Linking Traffic Volume and Weight Data for Mechanistic-Empirical Pavement Design

Abstract

Pavement engineers are now practicing the cluster method to categorize routes by grouping them based on the similarity of their axle load spectra (ALSs). Ideally, once clusters are formed, a classification model needs to be developed with the help of site-specific attributes to assign a new site to an appropriate cluster. Still, there is no straightforward model that can relate the clustered ALSs with the easily collectible information, such as routes’ types or traffic volume data. To this end, this study developed a new ALS database based on the routes’ functional classes, locations, and vehicle class distribution (VCD) data. In the beginning, the routes were divided into six groups on the basis of their locations and functional classes. After that, this study performed a cluster analysis to split the routes of each group into an optimum number of subgroups considering both ALSs and VCDs of those routes. Finally, the representative ALS of each subgroup was calculated by averaging the site-specific ALSs of routes belonging to that subgroup. It is found that the absolute error computed for the proposed ALS database with respect to site-specific ALSs is lower than the existing methods. The observations and findings are based on the tandem axle of the single-trailer truck, which can be applicable to other axle types.