Understanding Brake Specific Fuel Consumption (BSFC)
Brake specific fuel consumption measures how much fuel an engine burns to produce a given amount of power over a given time. It is the standard way engineers compare the efficiency of engines regardless of their size: a bigger engine will naturally burn more fuel in absolute terms, but BSFC normalizes that by the power it actually delivers.
The formula
BSFC is defined as the fuel mass flow rate divided by the brake power output:
- BSFC = ṁ / P, where ṁ is the fuel mass flow rate (mass of fuel consumed per hour) and P is the brake power measured at the crankshaft or dynamometer.
- In imperial units this is usually expressed in pounds per horsepower-hour (lb/hp·hr): fuel flow in lb/hr divided by power in hp.
- In metric units it is usually expressed in grams per kilowatt-hour (g/kWh): fuel flow in g/hr divided by power in kW.
- The two units convert directly: 1 lb/hp·hr ≈ 608.3 g/kWh, since 1 lb = 453.6 g and 1 hp = 0.7457 kW.
Typical BSFC values
Naturally aspirated gasoline engines commonly run around 0.45-0.60 lb/hp·hr at a steady operating point. Turbocharged or supercharged gasoline engines often run a little higher because of fuel enrichment used to control combustion temperatures under boost. Diesel engines are generally the most efficient, typically around 0.30-0.40 lb/hp·hr, thanks to their higher compression ratios and lean, unthrottled operation. These are rule-of-thumb ranges, not fixed limits — actual BSFC depends heavily on engine design, tuning, and the specific speed and load at which it is measured.
Why BSFC matters
Engine builders and tuners use BSFC to size fuel systems (injectors, pumps, fuel lines) for a target power level, to compare how efficiently different engines or tunes convert fuel into power, and to spot problems — a BSFC that is unexpectedly high for a given engine type often points to a rich air-fuel mixture, ignition timing that is off, or a mechanical issue. Because BSFC varies with engine speed and throttle position, meaningful comparisons should use readings taken at similar power and RPM.