Research on Damage Identification of Truss Structures based on Bayesian Updating and Rejection Sampling Method

Abstract

The research on the early damage detection method of truss structure has important practical value for taking remedial measures in time and reducing the hidden danger of safety. In this study, a specific structural truss beam was designed and the seismic action was simulated. Subsequently, specific nodes are analyzed to assess the resulting structural damage. Bayesian updating method solves the uncertainty problem effectively by introducing the prior distribution and updating the posterior distribution with new observation data. This allows the model to infer global patterns from local information and represent the predicted outcomes in probabilistic form. Compared with the traditional parameter estimation methods, Bayesian updating can make full use of the existing observational data and obtain more reliable and accurate model prediction results in the case of limited data. The Bayesian method of updating the truss structure is flexible and adaptable, allowing updates and adjustments to be made at any given time based on new observational data. As a result, it has significant advantages in analyzing and predicting real-time data and is suitable for scenarios where models need to be dynamically adjusted.