Abstract: Bayesian methods can fully utilize the monitoring data to update structural parameters and their probability distributions.The Bayesian Updating with Structural reliability (BUS) method transforms the complex inverse analysis of structural parameters into an equivalent structural reliability problem,which can then be solved using subset simulation to infer the posterior distribution of the structural parameters.However,when integrating a large amount of monitoring data to update the posterior statistics of structural parameters,the failure region formulated via the BUS method grows exceedingly narrow and intricate.This affects the generation of conditional samples in the subset simulation and further results in a significant sample impoverishment issue during the posterior distribution inference.To address this issue,a Bayesian updating method with adaptive dimensionality reduction based on Principal Component Analysis (PCA) is proposed.By applying PCA to reduce the dimensionality of seed samples at each subset simulation level,the proposed method reduces the generation of duplicate samples during the posterior distribution inference,increases candidate sample acceptance rates,and alleviates the sample impoverishment.Finally,the effectiveness of the proposed method is validated through the case study of a structural beam considering corrosion deterioration.The results indicate that the proposed method not only can effectively solve the Bayesian updating problem of structural parameters with the massive monitoring data,but also can overcome the sample impoverishment issue commonly encountered in the high-dimensional Bayesian updating,which offers an effective way to update the engineering structural parameters and their distributions with the integration of extensive monitoring data.