Scholarlink

<p>Chloride ingress in reinforced concrete (RC) structures deteriorates the structural service performance, seriously threating to the structural safety. In real-world, uncertainty and stochasticity are inevitably involved in testing and modeling of the diffusion profile of chloride ingress concrete (CIC). In this paper, a hierarchical Bayesian estimation framework is developed to predict the chloride concentration profile in concrete. Firstly, a CIC stochastic model based on the Fick’s second law of diffusion is constructed to characterize chloride concentration profile. Error terms are included in the stochastic model to capture measurement errors associated with inspections. Secondly, by an instruction to a universal parameter vector and an equivalence of lower order moments, a hierarchical Bayesian estimation method based on trans-distributional reversible jump Markov chain Monte Carlo algorithm (RJ-MCMC) is developed to select a single best distribution-form of the stochastic model parameters from among their multiple distribution-forms or structures. Thirdly, given that distribution-form was specified, the hierarchical Bayesian estimation method and Hybrid Markov chain Monte Carlo (H-MCMC) algorithm are incorporated to update the parameters in the CIC model based on measurement data from inspections. An example involving an in-service RC bridge was employed to validate the developed CIC model and demonstrate the proposed Bayesian estimation framework for the analysis of chloride concentration profile. Results of the analysis indicate both the uncertainty and stochasticity in the parameters of the CIC model as well as the uncertainty in distribution-form must be accounted for in the prediction of chloride concentration profile. The proposed framework will facilitate better decision-making for maintenance and repair activities.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Bayesian stochastic model for prediction of chloride content profile in concrete under marine environment

  • 期刊论文

  • 发表日期：

    2025年12月5日

查看全文
  • Meng-Cheng Chen,
  • Li-Qin Guo,
  • Wei Fang,
  • Ming-Yang Zhang,
  • Kai-Cheng Xu,
  • Hong Huang,
  • Li Xie

摘要

Chloride ingress in reinforced concrete (RC) structures deteriorates the structural service performance, seriously threating to the structural safety. In real-world, uncertainty and stochasticity are inevitably involved in testing and modeling of the diffusion profile of chloride ingress concrete (CIC). In this paper, a hierarchical Bayesian estimation framework is developed to predict the chloride concentration profile in concrete. Firstly, a CIC stochastic model based on the Fick’s second law of diffusion is constructed to characterize chloride concentration profile. Error terms are included in the stochastic model to capture measurement errors associated with inspections. Secondly, by an instruction to a universal parameter vector and an equivalence of lower order moments, a hierarchical Bayesian estimation method based on trans-distributional reversible jump Markov chain Monte Carlo algorithm (RJ-MCMC) is developed to select a single best distribution-form of the stochastic model parameters from among their multiple distribution-forms or structures. Thirdly, given that distribution-form was specified, the hierarchical Bayesian estimation method and Hybrid Markov chain Monte Carlo (H-MCMC) algorithm are incorporated to update the parameters in the CIC model based on measurement data from inspections. An example involving an in-service RC bridge was employed to validate the developed CIC model and demonstrate the proposed Bayesian estimation framework for the analysis of chloride concentration profile. Results of the analysis indicate both the uncertainty and stochasticity in the parameters of the CIC model as well as the uncertainty in distribution-form must be accounted for in the prediction of chloride concentration profile. The proposed framework will facilitate better decision-making for maintenance and repair activities.