How many times have you heard, “the devil is in the details”? Nowhere does it apply more than in fuel-economy testing. Accurate measurements are important, but controls are even more important. The most subtle differences, like tire pressure, tread depth, or the gap between the trailer and the cab can affect the outcome of a test. Even wind and traffic conditions can skew the results.

The engineers at Kenworth’s Research and Development Center in Renton, Wash. have it all under control, and they produce some of the most accurate fuel-economy data you’ll ever likely come across.

That’s what made this particular exercise so intriguing. Kenworth recently invited two truck writers, myself and Steve Sturgess of the American magazine roadSTAR to participate in a fuel-economy challenge. We drove the same T600 over the same route on two different days. Data was filtered, measurements adjusted, prevailing conditions compensated for, and the winner was announced a week or so later. Sturgess won, I’m humbled to admit, but only by 0.11 mpg.

That the difference was so small over nearly 500 miles was interesting enough. But what was even more interesting was the reason for the difference and how they measured it.

Test Parameters

Kenworth struggled for a long time to come up with a test route that reflected “real world” conditions for a wide range of applications. From the Seattle area, they couldn’t go far north because of the Canadian border.

Running east from Seattle put them into the mountains, and then onto flat rural highway, unlike conditions found in much of the country. The chosen route runs 496 miles, south down Interstate 5 to Portland, and then east along a hilly and winding section of Interstate 84 to The Dalles, Ore. The return route is the same.

While the route is mostly four-lane highway, there’s some two-lane as well as some heavily trafficked areas-road conditions that reflect the diversity customers are likely to encounter.

To maintain consistency, the trucks stop at all the same places each trip, and engines are started and shut-down simultaneously to keep idle time the same. They run through the weigh stations even when they are closed for more repeatable results.

The T2000 control truck is equipped to monitor temperature and humidity along the route, and after the trip, Kenworth’s R&D engineers download weather records from several stations along the route to factor in wind speed and direction.

Post-trip refueling is a process in itself. Trucks are parked on a specially constructed flat pad and refilled to a prescribed level. Fuel temperatures are recorded to compensate for expansion of the fuel at different temperatures. The number of gallons (measured to within four ounces), the final fuel temperature, and the ambient temperature and relative humidity are recorded.

Data Downloads

Each truck on our test drive was equipped with a “Silverleaf” data recorder, an instrument spec’d by Kenworth for fuel-economy testing. The device continuously records all sorts of engine operating conditions and it also displays real-time information for the driver, such as rpm, speed, engine load, and most importantly, turbo boost.

All the data is downloaded following the trip and it’s compiled and analyzed by Moses Luyombya in the lab. He searches for anomalies and unusual conditions that could skew the data, so what is produced on a given trip is an incredibly accurate record of the conditions under which the fuel was consumed. This accuracy makes comparative analysis both easier and more precise.

With all the control data already in place, testing for a new device, say a redesigned roof fairing, would require a minimum number of runs while maintaining integrity comparable to that of an SAE (Society of Automotive Engineers) Type II fuel test.

The R&D crew at Kenworth have run the route so many times that they can actually compare data from a selected portion of the trip, such as a climb on a given hill, under different temperature, humidity, or road conditions. They’ve compiled elevation profiles of every mile of the route so the hills are easy to identify.

The Final Outcome

The T600 Sturgess and I drove was a customer spec’d truck, loaned to us for this test. Brand new — save a thousand or so test miles the R&D people put on the truck-it showed 1,298 miles when I drove it, and 1,988 miles when Sturgess had a go. (The fuel economy numbers would likely be significantly better with a broken-in truck.)

I burned 74.898 gallons (US), for a raw mpg of 6.62. Sturgess burned 72.035 gallons for a raw mpg of 6.89. My fuel mileage was adjusted upward to 6.78, to account for the rainy weather on the final third of the trip, and for the relative newness of the truck compared to Sturgess’ run. It works a bit like a golf handicap. All the adjustments were based on data gathered from the control truck on previous runs.

For this test, R&D engineers, Gary Ziebell and Tom Liethen drove the T2000, each driving one leg of the trip on both runs while the other rode with the test driver.

During my run, the Silverleaf showed my trip fuel economy to be 7.1 mpg, and as soon as it began raining, Liethen told me to watch the economy drop by 0.2; and sure enough it did, exactly by 0.2 mpg. These guys know their stuff.

The “Key Numbers” (sidebar graphic above) reveals some interesting differences in the way Sturgess and I managed the trucks. You can see his average speed was slightly higher than mine, as was the maximum speed he attained on the run. His average engine speed was 10 rpm lower than mine, suggesting more time overall in top gear. His average acceleration and turbo boost numbers were both lower than mine, and therein lies the key to his victory in the fuel-economy challenge.

“Manifold pressure represents power demand on the engine, and aggressive throttle application results in correspondingly high manifold pressure,” explains Ziebell. “Using less throttle — sacrificing a little acceleration — uses less fuel overall at a given speed, so the driver who is gentler with the throttle pedal is going to get better mileage.”

The “Acceleration, true mean” value in the chart refers to the average percentage of throttle used during the course of the entire 496-mile trip. My value is 11.742 percent compared to Sturgess’s 7.289 percent, which means he used an average of 4.453 percent less throttle than I did to complete the trip.

In contrast, Jim Booth, a former “test driver” with Caterpillar who still runs a fleet of trucks in revenue service hauling for Cat. Believe it or not, Booth manages better than nine miles per U.S. gallon with a similar engine in a well broken-in T2000. His throttle use numbers are in the 3-4 percent range.

This kind of data is invaluable in understanding fuel economy. Certainly speed is a factor, but so is driver performance. And from the engineering perspective, collecting data the way Kenworth’s R&D people do, they can experiment with different driving styles as well as hardware add-ons to improve fuel economy. They have invested heavily in this program, and looking forward, the next change they want to test can probably be accurately measured in just a few runs rather than dozens, because they’ve already got the baseline data to compare.