North Carolina’s Experience in Development of Pavement Performance Prediction and Modeling

Abstract

Development efforts in pavement performance prediction by the North Carolina Department of Transportation are described. Research into other states’ approaches was also conducted. The initial idea was to use family curves. However, because of a lack of data in key areas, it was decided to use an individual section’s pavement condition rating (PCR) data for performance prediction. The process of selection and justification of a functional form for curve fitting is detailed. An adaptive scheme to accommodate a realistic PCR history containing cycles of decline and improvement in the ratings is detailed. Abnormal sections that did not fit the models developed for individual sections were identified. These were either ( a) section with too few datum points for modeling or ( b) sections in which the last few ratings leveled out, resulting in a prediction of an unreasonably long life span. The development of family curves and their application in the processing of abnormal sections are also discussed. The developed models were then evaluated by comparing the predicted rating with the actual rating.