Mechanistic–Empirical Modeling in Network-Level Pavement Management

Abstract

Researchers and practitioners have long recognized the advantages of mechanistic modeling. As agencies implement the Guide for Mechanistic–Empirical Design of New and Rehabilitated Pavement Structures [referred to as the Mechanistic–Empirical Pavement Design Guide (MEPDG)] in the local project-level pavement management system (PMS), its potential as a planning tool in network-level analyses remains unrecognized. A unique, large-scale application of MEPDG is presented. The Nebraska Department of Roads uses a decision tree that systematically screens and identifies candidates for entry into a multiyear optimization program. The process uses linear deterioration rates that have been shown to provide an R2 value as low as 14%. In a retroactive 5-year analysis, 86 sections were prescreened by using the existing models. Predictions indicated that 85 of the 86 sections would have been candidates for maintenance in the first 5 years, whereas field data suggested that only 23 should have been targeted. Calibrated MEPDG distress models calculated with local performance indices replaced the existing linear models. When the analysis was repeated with the mechanistic–empirical models, 35 of the 86 sections qualified for entry into the 5-year analysis, a 70% improvement in the accuracy of the forecasted cost. This research fulfills a timely need for the transitioning of network-level PMS toward mechanistic practices. Consideration of fundamental material properties, climatic conditions, and the structural response to traffic loading provides improved accuracy in planning. Furthermore, production variability and prediction uncertainties can be quantified and used as an additional probabilistic decision-making parameter. Detailed results of the implementation of MEPDG in the local network-level PMS are presented.