Updating Traffic Inputs for Use in the Pavement Mechanistic-Empirical Design in Michigan

Abstract

A study, completed in 2009, characterized traffic inputs in support of implementing the new Mechanistic-Empirical Pavement Design Guide for the state of Michigan. The inputs included monthly adjustment factors, hourly distribution factors, vehicle class distributions, axle groups per vehicle, and axle load distributions for different axle configurations. During the last 7 years, new traffic data were collected that reflect recent economic growth, and additional and downgraded weigh-in-motion (WIM) sites. Performance models for the Pavement mechanistic-empirical (ME) design were recently calibrated to local conditions in Michigan. Hence it was appropriate to incorporate these changes in the re-evaluation of traffic inputs. Weight and classification data were obtained from 41 WIM sites throughout the Michigan to develop Level 1 (site-specific) traffic inputs. Cluster analyses were conducted to develop Level 2A inputs. Classification models (decision trees, random forests) were developed to assign a new site to these clusters; however, this proved difficult, hence, an alternative, simplified method to develop Level 2B inputs by grouping sites with similar attributes was also adopted. The optimal set of attributes for developing Level 2B inputs was identified by using an algorithm developed in this study. The effects of the developed hierarchical traffic inputs on the predicted performance of rigid and flexible pavements were investigated using Pavement-ME. Based on the statistical and practical significance of the life differences, appropriate levels were established for each traffic input. The methodology for developing traffic inputs is intuitive and could easily be adopted by Michigan Department of Transportation for future updates.