formula79

There is a lot of the Regal in the new Malibu

AdioSS

I know that the City Fuel Economy rating includes accelerating at a set rate & other changes in driving, but I am not sure how the highway MGP rating is tested. Do they set the cruise control at a specific speed & measure the fuel consumption? Does it include changes in speed to compensate for different conditions? At what speed is the highway MPG figure rated?

It is good to see a Camaro that is officially rated at 30MPG. But you just know that they would have loved to have introduced the car with that number instead of 29. I would have expected the Malibu Eco to get a 40mpg or better rating, but this car looks like the frontal area was increased over the previous model due to increases in width & height. The Cd is lower, but it punches a bigger hole in the wind.

My Dad's LS2/4L65E/3.15 performance geared C6 is able to get 30MPG with the cruise set to 65 on a flat road, but was rated less than that. The standard 4 speed auto cars were equipped with 2.73 gears and the newer 6 speed autos start off with 2.56 gears, but the fuel economy rating went down with the larger LS3 engine.

Threxx

Pretty much all cars are capable of exceeding their EPA ratings depending on how they're driven. I've found some cars are more readily able to exceed those ratings than others, though.

Click on 'highway' and 'high speed' on this chart to see what the EPA's test cycle is like:
http://www.fueleconomy.gov/feg/fe_test_schedules.shtml

I really wish they'd have something like a set 75 mph cruise test where they just set it on cruise control over flat ground and see what the result is. That'd be a lot more indicative of real highway 'road trip' driving.

I recently learned that the EPA has manufacturers do their own testing per the EPA's guidelines, and only double-checks the manufacturers with EPA in-house testing around 10 to 15% of the time. That surprises me. Sure seems like manufacturers would be doing anything they could possibly think of to swing results a bit in their factor.

Z28Wilson

The problem is that unless you're road-tripping through Nebraska every time you're not going to cruise hundreds of miles over flat ground either.

When people say "real world" driving it is still open to a lot of interpretation, since the world you drive in and mine may still be a lot different.

The EPA testing methods are still far from perfect but I think they are a bit better than they were. I agree with you though that when it comes to highway ratings most people are interested in how efficient a car is 5-10 mph above posted highway speed limits, on flat or slightly undulating roads.

Threxx

When I first wrote what you replied to I wanted to write something like 'on slightly undulating roads' like you were talking about, but then I realized for nationally standardized testing purposes, that's just not a realistic request, unless the EPA were to come up with their own test course and stop asking manufacturers to do most of their own EPA ratings. So that's why I was talking about flat ground just since I guess that would be easier to standardize at dissimilar testing grounds.

Z28Wilson

Agreed. The most important thing is that every manufacturer is testing the same way, which it seems like the new standards require them to, and then people need to understand that they may do slightly better (or slightly worse) than the indicated mileage. You're never going to get every manufacturer to give every one of their models to the EPA to test.

Threxx

Why is that? I'm seriously curious since until today when I read the EPA's site say they only verify 10-15% of vehicles, I was under the impression that they actually did test all vehicles.

It just doesn't seem like, logistically, it would be that difficult of a requirement.

"If you sell a vehicle in the US, you have to lend one of each powertrain configuration to the EPA for testing."

Honestly I feel like I, personally, could run a much more useful operation for fuel economy testing than the EPA currently is.

Koz

I work in the emission/fuel economy lab at GM, which (like you mentioned in an earlier post) is where General Motors does all the emission/fuel economy testing according to EPA, CARB, European, and Japanese regulations (we can certify a car for anywhere in the world in our lab).

EPA does do random confirmation testing, and if your results meet certain criteria it becomes automatic confirmation testing (not random). These criteria include major powertrain/technology changes to your product, or class-leading results. When we certified the Volt last year, it was immediately called into the EPA lab for confirmation testing. Your results must agree with EPA results, and EPA results are used if there's a disagreement. A manufacturer can not legally sell a car in the United States unless it is tested according to EPA/CARB requirements. Monetary penalties would be huge, on a per car basis, if rule-breaking/noncompliance were discovered. Even jail-time is a possibility. Emissions compliance is federal law, so it's not something a large company takes lightly.

As far as testing every car from every manufacturer, sure it makes common sense, but it's not as simple/straightforward as you'd imagine. Just for fun, take a guess as to how long you think it takes to run the emissions & fuel economy tests on a single vehicle.
















~7 days for a single vehicle. 7 days the vehicle must be inside the lab (controlled environment within temperature and humidity limits), either on the test site, hooked up to a piece of test equipment, or in climate controlled soak areas in-between tests. This doesn't include any preparation, this is 7 days of time that it's on the lab floor going through the certification process. Our lab runs 3 shifts, 5 days a week (7 days a week more often than not) to help with throughput.

And it's not just new cars that you have to test. Every new model year has to be certified before it can be sold in the U.S.; even if nothing has changed from the year before.

Granted, you can test multiple cars at the same time if you have multiple test sites, but when it's all said and done, the equipment required to test a single car costs millions of dollars (I'd guess ~10M for the equipment to test one car at our lab), so depending on your pocketbook, you're only going to be able to have enough equipment to test X number of cars at any time. And the equipment must undergo monthly, quarterly, semi-annual, and annual calibrations according to federal regulations; so there's mandatory 'down-time' while everything gets checked out that'll hurt test throughput too.

It's a big process, and since it's federal law, the only way to change something (for the better or worse) is by an act of congress. That being said, between new vehicle technologies and new regulations the testing is always becoming more and more complex and advanced. It makes for some interesting engineering projects

Threxx

Great post. Thanks for the info.

Do they actually run these cars on the road for these mileage ratings? If not, how do they appropriately adjust for road and air friction?

Koz

Official fuel economy and emissions testing is done in the lab on a chassis dynamometer, but there are companies that make portable emissions sampling equipment that you can hook up in the bed of a truck, or in a small trailer. These aren't legal for use for official certification for light duty vehicles; but I think they are for heavy duty stuff (like Caterpillar, maybe some semi-truck type stuff, off-road equipment, etc. The standards and driving/operating cycles for heavy duty stuff is completely different than light-duty passenger vehicles though). Some of the engineers that work on the HD trucks at GM use this portable equipment for road trips, but as far as I know GM "HD" trucks like the 1500-3500 series are still considered 'light duty' in regulatory terms, so they're tested to the same drive cycles that you posted earlier.

(edit) I asked around today and it turns out there is some regulatory testing that requires the portable emissions measurement systems for our diesel trucks. There's some sort of high-load conditions (like trailer-towing) during which the trucks aren't to exceed certain values of NOx emissions. They do this testing on road with the portable stuff. Everything else though is done in the lab. (/edit)

Even though the testing is done in a lab, the dynos are programed to simulate real-world road load. If you think about it, in terms of a test robustness you're much better off testing on a dyno where things are controlled and repeatable versus testing out on a road where you're at the mercy of the weather (temperature, humidity), traffic, road conditions (pot holes, ice, grade, whatever). So the emission dynos are able to reproduce accurate road loading for the vehicle, but in an accurate and repeatable manner. There are two SAE procedures that are followed, the first one, J2263, outlines how to measure the roadload acting on a vehicle when it's drivong on a road. It's recommended that a long, straight, smooth, level track is used. Roadload (vehicle drag) measurements are taken and corrections are made for headwind/yaw, air drag (depending on air temperature and humidity), road grade, etc. The outcome of this procedure is knowing the roadload that the vehicle experiences as a function of velocity (roadload or drag changes as you change speed). If you plot road load verses speed you get a somewhat quadratic looking curve, low roadload at low speeds, increasing exponentially as speed increases.

The next SAE procedure, J2264, outlines how to measure the load that the vehicle is subjected to when it's put on a chassis dynamometer. This 'lab load' is less than what's on the track since there's no air drag, much less inertia (vehicle doesn't move, but drivetrain & dyno both rotate so there is rotational inertia), different friction at tire contact patch, etc.. The outcome of this procudure is knowing the load that the vehicle sees when it's strapped to a dyno.

Once you know the real-world road load from the 2263 procedure and the laboratory loading conditions from the 2264 procedure (which is less than the real-world loading), all you need to do is figure out the difference between what you measured in the real world, and what you have in the lab. Once you know the difference, you just program the dynamometer to compensate for that difference. When you're done you have lab loading + dyno simulation = real world loading.

Lab loading is primarily vehicle & dyno parasitics. Drivetrain friction, bearing friction, etc. The dyno simulation compensates for air drag, tire deformation effects (tire contact patch is different on a dyno roller than on a flat road), linear inertia, etc. Things that aren't present once you install a vehicle in a test site, but are present in the real world.

When it's all said and done, you have to be able to reproduce the same loads that the vehicle would see out in the real world to within 2lbs at any speed between ~10-80mph. Our dynos are accurate to 0.67lbs. Unfortunately they don't have the necessary software to produce HP/torque curves

There are also non-certification tests that get used for general development/driveability type tests. These non-certification tests include grade simulation, going uphill and downhill, etc. The dynos can simulate grade, towing a trailer, all kinds of stuff. It's pretty neat.

96_Camaro_B4C

This thread should be stickied because of Koz's replies above.

Threxx

Wow. Again, great info. Thanks!

I'm beginning to understand why OEMs might prefer to do this testing in-house. It probably gives them a lot of great data to make their own tweaks to the final product to balance out MPG ratings and general driveability/durability.

Whereas if the EPA did all the testing, OEMs would still probably want to buy all that testing equipment anyway so they could tweak their final product before handing it over to the EPA. So I guess in the end the EPA is just reducing their costs by doing this semi-random testing and letting the OEMs do the rest.

It seems to me, though, that with these crazy complex measurements they do, that it wouldn't be that hard for the EPA to request extra data, rather than use these composite test cycle scores. Maybe not for stuff that would show up on the window sticker, but why not just ask for data for 'steady 75 mph on flat terrain, xyz weather with no wind'? Seems like that sort of rating could be really useful to car shoppers looking at the EPA website or GM.com or whatever, even if it's too much info for a window sticker.

Koz

This is exactly what most OEMs think. The cost is extremely high, but it's definitely worth knowing what numbers you'll get from EPA if they test it. Most of the engineers who get their cars tested at the lab are calibration engineers who impact emissions or fuel economy, so they test their cars as they're developing new engine and transmission calibrations, or new hybrid battery management calibrations, or start-stop, etc.

I think last year our lab ran something like 50,000 tests, so there's a lot of work that goes into developing the calibrations to be able to pass emissions standards, give good fuel economy, and maintain decent drivability for the customers.

I agree it would be useful. My guess as to why they don't do something like that is: the different tests in the link you posted really weren't designed with fuel economy as the main concern, they were designed to judge emissions-based performance. The same technologies that measure emissions are used to measure/calculate fuel economy, so you get both results from the same tests, but ever since the early '70's up until very recently the main concern the government had was with what pollutants are coming out of the tailpipe.

So testing fuel economy for a steady state highway cruise would be good info for the customer, but it'd probably make almost no emissions so the government wouldn't have much to regulate.