I rent cars several times a month for my job. Today Hertz delivered me a shiney new Chevy Monte Carlo LT (my commentary on the design of this car could be a whole seperate thread). I went to the Chevy website and noted that the car has a 211 hp V-6 with variable valve timing and claims 31 mpg highway. Upon returning to Indianapolis today after a 250 mile roundtrip to a client meeting, I was shocked to see that the tank was nearly dry. I calculated an average FE of 20.1 mpg. The driving was a 50/50 mix of state highways and Interstate highways. I usually rent cars such as Sonatas, Camry's, Altimas, etc... I will start paying better attention to their real world mileage vs EPA.

 

Does the EPA test vehicles on a treadmill?

If any of you do much treadmill running, it's not the same. The incline to simulate pavement running needs to be about 1.5%.

Aside from a dated test that does not consider the use of A/C, higher speed limits, more aggressive driving, this in perhaps another factor.

 

As far as I know, EPA tests are still dynometers (or whatever the fancy name for the car treadmill is.) As I recall, that's also with a constant speed of 65 MPH.

As always, the fast you go, the worse your FE.

But yeah, after a hybrid, most cars suck.

 

Dynometers and running treadmills don't increase the wind as you go faster.

Someone stationed in Afghanistan trained for a marathon on a treadmill. Sounds like wimpy until you consider all the land mines still out there.

 

I spent the last weekend on the trails around Moab, Utah in a 2004 Jeep Wrangler 5-speed. They're EPA rated at 16 city, 19 highway for the 4-liter version, like what I was riding in. Although we got a very respectable 12.25mpg for our 164 miles off-road, most of which with the tires at 20psi for compliance/traction, the highway fuel economy was around 15mpg averaging 75mph or so. While that's not too bad to achieve about 80% of the EPA highway rating, it's still pretty disapointing to realize there's still vehicles out there that can only get 15mpg on the highway.

 

The EPA does NOT test at a steady 65mph, they never even reach 65mph. They have a set program they run each car through on a dynometer and they measure the fuel it uses by the exhaust gases. The average speed for the highway portion of the test is 47mph and their maximum speed is 60mph. They do use load and calculate for inclines and wind, and even after that they divide the number by something to try and simulate more real world numbers...and they still fail pretty bad. However, if you drive the cars exactly like the EPA does in their tests, you'll get the same milage and perhaps even better than the two big numbers on your window sticker.

I use to have a great website marked that talked about the testing in detail with pictures and everything, it was a government website too. I remember they said something like with the acceleration rate they use during the test it would take them 18 seconds to reach 60mph. 99.999% of the population (and I'm probably rounding down there), accelerates faster than that.

 

Quite a bit of info on the EPA tests is on www.fueleconomy.gov.

Here are a couple of links of interest in their site:

http://www.fueleconomy.gov/feg/how_tested.shtml - overview of the tests.

http://www.fueleconomy.gov/feg/fe_test_schedules.shtml - some higher level info and graphs of the tests.

http://www.epa.gov/otaq/testproc.htm -EPA sight that has links to .pdf files on the procedures (- way more involved than I want to get!)

So, armed with this info, anyone want to venture an analysis of why the various hybrids get the EPA mileage they do?? I look at the graph of the city portion and can kind of see why the Prius gets 60mpg city, but I'm curious if the test drains the battery at the expense of mpg? You'd think the EPA would account for that, though?

 

I'd always figured EPA numbers were a guide for relative vehicle performance and not necessarily linear to real world numbers. That is why plotting the MPG vs MPH performance curve is important and the methodology is fairly straight forward:

- hold as many variables as constant as possible (route, speed, vehicle status)
- drive far enough to even out fill-up errors (at least 2-3 gallons / 8-12 l.)
- repeat tests later after making changes (tire pressure, oil change, e.t.c.)

Each data point took about 4-5 days, the blink of an eye over a year. But having performance curves, allows informed choices and achieving predictable MPG goals. More importantly, it revealed "don't go here" speeds. For example, there is a transition area at ~42 MPH where the Prius has to begin running the ICE and this impacts MPG.

My Prius fuel economy is excellent, +50 MPG, at speeds at least 4-5 MPH above and below the 42 MPH transition speed. The worst MPG occurs when the Prius is kept right at the 42 MPH transition speed and this is reproducable. So I drive following these rules:

- 38 MPH and below
- 51 MPH and above (*)

The result is achieving hypermiler status this winter. It also makes possible studies of secondary factors such as warmup and last-mile protocols and urban pulse-and-glide.

My data is for an 01-03 model Prius and other hybrid models will have different curves. Yet nothing like it exists in the owner's manual or the Toyota web site.

Bob Wilson

* - Double the margin above the 42 MPH transition to deal with traffic slow-ups. Be sure to run several miles further than the accelleration distance.