Damage Identification in Truss Structures Under Different Loading Conditions Based on Bayesian Updating

This study establishes a structural damage identification method based on Bayesian updating, targeting Pratt trusses under static loading, to investigate the influence mechanisms of vertical, shear, and horizontal loading conditions on parameter updating efficiency. By constructing likelihood functions from displacement responses and integrating prior distributions, the axial stiffness (EA) parameters of the truss are updated for structural damage identification. Results indicate that under vertical loading, the coefficient of variation (COV) of updated chord member parameters decreases significantly by 84% compared to the prior distribution (1.6% vs. 10%), forming an efficient chord damage identification mechanism; horizontal loading achieves a 56% COV reduction for web members (4.3% vs. 10%), demonstrating localized damage-specific identification capability; shear loading synchronously optimizes the COV values of all members, providing global rapid screening potential. A load-combination strategy is proposedvertical loading for chord members, horizontal loading for precise web member localization, and shear loading for preliminary global screeningvalidated through 10,000 sampling iterations, enhancing system identifiability by 63%.