Effects of Test-Specimen Boundary Conditions on the Interpretation of Stress–Strain Results for Hydraulic Asphalt Concrete

Determining the realistic shear stress–strain–strength behaviour of asphalt concrete is important when designing asphalt pavements and impervious asphalt linings, facings, and cores for dams and dykes. The behaviour is often determined by applying different test methods to specimens of different sizes in the laboratory. However, the results of these different test methods can vary significantly even when evaluating materials produced using the same recipe under the same test conditions, which causes uncertainty. This paper presents the results of a study on hydraulic asphalt concrete obtained through triaxial compression (TC), slant shear (S), modified slant shear (MS), and direct simple shear (SS) tests under the same shear strain rate and temperature conditions. The stress–strain distributions in the specimens were numerically analysed. The results showed that the S, MS, and SS tests provided very different material shear stress–strain–strength behaviours compared to those derived from the TC test. The computed results showed that these significant differences were caused by the non-uniform stress and strain distributions, and significant stress concentrations, in the specimens, mainly owing to the test boundary conditions. Therefore, the results of the S, MS, and SS tests may have been misleading. The test results had to be corrected by considering the non-uniform stress and strain distributions in the specimens to obtain realistic results.