Coastdown procedure to extract road loads

[murpia]

Can anyone supply a reference to a suitable coastdown procedure to determine road loads, specifically aerodynamic drag and tyre rolling resistance.

I have access to some data already but my own analysis so far drastically underestimates engine power at terminal speed so obviously I have made some errors.

There are a few SAE papers I know, but before I stump up the cash to purchase them I wondered if there were any good public domain references.

Regards, Ian

 

[DaveBidner]

Hi Ian,
The Bosch Automotive HB has something basic on calculating them, Fd=A+Bv+Cv^2. You could do a little equation solving Coast down 80 to 60 time and 30 to 20, two equation, three unknowns. I would set aero to zero at low speed and substitute that into high velocity. You'll have a lot of noise, grade, wind, basic measurements. If your resistance seems low then you have know how your engine/ transmission might be influencing it. Just thought, your terminal speed is also a good point, if you know the power and speed you have the drag force!
Good luck, let me know if the data cleans up.
Dave B

 

[Rob45]

Rather than calculate, why not just measure?

That's what the OEM's do, and it's fairly simple.
You coast down and plot time to speed.
Then you turn around and do the same thing in the opposite direction, hoping to minimize the effects of wind.
A sufficient number of tests and close attention to reducing measurement errors will give surprisingly good results with even a stopwatch.
Of course, you can't separate tire rolling resistance from aero drag by testing alone, and as well you have wheel rotational inertia to contend with, but that can also be measured by testing a wheel/tire assembly.

Fifth wheel assemblies are available to rent if your onboard speedometer isn't accurate enough, and you can automate the plotting of sped to time, minimizing the number of runs required.

 

[CentralCoaster]

I have an excel spreadsheet from school where we put in coastdown data to predict a Cd. It was hardly straightforward to get from real data to a useful Cd.

I did make a sheet to predict acceleration, aero limited topspeed, from a dyno curve. It was surprisingly accurate, but I'm struggling on predicting the launch. Maybe too low of resolution?

 

[Rob45]

It is in fact fairly basic physics to get from coastdown data to road load; getting from there to a "prediction" of drag coefficient might be another matter, but I'm curious as to why you need to separate out the various components of road load?
You know of course that the slope of the coasting curve at any given roadspeed, i.e., the deceleration at that speed, gives the roadload from F=ma.
This can be expressed as the roadload horsepower.

You should also factor in the angular acceleration of wheel/tire assemblies for a more precise result; this requires measuring the moment of inertia of a wheel/tire assembly, from which you can add an "equivalent mass" to the mass of the vehicle itself.
For cars up to about 3000 lbs, 150 lbs might be added for wheel inertia.

Now if you repeat the testing in reverse and do a series of acceleration tests, add the two HP figures, you've got rearwheel horsepower to a fairly accurate degree.