this magnitude are anywhere near the fuel control system.

Acoustical piping resonance
 

Probably off topic and very unlikely, shrug.
Unusual acoustical standing waves in piping can restrict fuel flow. These are very difficult to recreate or analyze, as they leave little evidence. These standing waves generally depend on a combination of fluid flow, pump speed, pipe geometry, external vibration and fluid density.

Also, the investigation board mentioned cavitation damage to the fuel pumps.
In your collective opinion, how long would it take for cavitation to damage a 777 fuel pump? Could the damage could happen in the time preceding this accident or would it take longer?

would you care to restate the question (in bold) again

Hi lomapaseo,

Your Ameican colleague shawk has logged off. Think what he/she may be asking is:

Do we infer that the HP fuel-pump cavitation damage - as found by the investigators - all took place on the accident flight? Or had it been building up over many flight-cycles, i.e., had there been undiagnosed fuel starvation on previous flights?

Good night, Chris

For those who are looking for a cause of the accident to be the fuel pumps. I just wanted to say the 777 engines will still run quite happily on suction feed with all fuel pumps switched off provided you are not at a high FL.

In fact if you have to do a partial gear landing the checklist instructs you to switch off all fuel pumps on approach.

It was the Engine HP fuel pumps (mechanical, from an accessory gearbox, and unswitchable) that were cavitated, not the (LP) Tank pumps...

Pumps & Cavitation
 

Christopher Brennen's books
are available online in their entirety courtesy of the author and CalTech.


Have at it!



-=MachacA=-

Relays
 

Some basics might help. An electromechanical relay is a coil of wire around a magnetic (iron) core, and a magnetic armature that moves one or more electrical contacts when the coil is energized.

A typical 28 volt DC relay needs at least 18 to 20 volts DC applied to its coil in order to operate. There is no way any radio frequency signal short of a lightning hit will operate or release the relay directly.

Here's the fine print part: relays often have a diode wired across their coils for the purpose of suppressing the high voltage pulse that the coil generates when you suddenly de-energize it. Depending on the speed of this diode, and the topology of related circuits, the diode could turn RF energy induced into the wiring to the coil into enough DC to operate the relay, or drop it out if it is already energized.

The problem is that you would have to induce some 20 volts of stray RF directly into the relay coil wiring, for a significant period of time, for this scenario to work. This RF would have to penetrate both the acft metal hull and any shielding on the wires. I have no idea how much RF power would have to be delivered to the outside of the hull, but my guess is it would be upwards of tens of kilowatts. This massive amount of RF power would probably disrupt other electronics in the acft long before it acted directly on any relay.

(DC = direct current, RF = radio frequency energy)

http://www.alliedpilots.org/Public/P...ne/hotline.asp

This is APA Communications Director Gregg Overman with the APA Information Hotline for Thursday, February 28.

777 SAFETY ALERT: This afternoon American Airlines Flight 229, a Boeing 777 bound from MIA to LAX, had the left engine hang up on approach to LAX at approximately 2,000 feet. The auto throttles were on and the left engine hung at approach idle as the right engine accelerated normally. It is believed that the left engine would not respond to throttle inputs for 10-15 seconds before finally responding and accelerating to the commanded thrust. The right engine performed normally. Maintenance has downloaded the DFDR data and Maintenance action to be taken (ATBT) will be to sump the fuel tanks, pull the engine fuel filters and check for contamination, BITE check the EEC and check the MAT for any existing faults. We will continue to update you as we get more information.

RFI
 

I think you are barking up the wrong tree here! I just did a quick calculation using 28 Volts across a 1k ohm load (the coil at RF) and used 140dB for space loss beyond the e-field.

I came up with 78 TeraWatts!

If someone was using that kind of RF power, you would know. All the lights in and around the airport would dim!

Halfnut you don't specify engine model; is AA using Rollers?

I've seen RR engines on some AA 777's, Glueball

Grounding Time?
 

Will the authorities ground the 777 fleet now? They did it to the worldwide DC-10 fleet in 1979, wrongly as it turned out.

GB

Another Acceleration Hangup on a B777.

This may be a blessing in disguise!!

Remarkable and let's hope 'they' can identify the cause. Presumably the fuel system and FADEC is standard to whatever engine leaving only the engine driven High Pressure fuel pump/s as a variation.

What doesn't seem to have been covered so far on this thread are the ACUs - Acceleration Control Units. Are these part of the FADEC and hence computer controlled and are there variable compressor guide vanes and compressor bleeds in the engines also being controlled by the ACUs?

Remember the RR Avons and the gubbonry necessary to have them accelerate cleanly what with guide vanes and bleed valves inter related. Was it 18 or 28 seconds from idle to max following a throttle slam? Once had one of these in an F-86 Sabre maladjusted under investigation hicoughing around a circuit with some of the surges putting out the fires. Saved from deadsticking with easy HOT relights.

For the uninitiated a HOT Relight is defined as one that can often reliably be achieved within 10 seconds of a flame out.

I can confirm Jetdoc's(?) statement that the (2) surge tank pressure relief valves have both positive and negative pressure relief. There is a description in the AMM on how to check the valve using positive and negative pressures. So the vacuum in tank theory seems to have lost its momentum :p

On being sanctimonious...
My company does not allow me to take information from company computers without permission.
All the Boeing manuals on my company computers have very prominent copyright statements.
No doubt similar restrictions are placed on those posting pictures here.

Also...
Some of the technology used on the 777 is classified material. Boeing recently got into trouble for selling airplanes to a foreign country.

What copyright de-restrictions are placed on Americans are not necessarily the same as those placed on other nationalities.

There is a difference between being sanctimonious and worrying about fellow PPRuNers losing their jobs and being punished financially and/or being imprisoned.

Rgds.
NSEU

Note however...
The trigger point for fuel pump low pressure switches can be as low as 4psi... The output of the boost pumps is, nominally, 12psi... so the pumps can, in theory operate down to almost 1/3rd the normal pressure before a warning is triggered.

Also, flow rate can vary depending on the type of 777 boost pump ...even though the output psi is the same. One 777 airline AMM shows the data for 2 pumps. Flow rate for one is 16,000kg/hour. Another is 19,320 kg/hr.

I'd be interested to know what sort of pressures are required to open the check valves on the boost pumps.

Rgds.
NSEU

Ref the AA 777 incident yesterday, I understand the problem was probably with the autothrottle rather than the engine. Entirely different to the BA 777 incident.
Hope this stops the thread diverging.....I'm learning a lot about cavitation and bubbles here!

Just a follow up of the AA :

http://www.flightglobal.com/articles...-throttle.html

Rather a strange statement, let me explain.

The auto throttle drives the throttles>>> the throttle position (angle) is sent to the Fadec (EEC)>>>the Fadec sends demands to the fuel metering valves to give the appropriate power.

For the problem to have been due to the autothrottle system it would have needed a failure that, when power was demanded, only moved one of the throttles forward (which could have simply have been rectified by moving the lagging throttle forward).