Fuel consumption and fuel strategy per stint
Know your fuel per lap to the decilitre, validate the ECU figure against measured fuel by mass and temperature, and turn it into laps per tank, start fuel, stint lengths and fuel-saving targets.
Fuel consumption and fuel strategy per stint: L1 Foundation explanation
Three questions before every race (Segers ch. 16):
- How much fuel is used per lap? (Racing thinks in litres or kg per lap, not L/100 km.)
- How many laps can be done on a tank? That sets the minimum number of stops and the pit window in refuelling races – and the start fuel in sprint races.
- What is the lap-time cost of fuel mass, and how does balance change as fuel burns (CoG drops; if the tank isn't central, weight distribution shifts)?
How fuel use is measured:
- Tank level sensor – simple, but sloshing and tank shape make it imprecise.
- Flow meters – supply minus return flow; must be temperature-compensated because returning fuel is warmer and less dense.
- ECU calculation – sums injector opening time × injector flow rate. Requires a scaling factor that you calibrate.
Basic stint arithmetic:
Example: 96 L tank, 3.57 L/lap → 26.9 laps; you plan 26 (Segers' GT1 example). In a sprint race without refuelling, start fuel = race laps × consumption + formation/out laps + safety margin + the fuel sample the scrutineers require (often ~1 L) + unusable fuel.
Typical consumption: GT3 roughly 2.5–4 L per lap on a 4–5 km circuit; lighter single-seaters much less. Always measure at your event.
Fuel consumption and fuel strategy per stint: L2 Practitioner explanation
Validate the ECU fuel figure by mass, not volume. Fuel expands roughly 0.1 % per °C. Fuel put in cold from a drum and drained hot after a session differ in volume even at equal mass. Weigh fuel in and out (scale accurate to ~±0.01 kg) and record its temperature. The volume at reference temperature is simply ; the volume the fuel actually occupies at its temperature is
with ≈ 0.72–0.76 kg/L for race gasolines (check the supplier's sheet) and /°C.
Worked example (method after Segers §16.1, new numbers): kg/L, 30 laps including in- and out-laps.
- In before the session: 60.0 kg at 25 °C → 80.0 L × 1.010 = 80.8 L
- In during the session: 45.0 kg at 30 °C → 60.0 L × 1.015 = 60.9 L
- Drained after: 25.0 kg at 40 °C → 33.3 L × 1.025 = 34.2 L
Used: 80.0 kg (2.67 kg/lap) or 80.8 + 60.9 − 34.2 = 107.5 L at the fuel temperatures (3.58 L/lap); at 15 °C the same mass is 106.7 L. If the ECU reported 106.0 L, correct the injector scalar:
Compare like with like: the injector model integrates volume at the temperature of the fuel in the rail, so compare it with temperature-corrected volumes (as Segers does). If your ECU reports mass or a reference-density volume, compare with the weighed mass or instead – here the choice of reference alone is a 0.8 % difference. Once calibrated, use the ECU's lap-by-lap figure (which excludes in/out laps) rather than session totals.
Fuel effect on lap time. From lap simulation or back-to-back runs, on a 4–5 km circuit: GT3/touring cars (1250–1400 kg) ≈ 0.015–0.025 s per kg per lap; lighter single-seaters (F3/F2/F1-type) ≈ 0.025–0.035 s/kg. It scales roughly with lap length. Segers' touring-car simulation gives 0.014 s/kg on a ~103 s lap. For a GT3, 60 kg of fuel ≈ 0.9–1.5 s per lap at the start of a stint.
Fuel saving tools: lift-and-coast before braking (most efficient – saves fuel where the car would decelerate anyway, costing little time), short-shifting, leaner/lower-power maps where allowed. Express saving targets as a per-lap number on the dash ("target 3.45 L, delta −0.05").
Running out of fuel isn't the only risk. Low fuel causes pickup starvation under sustained lateral or longitudinal g (surge) if the collector isn't full – the last few litres may be unusable at high g. Know your car's minimum safe level.
Fuel consumption and fuel strategy per stint: L3 Advanced explanation
Optimising fuel load. Every kilogram carried costs time on every lap until it's burned. For a stint of laps with burn (kg/lap) and starting reserve , the time cost of the reserve alone is . 5 kg extra over a 30-lap GT3 stint at 0.02 s/kg/lap = 3 s – real track position. Margins should therefore be set from measured consumption scatter, not guesswork. Random lap-to-lap scatter partly cancels: the margin for it is ≈ , not (30 laps, σ = 0.05 L/lap → 0.55 L, not 3.0 L). Systematic uncertainty (ECU calibration error, a hotter day, a different map) adds linearly: . Then add fixed allowances for formation and safety-car laps (SC laps burn roughly 40–60 % of a racing lap).
Fuel saving vs lap time. Define the marginal cost of saving: (seconds lost per kg saved). Lift-and-coast typically has the lowest λ; short-shifting next; lean maps vary. Each tool has a limited capacity per lap, so use them in order of λ until each is exhausted. Within a tool, apply it in the braking zones where it is cheapest, not uniformly round the lap.
When to save. An SC gives free saving, so with a high SC probability it pays to defer part of the saving and keep the option. In a green race, spread the saving over many laps rather than bunching it at the end: the time cost per kg rises steeply with lift length (λ grows with the amount saved per lap), so the cheapest plan has equal marginal cost on every lap, and bunching also risks missing the target. Saving early means carrying the already-saved fuel for longer, which costs only per lap – a second-order effect that, if anything, slightly favours saving late. Strategy tools model expected value across SC scenarios.
Energy-limited regulations. In WEC Hypercar (and LMGT3), BoP defines an energy allowance per stint (virtual energy tank, MJ) plus power limits; the "fuel" problem becomes an energy-per-lap target measured by homologated sensors. Exceeding the stint energy is penalised, so teams manage energy per lap via lift-and-coast, deployment and maps. IndyCar fuel races (ovals) use fuel numbers per lap with drafting effects – running in a tow reduces consumption significantly.
Calibration hygiene: recalibrate after injector, fuel-pressure or map changes; check fuel density for each fuel batch; record fuel temperature at refuel in hot conditions (fuel left in the sun).
Fuel consumption and fuel strategy per stint: L4 Expert explanation
Timed races: the lap count is itself uncertain. In a race run to a clock, the leader takes the flag on the first crossing after time expires, so
A 60-minute race at a 1:50.0 average gives 3600/110 = 32.7 → 33 laps. At 1:49.0 it gives 33.03 → 34 laps. One second per lap of leader pace near an integer boundary adds a whole lap of fuel, so fuel for the fastest credible leader pace, not your own. If you will be lapped you run one lap fewer; you cannot count on it. SC laps lower the lap count but each still burns ~50 % of a racing lap, so model the fuel and the lap count together across SC scenarios.
The fuel-saving optimum is a convex problem. Let the lap-time cost of saving kg on lap be , convex because longer lifts start at higher speed and cost more per kg. Minimise subject to . The Lagrange condition is on every lap: equal marginal cost everywhere, i.e. equal saving per lap if the laps are alike.
Illustration with (s, x in kg): saving 2.0 kg over 20 laps at 0.1 kg/lap costs 20 × 0.05 = 1.0 s; saving it in the last 5 laps at 0.4 kg/lap costs 5 × 0.8 = 4.0 s. The mass effect works the other way but is small: at 0.02 s/kg, carrying the already-saved fuel for longer costs only ~0.3 s here.
Option value of an SC. If deferring the saving means an SC will make it free with probability , and otherwise you must bunch it, deferral wins only if , i.e. . The practical answer is usually a hedge: save a baseline rate now and keep part of the target in reserve for the SC that may come.
Limits of the model. is not identical per lap (traffic, tow, tyre state). The tow effect can be larger than any tool: running in a slipstream can cut consumption by several percent, which is why IndyCar oval fuel races are won in the draft. In energy-limited classes the constraint is MJ per stint measured by the organiser's sensors, but the same equal-marginal-cost logic applies.
Key formulas
- usable tank volume (L)
- safety margin + sample (L)
- consumption per lap (L/lap)
- m
- fuel mass (kg)
- density at 15 °C (kg/L)
- expansion coefficient (≈0.001 /°C)
- T
- fuel temperature (°C)
- sensitivity (s/kg per lap), 4–5 km lap: GT3 ≈ 0.015–0.025, single-seaters ≈ 0.025–0.035
- fuel mass on board (kg)
- lap-to-lap consumption scatter (kg or L)
- N
- laps
- b
- possible systematic bias per lap
Fuel & stint planner
Plan stints, pit stops and start fuel for a race, and find the fuel saving per lap needed to drop one stop.
| Stint | Laps | Fuel used [L] | Fuel added before [L] | Refuel time [s] |
|---|---|---|---|---|
| 1 | 31 | 111.6 | 113.6 (start) | — |
| 2 | 30 | 108.0 | 108.0 | 36.0 |
| 3 | 30 | 108.0 | 108.0 | 36.0 |
Stints split evenly; start fuel is the first stint plus margin (lighter car), each stop adds the next stint's fuel (car arrives with the margin left). Pit-lane time loss and tyre changes are not included.
Trackside takeaways
Calibrate the ECU fuel figure against fuel weighed in and out, with temperature correction.
Fuel expands ~0.1 %/°C – never trust volume-only measurements with hot drained fuel.
Every extra kg costs ~0.015–0.025 s per lap in a GT3 (~0.03 s in a single-seater) until burned; set margins from measured scatter.
Lift-and-coast is usually the cheapest way to save fuel.
Remember the fuel sample, formation laps and unusable fuel at high g.
On your car
Endurance (GTWC Endurance, IMSA, 24h races): fuel per lap sets stint length and pit windows; SRO/IMSA BoP can define tank capacity or refuelling time. Sprint races: start fuel minimised for pace.
Check yourself
A car uses 3.4 L per lap and has a 100 L tank. You need a 1.5 L reserve for margin and the fuel sample. How many full racing laps can you plan per tank?
- Segers – Analysis Techniques for Racecar Data Acquisition — §16.1 Fuel Consumption
- Segers – Analysis Techniques for Racecar Data Acquisition — §16.2.1 Fuel Load Variation
- Segers – Analysis Techniques for Racecar Data Acquisition — §15.4 A Worked out Example · lap time vs vehicle weight