Traffic Input for Mechanistic-Empirical Pavement Design in Tennessee

Abstract

Truck traffic is an important input for pavement design and analysis. Proper characterization of traffic patterns contributes to the design of reliable and cost-effective pavement structures. In this study, data from nine weigh-in-motion (WIM) sites in the Long-Term Pavement Performance database were utilized to develop the regional traffic input (level 2) for the implementation of mechanistic-empirical pavement design in Tennessee, U.S. Hierarchical clustering was performed to characterize the traffic patterns among the analyzed WIM sites, followed by the sensitivity analysis to evaluate the impact of generated traffic inputs. Three levels of traffic inputs—level 1 (site-specific), level 2 (statewide average and specific cluster-based), and level 3 (nationwide default)—were compared on both new pavement and pavement overlay design for typical pavement structures in Tennessee. Results demonstrate that the cluster-based level 2 provides the closest performance predictions compared with that from the site-specific level 1 data. Specifically, for new flexible pavements across eight WIM sites, the root mean square errors (RMSEs) achieved using level 2 data for fatigue cracking, asphalt concrete (AC) rutting, and International Roughness Index (IRI) are 0.491%, 0.023in., and 2.452 in./mi, respectively. For pavement overlays at nine WIM sites, the RMSEs for fatigue cracking, AC rutting, and IRI are 0.011%, 0.027 in., and 1.5 in./mi, respectively. It is recommended to adopt the cluster-based traffic input for pavement design when site-specific level 1 data is unavailable.