Study of stress-strain relationship of hydraulic asphalt concrete at low to intermediate temperatures based on experimental data regression

Abstract

The conventional principal structural theory modeling is constrained by the underlying assumptions that make the established constitutional model incapable of accurately describing the stress–strain relationship of asphalt concrete. Subsequently, in this study, the traditional modeling idea was changed and the projection pursuit regression (PPR) method was used to analyze triaxial test data obtained from asphalt concrete. The PPR models utilized to calculate deviatoric stress and volumetric strain were developed and the validity and applicability of the PPR model and this method were verified. The results show that the modeling method avoids the problem of inconsistency between a priori model assumptions and objective reality. Asphalt concrete has complex strain softening, strain hardening, shear shrinkage, and shear expansion characteristics. The model employed to calculate deviatoric stress and volumetric strain based on PPR method is superior to the E–B and “Nanshui” models in its capacity to reflect the characteristics of strain softening and shear deformation. The results calculated using the PPR model can accurately describe the stress–strain relationship of asphalt concrete. Studying the stress–strain relationship of asphalt concrete from the perspective of the examination of data breaks through the constraints of traditional modeling ideas, which has certain theoretical significance. The research results can provide a new solution for non-hypothesis modeling of nonlinear data and accurately describing the stress–strain relationship of asphalt concrete.