737 Fuel anomaly.

[BARKINGMAD]

Whilst we bat the issue of refuel error tolerances about in another thread, has anyone with a direct line to Mr Boeing asked about why the 737 indicated tank total fuel seems to reduce by 100-200 kgs at the time the centre tank empties and those pumps are switched off?

As occupational therapy on those long trips, I once spent a few sectors with a piece of graph paper plotting FMS, accurate tank total (to nearest 10kgs), and ramp minus engine totaliser readings at the same waypoint, and got 3 lines wobbling down the page til the centre tank INDICATED empty with flickering lo-press annunciators.

Subsequent checks showed the ramp minus totaliser readings had crossed over to above the tank total line, which eventually managed to recover some of its "lost" fuel as more was consumed. The FMS figure always stays optimistically above the others, presumably due to the comms delay between that display and the fuel summation unit. I never use the FMS reading in my decision making, unless it's less than the others.

Is there fuel hiding in the centre tank somehow, which should be scavenged at the half-wingtank mark, but possibly too late to prevent a fuel diversion at approx 4 tonnes and upwards?

And why do the centre tank lo-press annuns come on almost at the same time in the shiny new 'frames, but amost always the left one leads the right in older 'frames? Are they mounted at different levels in the centre tank, assuming the craft is trimmed wings level.

I know I should get out more and meet interesting people, but as I'm stuck looking at and plotting these readings for a job I thought I'd wave the topic as part of my further education in the mysteries of the (now) 73NG. :confused:

[BOAC]

Cannot answer part I, nor why the 'shiny's' do not exhibit this, but in the ones I flew 2/3/4/7/8 (older) the left ctre pump draws from further forward than the rear so will probably come on first. Maybe Boeing moved the pumps in later models.

[Skyjob]

BOAC, you are correct in the fact that the location of the pumps is the reason for the difference in lights to come on, as it is a function of body angle (angle of attack), this is true for all models.

Regarding the difference between Classic and NG, there is the fact the the pumps are indeed in totally different locations, the Classics in a dry area of the wing root, the NG's inside the centre fuel tank, as part of the redesigned fuel cell structure.

The scavenging function is also completely different, Classics operate the scavenge pump for 1-20 minutes after the center tank pumps are switched off, but NG's start the scavenge pump when main tank 1 is approximately half full and its FWD pump is operating, and the pump will continue throughout the remainder of the flight.

The following image will provide you the clues you are looking for:

http://www.b737.org.uk/images/fuelpump_locations.gif

Additional differences of the Fuel quantity indicator system accuracy:
Classics analog dials: +/-3.0%
Classics digital dials: +/-2.5%
NG without densitometer: +/-2.0%*
NG with densitometer: +/-1.0%*
*) with main tanks >50%, roll +/-1 deg and pitch -1 to +5 add 0.5%

Regarding the FMS quantity, the anomaly you described can be explained by the fact that the fuel upstream of the fuel pumps, downstream of the spar valve is taken into account when calculating the remaining fuel on board, hence always reading a slightly higher value. This is substantiated by closely observing the fuel gauges after the engines have been shut down and the fuel in the lines returns to the tanks, which is seen by a slow increase in fuel tank quantity in the approximately 1 minute after shutdown. Actual arrival fuel read from the gauges should therefore never be done prior to this time, as fuel is still returning to the tanks.

[ImbracableCrunk]

Skyjob,

How does any of that cause the M.I.A. 500lbs of fuel after the center tank is emptied?

[BOAC]

Skyjob - I have a better diagram somewhere but cannot locate it at the moment, but yours is 'missing' one off ctre tank pumps for each variant! The other important factor your diagram does not show is that it is not where the 'pumps' are that matters, but where the draw pipes lie, which is what causes the 'event'.

[Skyjob]

The positions of the pumps are as follows

On the Classics:

Center Right Pump most forward, thus exposed first to run dry;
Center Left Pump middle of tank, thus when exposed some aft fuel in center tank remains;
Scavenge Pump Inlet located most aft of tank, this should (over time) be drawing all remaining center fuel out, but this takes time, and sometimes this pump is not effective enough to drain the tank remains before landing, then resulting in an apparent fuel quantity in the center tank which is yet to be used. Priming the center pumps, is a method to re-start the scavenge function.

On the NG:
Center pumps are both located on the rear spar.

The missing pumps are not shown as one half of diagram shows the Classic and the other the NG. Here are two other diagrams showing each type:

Classic Schema (https://docs.google.com/file/d/0B_b6TyjPLrfBUEl5eTQzNHByeGM/edit?usp=sharing)

NG Schema (https://docs.google.com/file/d/0B_b6TyjPLrfBS3FOM1Q5WGFManc/edit?usp=sharing)

[BARKINGMAD]

Thanks folks for locating the ctr tank pumps and explaining the secondary question in the original post.

But has anyone found my missing 100-200 kgs of fuel yet, or have any of the multitude of 737 drivers out there even noticed??

Bearing in mind the indicated contents are an important part of our diversion criteria, I would have thought there'd be enthusiasm for some sort of explanation?

I don't have a direct line, nor a proxy whom I could use, to ask the Seattle/Everett team so I live in hope..................:confused:

[framer]

Everyone I fly with definitely notices, and then accepts it as completely normal.
I too would like to know where it goes.

[737ngpilot]

I have seen this as well, and I have found a 300 to 700 lbs loss if you will depending on the airplane I am in.
Hope somebody can get us an answer

[BARKINGMAD]

During todays TFS and return, I noticed with some enthusiastic fuel monitoring that the decay in indicated tanks total started, on this particular 'frame, before the switchover to wing tanks.

The normal "recovery" did not occur, and the external factors such as winds and temps did not account for the decrease.

Over to you 73 all types drivers, it might provide something to alleviate the routine boredom with which some are afflicted?! :confused:

[BOAC]

Can I confirm where you see the "indicated tanks total"? You 'dismiss' the FMS quantity but I assume this is where you see the difference - or is there an actual reduction in wing quantities?

I always used the FMS as 'advisory' only and mostly ignored the odd 'up-see and down-see' there and took the PLOG fuels as master since they will probably be more accurate. How does tanks compare to PLOG?

[spannersatcx]

I don't know the 737 so speculative on my part, fuel transfer is normally done through the refuel gallery/lines, this holds 100kgs or so, when it is in the gallery it is not detectable and therefore gets lost for a while.

[BARKINGMAD]

BOAC I'm reading the 3 individual digital tank contents and adding them to the nearest 10kgs, the minimum the NG will display. This is the source of the reduced fuel total.

Yes, I'm ignoring the FMS total, as it lags both the engine display total and my 3 tank addition total.

When compared with our very accurate PLOGS, the comparison shows a consistent reduction, over many 'frames and in 4 different companies.

As I hope I explained before, ramp total minus the engine totaliser readings starts out less than tank totals and then becomes more, all taken at exactly the same waypoints.

My next task is to concentrate (keeps me awake and amused!) on the changeover of these readings and try to find a factor which affects one or both.
I suspect that ramp minus totaliser may produce a more consistent line on the graph.

I really must get out more.....................................:)

[BOAC]

So you must be seeing a 'snap' REDUCTION in wing tank quantities at ctre tank empty?

Never too sure about the totaliser accuracy since I do not know how it works out SG and ctr tank fuel will be warmer than wing after a while.

[ImbracableCrunk]

I wouldn't call it a 'snap' reduction. It's rather slow. On a trip from the US West Coast to Hawaii, you'll be 500lbs under burn until the center tank is emptied and then a fix or two later you'll be even on burn.

[BOAC]

I wouldn't call it a 'snap' reduction. It's rather slow. the 737 indicated tank total fuel seems to reduce by 100-200 kgs at the time the centre tank empties and those pumps are switched off? - looking at two different events?500lbs under burn until the center tank is emptied - you mean by that you seem to be burning less?

[Skyjob]

"Regarding the FMS quantity, the anomaly you described can be explained by the fact that the fuel upstream of the fuel pumps, downstream of the spar valve is taken into account when calculating the remaining fuel on board, hence always reading a slightly higher value. This is substantiated by closely observing the fuel gauges after the engines have been shut down and the fuel in the lines returns to the tanks, which is seen by a slow increase in fuel tank quantity in the approximately 1 minute after shutdown. Actual arrival fuel read from the gauges should therefore never be done prior to this time, as fuel is still returning to the tanks."

Basically people we need to look at the difference between the fuel gauges which indicate what the fuel quantity is in the tanks and the FMS which calculates fuel on board.

Fuel gauges read the fuel tank quantity only, when engines are shut down, some fuel pumped towards the engine with engines running returnes to the tank, increasing it quantity after shutdown in the first ~60 seconds. This amount is expressed to rounded nearest 10.

FMS calculates the quantity on board, taking into account fuel in tanks and fuel pumped towards but not yet used by the engines, thus always indicates a higher value then fuel quantity gauges. This amount is rounded to nearest 100 as expressed in tonnes.

The two are fundamentally different, with the gauges always under reading actual total fuel on board until engines have been shut down and system returned to its rested state when fuel has returned to the tanks.

Best way to check fuel used is the using the fuel used indicator, expressed to nearest 10. Using fuel on board before departure reduced by fuel used gives most accurate reading to 10.

And as this accuracy is not really ever needed the FMS quantity should be used as it calculates exactly the fuel on board and the fuel used and the fuel not in tanks anymore but not yet used by the engines. It gives accuracy to 100, close enough on an aircraft burning ~37kg per minute, as closer accuracy to the 10 would change every 20 seconds!

[BOAC]

Skyjob - you are on the wrong tack, I think? BM posted
"Subsequent checks showed the ramp minus totaliser readings had crossed over to above the tank total line" and did not use FMS figures, just gauges..

[ImbracableCrunk]

BOAC,

I have seen gauge issues where the quantity changes rapidly, and this isn't so with the center tanks being empty.

In my experience, we push with the correct fuel indicated on the FMC. At cruise, the FMC shows 500 pounds more than you'd expect to have before the center tank is emptied. Once on the main tanks, the FMC shows the expected total again.

Haven't checked it via the totalizer, but I guess it's not as much "under burn" as it is extra fuel being "created" somehow.

[737ngpilot]

I have checked the burn and it's very close to flight plan, as have been mentioned here, the FMC will show entra fuel, and magically as the center tanks empty, we are back to even. I have notices this "extra" fuel is not the same on all tail numbers it varies from 400 to 700 lbs

[MZ73]

The only explanation to it after monitoring for a long time is that it isn't about what reads the fuel it is about how it is read and measured by the corresponding capcitators , so just consider the body angle and the attitude with the existing CG and what phase of flight you are in? And you will figure out why with more nose up attitudes you get more fuel indicated and vice versa , effect of center tank fuel consumption on the above factors and why we have more fuel indicated by the begining of the flight and less than what was indicated by it's end .

[HazelNuts39]

Subsequent checks showed the ramp minus totaliser readings had crossed over to above the tank total line, which eventually managed to recover some of its "lost" fuel as more was consumed.
Scavenge Pump Inlet located most aft of tank, this should (over time) be drawing all remaining center fuel out, but this takes time, and sometimes this pump is not effective enough to drain the tank remains before landing, then resulting in an apparent fuel quantity in the center tank which is yet to be used.
Doesn't that explain it? When the center pumps switch off, the fuel remaining in the center tank is temporarily 'unavailable'. It is available again after the scavenge pump has transferred it to the wing tank.

[BOAC]

Doesn't that explain it? HN39 - no.total fuel seems to reduce by 100-200 kgs at the time the centre tank empties - I had followed that line too but BM is quite clear - the 'usable' ctre tank fuel is ZERO at that moment - gauge shows ZERO. Yes, we know there IS fuel to be scavenged but it is not 'usable' as far as the gauging is concerned - at least until it arrives in Tank 1.

[BARKINGMAD]

Skyjob, Many thanks for your info, it increases our knowledge of how the system works but my observations have been made in the cruise, with steady pitch/roll attitudes, no climbing and obviously without the complications of engine shutdown, fuel sloshing around galleries and indicated values changing after shutdown.

Alas my recent flights since the thread started have been with lower fuel amounts, with centre tank switchoff occurring in the climb, thereby invalidating any attempt to record the STEADY STATE fluctuations which are the main point of the posting.

My proposed datums are the PLOG/OFP stated figures, whose accuracy I have no info about but have to assume they are good and linear, and the ramp-minus-used figure which I assume is more accurate, though it might show a variation due to the decreasing fuel temp. I accept that APU useage since refuel is a constant for that sector and therefore will not change with subsequent readings.

That said, anyone trying this monitoring should notice the consistent and possibly alarming drop on tank totals at or near the changeover to wing tank useage.

It may be that the 2.5% tolerance kicks in and changes to minus rather than plus, equalling virtually 100kgs per wing tank, but this in itself would be an odd characteristic?

Longer flights permitting, I will restart plotting a graph of the relevant values and report back in due course.

A colleague has told me that tolerating flashing ctr pump lo-press annunciators until they illuminate steady lasts for up to 6 minutes and reasons that if the lo-press light extinguishes even momentarily, then there is fuel present to provide pressure and 6 mins at cruise settings approximates to 200 kgs.

I notice the auto shutoff logic requires 15 secs of continuous illumination, so would this be where the "missing" 200kgs are hiding and the ctr tank indication indicates zero early? Subsequently this fuel will become "available" at scavenging stage (approx 1900 per wing tank), but by then the diversion decision may have been made with a 4+ tonne reserves figure? :confused:

[BARKINGMAD]

More graphs later and it now transpires that the apparent decrease does manifest itself BEFORE the CTR pumps are turned off.

Latest expanded scale plot shows the "dip" started at 8400kgs, so back to the drawing board for me.

Curiously some of my RHS colleagues had observed this apparent dip, others who look at the FMS figure minutes before the waypoint and record it as such, have not noticed.

I really must get outa the cockpit more..............................................:confused: :confused: