Prediction of Friction Degradation in Highways with Linear Mixed Models

The development of a linear mixed model to describe the degradation friction on flexible road pavements be included in pavement management systems is aim this study. It also aims at showing that, network level, factors such as temperature, rainfall, hypsometry, type layer, and geometric alignment features may influence throughout time. A dataset from six districts Portugal with 7204 sections was made available by Ascendi Concession highway network. Linear models random effects intercept were developed for two-level three-level datasets involving time, section district. While are region-specific, offer possibility adopted other areas. For both levels, two approaches made: One integrating into only variables inherent traffic climate conditions including intrinsic characteristics. prediction accuracy improved when geometrical features, layer considered. Therefore, accurate predictions evolution time assist manager optimize overall level safety.