Impact of Traffic Data on the Pavement Distress Predictions using the Mechanistic Empirical Pavement Design Guide

Abstract

ABSTRACT This study examines the adequacy of using conventional traffic data and national default values in the absence of weigh-in-motion (WIM) data for pavement design. A comparative study was conducted on 14 unique sections in Arizona (AZ), where WIM data are available through the Long Term Pavement Performance (LTPP) program. The study consists of two parts: 1) comparisons of input traffic data and 2) comparisons of pavement distresses predicted by the Mechanistic Empirical Pavement Design Guide (MEPDG). The traffic related input parameters include average design-lane truck volumes, Vehicle Classification Percentages (VCP), Monthly Adjustment Factors (MAF), axle load distribution factors and the number of axles per truck. The truck volumes and VCPs are available through the Arizona Department of Transportation (ADOT) while only national average values are available for the other traffic inputs in the absence of WIM data. The comparisons of the input variables showed that the truck volumes for a design lane estimated from the ADOT database and default MAFs differed significantly from those in the LTPP database. The national default axle load distribution factors differed somewhat from the site-specific values. The differences in input data were reflected in the pavement distress values that were predicted by MEPDG. The outputs of the design guide reveal large prediction errors, particularly for longitudinal cracking, exceeding 40 percent in absolute percent error on average. The large difference in design-lane truck volume was the major source of the large prediction errors. The national default factors also generated moderate prediction errors, and the performance improved slightly with the use of the AZ average factors. Finally, the differences in MAF had little impact on predictions of pavement distress.