Scholarlink

<p>Corrosion significantly impacts interfacial bonds in steel-reinforced concrete composite systems, leading to potential structural failures if not adequately addressed. Accurately predicting this strength is essential to assess structural integrity and improve structures’ longevity in marine environments. This paper deals with the experimental verification of the corrosion phenomenon of embedded steel within concrete. An accelerated corrosion setup has been put in to perform long exposure of laboratory-sized concrete cylinders to an aggressive environment (3.5% NaCl saline solution). The monitoring matrix (half-cell potential, relative humidity, pH, bond strength) was obtained through direct and indirect assessment. Later, the researcher extended the study to develop a physics-involved deep learning algorithm to understand the mechanism of interrelated parameters, which influence the criticality/severity of the corrosion process in reinforced cement concrete (RCC) structure. The accuracy of the proposed algorithms was carefully analyzed using statistical metrics. The comprehensive comparisons of different deep learning models show that the CNN-LSTM hybrid model accurately predicts bond behavior under varying corrosion scenarios with lower error. In the future, the reliability of the proposed model will help designers or field engineers further enhance its predictive capabilities in real-world applications.</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

The Effect of Corrosion on Reinforced Cement Concrete (RCC) Structure for Long Exposure to Acidic Environment: A Deep Learning-Based Approach

  • 期刊论文

  • 发表日期：

    2025年7月5日

查看全文
  • Lukesh Parida,
  • Sumedha Moharana

摘要

Corrosion significantly impacts interfacial bonds in steel-reinforced concrete composite systems, leading to potential structural failures if not adequately addressed. Accurately predicting this strength is essential to assess structural integrity and improve structures’ longevity in marine environments. This paper deals with the experimental verification of the corrosion phenomenon of embedded steel within concrete. An accelerated corrosion setup has been put in to perform long exposure of laboratory-sized concrete cylinders to an aggressive environment (3.5% NaCl saline solution). The monitoring matrix (half-cell potential, relative humidity, pH, bond strength) was obtained through direct and indirect assessment. Later, the researcher extended the study to develop a physics-involved deep learning algorithm to understand the mechanism of interrelated parameters, which influence the criticality/severity of the corrosion process in reinforced cement concrete (RCC) structure. The accuracy of the proposed algorithms was carefully analyzed using statistical metrics. The comprehensive comparisons of different deep learning models show that the CNN-LSTM hybrid model accurately predicts bond behavior under varying corrosion scenarios with lower error. In the future, the reliability of the proposed model will help designers or field engineers further enhance its predictive capabilities in real-world applications.