Calculating Mark V Fuel Flow from KFQLM1 and Flow Divider

Where the Missing Liters Come From

A liquid fuel flow divider is a positive-displacement machine. Fuel forced through it turns a gear set, and shaft speed is proportional to delivered volume per unit time. The control system never sees liters — it sees rpm from the divider speed pickups, normalizes that to percent of a rated speed, multiplies by a mass-flow scaling constant, and integrates. Every liter in the report is three readable numbers and one integration you cannot see.

That integration is where the shortfall lives. Counters, accumulators and timers are slow-voted values in the Mark V; the write into the accumulator memory area does not happen at the frame rate that executes the control sequence. Each update turns a sampled rate into an increment, and truncation in that increment only ever removes volume — it never adds it. Add the flow that passes during starts, stops and fuel transfers between updates and the on-board total sits consistently below a temperature-compensated meter. This is arithmetic, not a fault. The Mark V liquid fuel total was built for control and trending, not as a revenue-quality totalizer, and no constant tuning converts it into one.

Three Sources for a Liquid Fuel Total

For a plant report, take the second path. It uses the same primary signal as the accumulator but removes the one error you cannot influence — the write rate into accumulator memory — and puts the specific gravity in your hands where it belongs. Reserve the meter for anything that turns into money.

LIQ_GAIN is not part of that chain. The reporting path is PR_3_MAX → percent speed → KFQLM1 → mass → volume, and none of it needs the liquid fuel control gain. If the divider rating or the fuel density changed, re-derive the whole liquid fuel constant set by the frame-specific procedure in the Control Specification issued for that unit rather than back-solving a gain from a fuel report.

The Flow Divider Calculation

1. Read the flow divider rating: PR_3_MAX is the rpm at 100 % flow divider speed. Take it from the Control Specification or from CONST.SRC.
2. With the unit running on liquid fuel, read the flow divider speed and convert it to percent: %speed = N_fd / PR_3_MAX * 100.
3. Read KFQLM1 from the scaling file in use — METRIC.SCA carries kg/sec per %, ENGLISH.SCA carries lb/sec per %.
4. Compute mass flow: FQLM1 = KFQLM1 * %speed, in kg/s or lb/s.
5. Convert to volume with the actual specific gravity of the fuel, not a book value.
6. Integrate over the reporting window at your logging interval.

%speed        = (N_fd / PR_3_MAX) * 100
FQLM1         = KFQLM1 * %speed              ; kg/s or lb/s
FQLM1_kgs     = FQLM1_lbs * 0.45359          ; if using ENGLISH.SCA
mass_kg       = FQLM1_kgs * t_seconds
volume_L      = mass_kg / SG                 ; SG in kg/L; 0.83 -> 1.2048 L per kg
volume_USgal  = volume_L / 3.785412
US_gpm        = (FQLM1_lbs * 60) / (8.3454 * SG)

The number that matters before any of this is PR_3_MAX. If a divider was swapped for a different rating and the constant was never updated, every total scales by the ratio of the two ratings and no amount of specific-gravity work will close the gap.

US Gallons in Mark V Scaling

When the scaling is English, the gallon is the US gallon — 3.785412 L. A flow divider rating written in the Control Specification as "x gpm at y rpm" is US gpm as well. The imperial gallon is 4.54609 L, so treating one as the other moves the total by a factor of 1.201, a flat 20 % error that dwarfs everything the accumulator sampling costs you. If a report is off by roughly a fifth and the ratio holds at every load, stop looking at constants and look at the spreadsheet conversion.

A quick consistency test settles the question without waiting for a fuel delivery: take KFQLM1 in lb/sec/%, multiply by 100 to get lb/s at full divider speed, convert with US_gpm = lb/s * 60 / (8.3454 * SG), and compare the result against the gpm rating printed in the Control Specification. Agreement within a few percent confirms both the constant and the gallon. A 20 % gap names the unit error outright.

Specific Gravity as the Dominant Error Term

Mass-to-volume conversion propagates density error one-for-one: a 1 % error in specific gravity is a 1 % error in liters, every hour, in the same direction. Diesel at SG 0.83 gives 1.2048 L per kg, but the SG stamped on a data sheet is a reference-temperature value, and fuel at the forwarding pump discharge on a hot day is less dense than that. The Mark V applies no volume correction; a custody meter applies one, using the correction tables in ASTM D1250 / API MPMS Ch. 11.1. Two instruments correcting differently will never agree, so log fuel temperature alongside the totals and use the SG at that temperature in the calculation.

Divider behavior moves the same direction. Positive-displacement dividers leak internally, and slip rises as viscosity falls with temperature and as differential pressure across the divider rises. At low flow and hot fuel the indicated speed under-reads the volume actually passed, which shows up as a deficit that worsens at part load rather than a constant percentage.

Verification and Error Signatures

1. Log flow divider speed at one-minute intervals from the historian or the operator display.
2. Compute FQLM1 per sample and integrate to a volume using the measured SG.
3. Record the Mark V accumulator at the start and end of the window and take the delta.
4. Take the meter delta over the identical window, plus fuel temperature at both ends.
5. Compare all three. Steady flow, hand calculation and accumulator within a few percent means the constants are correct and the residual is sampling.

Stop when the hand calculation and the accumulator agree at steady flow but both sit outside the meter by a stable offset — at that point the discrepancy is in the fuel system topology or the meter's temperature correction, not in the control constants, and it is a mechanical and fuel-accounting problem. Escalate to GE technical support with the Control Specification, CONST.SRC and the scaling file when PR_3_MAX or KFQLM1 do not match the installed flow divider nameplate, or when the liquid fuel constant set was never re-derived after a divider or fuel change; those are unit-specific issued values and the derivation for your frame belongs with them.

FAQ