The emissions are measured using a Portable Emissions Monitoring System (PEMS). This directly measures the pollution from the exhaust. The PEMS also measures the speed, acceleration and position of the vehicle to ensure the car is driven in a style typical of the driver of the model. There is no electronic connection between the PEMS and the engine management system of the vehicle so the car cannot detect it is being tested. This makes it much more difficult to deploy a defeat device to trick the test and put the car into a low emission mode.
The test itself is conducted on roads to the west of Paris. It has 3 parts, urban, suburban/rural and high speed. Although the route driven is the same the CO2 is calculated for each phase of the test. After the test the overall CO2 figure is calculated using a weighting typical of the average use of the car derived from a PSA database of car usage. We also ensure the vehicle is appropriately loaded based PSA data which shows variations between 1.6 and 2.1 for different models. The test result is also adjusted to account for varying weather conditions, cold starts and diesel particulate filter regeneration events.
The way the car is driven is crucial to the test results. A mix of profession and amateur drivers are used. If the driving style is unrepresentative of the average use of the car the test is rejected and repeated. Several legitimate tests are conducted and early indications are the results reproducible to within 0.1-0.2 l/100km. During the development phase we have been checking the PEMS measurements in a lab using the WLTP cycle and achieved an excellent correlation for diesel cars – further work is ongoing to finalise measurements for petrol vehicles.
To summarise – the test is best practice, scientifically robust, representative of an average driver of the model and conducted in a way to be sufficiently reproducible. The results provide customers with the information they need to select the most efficient model for their use. It is more representative of real world emissions than the new WLTP lab test. We hope other manufacturers take up the opportunity of apply the method to their own models in the future and this becomes the default approach for reporting fuel economy.