Mechanistic investigation of grid-reinforced asphalt pavements with measured interface properties

Abstract

The bond properties at layer interfaces are a required input in mechanistic asphalt pavement models. For design and analysis procedures, involving the calculation of key responses that are commonly linked to performance, perfect bond conditions are regularly assumed between layers. One emerging approach for prolonging the service life of asphalt pavements, either new or rehabilitated, is installing asphalt grid reinforcement (AGR) products between paving lifts. In these cases, the perfect bond assumption may not hold – undermining the reinforcement potential. Accordingly, the objective of this study was to investigate the effects of layer interface properties on key responses in pavements with AGR. The investigation was carried out by combining results from a full-scale construction, laboratory tests on asphalt concrete (AC) cores, and synthetic simulations. The latter were performed with a recently developed semi-analytic mechanistic code that can accept AGR products. This code can handle time- and temperature-dependent layer and interface properties, as well as moving loads. From the laboratory tests, it was found that the bond between AC lifts, with and without an AGR, is time- and temperature-dependent – best characterised by a relaxation interface stiffness function. This relaxation function was measured to be consistently larger without an AGR than with an AGR. Nonetheless, pavement simulations showed that including an AGR has no pronounced effect on strain magnitudes under high-speed/low-temperature conditions, can lessen horizontal strain magnitudes at the reinforced interface and at the AC bottom (mainly under slow-speed/high-temperature conditions, and depending on the installation depth of the AGR), and has no pronounced effect on deflections regardless of the loading speed and temperature level. It is concluded that even if laboratory measurements display a reduced interface stiffness when including an AGR, the reduction may not have any practical implication on key pavement responses, and the reinforcement potential is not necessarily undermined. The modelling and experimental approaches outlined and utilised in this study offer a rational tool for analysing this matter/question on a case-by-case basis.