Summary

As for light duty vehicles, heavy-duty vehicles are subject to the introduction of new technologies to improve fuel efficiency and to decrease exhaust gas emissions. For this reason, demonstrations have been set up all over the world. It appeared however that the results of heavy-duty demonstrations are difficult to compare, mainly due to the variation of technologies, vehicles and fuels used. This is especially true for the emission measurements. Emissions are measured over a variety of test cycles, on engine level and on vehicle level. The relation between the results of these test procedures and the emissions actually produced on the road is largely unknown, as well as the relation between the test procedures among each other.

Within this project, three city bus technologies were selected to compare emissions and fuel consumption in real traffic (city and rural), in several vehicle test cycles (CBDC, DUBDC, De Lijn) and in the main official engine test cycles (ESC, ETC, US-FTP, Japan 13-mode). The purpose was to look for clear relations between these test procedures and to see in what way they reflect real traffic emissions. The three buses were a Euro-2 diesel bus, a natural gas bus with stoichiometric fuel control and three-way catalyst and a natural gas bus with lean burn fuel control.

When comparing fuel consumption of the three technologies in real city traffic, it can be concluded that the natural gas buses both had clearly higher fuel consumption (in diesel equivalents) than the diesel bus, caused by a lower average engine efficiency and a higher curb weight (due to heavy CNG cylinders). The stoichiometric natural gas bus reached very low exhaust gas emissions compared to the diesel bus (most regulated emissions were around 10 times lower). The lean burn natural gas bus needed adjustments in its lambda control setting in order to lower its high NOx emissions.

The relation between real city traffic and the simulated city cycles differs from technology to technology, also depending on the acceleration capabilities of the buses.

The DUBDC cycle was very comparable to the real city tests (in Brussels) for the diesel bus, but was too demanding for the natural gas buses, leading to a higher average engine load compared to real traffic and consequently also to differences in emissions. The CBDC cycle was very comparable to the real city tests of the lean burn natural gas bus in Hamilton , Canada . For the other two buses the engine load distribution in the CBDC cycle differed more from the real city tests.

As for the simulated city tests, there is no unique relation between the real city traffic tests and the different engine test procedures. To some extent this is also due to the different gearboxes in the three tested city buses. The US-FTP test and the ESC test clearly have a different engine load distribution compared to the real city traffic tests. For the US-FTP this leads to a serious underestimation of NOx emissions of the lean burn natural gas engine. The ETC test (for the diesel engine) and the Japanese 13 mode test (for the lean burn natural gas engine) have a more realistic distribution of engine load. Because of the lack of transient effects, stationary engine tests (ESC, Japanese 13 mode) lead to an underestimation of THC and CO emissions.

The results of this project show that comparison of different test procedures is not straightforward. While specific test procedures do reflect real traffic emissions for one technology, this is not a guarantee to get acceptable results for other technologies.

Participants

• Belgium
• Finland
• Italy
• Japan
• Netherlands
• Sweden
• USA