Comparative Analysis of XFEM and Phase Field Approaches for Fracture Prediction in Flexible Ti-6Al-4V Thoracic Implants

The scientific literature increasingly supports the use of computational models to predict fracture across a wide range of applications, which, when calibrated with experimental data, can yield highly consistent results. Although the extended finite element method (XFEM) is widely used in commercial packages, phase field (PF) methods have emerged as a robust alternative. In this study, a cohesive zone model (CZM) was implemented using both approaches (a PF model with an implicit damage initiation criterion and a standard commercial XFEM solver with an explicit damage initiation criterion) to analyze their robustness and computational efficiency. First, a standardized fracture test of a compact tension (CT) specimen was simulated and compared with experimental data to validate both methods, achieving accurate predictions under plane strain conditions with a dominant mode I fracture behavior. Subsequently, the application of both fracture models was extended to flexible thoracic prostheses across two distinct chest wall reconstruction scenarios: a single-rib unilateral model and a multi-rib bilateral configuration. An extreme-case compressive displacement was assessed to identify critical regions susceptible to fracture initiation and to evaluate the structural limits of the proposed designs. The results showed that the PF approach required a higher computational time, but exhibited more stable convergence. In contrast, the XFEM-based solver required careful mesh calibration to ensure convergence under complex conditions. These results highlight the potential of the PF approach as a practical tool for identifying and improving critical regions of implants, overcoming the limitations of commercial XFEM implementations.