Emissions Scandal Expands: Mercedes-Benz, Honda, Mazda, and Mitsubishi (theguardian.com)

An anonymous reader writes: Volkswagen has taken some serious heat for deliberately circumventing emissions tests with "defeat devices" in some of their vehicles. While no other cars have been found to use specific devices to fool tests in the same way, we're now learning that many manufacturers still mysteriously perform worse in the real world. Last week, the Guardian revealed that diesel cars from Nissan, Hyundai, Citroen, Fiat, Volvo, and Renault emitted significantly more pollution in realistic driving conditions than the tests supposedly allow. Now, we learn that vehicles from Mercedes-Benz, Honda, Mazda, and Mitsubishi emit substantially more than they should as well. For example: "Mercedes-Benz's diesel cars produced an average of 0.406g/km of NOx on the road, at least 2.2 times more than the official Euro 5 level and five times higher than the Euro 6 level. Honda's diesel cars emitted 0.484g/km of NOx on average, between 2.6 and six times the official levels." This provides clear evidence that the automotive industry is designing its cars to follow the letter of the law (passing tests), but not the spirit (actually reducing pollution).

Comments

Maybe

[Anonymous Coward]

Or it reveals that the testing mechanism was always wrong. It's a leap to say that differences between the tests and "real driving" represent fraud, until it's proven that the cheating mechanism is actually there (as it is in VW).

Re:Maybe

[MatthiasF]

Wish I had mod points, but I agree with this sentiment. Tests present a statistical average but real-world terrain and human driving is all but average.

Re:Maybe

[tripleevenfall]

Obviously, it's important for a test to always be uniform. If you tried to test cars under "realistic" driving conditions the tests would all be different.

Realistic driving conditions are variations in temperature, terrain, traffic flow, etc.

Realistic driving conditions vary based on the habits of the driver.

Realistic driving conditions vary based on the condition of the car over time.

Maybe instead of ballyhooing these tests, we should apply common sense to them. Maybe we should see them as a group of data points and not a limits, guarantees, or absolutes?

Re:Maybe

[Rei]

Indeed. VW did very egregious cheating, deliberately detecting tests and then optimizing for them. It sounds like these others are not engaging a "test mode"; but have optimized themselves for conditions that are tested for (at the expense of power and fuel efficiency) while optimized themselves for power and fuel efficiency in conditions that aren't tested for. Not as egregious, but still clearly problematic. There's clearly gaping holes in the system.

It also puts to lie this massive increase in diesel cleanliness over the years. It's improved, no question, but not nearly as much as has been marketed, particularly in smaller, cheaper vehicles. The same old choice remains: you can get a ~15% increase in fuel efficiency by mass (~30% by volume), and thus ~15% reduction in CO2 emissions, by going with a diesel, but it'll come at the cost of a more expensive engine (has to be built stronger to handle the higher compression, all issues of additional pollution control systems aside) and will kick out more health-impacting pollutants. And it just comes down to chemistry: if you burn fuel in air at hotter temperatures and/or higher pressures, you favor the production of chemicals like NOx - high temperatures and pressures make nitrogen more reactive. And you're going to kick out more PM for similar reasons. The higher temperatures and pressures help with CO and unburned hydrocarbons (they favor more complete combustion), but the scale of the added NOx and PM problems are much greater.

Contrary to what they've been pretending, a major way that car manufacturers appear to have been reducing NOx emissions in diesels is simply by burning their fuel cooler / less efficiently in conditions that are being tested for, and hotter the rest of the time to keep their performance and efficiency numbers up.

Re:Maybe

[Anonymous Coward]

That makes zero sense. What you're suggesting is that we let automakers continue to destroy our environment, just so they can make profits, like they have some sort of god-given right to make a profit. It's this kind of whoring that has put the planet in the state it's in now. All automakers should be moving to 100% electric at this point anyway. Lowering the bar for emissions would be a step backwards.

Re:Maybe

[DarkOx]

It sounds like these others are not engaging a "test mode"; but have optimized themselves for conditions that are tested for (at the expense of power and fuel efficiency) while optimized themselves for power and fuel efficiency in conditions that aren't tested for.

Or the tests simply don't reflect the typical driver at all. Look the EPA sticker on my car says 39mph high way. Where I live there isn't much other traffic most of the time. Its pretty rural so the determining factor is more my driving than anything else. Weather and time of year in terms of summer blend vs winter gas probably has an effect as well. If I moderate my driving I can get well over the sticker, I have averaged as high as 42mpg over a tanks. The way I usually drive, I get about 27.

I don't suspect cheating either because I can reproduce the results to my satisfaction. It don't even think its a case of optimizing to the test. I think its just a case of optimizing period. The best fuel economy is observed by traveling at a constant speed and accelerating slowly when that is required. Highway travel is usually constant speed over several miles or more @55-75mph, so that is the behavior that should be targeted. Its pretty easy to work out that is what the car is optimized to do just using your trip odometer and recording how much fuel you buy. Sure you could tune it to deliver better acceleration (just change the gear ratios would be one obvious way) but the cost would probably be economy cruising, anyway you go about it.

As other have pointed out the test has to be specific and control for as many variables as possible, otherwise we can't use it for comparative use cases. So add additional test conditions, maybe publish results and set standards for different driver profiles, aggressive, nominal, hyper miler, and publish them all. Don't develop a new single test case because you will create a perverse incentive to target that use case and it will likely be more distant from actual use than city/highway profiles they have now.

Re:Maybe

[budgenator]

All manufacturers designed their engine/transmission management systems to pass the test, what happened outside the test conditions were in the "your actual mileage may vary" zone; It's just like "common core" where they teach the kids to specifically pass the test. If you want cars to perform under realistic driving conditions the same as they do under emissions test, you have to make the test conditions as close to realistic as possible. The big three used to test their suspensions by driving down a particular bumpy section of Woodward avenue in Detroit, when they announced that that section of road was going to be repaired and repaved, the manufacturer surveyed the road and duplicated it at their test tracks to maintain continuity, and keep the test as close to realistic as possible.

Volkswagon's mistake was they actually change the operating parameters based whether the vehicle was being tested for emissions or for mileage or under normal conditions, the next step is for the bureaucrats to realise that they can have vehicle emit differing levels of emissions based on location and weather; your car may suck donkey balls in San Francisco, but run like a champ in Montana.

Well, goodbye passenger car diesel!

[RoTNCoRE]

I don't think diesel passenger cars will be a thing much longer in North America after this. And time to change the tests to measure results in real world usage conditions.

Re:Well, goodbye passenger car diesel!

[Anonymous Coward]

I took autoshop in the late 90s. I remember my autoshop teacher swearing up and down that diesels could not be cleaned up enough to make a good passenger car engine. He swore up and down that they were cheating on emissions tests back then. Never underestimate the "shadetree mechanic" type of person.

Re:Well, goodbye passenger car diesel!

[aaarrrgggh]

The standby generator market which I am very familiar with half jokingly refers to the next round of emission standards as "diesel air cleaners"-- the standby generators will be required to have cleaner air out of the exhaust pipe than what comes in.

Emission standards have become very strict; while the objectives are good, they are pushing the realm of what is viable.

For standby generators, continuous monitoring is practical, but when it takes 20-30 minutes to come up to full temperature and the normal run time is less than an hour, actual emissions are going to be much worse than the idealized continuous running state. We actually need to add artificial load to make the emissions control systems work properly-- increasing CO2 and NOx, and DPM t(o some level) in absolute terms, but reducing them relative to the engine size. I imagine cars are in a similar condition; "real world" is an ambiguous design condition.

Not surprising and can you blame them?

[flappinbooger]

The law says "pass this test" so they pass the test.

How is this different than standardized testing in schools? The state says "pass this test" so the teachers train the kids to pass the test.

Do they actually LEARN anything useful for the real world?

Do these cars actually have low emissions when driven in the real world??

You be the judge.

Re:Honda Diesel?

[beelsebob]

Why? It's not like Slashdot is a US centric web page... Why does it have to make special mention of the status in the US?

Re:Honda Diesel?

[Anonymous Coward]

Obviously anything not concerning US citizens should be labeled as such. After all, the site's motto is "News for US Nerds, Stuff that matters to US only".

Why no diesel hybrids?

[Fencepost]

I don't understand why we're seeing all these gasoline hybrids instead of diesel ones. Aren't diesels running in their optimum range much more efficient? And with all these emissions issues turning up, isn't it feasible to set up diesel hybrids to basically always run in a narrow range with the best emissions and efficiency possible?

Re:Why no diesel hybrids?

[bre_dnd]

Peugeot makes them: http://www.peugeot.co.uk/hybri...

Re:Why no diesel hybrids?

[Frederic54]

In the US you mean? In Europe there is diesel hybrid, in Korea there is even a Hyundai Elantra LPG hybrid.

Re:Why no diesel hybrids?

[Solandri]

I don't understand why we're seeing all these gasoline hybrids instead of diesel ones. Aren't diesels running in their optimum range much more efficient? And with all these emissions issues turning up, isn't it feasible to set up diesel hybrids to basically always run in a narrow range with the best emissions and efficiency possible?

Diesels already almost always run in their optimum range. A car engine basically has three operating states that are important. Accelerating from a stop, cruising (usually at highway speeds), and accelerating at highway speeds (to pass).

Gasoline engines hit peak power and torque at the high-end of their RPM range. That's great for accelerating at highway speeds, not so good for cruising and accelerating from a stop. Because most of the engine's time is spent cruising, that's where you need to optimize fuel burn rate to improve overall fuel efficiency. Gas engines have a lot of problem with this because it's not coincident with their peak power and torque production. Consequently you're having to optimize the engine's performance at two hugely different RPMs. The hybrid helps a lot here because the electric motor provides a lot of torque at 0 RPM for accelerating from a stop (power = torque * RPM * a constant),and allows the gas engine to be shut off completely for a while during cruising. So now you can optimize the gas engine for high-RPM efficiency, and rely on the electric motor for what would normally be low-RPM operation.

Diesel engines have a higher compression ratio so hit peak power and torque at the low-end of their RPM range. That's great for cruising and accelerating from a stop, not so great for accelerating at highway speeds. This is why they're so common in tractor trailers - it's OK if the truck takes a long time to accelerate at highway speeds, but you want good power and fuel efficiency during cruise. Since the diesel engine's peak torque and power happen close to cruise, they're a lot easier to optimize for fuel efficiency.

A hybrid won't actually help much here because it doesn't add much - the diesel engine already has lots of torque close to 0 RPM, and is fuel efficient during cruise. About the only thing a hybrid would add would be regenerative braking. While that's a big deal in city driving, the vast majority of the driving tractor trailers do is on the highway, so again there's little benefit from the hybrid. The best thing to add to a diesel is actually a turbo. Their weakness is power output at higher RPMs, and a turbo provides extra power at the high-end of the RPM range, which improves accelerating to pass at highway speeds - precisely the driving stage diesels normally have problems with.

Re:Honda Diesel?

[damn_registrars]

Slashdot is hosted in the US. The employees for Slashdot all work in the US. The majority of slashdot readers are in the US. The majority of slashdot commenters are in the US as well. If they want an informed audience and an informed discussion it would be worth pointing out that three of the four brans listed don't sell Diesels in the US.

Re:Honda Diesel?

[halivar]

Why? It's not like Slashdot is a US centric web page...

In all actuality, it really is. Slashdot covers US politics to an extent that it covers no other country (or even perhaps all of them combined). And it's not "politics in America affects everyone", either: I can't for the life of me figure out why, say, a Scandi cares about H1B tech hires in California.

Couple of Engineers

[Jason Levine]

The VW boss recently said "It's the decision of a couple of software engineers, not the board members." It looks like those two software engineers snuck into all these other car companies and altered their systems also! How nefarious!

Re:Diesel are more eco-friendly than gasoline

[0123456]

Yes. Blasting a black cloud of carcinogenic particulates in my face every time you accelerate is definitely eco-friendly.

Cheating on the Test

[e4liberty]

There was an article in last week's Economist on this. From recollection... in Europe, the testing is not done by an EPA-equivalent government agency, but by third party test labs. There, to get the business, the testers allow the auto manufacturers to rig the test: remove mirrors, remove all weighty optional equipment, remove seats, tape the door and window cracks, etc., etc. In other words, they are not testing the same car that they are selling.

Re:Honda Diesel?

[grimmjeeper]

It's a Jeep. If there ain't no oil under 'em, there ain't no oil in 'em.

Re:Diesel are more eco-friendly than gasoline

[mrchaotica]

Here's a newsflash: particulate emissions are regulated by mass, but biological harm is proportional to the number of particles. The fact that those carcinogenic particulates from Diesels are big enough to form a visible cloud means they're less dangerous than the much larger number of tiny invisible particulates that gasoline engines emit.

(Not to mention, the modern Diesels being discussed have particulate filters -- which do actually work; the "emissions cheating" is about NOx, not particulates -- but modern gasoline engines still don't. And by the way: gasoline engines emit a fun mix of toxic substances such as benzene and formaldehyde that are much lower in Diesel emissions, and which are totally unregulated. New gasoline engines are way more carcinogenic than Diesels, even by a wider margin than they used to be.)

News reports: Volkswagen used special hardware.

[Futurepower(R)]

"Volkswagon's mistake..."

Apparently it wasn't a "mistake". Apparently Volkswagen used special hardware and software to break the law.

Yesterday on PBS NewsHour the CEO of Volkswagen said the dishonesty was the fault of unknown rogue software engineers, and no managers knew about it. However, special hardware was designed into the system; that couldn't have happened without help from other people in the company, including hardware buyers.

See this article: Older VW diesels will need software and hardware fixes, Horn tells lawmakers.

The CEO seems to be lying deliberately. He says "software". Then later mentions "hardware".

That Auto News article was apparently written by someone who doesn't understand that, if hardware is required, the dishonesty must have been approved by Volkswagen management.