Time Resolution Effects on Accuracy of Real-Time NOx Emissions Measurements 2005-01-0674

The advanced development of a previous in-use emissions measurement system developed at the University of Alberta, has been used to illustrate the challenges in accurately measuring real-time mass emissions of NOx, with specific attention given to the issue of sensor time resolution. An analysis of the alignment of vehicle and emissions data has shown constant value time shifting of remote emissions sensor data, to match vehicle data, as the most accurate method for synchronization. Although variable time shifting routines theoretically determine alignment time more accurately, the variable shifting of slow response sensor data has shown an added smearing effect to time shifted remote analyzer data.

The effect of sensor response time on accuracy of mass emission rates, has shown that slow response remote emissions sensors are under predicting the total emissions produced by vehicles. Using data post processing to correct first order time response characteristics of the NOx sensors, resulted in improved time alignment of NOx spike peaks with mass air flow (MAF) peaks, faster response times, and larger peak concentrations. The response corrected g/km emission rates ranged from negligible change to 2.5 times original uncorrected results, indicating the dramatic range with which signal processing can effect emissions results. These results indicate the importance of sensor time resolution, and the benefits which can be realized when utilizing data post processing to correct slow responding sensors. However, limitations in the form of signal noise amplification, uncertainty in peak concentration values due to slow sampling rates, as well as peak NOx catalytic reduction characteristics, introduce uncertainty into its definitive results.