The new mechanistic-empirical model (M-E) developed under the NCHRP 1-37A initiative is a significant advancement in pavement design mainly because it addresses industry needs to switch to a performance-based practice. However, achieving this performance-based design goal with the new M-E guide depends substantially on employing mechanistic material characteristics in the analysis to produce accurate predictions of a number of distress types and smoothness. Material properties needed as input by the new M-E model for conducting an advanced design exercise (level 1) are beyond what is available from conventional characterization techniques adopted in the current practice. Laboratory facilities are not yet ready to meet testing requirements related to the mechanistic characterization scheme. Review of alternatives introduced to circumvent the need to perform elaborate mechanical tests revealed that the simplified approach adopted in the lower input levels 2 and 3 undermines the impact of material properties on performance. Results of the review suggest that the potential for accumulation of permanent deformation may be underestimated by 45% as a result of overestimating the dynamic modulus of asphalt concrete (AC) layer. This paper describes a proposal for using generic mechanistic material properties to perform design based on level 1 for AC and level 3 for unbound materials. These properties include the dynamic modulus of typical HMA mixes prepared according to a local standard and the resilient modulus of commonly used unbound materials (granular aggregates). Preliminary results indicate that the proposed approach is more reliable and results in lower design errors compared with the process incorporated in the current version of the M-E guide. The generic properties may be used to expedite implementation of the M-E guide in Canada until adequate mechanical testing capabilities are established. The generic pavement material properties produced at NRC are now stored in a database “Material Library”, which could be further populated using results from new tests that cover a wider range of material and construction variables; mainly to reduce the margin of error. The proposed database will not only provide the input properties needed for applying the M-E design model but will also support tasks associated with the calibration of the M-E component used to predict performance to make the model more sensitive to unique local conditions and construction practices.
