Automotive e-ssentials

When referring to the report published by PANORAMA on ARD (German TV channel) on 05.10.2017,
a feeling arises that car manufacturers are now allowed to cheat at type approval tests with the authorization of the type approval authority.

To understand the whole procedure of type approval and emission testing, certain knowledge of the applicable regulations is required. Even experts, who are confronted with the regulations on a daily basis, sometimes struggle with the new emission regulations, because the whole procedure has become quite complex. This article provides a slight insight into the new emission regulation, with a strong focus on the transfer time from NEDC into WLTP – the so-called correlation period.

The CO2 fleet targets (EC 443/2009) have been set based on values determined by driving the NEDC. Now that the new worldwide harmonized light vehicle test procedure – WLTP is introduced, where not only the cycle but the procedure as a whole is amended, new values for the CO2 fleet targets need to be estimated.

Since experience with the new procedure is missing and new targets cannot be estimated easily, a procedure has been created for the so-called transition phase where the “new” WLTP CO2 values are translated into “old” NEDC CO2 values. Both values are mentioned in the CoC document in the end.

For the “translation” the CO2MPAS tool has been created by the JRC (Joint Research Centre, hosted by the European Commission). The tool itself is nothing less than a complex engine/vehicle simulation model which is self-adapting to given input data. In order to gain resilient results, it need to be fed with several information on the engine and the vehicle, as well as information on the driven WLTP cycle of the vehicle dedicated to be simulated. Generating accurate results requires the input data to be of a certain accuracy, otherwise the simulation could lead to an inconsistent result.

In general, the new regulation (EU 2017/1151) introduces plenty of additional calculation routines into the type approval process. Calculation is not only requested in the estimation of CO2 Emission and fuel consumption, but also in the field of RDE (real riving emission) measurements. For example, RDE test results are assessed by mathematically correlating the RDE test results with the emission results of a chassis dynamometer. Keeping in mind, that all mathematical models strongly depend on the accuracy of the input parameters, special attention has to be paid. As mentioned before, the results are only as accurate as the input data.

Therefore, a certain accuracy of the input values used within the CO2MPAS tool is demanded. Otherwise the tool will not work properly and the created results are not resilient.

To understand the position of the CO2MPAS tool it is essential to know where in the type approval process this tool is situated.

This graphic seems really easy to understand, but the part “CO2MPAS” within this process is not only the simple simulation, but also covers several more steps before and after the simulation.

Each step is explained in some detail in the following.

Pre-Processing

The required input data file for the calculation consist of certain information on the engine and the vehicle, such as the following:

Furthermore, information on the driven WLTP test cycle need to be provided such as the following:

• Dyno velocity
• OBD velocity
• OBD engine Speed
• Engine coolant temperature
• OBD engine calculated load
• Alternator current
• Battery current

All this information needs to be provided by the manufacturer except for the information on the NEDC road loads, which need to be determined by the technical Service. In case of an interpolation family, the information on road loads and CO2 values need to be provided for both vehicles, H as well as L.

The system of building families is one more subject introduced with the new regulation. Though, the subject of building test families is not really something totally new, as it has been introduced in the field of RDE measurements already. But, now the system of building test families gets extended and several more test families are created. One of these new families is the interpolation family.

In contrast to the old regulation, vehicle types with regard to emissions are no longer separated by their mass and road load only, but the whole cycle energy demand is the new criteria. Vehicle H in this sense refers to the vehicle with the highest energy demand, while vehicle L refers to the lowest cycle energy demand. Cycle energy demand means the total consumed energy when driving the WLTP.

Simulation

If the technical service has approved the provided data, the simulation is performed. The simulation of the vehicle H and L are performed by the technical service.

Like with the WLTP testing, the vehicle is tested against the NEDC specific manufacturer declared value. There are two possibilities for the simulated NEDC result:

• CO2MPAS approves the declared value
• Simulated value exceeds the declared value by more than 4%

In the first case, everything is fine. The declared value is approved as type approval value. In the second case, the manufacturer has two options:

• Accept the higher value as type approval value
• Demand the vehicle to be tested on a dyno

No matter what, each simulation result is to be time stamped and get diced. Yes, with the new regulation, results are about to be diced, where in this case results refer to a decision if an additional physical is to be performed.

Dicing means, the results are sent to a time stamp server, which stamps the results, for having a reference when the simulation has been performed. The response from that server is a random number between 1 and 99.

Post-Processing

If that number is between 90 and 99 the vehicle is selected to be tested in NEDC. In case that the number is between 90 and 94 vehicle L is tested and in case that the number is between 95 and 96 vehicle H is tested. These tests are the so called “Random Tests”. As it is obvious, 1 of 10 vehicles is about to be tested additionally under NEDC conditions.

In case that CO2MPAS did not approve the declared value and the manufacturer demanded the vehicle to be tested on a dyno, the response from the server can be ignored.

If the vehicle is tested on a dyno, the declared value has to be approved. Up to three tests are performed. If the first test measurement exceeds the declared value by not more than 4%, the declared value is approved as type approval value. If the first test measurement exceeds the declared value by more than 4%, another test is performed. If the average of the two test does not exceed the declared value by more than 4% the declared value is approved. If the average exceeds the declared value by more than 4% a third test is performed and the average of the three tests will be the type approval value. These tests are the so-called “Double Tests”.

If a vehicle is physically tested in the NEDC the deviation factor is determined. This factor displays the difference between the declared value and the test result and is mentioned in the CoC documents.

To conclude, with the new regulation a whole lot more of calculation is performed within the type approval process, including many possibilities for manipulation. But, is CO2MPAS a tool to perform authorized cheating as portrayed by PANORAMA?

NO!

Because, not only the new value is documented, but also the sources where it is coming from. Furthermore, when a physical test is performed the verification factor is determined as well. Despite, it needs to be mentioned, that only within the phasing-in of the WLTP the CO2MPAS tool is used.

The following timeline provides an overview of the usage of CO2MPAS for new types, existing types and end of series types.

So, for the next three years, calculation, simulation and dicing will play an important role in the procedure of type approval.

For further information, please contact our expert Frank Exner .