The models of traffic-induced pavement wear, which form the basis of current pavement design and management practices, are based on static axle loads. Real vehicles, however, are dynamic systems and generate dynamic loads, which depend on the vehicle characteristics (mass, load distribution, suspension stiffness and damping, wheelbase, tyre pressures), its speed, and the geometry of the pavement.
This report describes a series of three accelerated pavement tests undertaken, between 1993 and 1998, at the Canterbury Accelerated Pavement Testing Indoor Facility (CAPTIF) in Christchurch, New Zealand, to determine the relationship between dynamic wheel loads and pavement wear, particularly for the thin-surfaced unbound granular pavement structures widely used in New Zealand.
The first test used a New Zealand-style thin-surfaced pavement structure with a nominal design life of 350,000 load cycles trafficked by one SLAVE (simulated loading and vehicle emulator) unit fitted with a multi-leaf steel spring without viscous damping, and another fitted with a more modern parabolic leaf spring with viscous damping. The pavement failed prematurely after only 35,000 load cycles.
The second test was a key part of the OECD DIVINE (Dynamic Interaction of Vehicles and Infrastructure Experiment) project, comparing the effects of loads generated by an air spring suspension with viscous damping with those from a multi-leaf steel spring. An asphaltic concrete surfaced pavement, with an expected design life of about 500,000 load cycles or less, withstood 1,700,000 load cycles without reaching the failure conditions.
The third pavement test used the same vehicle loading configurations as the second test but on a thin-surfaced pavement structure similar to the first test. This pavement had a design life of 94,000 load cycles, and lasted 300,000 load cycles without reaching failure conditions. This test provides a useful cross-link between New Zealand-style pavement designs and thicker asphalt concrete pavements used in Europe and North America, on which much of the international research is based.
Conclusions are that the wheel path subjected to higher dynamic loading showed a significantly greater variation in wear than the wheel path with lower dynamic loads. Reducing the structural variability in the pavement structure and reducing dynamic loading will result in a more uniform distribution of wear and will lower the maintenance requirements. Both pavements withstood considerably more load applications than predicted by most design models, though the AUSTROADS pavement design guide provides the most accurate predictions of pavement life. If the design models accurately reflect the performance of in-service pavements, construction variability, aging and environmental influences are suggested as significant contributors to pavement wear.