Asphalt mixtures belong to a type composite materials widely used in road engineering. This composite has intrinsic self-healing capability that makes it possible to partially restore its properties after damage. This research aimed to understand, predict and utilize this capability to improve material efficiency, enhance pavement performance predictions, and extend road service life.

This study investigated the fundamental mechanical properties of asphalt mixtures, focusing on the nonlinear behavior of the basic constituent, asphalt matrix. Through a series of experiments, a three-dimensional nonlinear constitutive model was developed, accounting for compression-tension asymmetry and rate-dependency. Besides, a multi-purpose self-healing model was proposed enabling the prediction of healing efficiency as a function of the level of damage, resting time and loading rate.

To characterize material performance, a modified Nelder-Mead optimization approach was developed fully parallelized capable to provide very accurate identification of the entire set of nonlinear parameters with unprecedented efficiency. Finally, a high-fidelity three-dimensional finite element model of asphalt mixtures was developed, verified through experimental results. This model predicts damage, healing, and fracture processes under different loading rates and at global and local level.

The research provides valuable insights into asphalt materials, paving the way for more efficient and durable road construction. The computational framework can be extended to analyze other engineering materials with similar properties.