Quantcast

How independent researchers tracked down VW’s diesel software hacks

Volkswagen-Car-Logo

Ever since news broke that VW had cheated on its emission controls, there have been nagging questions about how the company got away with it, and why it took an independent research team to find the errors in the first place. Here’s what we know so far.

VW was able to hide its fraudulent settings by programming its vehicles to detect when they entered what’s known as “test mode.” All modern cars have test modes that they can engage for fuel efficiency testing, because fuel efficiency and emissions are typically tested on a rolling device called a dynamometer. This allows the EPA or any other organization to monitor every car in the same way, and ensures an equivalent test environment across devices. Modern cars, however, don’t like having just their front wheels rolling while the back wheels are stationary, and will interpret this as a traction or stability control problem. In order to prevent the car from sabotaging its own test, the vehicle is placed in this test mode. There’s nothing intrinsically sinister about it; it’s part of the entire process.

In VW’s case, setting the car to test mode is what explicitly told the vehicle to use its full emissions control technology. Set the car to normal mode, and it went on happily spewing out particulates and nitrous oxide, both of which contribute heavily to smog. There’s been some chatter on the Internet about how this relates to greenhouse gases and CO2 emissions, but this is based on a misunderstanding of the facts. CO2 emissions in vehicles aren’t monitored and weren’t the problem here. Smog, meanwhile, is a major problem in industrialized cities.

How VW got caught

VW’s downfall came at the hands of a small West Virginia University team working for CAFEE (Center for Alternative Fuels, Engines, and Emissions). Far from planning a hit job, CAFEE was contacted by the International Council on Clean Transportation to conduct real-world tests on light-duty diesel vehicles. The ICCT had observed that European vehicles were generally emitting 4-7x more pollution than was permitted by European standards, but knew that American vehicles were somehow meeting much stricter test conditions established by California. The point of the study was to gather data about US vehicles under US conditions.

VehicleA-640x495

After some digging, we’ve located the original report prepared by CAFEE — all 133 pages of it. If you’re curious about how the test criteria were selected and how the test vehicles were configured, all of that information is available online, along with maps of the test routes, explanations for why these routes were chosen, and close-up photos of the three tested vehicles with their mounted hardware. Testing vehicles under these kinds of conditions is anything but easy, as the document explains:

“Table 3.9 lists all the parameters and emissions constituents collected during on-road testing for this study. Emissions parameters were sampled and stored continuously at 10 Hz frequency, whereas GPS and ECU data were updated at 1 Hz, but stored at the same frequency as emissions data (i.e 10 Hz) by the data acquisition system. An external sensor was used to measure ambient conditions, including temperature, barometric pressure and relative humidity, feeding data directly to the OBS data acquisition software. Vehicle position (i.e. longitude, latitude and altitude) and relative speed were measured by means of a GPS receiver, allowing for subsequent calculation of instantaneous vehicle acceleration and distance traveled.”

Page through the report, and it’s striking just how meticulous these researchers were. These vehicles weren’t just driven in a circle or across an arbitrary route; the researchers logged thousands of miles of test driving in monitored conditions, and in ways designed to ensure that the appropriate metrics were being gathered. The chart below shows how badly VW’s two test vehicles (Vehicle A and B) blew their NOx emissions rates, while Vehicle C (a BMW) conformed to the expected standard, save for the uphill / downhill test route. Because some routes produce far more pollution than others, exceeding the test rate in this area isn’t a strike against BMW in the same way.

GraphVW-640x370

Also note the VW results when their vehicles were tested in the FTP-75 ‘Bag 3′ chassis dyno test. This is the “test mode” scenario described above. When the vehicles knew they were undergoing emission testing, they were fully capable of conforming to the strictest Tier2 Bin5 standard.

This ocean of data was later turned over to authorities who used it to begin questioning VW for answers — and after stonewalling and lying for more than a year, VW eventually came clean and acknowledged that they’d been cheating all along. We still don’t know why VW chose to cheat, or how the company thought it could get away from this kind of disaster in the long term — but the costs are likely to far outweigh the benefits by the time all is said and done. VW’s cheating could effectively kill the diesel market in the US and damage it in Europe, where lawmakers have called for strict limits on diesel emissions to improve air quality. via – extreme tech

Comments

comments