Scholarlink

<p>This study investigates the effect of fly ash replacement on the split tensile and flexural strength of concrete at different water-cement ratios (0.35, 0.45, and 0.50) over multiple curing periods (7, 28, 56, 90, and 120&#xa0;days). Increasing fly ash replacement reduces both split tensile and flexural strength, with the decline being more pronounced at early curing stages. However, extended curing significantly improves strength recovery, mitigating initial losses. A lower water-cement ratio helps retain higher strength levels, whereas a higher ratio accelerates strength reduction over time. Support Vector Machine (SVM) models with polynomial, radial basis function (RBF), and linear kernels were employed for predictive modeling. Performance evaluation via various metrics confirms that the SVM-RBF model outperforms others, achieving an R2 of 0.992 for both split tensile and flexural strength predictions during training, and 0.939 and 0.918 respectively during testing. Other performance metrics, such as Mean Absolute Error (MAE), were lowest for the SVM-RBF model, with values of 0.073 and 0.137 for split tensile strength and 0.105 and 0.208 for flexural strength during training and testing, respectively. Sensitivity analysis highlights curing time as the most influential factor.</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

Support vector machine-based models for predicting flexural and tensile strength of cement concrete incorporating fly ash with varying water-cement ratio

  • 期刊论文

  • 发表日期：

    2026年2月16日

查看全文
  • Rohit Kumar Mishra,
  • Arun Kumar Mishra,
  • Anish Kumar

摘要

This study investigates the effect of fly ash replacement on the split tensile and flexural strength of concrete at different water-cement ratios (0.35, 0.45, and 0.50) over multiple curing periods (7, 28, 56, 90, and 120 days). Increasing fly ash replacement reduces both split tensile and flexural strength, with the decline being more pronounced at early curing stages. However, extended curing significantly improves strength recovery, mitigating initial losses. A lower water-cement ratio helps retain higher strength levels, whereas a higher ratio accelerates strength reduction over time. Support Vector Machine (SVM) models with polynomial, radial basis function (RBF), and linear kernels were employed for predictive modeling. Performance evaluation via various metrics confirms that the SVM-RBF model outperforms others, achieving an R2 of 0.992 for both split tensile and flexural strength predictions during training, and 0.939 and 0.918 respectively during testing. Other performance metrics, such as Mean Absolute Error (MAE), were lowest for the SVM-RBF model, with values of 0.073 and 0.137 for split tensile strength and 0.105 and 0.208 for flexural strength during training and testing, respectively. Sensitivity analysis highlights curing time as the most influential factor.