Estimating horsepower from airflow

An engine is an air pump. It makes power in proportion to the mass of air it can move, because fuel is only useful in the ratio the air allows. Every honest power estimate is really an airflow estimate with a conversion on the end.

air mass flow = VE × displacement × (rpm ÷ 2) × charge density
fuel flow     = air flow ÷ AFR
power         = fuel flow ÷ BSFC

The rpm ÷ 2 is because a four-stroke fills its swept volume once every two revolutions. Everything else is a number you can look up, measure or argue about — which is the point.

The four things you are really guessing

Volumetric efficiency is how completely a cylinder fills, as a fraction of its swept volume. A stock SOHC is around 0.82; a DOHC VTEC on its second cam lobe reaches about 0.95; a well-developed race engine with individual throttle bodies exceeds 1.0, because a tuned intake arrives with momentum. This is the number that head work, cams and manifolds actually change, and it is where every “port work = ×1.10” multiplier was really pointing.

Charge density is where boost and weather live, and it is just the ideal gas law: ρ = P ÷ (R × T). More pressure is more air; more heat is less. Both matter and only one of them is on a boost gauge.

AFR, around 12.8:1 naturally aspirated at full throttle and 11.8:1 on boost, because a boosted engine is run rich to keep it alive.

BSFC — brake specific fuel consumption — is how much fuel it takes to make a horsepower for an hour: about 0.48 lb/hp/hr naturally aspirated, 0.55–0.60 boosted. Lower is more efficient.

Get those four roughly right and the answer lands within about ten percent, which is as much as any closed-form estimate can honestly claim.

Why boost is not a multiplier

The tempting shortcut is that ten psi on a 14.7 psi day is a pressure ratio of about 1.68, so power goes up 68%. It does not, because compressing air heats it:

T₂ = T₁ × (1 + (PR^((γ−1)/γ) − 1) ÷ η)

At a pressure ratio of 2, from 20 °C, with a compressor running at 72% efficiency, the charge leaves at about 109 °C. Hot air is thin, so a good part of the pressure gain is handed straight back. That is the entire argument for an intercooler, and it is why intercooler effectiveness is an input here rather than an assumption — the difference between none and a good air-to-air is worth more than fifteen percent.

Compressor efficiency is also the honest version of “turbo sizing”. A turbo’s size does not multiply power at a given boost; it decides where in the rev range you get it, and a well-matched compressor helps only by running cooler. That effect is real and small, and it is already in the equation above as η.

Why altitude is not a flat percentage

The rule of thumb is three percent per thousand feet, and for a naturally aspirated engine it is about right: a mile up, ambient pressure is 83.4 kPa against 101.3 at sea level, so the engine loses 17.7% — 3.35% per thousand feet.

A turbocharged engine is a different case, and this is the one the old calculator got backwards by applying its altitude penalty after boost. A turbo compensates: holding the same gauge boost, the same engine loses about 9.9% at that altitude rather than 17.7%, because it is adding pressure on top of the thin air rather than starting from it. Set up to target absolute manifold pressure it loses less still — at the cost of a compressor working harder, running further up its map, and making more heat.

The quarter mile

Two standard equations, both taking flywheel power and weight in pounds including the driver:

ET  = 5.825 × (weight ÷ power)^⅓
mph = 234   × (power ÷ weight)^⅓

Feeding them wheel horsepower is a common slip and makes a car look about half a second slower than it is. They also assume the car can put its power down; no closed-form equation knows about your tyres.

On the rule of thumb

The calculator this replaces estimated power as displacement ÷ 10, adjusted by rpm and a set of multipliers. For the engine it was fitted to — a stock naturally aspirated Honda VTEC at 7,600 rpm — it is genuinely close, because those engines really do make about 100 hp per litre. A 1,797 cc B18C1 gives 179.7 by that rule against a published 170.

It has nothing to say about a 40 °C day, a missing intercooler, a compressor off its map, or a head that flows differently from the one it was calibrated against. The calculator below shows both numbers, so you can see where they agree and where they stop.