Modeling On-Road Particle Number Emissions from a Hybrid Diesel-Electric Bus: Exploratory Econometric Analysis

Document Type

Journal Article

Publication Date


Subject Area

infrastructure - station, infrastructure - vehicle, mode - bus


Vehicle exhaust, Regression analysis, Regression, Particulates, On-board monitoring, Number of particles, Nonstationarity, Multicollinearity, Motorways, Hybrid vehicles, Heteroscedasticity, Freeways, Exhaust gases, Exhaust emissions, Econometric models, Dual fuel vehicles, Diesel electric buses, Controlled access highways, Automobile exhaust, Autocorrelation


The purpose of this econometric analysis is to model the concentration of particle number emissions from a hybrid diesel–electric bus in terms of operating characteristics. Particle number concentrations are modeled instead of particle mass, and the emissions are recorded by using on-board instrumentation in real-world driving conditions. The operating characteristics included in the final models are two engine parameters (fuel rate and engine speed) and two vehicle parameters (velocity and acceleration). The emissions data possess properties that frequently cause problems in regression analysis: nonstationarity, multicollinearity, heteroscedasticity, and autocorrelation. Methods for overcoming and minimizing the effects of these properties are implemented. Aggregation is used to reduce the nonstationarity of the data, and one of the potential operating characteristics is removed to decrease the multicollinearity. The Newey–West autocorrelation consistent covariance estimator is implemented by using ordinary least squares to produce a model that accounts for heteroscedasticity and autocorrelation, without requiring assumptions to be made about the structure of the model disturbances. A first-order autoregressive process is used with feasible generalized least squares as a comparative model. The two models have similar coefficients, fit, and predictive capability. However, the models are specific in scope to typical freeway driving conditions. In future studies, the researchers anticipate applying econometric analysis to model particle number emissions among different routes, bus technologies, aftertreatments, and atmospheric conditions.