Portable Emissions Measurement System - Accuracy of PEMS

Accuracy of PEMS

The question often arises as to the target accuracy of PEMS. As PEMS are typically limited in size, weight and power consumption, it is often difficult for PEMS to offer the same accuracy and variety of species measured as is possible with top of the line laboratory instrumentation. For this reason, objections were raised against using PEMS for compliance verification.

On the other hand, fleet emissions deduced from laboratory measurements can be subject to significant inaccuracies if the selected engines and operating conditions were not representative of the fleet, or if deliberate anomalies (i.e., dual mapping of the ECU) were not demonstrated during laboratory testing.

The question of how accurate a monitoring system needs to be therefore cannot be objectively answered, neither can a monitoring system be easily designed, without first considering the intended application of the system and the errors associated with different approaches.

It is expected that a variety of on-board systems will be designed, ranging from suitcase-sized PEMS to instrumented trailers towed behind the tested truck. The benefits of each approach need to be considered in light of other sources of errors associated with emissions monitoring, notably vehicle-to-vehicle differences, and the emissions variability within the vehicle itself. In other words, one needs to consider the total of:

1. The difference between what is measured and what is actually emitted during a test
2. The difference between what is emitted during the test and what the vehicle emits during its everyday duties
3. The difference between the emissions characteristics of the tested vehicle and the overall emissions levels of the entire fleet.

For example, when evaluating a benefit of cleaner fuels on a fleet of city buses, one needs to compare taking a bus out of service, installing a laboratory-grade monitoring system, loading it with sandbags and driving it on a simulated route against testing several buses on their regular routes, with passengers on board, using a simpler (and possibly less accurate) monitoring system.