A 20% reduction in drag = a 10% increase in fuel economy ?

[aerohead]

Originally Posted by UltArc

The math I am not sure about, but it seems to be a quadratic equation rather than linear.

I do not THINK it is as easy to compute as that. I am not an expert, nor do I know too much at all about the subject, I just can't believe it to be so easy.

General Motors came up with the Cd vs mpg relationship and published in an SAE Paper.
As of 1990,the GM Aero Lab said the relationship was still valid.
It does assume a constant BSFC and is not intended to supersede wind tunnel testing or official coastdowns,but it can be used with high confidence.

[aerohead]

Originally Posted by KamperBob

Your math misses other variables you don't know: rolling resistance of tires, engine mods, driving habits, etc. A-B testing risks time correlation. A-B-A testing rechecks baseline. However, once a car is supermodded another "A" run isn't very likely. That's not necessarily a bad thing. Real gains are sticky. Why undo successful mods to prove they helped? Human nature... (shrug)

The relationship is dependent on a smooth,level road.No curves,no wind,no rain,and a constant velocity.
It's a tough trick to pull off.The best I can do is to compare long trip data from year to year with varying mods,and document everything I safely can without crashing.
Man,that A2 Wind Tunnel is looking mighty fine!

[jime57]

Originally Posted by aerohead

The relationship is dependent on a smooth,level road.No curves,no wind,no rain,and a constant velocity.
It's a tough trick to pull off.

Yes it is a tough trick. But it isn't hopeless. I found I could get pretty tight baseline runs by running two directions(nulls wind), constant ave. speed via Scangauge, hills become insignificant if the baseline and change are done back-to-back, and driver variability "smooths" out if the course is long, say 10 miles or more.

Takes lots of time and fuel, unfortunately.

A cruise control would help maintain the constant ave. speed, but unfortunately that doesn't work very well on the MT Insight

[Cd]

Originally Posted by MetroMPG

Also, why did you choose 45 MPG as the "original" 65 MPH fuel consumption figure? There's no rule that says the EPA figure necessarily equals MPG @ 65 MPH. Depending on the car, real world MPG at a steady 65 could be higher or lower than the highway rating.

I chose 45 MPG as the original figure because i seem to remember that is what a manual 1992 Civic CX hatchback is rated at on the highway, and that is the number that i seem to remember Bassjoos said was his baseline before the mods.

Fuel Economy of the 1992 Honda Civic

[Cd]

Originally Posted by Diesel_Dave

It's all speed dependent. I found these graphs from a Cummins white paper. It's geared towards semi trucks, but it illustrates the principle.

This second plot (from the same white paper) shows % mpg benefit vs % aero improvement for different duty cycles:

It looks as if a 20% decrease in drag equals just over 12 % increase in FE ?

So would a small car have more to gain, or does everything scale up or ...

[freebeard]

My guess would be that the the original '20% gets 10%' conjecture is a 'rule of thumb', based on the total resistance being half aero and half rolling resistance; but with rolling resistance being linear and aero being exponential—not always accurate in all cases.

Edit: So I see it's not that I'm a slow typist, there were three pages to the thread, not one.

So I'm going to fall back to Your Mileage May Vary.

[aerohead]

The source of the CdA/mpg relationship is:
"Aerodynamics for Body Engineers"
by Kent B. Kelly and Harry J.Holcombe
General Motors Styling Staff,General Motors Corporation
SAE Paper No. 649A,January 1963,at the Automotive Engineering Congress
SAE Transactions,Volume 72,1964,page 571
"... a 10% reduction in aerodynamic drag makes a 5% reduction in fuel consumption possible at 55 mph,and a 6% reduction of fuel consumption possible at 70 mph."
--------------------------------------------------------------------------
This relationship has absolutely nothing to do with the EPA.
It is for constant-speed top-gear cruising on a straight,level,dry,paved road with zero wind.
Typically,a small 'test' tank is in parallel with the main fuel tank,with a transfer valve operated by the test driver(s).As the car enters the test section of the 'track' the fuel valve is diverted over to the smaller test tank for the duration of the test period,then switched back to the main tank upon conclusion of the test.
By measuring the mass of fuel,before and after the runs,and knowing the constant velocity and distance,the mpg can be computed on a mass/volume basis.
-------------------------------------------------------------------------
Today,with Scanguage and other 'instantaneous' mpg displays,anyone who can access a good stretch of highway ought to be able to construct a complete velocity/mpg baseline for future comparisons to deduce drag coefficients after mods.

[Vekke]

If I remenber correctly you also need to make your gearing same amount taller that you can save fuel on your aero mods. Secret for this relates that BSFC maps... Other option is to use P&G technigue and gearing is not that big problem.

[aerohead]

Originally Posted by Vekke

If I remenber correctly you also need to make your gearing same amount taller that you can save fuel on your aero mods. Secret for this relates that BSFC maps... Other option is to use P&G technigue and gearing is not that big problem.

Yes,both Gino Sovran,and Wolf Hucho emphasize proper gear-matching after streamlining in order to maintain the same load as before.It protects the engine from over-speeding,and keeps the engine operating on the same 'island' of it's BSFC map,yielding the biggest bang for the buck.
Sovran claimed that up to 60% of the streamlining benefit to mpg could be lost without it.
Carl Breer's streamlining team ran into this with the Cd 0.244 Chrysler DeSoto Airflow test mule.
I do believe that modern EFI technology may mitigate some of the loss,as manifold vacuum no longer controls fuel metering as in the carbureted-engine-cars with which early research was conducted.