Investigation of the Effect of Pavement Deflection on Vehicle Fuel Consumption: Field Testing and Empirical Analysis

The results presented in this report are part of Phase II of a two-phase study. Based on the results from mechanistic models of additional fuel consumption in vehicles due to the structural response of the pavement structure, Phase I of this study concluded that pavement has a small but important enough effect on vehicle fuel consumption to warrant field investigation. The goal of the Phase II study was to measure vehicle fuel consumption in the field on different pavement types in winter and summer and at different speeds, and to use the data collected to develop empirical models for this fuel-consumption effect. The field investigation presented in this report included 21 California pavement sections with different pavement types: flexible, semi-rigid, jointed plain concrete, continuously reinforced concrete, and composite structures. The vehicles selected and instrumented for the fuel economy measurements included a five-axle semi-trailer tractor, a diesel truck, a sports utility vehicle (SUV), a gasoline-fueled car, and a diesel-fueled car. Vehicles were run on cruise control and data were recorded at 45 and 55 mph on state roads and at 35 and 45 mph on local roads. The data from the field investigation were analyzed and used to develop an empirical modeling framework considering road geometry, wind, temperature, and pavement structural and surface (roughness and texture) effects on vehicle fuel consumption. Based on the final framework, a final empirical model was developed for each section. The report presents results of a factorial analysis of the effects of each variable using the final model for each vehicle type on each pavement type and in different California climate regions. The within-section variability is almost always greater than the variability between sections for a given pavement type and efficiency condition (tailwind, speed, and climate region) and the within-section variability is also usually larger than the variability between pavement types. Only the data for the heavy heavy-duty truck (HHDT) showed any meaningful difference in results between sections, but that variability is not tied to pavement type and is only present under certain conditions of speed, tailwind, and air temperature (tied to climate region). These results indicate that missing variables (or errors in the existing variables) need to be reduced in further experiments to observe measurable effects of pavements on fuel consumption in real-world driving. While air temperature interacted with cruise control speed for the HHDT, there was a lack of clear evidence that asphalt roads cause more fuel consumption for the HHDT even under the conditions where the most possible effect of pavement type was found. This suggests that pavement type is not the correct explanation for that variation. Instead, the variation in the effect of air temperature by cruise control speed for the HHDT likely has to do with differences in engine efficiency under different conditions.