Catching defeat devices: How systematic vehicle testing can determine the presence of suspicious emissions control strategies
While the Dieselgate scandal raised awareness of defeat devices—software calibrations that manipulate pollutant emission controls when vehicles are in the real world—third-party evaluation remains difficult because these devices are embedded in sophisticated computer code. To aid in detection by governments and third parties, this new white paper offers a practical, seven-step methodology that uses vehicle testing to determine the presence of an inappropriate calibration change. To illustrate, the authors test for devices that manipulate NOx levels in exhaust and use two Mercedes-Benz C-Class Euro 6b vehicles as demonstration vehicles—the diesel versions of a C180 and C200.
Testing revealed inconsistent emissions control and potential defeat-device triggers in the calibration strategies of both vehicles. The study categorizes the potential defeat devices as reduced exhaust gas recirculation (EGR) and reduced selective catalytic reduction (SCR) efficiency, both based on ambient temperature, around 12 °C; reduced EGR based on some measure of engine temperature; and reduced SCR efficiency based on some function of time, distance, or accumulated urea consumptions since engine on. Further, as the chart below illustrates, the impacts of the suspected defeat devices seem to be additive; combining several of the potential defeat devices and design limitations found via testing leads to NOx emissions of as much as 20 times the type-approval limit.
Summary of the type-approval test results and key testing conditions that led to elevated NOx emissions on the Mercedes C200 (first five tests) and C180 (last four tests).
Robust regulatory guidance is important for enforcing defeat-device provisions, but rigorous enforcement testing outside the type-approval process, like the procedure demonstrated in the study, is also needed to determine whether manufacturers are in compliance. Notably, the methodology presented here can be adapted to different regulatory contexts and research purposes.