Sardar, S., zhang, S., Larsen, L., Frodin, B. et al., "Evaluation of PM Measurement Precision and the Quivalency of the Single and Three Filter Sampling Methods for LEV III FTP Standards," SAE Int. J. Engines 9(1):342-354, 2016, doi:10.4271/2015-01-9045.
Present motor vehicle particulate matter (PM) emission measurement regulations (Code of Federal Regulations (CFR) 40 Part 1065, 1066) require gravimetric determination of PM mass collected onto filter media from dilute exhaust. To improve the current sampling and measurement procedures for TIER 3 PM emissions standard of 3 mg/mile, CFR part 1066 adopted five alternative PM sampling options. One option of great interest is sampling the entire test using a single flow-weighed filter rather than the conventional three-filter (one filter per test phase) approach. The single filter method could lessen the time needed for gravimetric determination by reducing the quantity of filters used for a test and possibly reduce the uncertainty in gravimetric measurements, particularly at sub 1 mg/mile PM levels. This study evaluates the single filter and, to a limited extent, the 2-filter alternatives adopted in 40 CFR Part 1066. The first part of this study comprised evaluation and equivalence assessment of a Horiba Quad PM sampler along with an AVL 472 SPC PM sampler. PM emission data were collected at Air Resources Board's Haagen-Smit Laboratory (HSL) over Federal Test Procedure (FTP), Urban Dynamometer Driving Schedule (UDDS) and US06 cycles for twelve vehicles. Measurement precision for both samplers was determined to be 11% for PM loadings in the 10 to 60 μg range. Furthermore, statistical analysis determined that the majority of CFR-compliant PM sampling instruments are also likely to have measurement precision close to 11%. The second part of the study assessed the equivalency of PM emission measurement using the single and 2-filter sampling options in conjunction with the conventional 3-filter method. A total of eight vehicles, which comprised a subset of the vehicles tested for the evaluation of the sampler equivalency and estimation of measurement precision, were tested over the FTP and UDDS cycles. Two test vehicles had FTP PM emissions of approximately 3 to 4 mg/mile and six vehicles had emission levels below 1 mg/mile. Results for the single filter method, when compared with the conventional 3-filter method, indicated very high correlation with R2 close to 0.99 and 0.92 for PM emission levels below 3 mg/mile and sub 1 mg/mile, respectively. The data also indicated that the single, the 2-filter, and the 3-filter methods had similar variability of measurements irrespective of the vehicle's emission level. Interestingly, the single filter method yielded slightly lower (less than 10%) PM emission compared to the 3-filter method. The difference is primarily attributed to background contributions (three vs. one) and/or likely the loss of the semi-volatile fraction of PM from the single filter because it is subjected to a vacuum environment for a longer period of time as compared to the conventional 3-filter method. Nevertheless, tunnel blank corrections demonstrated better correlation and convergence of data between the two methods. Statistical analysis (t-test) indicated no significant difference between the single and the conventional 3-filter method. The single filter method has the prospect of alleviating measurement challenges at the 1 mg/mile level to support implementation of the LEV III standards.