However, Pavement ME predictions did not improve when measured values for concrete strength or load transfer efficiency were used rather than design values. The RMSE of the Pavement ME predictions improved to 1.2% when constructed slab thickness measured with ground penetration radar was used instead of the design thickness. The root mean square error (RMSE) of Pavement ME predictions was 2.4% for the set of 13 sections. When design values were adopted for the different input variables, Pavement ME predicted less than 5% longitudinal cracking in 12 of the 13 sections, which agrees with measured cracking. None of the 13 sections had more than 3% of slabs with longitudinal cracking, despite four of them being subjected to relatively high traffic volumes (annual average daily truck traffic over 500 vehicles on the design lane) and having been in service between 9 and 19 years.

This technical memorandum assesses Pavement ME predictions based on the longitudinal cracking measured on 13 COA sections with half-lane width slabs evaluated as part of NCHRP Project 1-61. The NCHRP Project 1-61 has considerably expanded the range of climatic conditions for which reliable performance data are available by adding projects from Iowa, Kansas, and Philadelphia (in addition to Minnesota, Illinois, and Colorado). The thin concrete overlay on asphalt (COA) longitudinal cracking model of Pavement ME was calibrated with empirical data from COA sections with half-lane width slabs in Minnesota, Illinois, and Colorado.

Caltrans rigid pavement design software