The way the Environmental Protection Agency tests the emissions of a car looks a lot more like a videogame than driving.
The driver sits behind the wheel of a car on a treadmill and, using only the accelerator and the brake, tries to keep the car on a course that scrolls by on a computer screen outside the driver-side window.
Pressing the accelerator moves a dot on the screen toward the right edge; the brake moves the line to the left. If the driver keeps the dot in the lines for the four-minute test, the car has gone through a standard course and its emissions can be compared with the emissions from other vehicles and with the federal standard.
But the 20-year-old test bears about as much similarity to the real world of driving a car as the computer game Flight Simulator bears to flying an airplane.
Which may be why air quality has improved only slowly since the 1970s-though automakers boast that EPA tests show that tailpipe emissions of hydrocarbons and carbon monoxide are down more than 90 percent since then and nitrogen oxides are down more than 75 percent.
The Clean Air Act of 1990 instructed the EPA to devise a better test. After one extension, the agency had a deadline of the end of March to comply, but it says it cannot meet that either and is seeking another extension.
The EPA acknowledges that it needs to find a better way to test engine emissions.
“The old procedure assumes the world is flat,” said John M. German, an engineer at the EPA’s National Vehicle and Fuel Emissions Laboratory in Ann Arbor, who is helping design a new test.
Not only is the EPA’s world flat, but it also is traversed only by au tos that never exceed 57 miles per hour and carry precisely 300 pounds worth of passengers to destinations exactly 7.5 miles away. And when the engine is shut off, it is left off for 10 minutes-no more, no less-or is allowed to cool overnight before restarting.
All of this idealized motoring is done on a dynamometer, a treadmill connected to an electric generator, with the electric output directly proportional to the effort made by the wheels. Engineers also put a fan in front of the car’s radiator, to prevent overheating. The tailpipe emissions are captured in plastic bags for later analysis.
Engineers have long suspected that the tests are not realistic, missing such nuances as sharp accelerations and decelerations and frequent cold starts for around-town errands and the like.
On shorter trips, the catalytic converter and the engine do not get properly warmed up, and the engine does not have time to suck in and burn all the fumes that collect in the charcoal canister that captures evaporation when the car is not running.
But only recently, EPA officials say, has computer technology improved their ability to measure real-world conditions to the point that the problems with the official test have become glaringly apparent. In addition, as standards have become progressively tighter, German said, “what isn’t on the test becomes relatively more important.”
And the margin of error is not consistent. Running sample cars through a more realistic test shows that some kinds of engines emit huge amounts of pollution that go undetected at the EPA lab. But other engines show only modestly higher pollution levels. In both cases, engineers still are puzzling why.
One thing is clear: Cars on the road have been designed around the existing test. “The rules are-pass the test in the cheapest way possible,” said Charles L. Gray Jr., director of regulatory programs and technology at the EPA.
There is nothing sinister about this, agency officials say. But the result is a little like a high school student who cannot use a single four-syllable word in a sentence but knows 2,000 of them for the college entrance examination.
Automakers avoid criticizing the EPA the way high school students are circumspect about the shortcomings of their principals. But Dennis R. Minano, vice president for energy and environment at General Motors, acknowledged that the industry designs cars to beat the test.
The industry is hesitant to do better voluntarily, Minano said. If the government requires that emissions of a pollutant be reduced to, say, one gram per mile, a carmaker might aim for emissions of nine-tenths gram per mile, to make sure that the cars pass. But then the regulators will turn around and say, “Aha! You can do better,” Minano said, and will impose still stricter rules.
As the EPA moves toward improving its test drive, others are rushing to embrace it; beginning in 1995, several densely populated states. including Illinois, will beef up their emissions inspections by putting ordinary cars through the test.
Despite its shortcomings, it is far superior to tests in which a car idles and a machine samples for hydrocarbons and carbon monoxide, but not nitrogen oxides.
That will be an improvement for testing cars in use on the streets, but for testing prototypes, German said, “we need to expand the boundaries of the test, to capture more of what’s going on.”
Among the real-world conditions the new EPA test must provide for is the fact that today’s cars have a safety mechanism to prevent the catalytic converter from burning out under stress conditions, like driving up a steep grade.
