Numerical predictions of cracking evolution on RAP-recycled asphalt mixtures using viscoelastic cohesive zone model with Gaussian damage function

This study investigated the effects of Reclaimed Asphalt Pavement (RAP) on the fracture parameters and damage evolution pattern using an extrinsic nonlinear viscoelastic cohesive zone (NVCZ) model integrated with the Gaussian damage evolution function. The primary objectives were to evaluate the sensitivity of the model’s parameters to variations in the RAP content as well as to assess the effect of each parameter on the complex non-linear fracture process of bituminous materials. Two similar laboratory-produced asphalt mixtures with different RAP percentages (25 and 65 %) were assessed through dynamic modulus test to capture linear viscoelastic properties and the semi-circular bending test to obtain fracture test parameters by means of experimental-numerical calibration. A numerical parametric analysis was also performed to verify the influence of each NVCZ-Gaussian model parameter on the fracture response of RAP-produced asphalt mixtures. The calibration results revealed that two NVCZ model parameters, cohesive zone length (δ) and initiation stress ( σ i f ), could be identified without calibration, while Gaussian damage parameters (λ associated with mean value and RMS associated with standard deviation value in the damage-rate vs. displacement graph) varied systematically with RAP content. The variation in these parameters demonstrated a rational and consistent sensitivity to changes in RAP content, in line with the parametric trends observed in the study. To validate the outcomes, a third mixture with 45 % RAP (M45) was used. Its NVCZ parameters were predicted using an interpolation technique, eliminating the need for additional SCB testing or numerical calibration. The numerical predictions closely matched the experimental results, capturing both the initial and post-peak cracking behavior. Key cracking-related parameters (peak load, fracture energy, and flexibility index) further confirmed the strong agreement between numerical and experimental results. These findings demonstrate the potential of the NVCZ-Gaussian model as a promising mechanistic tool for predicting fracture behavior across mixtures with varying RAP contents and supporting more efficient performance-based mix design. • The viscoelastic CZ model with a Gaussian damage function, effectively predicted fracture behavior for varying RAP content levels. • Key model parameters of the Gaussian damage function demonstrated sensitivity to varying RAP content. • Mechanistic modeling with experimental data can be used to support more efficient RAP-recyclled mixture design. • Numerical modeling can provide additional insights into the damage evolution of RAP-recycled asphalt mixtures.