Scholarlink

<p>To address the growing concerns about increasing CO<sub>2</sub> emissions and the urgent need for recycling construction waste, this study proposes a method for predicting the compressive strength of CO<sub>2</sub>-mineralized recycled coarse aggregate concrete using a genetic algorithm-optimized backpropagation (GA-BP) neural network. A total of 52 experimental data sets were collected, with cement content, water content, fine aggregate content, replacement rate of mineralized recycled coarse aggregate, water absorption of coarse aggregate, and crushing index selected as input parameters, and the 28-day compressive strength of concrete cube specimens used as the output parameter. A GA-BP neural network model was constructed and Subjected to training, testing, and experimental validation. The results demonstrate that the optimized GA-BP model significantly improved prediction accuracy, achieving R-values of 0.99985 and 0.98018 for the training and testing sets, respectively. Additionally, the mean squared error and mean absolute error of predictions were both lower than those of the traditional BP neural network. Validation with 12 experimental tests revealed that the GA-BP model’s prediction error was within 0.77&#xa0;MPa. In the validation tests, the CO<sub>2</sub>-mineralized recycled coarse aggregate concrete (RCAC) demonstrated favorable mechanical properties, with an average 28-day compressive strength of 39.27&#xa0;MPa. Based on an estimated CO<sub>2</sub> sequestration rate of approximately 10.3&#xa0;kg per ton of RCAC, each cubic meter of RCAC concrete used in the experiment is capable of sequestering about 12.36&#xa0;kg of CO<sub>2</sub>, indicating significant carbon storage potential. This study confirms that the optimized model demonstrates strong robustness and high prediction accuracy, while the material itself exhibits significant CO<sub>2</sub> sequestration capacity, offering a promising pathway for integrating intelligent performance prediction with sustainable, low-carbon construction practices.</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

Prediction of the Strength of CO2-Mineralized Recycled Coarse Aggregate Concrete Based on the GA-BP Neural Network

  • 期刊论文

  • 发表日期：

    2025年9月4日

查看全文
  • Chunfeng Yang,
  • Dianwen Zhao,
  • Heng Zhang

摘要

To address the growing concerns about increasing CO2 emissions and the urgent need for recycling construction waste, this study proposes a method for predicting the compressive strength of CO2-mineralized recycled coarse aggregate concrete using a genetic algorithm-optimized backpropagation (GA-BP) neural network. A total of 52 experimental data sets were collected, with cement content, water content, fine aggregate content, replacement rate of mineralized recycled coarse aggregate, water absorption of coarse aggregate, and crushing index selected as input parameters, and the 28-day compressive strength of concrete cube specimens used as the output parameter. A GA-BP neural network model was constructed and Subjected to training, testing, and experimental validation. The results demonstrate that the optimized GA-BP model significantly improved prediction accuracy, achieving R-values of 0.99985 and 0.98018 for the training and testing sets, respectively. Additionally, the mean squared error and mean absolute error of predictions were both lower than those of the traditional BP neural network. Validation with 12 experimental tests revealed that the GA-BP model’s prediction error was within 0.77 MPa. In the validation tests, the CO2-mineralized recycled coarse aggregate concrete (RCAC) demonstrated favorable mechanical properties, with an average 28-day compressive strength of 39.27 MPa. Based on an estimated CO2 sequestration rate of approximately 10.3 kg per ton of RCAC, each cubic meter of RCAC concrete used in the experiment is capable of sequestering about 12.36 kg of CO2, indicating significant carbon storage potential. This study confirms that the optimized model demonstrates strong robustness and high prediction accuracy, while the material itself exhibits significant CO2 sequestration capacity, offering a promising pathway for integrating intelligent performance prediction with sustainable, low-carbon construction practices.