The United States Environmental Protection Agency (EPA) and California Air Resources Board (CARB) agencies are implementing a series of regulations that will control emissions, including oxides of nitrogen (NOx) and particulate matter (PM), from diesel engines during “in-use” conditions. The purpose is to ensure the emission standards can be maintained throughout the course of the engine’s useful lifetime. One of the most important regulations with respect to controlling in-use emissions is the Not-To-Exceed (NTE) regulation. This regulation sets limits for pollutants that are emitted during operation in a defined portion of the engine map and specifies the protocols required to make those measurements. Portable Emissions Measurement Systems (PEMS) are critical for the implementation of these in-use regulations. The EPA, CARB, and the Engine Manufacturers Association (EMA) formed a measurement allowance steering committee (MASC) to develop a “measurement allowance” to account for measurement error associated with the use of PEMS for in-use measurements. A comprehensive program was completed for the gas-phase measurement allowance (Miller et al., 2008; Buckingham et al. 2007; Fiest et al. 2007). This report presents the results of the in-use validation portion of the PM measurement allowance program. The MASC approach pursued an experimental program to try to use laboratory measurements and modeling to characterize errors that might be observed with in-use, and then validate those errors with in-use testing. The development of the MA involved a sophisticated Monte Carlo model that considered laboratory measurement errors associated with engine broadcast information, emissions measurements, affects due to ambient conditions, and affects that might be observed due to on-vehicle operating conditions. The in-use validation effort used the same PEMS systems used for the laboratory testing, but under real-world conditions. The idea was that the errors found in-use should be within the bounds of the proposed MA. For this program, comparisons were made between the PM PEMS and UCR’s Mobile Emissions Laboratory (MEL) reference laboratory under in-use on-road driving conditions. The MEL is unique in that it contains a full 1065 compliant constant volume sampling (CVS) system with gravimetric PM measurements, while being fully operational under on-the-road driving conditions. Two PEMS manufacturers were selected for the in-use validation testing, and multiple serial numbers of each of the selected PEMS were included in the in-use testing. Measurements were made from one class 8 heavy-duty diesel vehicle equipped with a original equipment manufacturer (OEM) diesel particulate filter (DPF). A bypass system was designed to simulate a failed DPF while maintaining the functionality of the diesel oxidation catalyst (DOC). The bypass was designed to target an in-use bsPM emission level of 25 mg/hp-h. In-use routes were designed and utilized to exercise the PM PEMS equipment over a range of environmental conditions, and included segments near sea level, in coastal regions, in desert regions, and longer uphill incline segments and segments at elevations up to 4500 ft. Prior to validation testing, the MEL underwent a series of 1065 audits and a cross laboratory correlation with SwRI. All PM PEMS and PM instruments tested showed a negative bias compared to the reference system. The PEMS2 non-regeneration mean bias at the 20 mg/hp-h bsPM emissions was -10 mg/hp-h, and at 30 mg/hp-h the mean bias was -18 mg/hp-h. When the intercept was forced through zero, the mean bias at 20 mg/hp-h went to -15 mg/hp-h. The PEMS3 non-regeneration mean bias at the 20 mg/hp-h bsPM emissions was -1.7 mg/hp-h, and at 30 mg/hp-h the mean bias was -2.2 mg/hp-h. When the intercept was forced through zero, the PEMS3 mean bias went to -1 mg/hp-h at the 20 mg/hp-h level. The other instruments used showed a higher mean bias than PEMS3 and lower mean bias than PEMS2. The proposed bsPM MA is 6 mg/hp-h at the 20 mg/hp-h in-use standard. For these results, the PEMS3 system was within the proposed MA, while the PEMS2 values exceeded the proposed MA. Several issues were discovered during testing, including problems related to testing under in-use conditions, operational issues, and post processing issues. The in-use issues ranged from electrical and mechanical connections, crystal usage from short NTE’s, valve switching, measurement signals, and crystal behaviors. Operational problems occurred during startup, commissioning, and with the systems prior to testing in-use. Typical issues include incorrect system configurations, procedures that don’t work, and issues with the startup software and other recommended practices that didn’t function according to the manual. The post processing issues ranged from data filtering, bsPM differences between processor versions, data identification, and method calculations not being available. In general, PEMS2 had more issues than PEMS3 for each of these categories.
