Development of Performance Curves for Pavement Rehabilitation and Maintenance Treatments

Using linear regression analysis, 16 flexible-pavement performance models were generated: 9 maintenance models and 7 rehabilitation models. The models related the pavement's Present Serviceability Index to its age, materials properties, traffic loadings, and environmental conditions. Eleven of the models have R-squared values exceeding 70%, indicating a good correlation between the regression models and the data used to develop them. All of the models have R-squared values exceeding 57%. To develop the performance models, data collected by Nevada Department of Transportation (NDOT) personnel over the life of 208 projects were used. Statistically significant samples were drawn from these projects for each of the 3 maintenance and 3 rehabilitation techniques that NDOT commonly employs. The maintenance techniques include flush seals, sand seals, and chip seals. The rehabilitation techniques include flexible overlays, mill/overlays, and roadbed modifications. To produce statistically accurate predictions, performance models for each technique were separately developed for each of NDOT's 3 districts (except that a statewide model was developed for milling/overlaying). In most cases where a large number of projects were available, some projects were set aside for a model verification study. Using the data from the set aside projects, 12 out of the 16 models were tested by comparing the predicted performance to the performance observed at the projects. These comparisons showed excellent correlations between the present serviceability index (PSI) values predicted by the models and those observed.