"Old" GPM (1/MPG) equation
 

Don't know how relevant it might be to *current* vehicles, but back in 1961-1967 a Ford Motor Company engineer came up with this logarithmic-form equation for estimating fuel consumption (161 car-sample):

c = K*(wt^0.3067)*(cid^0.3469)*(ar^0.3395); R^2 = 0.932

where:
c = gallons-per-mile
K = 5.248 x 10^-4
wt = weight in lbs.
cid = cubic-inch displacement
ar = axle ratio

• source: SAE Automotive Fuel Economy, Vol. 15 (PT-15), 1976, "Factors Affecting Vehicle Fuel Economy," by Clayton LaPointe, Ford Motor Company, page 105.

Related to Louis LaPointe?

Looks like a typical empirical equation. It should be close if you were to build a gas powered car with 60's technology... Changes in technology will alter the constants in the equation.

It comes out to 26mpg for my Civic, which is likely close if it were carbureted, had 60's aero, and were on crappy old tires.

You could probably generate a new version of one that would be more accurate. Or, likely a few of them to deal with different engine cycles (gas, diesel, hybrid) and aspiration (NA/turbo).

equation
 

It's cool that in it's day,they'd have a numerical algorithm which could,within an acceptable margin of error,predict performance with such few 'hard points' of data.
My college text for internal combustion engines and air pollution had a road load estimator equation which presumed a constant Cd and rolling force coefficient for tires for all cars.It only asked for weight,frontal area,and velocity.

The "hidden agenda" reasons I had for posting this "old" equation:

1) Shows how only WT, CID and AR are needed to estimate Fuel Economy (it's basically all about engine displacement and rpms).

2) Shows type (logarithmic) equation (product-of-powers): c = K*a*b*c

3) Shows quasi-cubic-root "power" for each variable: ie: WT^(≈1/3), CID^(≈1/3), AR^(≈1/3).

4) To provide *another* way to visualize/quantify how FE can be quantified...ie: "put into numbers"


...or, "rule-of-thumb" roughly: mpg ≈ 1900/(WT*CID*AR)^(1/3)