I would like to remind everybody that, according to a senior avionics manager at Honeywell in a paper published in a major conference, a certain aircraft type came within days of having its airworthiness certificate withdrawn in the early 2000's because of irresoluble problems with data transfer on critical data buses (known as "Byzantine faults").

That answers the question "how far do you go?": maybe all the way, depending.

PBL

Boeing doesn't know the answer, apparently, but it knows what it's looking for:

http://www.flightglobal.com/articles...h-landing.html

SF

From the URL above:

"Boeing's chief pilot flight operations safety division Capt Dave Carbaugh says the fact that the fuel pumps, according to the UK Air Accidents Investigation Branch (AAIB), had been cavitating not long before the impact"

Is my mind going? I only remember a statement in the latest AAIB report that there was damage caused by cavitation, not anything about when it might have occured.

Cats, be assured your mind's not going. Your link to the AAIB's statement proves that. But had you wondered whether Capt Carbaugh might know something you don't?

Question 1: If we cracked open a 100 average fuel pumps on a 777, how many would show signs of cavitating?

Question 2: How could one tell that the cavitation was recent, and if recent how recent (aka this flight vs flight before)?

My (layman's) understanding of the AAIB investigation process is that the manufacturer (and airline for that matter) will be closely involved at all stages. It seems likely, therefore, that Boeing and their chief pilot would have inside information about the direction the investigation is going.

What surprises me is that he "went public".

Bobbsy

To Question 2: perhaps the output pressure would show fluctuations which (depending on the sample rate) could indicate that the pump is cavitating?

Just remember that the cavitation is an indication that the pump had been operating with low inlet pressure, the pump itself would only have its performance mildly affected- ie not enough to cause the event

SuperT

I'm not sure we can imply that a problem with cavitation is only a function of low inlet pressure on a pump - with or without nucleation within the liquid phase, although I agree, nucleation will help considerably.

Presumably you're imply waxing or freezing of the fuel thereby restricting flow rates to the donkeys, lowering fuel pump inlet pressures and causing the cavitation. My own non specialist view is that the waving/freezing hypothesis died the death aeons ago on this most worthy of threads in the light of the published AAIB data showing fuel fpt to be < - 50 C - a suspiciously low figure.

The following, somewhat familiar by now hypothesis, alleges no failure of the pumps.

I'm rusty on this but cavitation can easily produce a mixture of true gas phase mixture fuel components (which may or may not contain air as well) from the mixture alone by mechanical induction in a liquid phase mixture if component(s) in the fuel mixture have a low enough Mr or levels of intermolecular bonding and the kinetic energy input is high enough.

I'm guessing, but I reckon any gas phase component's t½ at say +80C will be long enough for it to be a nuisance.

And and ……if the fuel has stratified in the tanks as well, with the low Mrs with their lower densities at the top of the tanks, they'll be fed to the donkeys at higher concentrations just when the FADECs go with throttle up so to speak, at the end of the flight.

As an aside, the pumps are defined as serviceable by the AAIB interim report. Now I take that to mean that they're within spec and I presume spec refers to volumetric efficiency against a liquid phase. If the poor things are trying to pump lots of eminently compressible of bubbles we will have considerably reduced flow rates and outa-gas symptoms.

Just what were the additives? What was their Mr?

It's already been said and it's still true, AAIB have done their job entirely to spec and we have to wait for the full report, but hey this is PPRuNe after all!

Quote:

What Boeing is doing

Reading the article, i assume the tests will not be limited to fuel properties alone, as the article suggests, but fuel system tests including failure simulations of fuel system components upstream of the engine pumps? I would not be surprised they will do tests equivalent to what i have posted in post #590. In this case for 2 donks instead of one.


Regards,
Green-dot

Just to build a little on the last few posts:

The investigation reports that the fuel pumps are serviceable. I take that to mean that the flow/pressure characteristics, when the pumps are bench tested on fuel, are within specification.

But when the cavitation damage occurred (and, depending on the causes, the design of the flow paths, and the metallurgy, that might have occurred relatively quickly), we can speculate that the output from the pump(s), would have been some dual phase mixture. If such a mixture was being passed by the pumps at the time the FADEC requested a higher fuel flow in response to the increased power demand, it is possible that the fuel flow rate within such a mixture would be insufficient.

You can bet your bottom dollar that the manufacturers (RR and Boeing) are throwing much bigger resources at this than the AIB ever can.

Being involved in automotive design for many years may I answer a question from a day or so ago, cavitation damage can appear very rapidly, even to the point of destroying a pump in a matter of minutes in extreme cases.

Sure, but remember that the pumps don't just have cavitation for no reason. There must have been an upstream flow restriction, and it is that flow restriction which caused the reduced flow out of the pump, not the cavitation at all.

smuff2000 wrote: "Being involved in automotive design for many years may I answer a question from a day or so ago, cavitation damage can appear very rapidly, even to the point of destroying a pump in a matter of minutes in extreme cases."


And I would expect that, since the pressures and flow rates generated by these pumps are many times that ever found in any automotive applications, the damage rate would be even faster.

Chris weston

We've been waiting for someone like you to enhance our knowledge of the 'phenomonen' of cavitation of our high pressure fuel pumps. But then you confuse us by introducing Mr and Mrs without explaining what that happens to be. Help!!

I know a little more than most about cavitation having been triggered into researching the subject after having an irretrievable flame out on a Canberra's left Avon engine followed by the loss of the remaining right engine on short final. Fortunately we just made it to the under run and rolled out on the runway, much to the relief of the nav who had been tempted to leave via his ejection seat. But that incident was caused by the failure of all four variable stroke piston pumps which rapidly disintegrated when they started to pump water instead of fuel. There were some signs of earlier cavitation damage discovered.

Experience in those days soon revealed that short periods of operation of the HP pumps at high altitude soon had HP pumps cavitating and then we were learning that jet fuel starts to boil/gassify/produce vapour as the suction side of HP pumps reaches some low value not far above a vacuum. Presumably there will be a temperature variable.

I asked in a previous post whether anyone could produce some feel or appreciation for the likely pressure below which one could expect cold fuel, with some water content, to start vapourising which would consequently mark the onset of HP pump inlet cavitation. Such a number will set a bench mark as to the extent of fuel flow restriction necessary to substantially reduce output flow.

I would expect that fuel cavitating at the pump intake would instantly return to total liquid at the considerably higher pressures at its outlet with its flow then moderated by the FADEC demand.

Can you elaborate and give us an enhanced comprehension as to the operation of the conventional HP pumps particularly under marginal conditions.

Apart from opening up further instructive speculation this will prompt pilots to a better appreciation of the train of events which result when they move those thrust levers and the booster pump switches.

A thought:

What is the situation with EROPS/ETOPS certification?

It would seem that there should be some reassessment in view of the uncertainty.

If both donks dead, forget about ETOPS.

ETOPS certification does not depend on guesswork. If and when the causes of the accident are known then it may or may not be looked at depending on the relevance of the accident causes.

Milt,

I must needs disappoint and apologise to you if you feel I have mislead you and others, engineer I ain't and I can't advise on something as safety critical as the operation of conventional HP pumps in likely cavitation situations. I can speculate that the 777 had a fuel mixture problem probably revolving around additives.

I have great respect for those people such as yourself for whom cavitation is a not merely an armchair and academic exercise.

I'm simply a pure-ish chemist who has the privilege to work with some competent Chem Eng people who, amongst other things, pump a lot of hydrocarbons around.

We agree that any gas phase hydrocarbons et al will go back to liquid phase more or less instantly and that their half life (t ½) will be short.

Mr is the Mass of a Molecule on the C scale and I am further to blame for not rendering it as such in my text. I typed the r s as the subscripts they should be at the time but on cutting and pasting, the site rendered them as "Mr" and in the plural as "Mrs" - I forgot the understandably limited word processing on the site, Word it ain't. I also typed +8 (superscript) O Centigrade which became +80C ………. a singularly startling LHR Jan temperature. My fault, not Danny's, I know the site does this, mea culpa.


CW