An accurate and efficient XFEM/Peridynamics coupling framework for simulating corrosion-fatigue crack growth

Corrosion fatigue (CF) presents a critical threat to structural integrity, yet its simulation is hindered by a persistent trade-off between crack-tip accuracy and computational efficiency. This study resolves this challenge by developing a novel coupled XFEM/Peridynamics (PD) framework that achieves both high accuracy and high computational efficiency in predicting CF crack growth. The framework strategically employs PD to capture the non-local, mechanochemically coupled physics at the crack tip, while leveraging the efficiency of XFEM to model both the macroscopic crack faces and the far-field elastic response. To fully account for the synergistic interplay between mechanical loading and electrochemical dissolution, a bidirectional mechanochemical coupling scheme is established. This explicitly integrates a stress-accelerated corrosion model with a degradation mechanism where accumulated electrochemical damage directly reduces the remaining fatigue life of local bonds. The model's performance was rigorously validated against experimental data for S355 steel. Results demonstrate a significant computational speed-up, with the proposed framework running 179-fold faster than pure PD and 2.8-fold faster than conventional FEM/PD coupling, without compromising accuracy. Crucially, the model accurately predicts crack growth rates under both pure fatigue and corrosion fatigue conditions, showing excellent agreement with experimental data. This work provides a powerful and computationally efficient tool for the integrity assessment of structures operating in corrosive environments.