New EPA lab gear will monitor mpg claims
By next fall, agency can verify all 5 automaker tests
Reprints
Print
Respond
Editor's note: An earlier version of this story mischaracterized the test cells being built at the EPA's vehicle laboratory in Ann Arbor, Mich. One of the two test cells can operate at 20 degrees Fahrenheit and the other one can operate at 95 degrees. The story also appeared on Page 6 of the Dec. 10 print edition.
As the EPA continues to investigate Hyundai and Kia for selling cars with false fuel economy labels, the agency is adding equipment at its Michigan laboratory to spot errors and discourage any automaker that might be tempted to cheat.
Fuel economy ratings have become a crucial selling point as gasoline prices have increased, raising worries that some companies are fudging the numbers. And the stakes are higher than ever with fuel economy standards scheduled nearly to double by the 2025 model year.
That's one reason the EPA has spent five years renovating the National Vehicle and Fuel Emissions Laboratory in Ann Arbor, Mich.
The renovation includes equipment that will enable EPA engineers to run a pair of fuel economy tests, for cold and hot weather, that they now can't verify at their own laboratory. The EPA says it does audit the tests at automakers' labs.
All the equipment should be operating by next fall, David Haugen, manager of the EPA's technology department, told Automotive News during a tour last week.
"This closes the big gaps," he said.
Under fuel economy laws, automakers test their vehicles to come up with city and highway mpg numbers, which they put on their window stickers. They also submit the data to the EPA, which audits a small percentage of vehicles.
The troublesome tests have been the "Cold CO" and "SC03" tests, two of the five tests that go into the standard formulas for city and highway mpg. They reflect that vehicles get fewer mpg when the driver starts the engine in cold weather or runs the air conditioning at full blast on a hot day.
The existing test cells at the laboratory cannot run those hot and cold weather tests. Some see that as a problem, because the results of Cold CO and SC03 tests have gone into automakers' fuel economy calculations since 2008.
Errors or falsified information on the tests could lead to inflated mpg numbers on window stickers, said John German, a fuel economy expert who worked at the Ann Arbor laboratory for 13 years between jobs at Chrysler and Honda.
"It is extremely important for EPA to upgrade its lab to be able to run these tests," said German, a senior fellow at the International Council on Clean Transportation.
Haugen said the EPA has audited automakers' results since 2008 during site visits to automakers' test labs.
But site visits are no match for what EPA testing engineers will be able to do after the renovation, he said.
"It will allow us better oversight," Haugen said. "We'll be able to test every manufacturer's vehicle in one site so we don't have test-site-to-test-site variation."
Even as they add equipment to check automakers' math and deter cheating, EPA investigators are still trying to figure out how Hyundai and Kia overstated the mpg of about 900,000 vehicles sold in the United States over the past few years.
Hyundai and Kia say they erred on the "coast-down" test, which provides data on aerodynamics, tire friction and engine resistance. The data are used to calibrate the dynamometers used to test vehicles' fuel consumption.
When they admitted to mistakes on Nov. 1, Hyundai and Kia said their technical center in South Korea made mistakes on the coast-down test because parts of it are open to interpretation.
The EPA allows automakers to do a mix of testing and computer modeling if they can come up with a model that replicates the results of real-world coast-down tests. That's where the test can be open to interpretation, Linc Wehrly, manager of the light-duty vehicle group at the EPA's Ann Arbor laboratory, told Automotive News.
"Some manufacturers do a lot of coast-down testing," he said. "Some manufacturers do a mixture of coast-down testing and modeling."
Auto industry sources say the EPA has had trouble auditing automakers' coast-down tests because the Ann Arbor laboratory lacks a straight, flat, two-mile-long track like an airport runway. The agency uncovered Hyundai and Kia's errors through a special round of audits that began in 2010.
The Cold CO and SC03 tests presented a similar problem to the EPA's engineers.
In these and other fuel economy tests, a car is placed on a machine similar to a treadmill, which is known as a dynamometer. The vehicle's wheels spin at various speeds for a length of time to simulate fuel use during typical city and highway driving.
To perform the Cold CO and SC03 tests, automakers must cool or heat their test cell to 20 degrees or 95 degrees Fahrenheit.
None of the EPA's existing test cells could replicate those conditions, so the company had to resort to site visits. That will change next year.
The EPA signed a contract in March with AVL North America Inc., a suburban Detroit supplier of emissions testing equipment, for two German-built test cells.
Haugen said they cost a combined $9 million.
One of the test cells can operate at 20 degrees Fahrenheit to run the low-temperature test and the other can operate at 95 degrees to run the high-temperature test. Each will be able to handle two to three vehicles per day while running those tests, said Jimmy Williamson, director of emissions at AVL.
EPA engineers already test 10 to 15 percent of new light-duty vehicle models each year using the three tests that are conducted at normal temperatures.
The new test cells would let them audit all five tests in the test cycle without help from an outside laboratory for the first time since the EPA changed the formulas in 2008.
Automakers still could try to game fuel economy tests by providing the EPA with vehicles that are different from the ones they actually sell, German said.
He says he doesn't think most companies would do that. If the EPA starts auditing all five parts of the test cycle, German said, he will be less concerned about cheating.
Larry P. Vellequette contributed to this report
• FTP: City test. A vehicle on a dynamometer, similar to a treadmill, makes 23 stops in 31 minutes. Average speed: 20 mph.
• HWFET: Highway test. 10 miles on the dynamometer. Average speed: 48 mph.
• US06: High-speed test, added for the 2008 model year. Various stops, accelerations up to 80 mph.
• SC03: High-temperature test, also added for 2008. 10 miles at 95 degrees Fahrenheit with air conditioner on full.
• Cold CO: Cold start test, also added for 2008. Identical to city test but at 20 degrees Fahrenheit.
You can reach Gabe Nelson at gnelson@crain.com.
