Scholarlink

Throughout the years, concrete structures like RC columns have been designed with improved efficiency and performance in order to satisfy serviceability, safety, and economic aspects. This was hugely influenced by the deterioration of concrete structures due to environmental factors and natural disasters that reduced the load capacity of the structures. In this study, the use of feed-forward backpropagation neural network in ANN was utilized in the prediction of confined compressive strength, fcc of circular concrete cylinders by using Shape Memory Alloy (SMA) strip, Carbon Fiber Reinforced Polymer (CFRP) strip, and steel strip. The fcc of circular concrete cylinders confined by these three materials was assessed by the use of the ANN toolbox in MATLAB 2016a software, where the performance of these models was then validated with the targeted value of fcc from the experimental works in the existing literature. The experimental results database contains 193 sample numbers of confined concrete cylinder resulting with 4 numbers of testing and validation data. Based on the analysis of the results, it is found that the SMA model has the closest prediction of fcc to the targeted value varying 0.9454 to 0.9783, compared to the other two models using CFRP strip and steel strip. The optimal neuron number is 4 of hidden nodes with MSE value ranging from 0.0021 to 0.0087. The result of this study verified the optimal prediction of compressive confinement strength of circular concrete cylinders by using SMA.

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

An Optimum Artificial Neural Network (ANN) Model for Prediction of Compressive Confinement Strength of Concrete Cylinders

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • S. N. I. Zulkifli,
  • M. A. Mohd Yunus,
  • F. A. A. Zakwan,
  • R. Ismail,
  • S. F. Senin,
  • M. H. Mohd Hisbany

摘要

Throughout the years, concrete structures like RC columns have been designed with improved efficiency and performance in order to satisfy serviceability, safety, and economic aspects. This was hugely influenced by the deterioration of concrete structures due to environmental factors and natural disasters that reduced the load capacity of the structures. In this study, the use of feed-forward backpropagation neural network in ANN was utilized in the prediction of confined compressive strength, fcc of circular concrete cylinders by using Shape Memory Alloy (SMA) strip, Carbon Fiber Reinforced Polymer (CFRP) strip, and steel strip. The fcc of circular concrete cylinders confined by these three materials was assessed by the use of the ANN toolbox in MATLAB 2016a software, where the performance of these models was then validated with the targeted value of fcc from the experimental works in the existing literature. The experimental results database contains 193 sample numbers of confined concrete cylinder resulting with 4 numbers of testing and validation data. Based on the analysis of the results, it is found that the SMA model has the closest prediction of fcc to the targeted value varying 0.9454 to 0.9783, compared to the other two models using CFRP strip and steel strip. The optimal neuron number is 4 of hidden nodes with MSE value ranging from 0.0021 to 0.0087. The result of this study verified the optimal prediction of compressive confinement strength of circular concrete cylinders by using SMA.