Real-time damage detection and localization in ship structures using iFEM and machine learning techniques

This paper presents an advancement in smart sensing and structural health monitoring (SHM) of large-scale ship structures, aimed at real-time detection and accurate localization of damage induced by extreme environmental conditions and high-impact events. This subject has attracted increasing attention in naval architecture, marine, and ocean engineering because such damage has critical implications for structural integrity and safety. In practical applications within harsh marine environments, the capability to rapidly and reliably identify the location of structural damage after an extreme event is essential. To address this need, the proposed approach integrates the inverse finite element method (iFEM), anomaly index formulation, and machine learning (ML) techniques. High-fidelity finite element (FEM) models are employed to simulate damage scenarios with high accuracy; these simulations are then simplified to enable efficient real-time analysis and seamless integration into the SHM framework. The methodology has been applied to a representative case study involving a portion of a containership, effectively overcoming challenges related to optimal sensor placement, environmental variability, and complex operating conditions. Ultimately, the study introduces an enhanced iFEM-based strategy combined with ML models for real-time SHM, providing a robust and scalable solution for damage detection and localization in large-scale ship structures under extreme conditions.