Monday, 12 October 2015

That Car Company and NOx emissions


I admit it. I own one of those cars that is affected by the scandal surrounding a certain German car company. I chose the car because the specification said it had the performance to do the job required as well as adequate safety systems and creature comforts. I trusted the manufacturer and the regulators to ensure that the car met current legal and environmental requirements. As far as I was concerned, it was a "black box”*.

On the face of it, it appears that the company stands squarely in the dock but I think the testing agencies are there too.


As identified in a recent WSJ report, the company has been hoodwinking the US (and possibly the European) test agencies. The EMS - a computer that controls the engine based on data from sensors littered all over the vehicle, including the driver’s controls - has been programmed to recognise when the car is in a test situation and alter the engine performance accordingly. Now that the US EPA has discovered what’s going on it is not difficult to imagine how it was done.

Several questions for the company arise from this discovery.

The first relates to the management “culture” of the company and, in particular, of that of the design team responsible. The CEO of the company in the US (Michel Horn) has stated to the US House Energy and Commerce Committee that it was the work of a few “individuals” and happened without the knowledge of the board.

This is strange, for it appeared that the company had made some significant break-through in Diesel technology when the new engine was announced. The brilliant engineers had managed to cajole their diesel engines to meet a tough environmental standard in a way that not only delivered good engine performance and fuel economy but also without the costly paraphernalia of Urea injection systems that other diesel engine manufacturers needed. This was not questioned at the time by the motoring journals, the public, other manufacturers or the Board of Directors of the car company. 

It appears that this Board had expressed no interest in how this feat had been achieved. Given the way this company has prided itself on both its technology and know-how, one would have thought that someone in the organisation would have questioned this and, had they been given a truthful answer, been suspicious. It's not that the senior management should know how their engineers achieved the result but that they should have ethics policies and audit procedures in place to make sure that the result had been achieved according to these policies. Or perhaps the company management style and reward mechanisms are so strong on achievement and conformity that no one dared to ask the question. The panel of a recent BBC podcast of “The Bottom Line” certainly seem to think so. 

Quite where the company goes from here is difficult to plot; they have to find a technical fix for the 11 million or so vehicles that are affected, one that will meet the demands of both the regulators and their existing customers. Neither will be easy. There’s talk of a “class” action being taken out by lawyers on behalf of (and crowd funded by) the "deceived" customers.

The regulatory test agencies should be in the dock too. They have two questions to answer. One is how they originally established the standards and the required tests; the second is how can they ensure that not only do the engines now meet the mechanics of the test, but that they meet its intention too – e.g. the emissions are controlled in all conditions.

The test standards and processes are developed jointly with the motor industry: this may compromise both the standards and the processes.

This is a regulatory system failure akin to (but perhaps not as economically serious as) the changes to bank regulations that are thought to have triggered the financial crash of 2008.

The second point, which is really a sub set of the first, is how the test authority can ensure that the cars meet the intention of the standard Perhaps they have to adopt the type approval testing applied within the aircraft industry - the test authority should insist on access to the computer code.

Would the authorities have the resources and capability to examine it in the detail required? If not, could the verification be “crowd sourced” by making the code “open access”? At present such code is the intellectual property of the company. In some jurisdictions, non-franchised repair organisations find it difficult to gain access to it for diagnostic use. Once the code is made public, clever programmers will be able to subvert it and provide customers with software to defeat the probable changes made to comply with the regulations.  David Golumbia discusses this issue in his article in Uncomputing .

Let’s assume that the car company can satisfy the regulators that the changes they plan to make to the affected engines are effective: it’s still not the end of the issue.  At present, there is no legal requirement (in the UK) for an owner of an affected vehicle to have it retrofitted. Including a test of these emissions in the mandatory annual test of roadworthiness (due after a vehicle is three years old) would force the issue. It is possible that such measurement may be required from some time in 2018. This gives the grey industry plenty of time to develop test beating “retrofit” kits that can be removed prior to the test - already, owners can “chip” their vehicles to improve performance. Such “test beating” might be discovered if would be possible to measure the emissions from specific vehicles from the roadside – so far, such measurement does not appear to be on the horizon local or government agencies within the UK.

Whilst on the subject of software in cars, another question arises. Computers are ubiquitous in cars – they control significant safety systems: traction control, anti-skid, and cruise control and navigation. There have been several demonstrations of hackers taking control, the most recent by a wire free remote connection.  

Computer control will soon be extended to the full “autonomous” vehicle. This poses the question of whether regulation can move fast enough to protect the public from the unintended consequences of the next technology. The UK government has made a start in this direction and published a set of guidelines under which such vehicles may be tested but these have yet to be translated into regulations.

The recent revelations also pose a moral question to be answered by users of the vehicles affected for there have been claims in the UK press that NOx emissions by diesel cars account for a number of deaths from pulmonary diseases each year. The sources for this information are unclear. The UK government committee on the medical effects of air pollutants in April of this year was uncertain and called for more research.

So what should the customer do now? Is continued use acceptable until the modification is applied? Is the modification just a technical fix with further unintended consequences? It might just push me (and the Company) into going electric.

But the regulation of test processes has to be resolved soon.

Notes


* It is black, as it happens but by "black box" I'm referring to is piece of finished technology as discussed in the book "Science in Action" by Bruno Latour, published by The Open University Press.


No comments:

Post a Comment