Axle Load Spectra Characterization by Mixed Distribution Modeling

Abstract

Traffic comprises a primary input to any pavement design methodology. As pavement design evolves from traditional empirically based methods toward mechanistic-empirical (M-E), there will be a greater emphasis placed on accurate characterization of the distributions of axle loads, because they are entered directly into pavement response models. Known as load spectra, these distributions of axle weight are often complex and not well represented by individual theoretical statistical distributions. This research proposes a mixed distribution model of two or more theoretical distributions to accurately characterize axle load spectra. This paper details the model development and demonstrates its predictive capability using weigh-in-motion data from 13 sites in Alabama. The data show the predictive capability of the mixed distribution model through high R2 values, exceeding 0.96 in all cases. The typical axle load distribution in Alabama is comprised of a combination of lognormal and normal distributions.