Uncertainties in emissions measurements in a partial flow sampling system

This paper investigates experimental uncertainties associated with gaseous and particulate emissions measurements in a partial flow emissions sampling system developed and built at the Larson Transportation Institute of the Pennsylvania State University. A small fraction of the tail pipe exhaust is diluted with dilution air and passed through a cyclone to eliminate particles bigger than 2.5 microns. The diluted exhaust is then passed through a 47 mm Teflon filter for gravimetric measurement of Particulate Matter (PM). Mass flow controllers operating at 5Hz are used to control the flow rates of dilution air, diluted exhaust, and proportional flow of diluted exhaust into a Tedlar bag. An ultrasonic flow meter is used to measure flow rate of tail pipe exhaust. At the end of a test, the concentration of gaseous emissions in the bag, namely CO₂, CO, HC, and NO_x are measured using a bag emissions analyzer. The mass of gaseous and PM emissions in the tail pipe exhaust for the duration of a test are calculated using the measured gas concentrations in the bag and weight gain of the filter respectively. An uncertainty analysis on the results from an actual test using this system is presented. The propagation of uncertainties in the subsystems used for gas concentration measurements, filter weight determination, and flow rate measurements are used to determine the uncertainties in the results, using the Root Sum Square (RSS) method. The measurement uncertainties from this partial sampling system are then compared to the measurement uncertainties from a full flow dilution tunnel equipped with emissions analyzers for the same test, when both systems are drawing samples simultaneously.