Predicting On-Road Particle Number Concentrations of Light-Duty Gasoline Vehicles from Gas Concentrations with Time-Series Cross-Section Regression

Abstract

There is a need to quantify real-world particle number emissions from light-duty gasoline vehicles (LDGVs) because of the known toxic properties of ultrafine and nanoparticles, the large number of LDGVs on the road, and the lack of field measurements of particle number (not mass) emissions during real-world vehicle operation. Onboard tailpipe gas and particle number concentrations and vehicle operating data were collected second by second with an instrumented 1999 Toyota minivan driven multiple times on a 17-mi test route by 22 drivers. Time-series cross-section regression analysis was applied to individual test runs to develop a suitable model for predicting particle number concentration based on gaseous pollutant concentrations and sampling conditions. The results indicate that particle number concentration can be effectively predicted from gas concentrations, ambient air temperature and relative humidity, and exhaust temperature. The model signifies the physical relationships between particle number and gas concentrations with respect to gas-to-particle formation processes and implies the possibility of future use of existing gaseous pollutant concentration databases for particle number prediction from a wider range of vehicles.