Fragman88

Just tripped over this in a US flight safety article:

`Or, as legend R.A. "Bob" Hoover has been quoted, "If you’re faced with a forced landing, fly the thing as far into the crash as possible.".

Probably not that relevant to these chaps as so little time to assess and act, but I will certainly listen to any thing said by Bob Hoover, and a nice refreshing reminder from an Ancient Aviator (with apologies for the `Ancient').

infrequentflyer789

Modern DFDRs can record hundreds (if not 1000+) of parameters - way more than the minimum required set, plus they have the QAR (which may record even more) for almost all of the flight (including through the onset of the problem).

I think it is highly likely that the AAIB have spar valve position data, their statement is that the data shows valve open, not indicated open or commanded open. They also state that movement (even uncommanded) "would have been recorded", which implies to me that there are FDR parameter(s) recording actual position as well as commanded.

They don't actually state that they couldn't make it fail, they state that they couldn't reproduce "uncommanded and unrecorded" movement (my emphasis) - which is actually quite a different premise. They then infer that there was no failure because none was recorded, rather than because it was failure proof.

I don't know enough about 777 instrumentation to be sure, but I believe that on older boeing types there was(is) a spar valve light that indicates closed and transit. It would be interesting if that has been lost (or relegated to a maintenance page) going to glass cockpit. However, I also think that the AAIB must have some basis for their statement that warnings would have occurred.
I agree - but on the other hand they may regard a flat line graph as superfluous if they believe they have already stated the same in words.

Don't get me wrong, I quite liked your spar valve theory, and the AAIB clearly also thought it was a line worth investigating - but I do think it is looking a little dead now.

[I have to admit I don't like the ice conclusion, I feel I need to re-read and digest the report a bit more, but it does seem to be what's left when they've eliminated other things, rather than driven by positive evidence. Feels very unsatisfying as a conclusion (even interim).]

infrequentflyer789

6.5 million is hours not flights (1.4M flights) - but it is all 777s (in fact all RR powered 777). That does cover several variants however, and fuel tank configuration is different between the variants, which may well be significant.

Of those flights, they've only looked at (only got?) data on 13,000 RR powered 777 flights, and the conditions are unusual within that set of data.

sydneymike

My sentiments entirely. And, therefore, in the meantime it COULD happen again. Even if the chances are only 1 in 13,000, why take it if alternatives are available. Most of my circle of frequent flying business travellers now selects flight carefully on longhaul NE Asian/Europe flights to choose the 744. Ice or otherwise. Maybe irrational, and was ridiculed earlier, but the fact is there is a PERCEPTION amongst frequent flyers (Who DO know one end of 744 from the other end of a 772/3ER) that this airplane has unidentified problems and should be avoided until otherwise.

snowfalcon2

With all respect,
Can someone confirm that the FARs would allow this condition? From what I've quickly gleaned (Gimli Glider) the RAT should always provide enough hydraulic pressure for the primary control surfaces.

L337

Hours not flights. I stand corrected. Thanks.

PS. No RAT on a 747-400.

PJ2

snowfalcon2;

It is true that the RAT will stall out if the a/c is permitted to fall below a certified airspeed which is about 140kts if I recall - I'll have to check the manual. The RAT on early 320s was subject to interference from the nosewheel slipstream - can't remember what was done to rectify the situation - perhaps the starboard offset, not sure.

dxzh

A few more thoughts related to centre tank pump inlets:

- Each OJ pump must be able to pump, while in climb and then cruise, 99% (as in (80t- 0.8t)/80t) of the fuel out of the centre tank before the inlets are free from the remaining fuel. The OJ pumps free of fuel, lose pressure and are subsequently switched off - the main tank boost pumps, which are permanently on, maintaining flow to engines in the meanwhile. In this case, the centre tank reduced to a little under 750kg indicated at its lowest at the time of the engines switch to main tank supply, before apparently increasing for some reason by 125kg or so to 800 kg indicated a few minutes later.

- Each OJ pump inlet is raised from the floor slightly and is in fwd part of the centre tank. What exact depth of liquid on top of floor (or liquid on top of ice on top of floor) immediately below the fuel pump inlet is required to submerge OJ pump inlet? How localised could that depth of water be given the geometry of the tank, position of ribs and the tank's shape during flight? Can anyone produce a diagram of exactly where everything is at the bottom of the tank near the OJ pump inlets?

- I assume that the fuel scavenge worked given drop from 800kg indicated to 0kg but ... given the knowledge that on cold flights like these that it is not uncommon for some of the fuel scavenge pipes to become blocked by ice, what would happen if the fuel scavenge lines local to the fwd positions on the left hand side and right hand side of the tank where the OJ pump inlets are were iced up, but the scavenge lines in the centre of the centre tank were working?

- I note from the rate of scavenge in Figure 1 the rate increased as the level of fuel indicated declined to 0kg. Could that suggest that as the level lowered: the rate of scavenge appeared to increase as less fuel/free water was accessible than expected as it was trapped where the scavenge lines were not working locally; and/or an expected increase given the vagaries and the geometry of tank?

Is it a red herring even to contemplate the boost pumps or centre tank pumps allowing the passage of fuel/water when they are selected OFF? I guess the premise of the suction feed test on the ground (as a pure test of the suction feed) would be flawed if it was a possibility that an engine-driven LP could draw fuel through any of the unpowered pumps from centre tank - and thus I would like to discount the idea completely. Yet as a Mad Hatter is drawn to a tea party, I am curious about the possibility given the suggestion in the interim report that air at least might possibly pass through the OJ/jettison check valve (and presumably therefore through the OJ pump when selected OFF too) ...

lomapaseo

Kind of like a computer with a hundred USB ports for data inputs. But do you have a hundred devices to hook up to these ports?

It ain't the DFDR recording limitation that just meets or slightly exceeds the regulated requirement, it's the availability of aircraft system sensors and wire routings that sets the actual limit.

With that said I really don't know what parameters were recorded for BA038

coool guy

I certainly believe the AAIB are very trustworthy and honest people simply trying to do thier best in what seems an unfathomable case.
However there are naturally those conspiracists who will always look for a cover up, bribery, corruption in any organisation and lets face it it can happen.
It should therefore be incumbent upon the board to be as transparent and straightforward as possible in all their reports and findings to the public.I have read all the 5 reports and all the posts on this thread. The reports are comprehensive and well written but that doesnt mean we cant criticise them!
My criticsim is this. Not once before did the AAIB mention the actions of the crew and I agree with Nod that the AAIB have never shown much interest in those actions, clearly focusing all their attention on fuel interuption and its causes.
So why do they now suddenly tell us that it was the first officer who disconected the Autopilot and selected flap 30? and why is this more relevant now that they have established that ice is the cause?
They must have known those facts from the very beginning but declined to tell us about it. I am sure it is not the case that they are trying to hide something but nevertheless it gives ammunition to those who think they are!
Why cant they just lay all the facts on the table from the start without worrying about whether it is relevant or not or who it might hurt or what the consequencies might be or the costs to the industry etc etc

dxzh

1. Main tank water scavenge

If centre tank fuel scavenge pipes can ice up, perhaps the main tank water scavenge ones could in certain conditions too? Maybe a scenario with cold-soaked fuel in each main tank at critical icing temperature at end of flight with a relatively high concentration of water after centre tank fuel scavenge and then, during cold stopover, a very cold water/ice mass accumulated in and around water scavenge inlets (near bottom of main tank), blocking them? As long as no more than 7 gallons of free water was in main tank at any one time, no excess water message would be recorded about this issue.

If each main tank's water scavenge lines became restricted by ice before the return flight, would that also explain how there was a very low level of suspended/entrained water to be found in fuel in main tank as it would not continue to be mixed into the fuel by the water scavenge and tend to separate out?


2. Icing

It almost seems easier to explain icing if there is some form of cascade effect where:

- first, relatively minor icing over time restricts narrower and colder scavenge pipes.
- then, unforeseen major condition such as lack of water scavenge and/or fuel scavenge develops.
- then, unlikely accumulation of free water/ice where it is not expected.
- then, accumulation goes on to cause bigger and worse things downstream.

And has to be readily duplicated on either side. And lack of warning messages and tell-tales such as water ppm accounted for.

Poof in Boots

Sorry Lads. Trawling through all these posts and your honest attempts to vindicate the AAIB's preliminary findings, seriously, what are the odds of ice causing both engines to fail at the same time? One I could understand, but not two.

If there was low fuel pressure to the engines caused by a build up of ice, why was this not noticed? Why didn't the PEK/LHR 777 that day also exposed to low temps did not reciprocate the problem. A mate of mine flying a BA 747 at the same time from HKG said fuel temps were in the amber for some time, but no engines stopped.

The fact that this debate continues proves the ice theory is not convincing.

It is something else.

Re-Heat

You are seriously trolling.

That is the whole point of the investigation

As high fuel flow was not demanded or required until final approach.

They don't know, but are probably considering the consequences of two back-to-back sectors in some of the coldest temps recorded for long durations on 777 missions to date.

Different plane, different route, different design. What is your point?

Poof in Boots

Well Re-Heat, many years ago flying back from ANC in very cold weather (-70f+) with US avgas, the main concern then was the fuel waxing, not any water freezing in it.

You would think a newer design aircraft would have better protection against deep cold for long periods. This is not a modern phenomenon.

To lose both engines at such a critical time is unbelievable. I landed 30 minutes before the BA038 but on a GE90 777 from the Gulf and saw it skid to a halt, as we were parked on a bussing stand at the Hatton Cross end of T4 on the opposite side of the runway to where YMMM came to rest.

On our approach quite high amounts of power were required as it was a gusty wind and we were shaken around a bit in our shoulder straps.

precept

See following link:

This is a final rule request for comments.

Justia Regulation Tracker Airworthiness Directives: Boeing Model 777-200 and -300 Series Airplanes Equipped with Rolls-Royce Model RB211-TRENT 800 Series Engines, - Federal Aviation Administration - 52909?52911 [E8?21138]

infrequentflyer789

I know, and I don't disagree, but the more modern types have moved away from point-to-point wiring to bus-based systems. That allows you to simplify the scarily complex wiring looms (or to do more with the same level of scarily complex). The 777 is ARINC 629 bus, which is multi-transmitter (and more advanced than anything other big jet until you get to 664 / AFDX on the A380 / 787), which should provide a lot of help with this issue.

Best public reference I've got for parameters recorded on a recent jet is for N651AA (flight 965, crashed in columbia). Quoted in various places as recording approx 300 parameters (see eg. Hall's Testimony on March 7, 1996). That is some way over the minimum (which is still 88 I think).

That flight was a 757 with previous generation (ARNIC 429) bus - so I would expect the wiring on a 777 to be able to route at least that many parameters to the FDR & QAR.

Nor do I, and I doubt we'll be told.

infrequentflyer789

The signal is slightly noisy - there are other blips - and that could account for this. Also, consider that when the OJ pumps switch off, does fuel already in the pipes / pump but upstream of the check valve fall back into the tank ? That might be enough to account for this signal.


I think there have been drawings posted way back in the various threads on this incident - but you'll have to trawl through a lot of posts to find them.

Could be, but I suspect it is more likely a result of the way it's controlled from float valve in main tank - I think the final scavenge rate works out higher than the fuel burn rate when scavenge started.

I think that was only if all the main tank inlets were blocked and suction feed woudl pull in either nothing or air from the CT. Nothing vs. air is not a great choice of fuels . Also, if this happens they've found different effects to those recorded on this flight, so it didn't happen in this case.

Overall, I would have thought that the important thing to do with the fuel system FDR data (ie. fig 1 in the report) would be to compare with a similar but incident-free flight. We don't have the data to do that but the AAIB will - I would be very very suprised if this hadn't been done already and anthing unusual in the fuel system data would have been looked into.

Backoffice

Whenever I read these posts I get the impression that contributors are talking about free ice floating around in the fuel.
I see this ice, whether it be from water in the fuel or the result of condensation being a sudden release, either due to temperature, flexing of the wing or vibration, of ice which has frozen and adhered to internal wing surfaces during the flight.
i.e. all the scavenge systems, fuel mixers and heaters could never touch this ice as it was attached firmly to a surface and released itself in quantity at roughly the same time, thereby clogging momentarily, the filters to the pumps.

infrequentflyer789

Thanks for the link, to quote the AD:

Hmmm. Is it me, or does that imply a far greater degree of certainty as to the mechanism than the latest AAIB report did ? I read the AAIB report as saying that this was one icing scenario (of two), that they couldn't reproduce it yet, but it was likely to have happened as it is the only hypothesis they haven't disproved.

dxzh

1. AD

AD pretty much as predicted last week.

As a stop gap measure to keep 777s flying (given regulatory imperative to avoid certain levels of risk for significant failure conditions hazardous to flight), it is definitely to be welcomed though it (almost inevitably due to its immediate nature) misses the opportunity to address some of the likely underlying contributory factors to "hazardous amounts" of ice accumulating in the "main tank fuel feed system".

I would highlight the AD does not reflect:
- the potential build-up of free water and ice in the 200-ER centre tank over two cold-soaked stages.
- the fuel in each main tank remaining cold soaked (together with its scavenge lines) on cold stopover.
- the timing of delivery of any centre tank water into each main tank fuel supply.
- the accretion of ice might not be of the "releasing" type in which case until it fully develops, full flow may not remove it or be affected by it.

In particular, while the in-flight run-up can happen 3 hours before TOD, any ongoing build-up of free water in the centre tank (not addressed by AD) may only be scavenged into the main tanks one hour or two before TOD as on BA038 - if run-up is done before fuel scavenge completed then arguably chain may not be broken unless require another run-up, say, 20 minutes after centre tank indicated as zero?

2. Location of "95%" restriction

If the restriction in each manifold is elsewhere than the "engine fuel" system, as could be suggested from the test rig results which showed significance of a time lag after the accelerations started but before the flow reduced, and operates as an effective restriction on 95% of the cross-section at a point in the "main tank fuel feed" system upstream of the engine, then this AD seems to miss the obvious risk that the restriction was never located in the "engine fuel" system! I assume though that (if I may try and be diplomatic) it is currently expedient to ignore that suggestion.

Ironically, pity the non-RR engined 777 crew and passengers - at least those flying in 777s equipped with RR should now face a reduced risk of icing-induced double rollbacks. Perhaps more risk-averse non-RR 777 operators though might see some advantage in being seen to implement the proposed AD on a voluntary basis, at least pending the next report?