Measuring Fuel Economy by Weighing the Car
 

Is this a good methodology for accurately measuring fuel economy? | TopGear.com.ph

Just wondering what everyone thinks.

The background... it was an economy test of around two dozen vehicles (only 19 showed up, but, whatevs). Previous testing had the technical crew spend two days filling the car before the test, just to make sure it was completely topped off.

Due to time constraints, this is what was done, instead.

Sounds stupid. Worked. I think. But still, a lot of questions. :p

Get a new technical committee. The way we do it here is better.

Getting to zero, filling, measuring the fill, then driving, then measuring what's left can be accurate. But every step is at best some added slop and a chance to get something wrong. At worst? A kindergartener could cheat and not get caught.

Better? Fill up, drive, then fill up again with as few variables as possible- same pump, conditions, method, etc. Measure fuel pumped on the refill.

You need a very precise scale and no fluid outflow from any driver or passenger. No adding of any weight during the test cycle (buy a soda or other drink-food).

Under those parameters I would prefer the weighing method.

regards
mech

Get on a Cat Scale and weigh at all four corners to be exact about tire pressures, plus correct any imbalances. Past that?

Scales were motorsports grade, unfortunately, they don't read to decimal places, even though they're accurate to fractions of a pound. I suggested correcting to whole numbers using small weights, but they didn't want to take the extra step... even though I tried to explain how quick and easy it would be.

Nothing in the cabin. No water, no snacks. No AC use, so water wouldn't condense around the AC evaporator, either. If the car got lighter due to evaporation from the radiator, then that's not our problem. :D

The first run was done that way. We needed several fills both pre- and post- to get exact numbers. Something that would be hard to do with two dozen cars.

-

The method I've been most satisfied with was doing multiple fill-ups over the course of the run... then you get the median fill-up, as the first few tend to be off due to issues with the tank.

-

Good if you don't have a line at the pumps. Last eco-run we had, the time spent waiting for the fill-up affected the numbers... which is how two diesel crossovers got 40-50 km/l on countback... because their high pressure fuel rails took a dump back into the tank while they were sitting parked.

-

Personally, I would love to do this with a fuel cell. Or even a plastic bottle or bag like Mythbusters uses. I've seen a run done with a bottle mounted to the hood on a modern diesel... so it's possible, even with a high pressure fuel rail, but that requires a modified hood, so not practical to do with a lot of cars at once.

Trying to calculate fuel consumption by weighing the difference of the entire vehicle is terribly inaccurate. Especially if only a few gallons are used. The Green Grand Prix used to rely on this using 4 race quality scales. Some of the cars actually made fuel.

Now that's interesting! And the kind of anecdotal data that would be useful to know. Have any links?

Also... What kind of run was it, and what were the parameters?

We avoided rain... all the cars were parked in the pit garages before the event, too.

Reading up... seems like a big issue with such small figures... 10 pounds versus 1 gallon is a pretty big level of error.

-

The big problem was the fuel sponsor would not allow refueling at the track... at all, which was maddening. In an ideal world, both methods would be used to audit each other.

So... I guess that leaves removing a washer nozzle, running a hose out to a bottle of fuel suction cupped to the hood, and running until the bottle is dry? :D

I think this would also be an issue for people who drive in areas where it snows a lot. I'm curious sometimes to how much mass my car picks up in slush and snow just on the drive to work each day.

Well this is only something that would be done in competition or one time test and is not even any good for that.

It's rather ridiculous to measure a small weight difference by measuring 2 heavy objects then subtracting one weight from the other.

True enough. But once you control for extraneous weight, it works... but with limitations.

Best really is to use fuel cells and weigh those, but that's a difficult ask with cars you can't modify.

I have a few questions... was the pre-comp weight taken the day before or right before each run?

Perhaps it was mentioned later in the thread, but I didn't catch it.

Why even consider this?

If a stupid plan is the best you're allowed to use, then that's the plan you use.

As said, and as cited by Charlie, we were prohibited from filling on-site by the sponsor. (the other year's sponsor allowed it, though)

-

All methods are inaccurate. What's important here is deciphering to what degree they are inaccurate, and why.

I've seen some wonky fill-ups in my time (and yes, the gas-bladder Prius was one of them).

I've seen cars take a gallon over "full" on fill-up, another quart or two after sitting for an hour, and then another half gallon the next day.

I've seen cars magically get a bonus of nearly a gallon at the pumps, even with all the controls, simply because they sat in line too long waiting for their turn to fill. And this is at an event organized by an old hand, who knew all the tricks and cheats. His next event closed some of those loopholes.

By and large, if done right, a fill-up can be accurate down to half-a-liter. But when you're measuring road cars, there are always jokers. If you've ever seen the obscenely complicated assembly that is a Honda Fit tank and filler tube...

-

We all know weight is inaccurate, but the question is why and how bad can it be?

If the problems at the race were due to rain, we can (and did) control for that.

If they were due to moisture condensing in the car due to AC use and sweaty drivers, we can control for that.

If they were due to overnight condensation, we can try to control for that.

If the natural variation in weight due to ambient humidity at the event (sans rain), or the natural loss of rubber, or even dust collecting in the sills over 100 km... if any of those is big enough to bugger the scales, then we can scrap the idea, altogether.

-

My personal preference is to go with manual fill-ups, with the car on crocodile jacks to eliminate most of the air bubbles, audited by weight, both times, using weights to round off the weight to the nearest half-pound.

I'd need evidence to show that this is necessary... and to call for re-validation of the method. (I didn't do the validation tests, and I have issues with them)

In an event where the best competitors only use 1 gallonUS/ 6.16 pounds of gas, an error on the scales of only 1 pound out of a 3,000 pound car makes a difference of 16% calculated fuel consumption. Weighing the whole car to determine how much fuel was used in a three hour event doesn't work.