Investigation of accuracy of pavement mechanistic empirical prediction performance by incorporating Level 1 inputs

The recently developed AASHTOWare pavement mechanistic empirical (M-E) design requires advanced characterization of pavement materials. In particular, M-E Level 1 analysis requires materials input parameters to be obtained from the laboratory test results. In cooperation with the New Mexico Department of Transportation (NMDOT), six pavement sites were selected to compare M-E Level 1 predicted performances from laboratory tested data with the field performances. Pavement materials such as subgrade, base, asphalt concrete, binder, aggregate, etc., were collected from these pavements sites during construction. Laboratory dynamic modulus, fatigue endurance limit (FEL), dynamic shear rheometer, and resilient modulus tests were conducted on the collected materials. Using these parameters, pavement sections, as constructed, were modeled and analyzed in M-E at Levels 1, 2 and 3. The predicted distress quantities for international roughness index (IRI) and rutting were compared with the field collected data as well as among the three design levels. The analysis results demonstrated that M-E overpredicts IRI and rutting in the pavement compared to the field performance. It was also observed that Level 1 showed the lowest rutting, whereas the maximum rutting was found at Level 2. The dynamic modulus of AC layer was underpredicted at Levels 2 and 3 resulting in higher rutting and IRI.