Despite Barra’s Denials, GM Diesel Test Results Indicate VW-Style Cheating

DUHzafiraWEB

We’ve suspected for some time that more automakers would be caught up in the Volkswagen emissions cheating scandal, and the first new perpetrator has apparently been identified: General Motors. GM CEO Mary Barra’s insists that VW-style software cheating on emissions tests “is not a condition that exists in our vehicles,”  but the German environmental group Umwelthilfe has sponsored tests that throw that claim into serious doubt [English press release in PDF format here].

In testing of the Opel Zafira 1.6 CDTi, performed at the Bern University of Applied Sciences, GM’s diesel engine passed NEDC cycle NOx tests performed on a two-wheel (single-axle) rolling road but emitted two to four times the Euro6 limit for NOx when the same test was performed on a four-wheel rolling road. This strongly indicates that a software “test mode” exists for this engine, although Opel insists that “The software developed by GM does not contain any features that can detect whether the vehicle is being subjected to an emissions test.” But, says International Transport Advisor Axel Friedrich,  “I have no normal, technically plausible explanation for the emission behavior of the Opel vehicle.”

The Opel Zafira in question was a front-drive 2015 model, with 6,000 km on the clock. The model officially meets the Euro6 standard, and in multiple NEDC-cycle tests performed on a two-wheel rolling road it came in under the 80 mg/km NOx limit. But testers noted emissions of up to 17 times that limit under certain driving conditions. And even when on a two-wheel rolling road, certain forms of acceleration not seen on official tests apparently caused the AdBlue exhaust treatment system to shut down:

With a continuous increase in the speed to 150 km/h, NOx emissions increased abruptly and exceeded the measuring scale of the analysis instrument. The test report states: “The behaviour could be explained by a shutdown of the AdBlue (Urea) dosing unit. Similar behavior was not detectable when operated in the 4-wheel drive [test] mode.”

Note that by 2WD/4WD, the testers are referring to the rolling road and not the vehicle drivetrain, which is only front-wheel drive. The test results led to the following conclusions:

“The measurements carried out show the following trend: The NOx emissions in the NEDC cycle depend on the test mode of 2WD/ 4WD. In the 2-wheel drive mode the vehicle met the NOx regulations. At low speeds, NOx emissions are not always identical, and are likely to depend on the activity or the storage effect of the SCR system. The behaviour of the SCR system seems to be dependent on the test mode, since the NOx trends are different in the two test stand modes of operation.”

The Swiss lab was lucky that Opel’s defeat device – if it exists – appears to be of the rather crude “Cheater 1.0” variety, described in our initial dieselgate story on September 21. From the fact that one set of wheel spins while the other set doesn’t, the device deduces that the car is being tested, and the exhaust is being kept clean. Volkswagen would not have been bagged. VW uses a slightly more sophisticated version, that looks at changing steering angles for cues of being under supervision. That  was hot seven years ago. What is the state-of-the=art of diesel cheating, only a few know.

Opel initially said that “the technology developed by GM Software has no features that can detect whether a vehicle happens to be in emission test cycles or not,” and when confronted with DUH’s test results the firm

stated, on 21 October 2015, that it had reconstructed and documented its own tests using an appropriate vehicle in accordance with legal regulations, “both on a two- and a four-wheel roller dynamometer. The emission behaviour determined in each case does not differ from one another.” Adam Opel AG reiterated: “The software developed by GM does not contain any features that can detect whether the vehicle is being subjected to an emission test.”

DUH has referred the matter to Germany’s equivalent of NHTSA (the KBA) and is demanding that the model be tested further, a process that could lead to a revision of its type-approval.

But although the German regulatory response is difficult to anticipate, this will doubtless be a PR disaster for GM’s CEO Mary Barra. Barra has insisted on numerous occasions that “I don’t think you’ll find something similar to that [VW’s defeat device]  emerging from GM,” and her leadership has been repeatedly held up as “the answer” to Volkswagen’s moral failings. With GM’s Opel division unable to explain why its diesel emissions are so different when tested on two-wheel versus four-wheel rolling roads, it seems that her statements came too soon. As GM enters its second scandal in as many years just as Opel is supposed to be returning to profitability, these statements will doubtless come to haunt Barra.

Meanwhile, the question becomes: what other GM vehicles, and what other automakers are still waiting to be caught?