Bayesian Updating-Based Damage Evaluation for Truss Structures

As critical load-bearing components in civil infrastructure systems including buildings and bridges, truss structures necessitate robust health monitoring for operational safety assurance. Conventional approaches, constrained by low temporal resolution and weak anti-interference capacity, pose significant challenges in accurate damage quantification. To address this, this study proposes a Bayesian probabilistic framework for structural damage assessment. A multi-condition experimental analysis is conducted to investigate the mechanistic influences of load patterns, sensor quantity, and spatial distribution under load variability and monitoring network configurations. The results demonstrate that the Bayesian approach achieves high precision and strong robustness in both damage localization and severity quantification, validating its methodological effectiveness. Notably, load uncertainty significantly alters the posterior probability distribution of damage identification, establishing it as a critical monitoring variable. Furthermore, sensor configurations with high stacked correlation coefficients exhibit low sensitivity to assessment outcomes, supporting the implementation of cost-effective sparse sensing systems. These findings provide a novel solution integrating rigorous theoretical foundation and practical applicability for dynamic damage diagnosis and risk management in complex engineering structures.