Probabilistic damage hazard analysis framework for crack detection by integrating Bayesian inference

Cracking is a prevalent issue in structural components. The extensive distribution of cracks can significantly accumulate failure, reducing safety and reliability while increasing maintenance costs. Accurately predicting crack locations and patterns is paramount for effective damage detection techniques. This study presents a novel framework, integrating Damage Hazard Analysis (DHA) and Damage Detection Element Method (DDEM). This unique approach, which has not been explored before, first employs the DDEM to discrete the system to several sub-systems regarding the crack type and criticality levels. Subsequently, the probability of crack locations and sizes are determined based on a new concept of crack-fragility analysis. The Bayesian inference is utilized to update the failure zone regarding the location and intensity of the cracks. Accordingly, Probabilistic Damage Hazard Analysis (PDHA) is introduced to assess the failure conditions using full probabilistic rules. Probabilistic Damage Hazard Analysis (PDHA) is introduced to assess failure conditions using a full probabilistic framework. The key innovations of this study lie in integrating structural reliability analysis, Bayesian inference, and probabilistic approaches for crack detection. The proposed framework improves damage assessment, optimizes maintenance strategies, and informs design codes by quantifying the intensity and location of detected damage and incorporating a reliability measure. The illustrative example demonstrates the framework's effectiveness in crack detection using PDHA and DDEM, encouraging further research and advancements in structural damage assessment.