Review and Prospects of Experimental Research and Phase-Field Modeling on Fracture Behavior of Composite Materials under Hygrothermal Environments

Abstract

This review systematically examines recent advances in understanding the fracture behavior of composite materials under hygrothermal environments, with a focus on experimental studies, phase-field modeling, and the emerging role of artificial intelligence. Experimental investigations have elucidated moisture diffusion mechanisms—ranging from Fickian to non-Fickian behavior—and identified critical degradation processes at fiber-matrix interfaces. The phase-field method has emerged as a powerful numerical tool for simulating complex crack propagation, with ongoing efforts to extend its capability to multi-physics scenarios involving thermo-hydro-mechanical coupling. Meanwhile, machine learning techniques, especially physics-informed neural networks (PINNs), are increasingly integrated into constitutive modeling and failure prediction, offering solutions to challenges such as computational cost and limited experimental data. Despite these advancements, key gaps remain, particularly in fully coupled hygrothermal-mechanical phase-field frameworks and the integration of physical mechanisms into data-driven models. Future research is expected to prioritize intelligent, multi-scale simulation platforms and digital twin technologies to achieve accurate lifecycle prediction and design optimization for composites in critical engineering applications.