Triaxial stress sweep test protocol considerations for permanent deformation characterisation of asphalt mixtures

Rutting is one of the most usual distresses in Brazilian asphaltic pavements and prediction of such distress is a major challenge. State of the art methodologies need consistency between laboratory test and field conditions. This paper investigates a potential modification of the triaxial stress sweep (TSS) test, according to deviator and confining stresses typically found in Brazilian roadways which have thinner surface courses than those for which the original protocol was developed. Different structures were simulated with typical material properties and layer thicknesses found in Brazil. A more realistic confining stress was adopted, and a deviator stress was calculated for each load level of the test protocol. A modified TSS test and the uniaxial repeated load test were performed in three asphalt mixes. One of them did not reach the minimum number of cycles, preventing rutting prediction, and suggesting a fragility of this particular mix with respect to the investigated distress. Such a result was consistent with field performance. A protocol modification resulted in a reasonable prediction for the two mixes, according to field observation. The final prediction methodology established through viscoplastic characterisation (TSS and shift model calibration) and simulation with the layered viscoelastic continuum damage program was coherent with reality, provided that an experimental protocol consistent with local particularities is performed.