Tuesday, 8 November 2016

Any ‘smoke and mirrors’ behind VW’s EA189 ‘fix’?

Does everyone understand what really lies behind VWs powertrain engineers’ implementation of their ‘fix’ for the EA189 engines? What is really implied in the ‘fix’?
Maybe everyone in the powertrain industry, and particularly those employed in the likes of various consultants, such as AVL, FEV and Ricardo, does fully understand the game VW has been playing when it allegedly contrived to allow cars fitted with its diesel engines seemingly to meet US Federal emission requirements.
However, there must be a groundswell of those who do not understand, yet want to know.
The latest pronouncement (4 November) from the company only clouded the issue; it did not give any in-depth explanation for the mysterious algorithm which had found its way into the software of those diesel engines powering vehicles subject to US emissions tests. And not just one model of engine, but various engine sizes.
The company merely stated, on 4 November 2016, that the KBA (the German Federal testing agency) had declared "without qualification that the implementation of the technical solutions for these models causes no detrimental changes to fuel consumption levels, performance data or noise emissions. The KBA had previously already affirmed this for all other models approved for the recall. After the modification, the vehicles also fulfil all statutory requirements and the duly applicable emissions standards."
Note those 12 key words: "no detrimental changes to fuel consumption levels, performance data or noise emissions."
How is this possible? How is it possible that both fuel economy and NOx levels can feel “no detrimental changes”? And while noise (NVH) is not a major factor it too does carry implications, certainly in terms of combustion-generated noise. So how is it possible that all three are not subject to "no detrimental changes"?
 And, if is possible, why was the algorithm installed in the first place? Yet, in effect, there must have been a need for it – because it was done. The VW vehicles seemingly could have been set up to pass the US emissions tests, using the latest ‘fix’, without the need for deceit.
So why were the algorithms installed?  Presumably to allow the cars to pass the US tests yet at the same time give sporty performance (but illegal emissions) when in the hands of customers. Who sanctioned the algorithm? And who created the algorithm? Was this devised by VW engineers? Or did a vendor write the algorithm and install it in its chip?
Volkswagen, in its 4 November 2016 statement talks of a software “update”. What is actually meant by “software update”? And what is actually entailed?
Does it imply partially rewriting the program or detecting the EA189's previous software (or programs)? Or will there be the physical removal of computer hardware (a chip, for example) when an owner’s car visits the VW service workshop??
The company statement also gives no insight as to what these software “updates” are designed to accomplish. What are the updates correcting? Which aspect of the engine management system is being affected? And why?
Is there, for example, an element of a requirement in the “update” to retard the fuel injection in order to lower peak combustion pressures and temperatures; the higher the combustion temperature the higher the level of NOx.
                               West Virginia University researchers
It may be recalled that way back in autumn 2014, researchers from the US’s West Virginia University received a grant to evaluate tailpipe emissions of diesel cars on America but made by European manufacturers.
The team consisted of university professors, Gregory Thompson and Dan Carder, and two students, Marc Besch and Arvind Thiruvengadam. All four aimed to collect as much data as technically possible.
The West Virginia University school’s Center for Alternative Fuels, Engines and Emissions were awarded the assignment as they possessed portable emission measurement system to stick in the car trunk, attached to a probe to shove up the exhaust pipe.
The grant for the tests came from the International Council on Clean Transportation (ICCT) a small non-profit organisation dedicated to helping to reduce vehicle emissions. It has an annual budget of $1 million and comes under the responsibility of John German who, in this case, was further prompted by Peter Mock.
Testers drove the monitor-equipped diesels from San Diego to Seattle. They were wary that Volkswagen may have rumbled that the vehicles were going to be subjected to testing. Indeed, VW may have in some way anticipated the emission test but the testers couldn’t be sure how so they took the vehicles on extended route testing.
It may be recalled also, that in another cheating case years ago, long-haul trucks were equipped with devices that allowed the engines to gradually discharge more and more harmful nitrogen oxides the longer the vehicle cruised at the same speed. The more emissions, generally speaking, the greater the engine power.
The researchers’ 1,300-mile trip under varying conditions was intended to expose any such scheme that VW may have implanted in the vehicles.
Meanwhile, the California Air Resources Board (CARB) tested similar vehicles in their laboratories and the vehicles passed the tests.
On the open road, the VW Jetta exceeded US nitrogen oxide (NOx) emissions standard by 15 to 35 times. The VW Passat was 5 to 20 times the standard. However, a BMW X5 passed the road test.
The California watchdog and the US Environment Protection Agency (EPA) opened an investigation into Volkswagen in May 2014. Talks between the parties went on for several months, with VW trying to replicate the West Virginia University results. The company said it had identified the reasons for the higher emissions and proposed a fix. That resulted in a recall of nearly 500,000 U.S. vehicles in December to implement a software patch.
                                Trying to confirm the software ‘patch’
The California agency continued to test VW cars after the recall began. It was concerned that real-world road tests couldn’t confirm that the software patch was working. Sure enough, nitrogen oxide emissions were still in violation of California and U.S. laws. The agency shared those findings with Volkswagen and the EPA on July 8 2014.
It was later discovered that an algorithm could detect when vehicles were undergoing emissions testing and thereby allowed the vehicle to deliver the correct result.
When the vehicle passed into the hands of the owner the algorithm knew this and allowed the software to return to a state where performance became preeminent possibly at the expense of fuel economy
Now magically, all is well, mainly by fitting an “air conditioner” and the software “update”. And, hey presto, fuel economy, noise and performance are not affected.
Is this a prime case of smoke and mirrors? Is there a smell at the bottom of the garden?
 If it is not ‘smoke and mirrors’, then what have VW's engineers come up with? What is their magic ‘fix’; the “update” – the “patch”?
The company should explain itself in greater detail. VW possibly believes, in its high-handed way, the buying public has no need for such detailed information. However, it has a duty to treat kits customers in an adult manner. It is up to the public to decide if they ‘need to know’ exactly what it is engineers have done and why.
Certainly, VW should explain why engineers fitted a special algorithm which could spot when an emission test was imminent.
It is akin to the airline pilot, blind in one eye but who, attending an eye test, beforehand manages to inject himself with a ‘magic’ drug which, for the duration of the test, gives him the necessary qualifying 20/20 vision in both eyes.
Is something going on with these engine which VW claims to be able to fix which the man is the street does not know about? The algorithm effectively deceived the testers to the point they allowed the vehicle to be certified when in effect they should not. Someone must have authorised the algorithm. Such elegant software does not find its way into an engine manage to system,
How is it possible that performance and fuel economy are not affected? Improving performance invariably means a knock on effect on fuel economy. There has to be a price to pay.
On 4 November 2016, Volkswagen Group said: "The technical solutions for 2.6 vehicles with EA 189 1.6-litre TDI engines have been approved by the Federal Motor Transport Authority (KBA included modifications that involves a software update.”
In addition, a so-called “flow conditioner” is being fixed directly upstream of the air mass meter. The implementation will take less than an hour of working time.
What is this “flow conditioner”, what does it do, how much does it cost and why was it not fitted before at the production stage?
And so here we are today. In one sense no further forward than two years ago. Algorithms do not happen of their own accord. Software is known to "crash" but not in such a well-orchestrated and convenient manner.

Interestingly, the engines affected all had to comply with Euro 4 and 5 and therefore do no employ SCR which gives the benefit of lower NOx levels and improved fuel economy but at the expense of owners having to top up with Adblue.

No comments: