NUMERICAL METHODS IN CRACK MATERIAL DETECTION AND ANALYSIS: A SYSTEMATIC REVIEW

Abstract

This systematic literature review examines recent advancements in numerical methods for crack detection and analysis, highlighting their increasing accuracy and computational efficiency. Using the PRISMA framework, an extensive search conducted in Scopus and Web of Science databases (2023–2024) identified 15 primary studies focusing on three main themes: numerical methods for crack detection, dynamic crack propagation analysis, and advanced material-specific fracture analysis techniques. Key findings indicate substantial improvements, particularly in hybrid numerical approaches like Extended Finite Element Method (XFEM), Physics-Informed Neural Networks (PINNs), and phase-field methods, which significantly enhance real-time crack detection capabilities in critical structural applications, including rotating machinery, fluid pipelines, and composite materials. Despite these improvements, several limitations persist, notably the difficulty in generalizing results across different materials, computational challenges when scaling models for large-scale industrial applications, and dependence on material-specific parameters requiring further experimental validation. Future research should prioritize developing universally adaptable numerical models that effectively integrate these advanced techniques to address these limitations, facilitating broader applicability and reliability in structural health monitoring.