Study: Control module MPG info not always accurate enough tool

"If you want to use ECM data, you have to do many tests for a long period of time, but at the same time, you will have other factors that will affect the accuracy or inaccuracy of your tests," says Yves Provencher, manager of PIT Group.

For a number of reasons, the data coming off a truck's electronic control module, or ECM, on mpg is only so close to the reality, a new study finds — so it's good in some ways fleets may be using it, not so good for others.

MPG is "a very complicated element to measure, especially when it comes to diesel engines," note the authors of the study, which comes from the Montreal, Canada-based nonprofit PIT Group (Performance Innovation Transport). The study conducted at a test track facility last fall examined 14 trucks that had four engines made by Cummins, seven by Detroit Diesel, one by Mercedes and two by Volvo.

In short, the study reports that ECM data on fuel economy/mpg varied in terms of accuracy and precision, and also varied from one engine type to another made by the same OEM. Further still, "for a given vehicle, ECM data precision and accuracy will even vary from test to test," according to the study.

To clarify that point, Yves Provencher, PIT Group's manager, describes to Fleet Owner the distinction between data "accuracy" and "precision" using an example of throwing darts at a dartboard. Accuracy refers to how close the dart actually gets to the bullseye, and precision means how dead-on each shot is aimed at the bullseye, or the variance in targeting precision.

"So the accuracy is how close you are from the real number," Provencher says regarding ECM fuel economy data. "You can be at 5% from the real number, and if you are precise, then you're going to be always at 5% from the real number." On the other hand, if your precision (or "targeting") is plus or minus 3%, for instance, it means your 5% accuracy could be that much farther off the mark.

The two measures are the result of the complex algorithms OEMs use to calculate fuel economy and report it through the ECM, factoring in things like engine speed, road speed, distance and fuel volume. "Because fuel consumption data in an ECM is derived from an algorithm and not from actual fuel flow, and does not account for fuel energy content, density or temperature, there is an inherent error with those calculations," the study notes.

Results from the study's 89 total tests include:

• The four vehicles with Cummins engines were tested a total of 24 times. Their ECM data showed accuracy of -5.4% to -6.2% and precision within 0.18% to 0.81%;

• The seven vehicles with Detroit Diesel engines were tested a total of 27 times. Their ECM data showed accuracy of -2.6% to 2.1% and precision within 0.37% to 1.09%;

• The single vehicle with a Mercedes engine was tested a total of nine times. Its ECM data showed accuracy of 0.9% and precision of 1.59%; and

• The two vehicles with Volvo engines were tested a total of 29 times. Their ECM data showed accuracy of -3.0% and 0.9% and precision of 0.25% and 0.84%.

Real-world application: fleet decisions

The study authors note that all the engines reported less fuel consumption via their ECM data than was measured by the study's gravimetric test methodology, with one exception: the Mercedes. However, Provencher points out, "We're not trying to say that one engine maker is better than another. We're just saying, 'Really look at how accurate this stuff is'" coming from the ECM.  

"So for example, if you're looking at a decision based on 1% or 2% difference between two technologies," he adds, "we're telling [fleets] that that's probably not the right tool to use." That's often the situation as tech-savvy fleets today work to optimize efficiencies across the board and leave no stone unturned, especially when it comes to fuel economy.

Provencher explains that ECM data is a better tool when considering measures that can involve larger differences in fuel economy to the tune of 5%-10%. That includes things like crafting driver profiles or monitoring driver activity such as idling, speeding, harsh acceleration and so on.

Also, looking at more ECM data over time will give a more accurate result — but be careful, since that also brings in other considerations. "If you do a large number of tests with the ECM, you're going to get closer to the reality. But at the same time, having a large number of tests means you may have different drivers, different weather; you may have different loads," Provencher tells Fleet Owner. "It's hard to do a large number of tests with the same conditions."

On the positive side regarding ECM fuel economy accuracy, he says that's probably not going to worsen with the engine's age — but beyond those driver-based and environmental conditions are others that change with a vehicle's age and will affect mpg, which also should be taken into account in the case of ECM data tests over time. 

"You have the exhaust system and all your filtering systems that will change with time; we also know that tires get better fuel economy with wear. So if you're going to say, 'I'm going to make sure I have a good test and I'm going to do tests over 15,000 km (approx. 9,320 mi.),' well, you're not going to have the same rolling resistance — the tires will be worn down and you'll see a huge difference in fuel economy, just because of the tires," Provencher says.  

So fleets need to keep all these factors in mind, he advises. "If you want to use ECM data, you have to do many tests for a long period of time, but at the same time, you will have other factors that will affect the accuracy or inaccuracy of your tests.

"We don't recommend to make a business decision like 'should I buy this [trailer] skirt or that skirt' or 'should I buy this tire or that tire' using the ECM," he continues. "It's not accurate enough to do that."


Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.