Estimation of energy consumption on the tire-pavement interaction for asphalt mixtures with different surface properties using data mining techniques

Abstract

The energy or fuel consumption of the millions of vehicles that daily operate in road pavements has a significant economic and environmental impact on the use phase of road infrastructures regarding their life cycle analysis. Therefore, new solutions should be studied to reduce the vehicles energy consumption, namely due to the tire-pavement interaction, and contribute towards the sustainable development. This study aims at estimating the energy consumption due to the rolling resistance of tires moving over pavements with distinct surface characteristics. Thus, different types of asphalt mixtures were used in the surface course to determine the main parameters influencing the energy consumption. A laboratory scale prototype was developed explicitly for this evaluation. Data mining techniques were used to analyze the experimental results due to the complex correlation between the data collected during the tests, providing meaningful results. In particular, the artificial neural network allowed to obtain models with excellent capacity to estimate energy consumption. A sensitive analysis was carried out with a five input parameter model, which showed that the main parameters controlling the energy consumption are the vehicle speed and the mean texture depth.