The EPA plans to audit more of the gas-mileage test data submitted by carmakers, and tighten its rules for that testing, after fuel-economy ratings on almost 20 recent models proved to be too high and had to be reduced.

The move follows the highly publicized rollbacks of gas-mileage ratings over the last two years by Ford, Hyundai, and Kia.

2013 Ford C-Max Hybrid, Los Angeles, August 2012

2013 Ford C-Max Hybrid, Los Angeles, August 2012

A dozen and a half models sold by the three makers were affected; their payments to buyers of those cars totaled hundreds of millions of dollars.

DON'T MISS: Ford Cuts Gas Mileage On 6 Models: What You Need To Know, And Open Questions (Jun 2014)

As reported Saturday in industry trade journal Automotive News (subscription required), the Federal agency will make a number of changes for 2015:

  • Issued revised guidelines on its "coast-down" test
  • Continue more aggressive auditing of data submitted by automakers
  • Potentially require some makers to verify data from pre-production test vehicles with actual production cars
  • Eliminate a rule that multiple vehicles in the same weight class that have identical powertrains need not be separately tested, but can use results from just one such vehicle

The coast-down test has been the cause of incorrect data for all three makers, as Ford was able to verify after a lengthy investigation that resulted in reduced ratings for six separate vehicles in June.

ALSO SEE: 2013 Ford C-Max Hybrid Owners Get Cash As Ford Lowers MPG Rating (Aug 2013)

Because fuel consumption is measured on a dynamometer, or "rolling road," inside a building, makers need to simulate the effects of aerodynamic drag as a car's speed rises and falls.

The test in question gathers data on the car's behavior as it coasts down from 80 miles per hour to a stop.

The automaker's test engineers then use that data to program the dynamometer to simulate the effects of real-world aerodynamic drag and other factors.

2011 Hyundai Elantra

2011 Hyundai Elantra

Errors in gathering that data can lead to incorrect assumptions for those simulations, affecting the amount of fuel consumed when the car is then dyno-tested.

Procedures for conducting such coast-down tests vary by manufacturer; the EPA plans to issue revised guidelines that more specifically spell out exactly how they are to be conducted.

Many manufacturers are presumed in the industry to optimize engine and control logic to consume minimal fuel under the specific and well-known test routines used by the EPA.

MORE: Hyundai, Kia To Pay Owners $395 Million For Overstating Gas Mileage (Dec 2013)

The problem of test results not matching real-world fuel efficiency is far worse in Europe, however, whose ratings can be 15 to 30 percent higher for the same vehicle than those in the U.S.

In 2012, the EPA added three more test cycles to its traditional "city" and "highway" routines, and emphasized the Combined rating rather on the window stickers of new cars.

EPA gas-mileage label (window sticker), design used starting in model year 2013

EPA gas-mileage label (window sticker), design used starting in model year 2013

In recent years, crowd-sourced data indicates that most gasoline vehicles appear to achieve within about 10 percent of their Combined EPA rating--although diesels are widely thought to do better on the highway than their ratings.

Meanwhile, it seems the EPA hasn't thus far discovered any vehicles whose EPA ratings were lower than they could have been.

[hat tip: Larry Nutson]


Follow GreenCarReports on Facebook, Twitter, and Google+.