Scholarlink

<p>In order to study the influence of large volume of mineral admixtures on the compressive strength and working performance of sea sand reactive powder concrete (SSRPC), Box-Behnken design response surface method (RSM) was used to optimize the experimental design. Based on the response surface test results, a high-precision prediction model of genetic algorithm optimized artificial neural network (ANN-GA) was constructed. Finally, the microscopic morphology and hydration products of the multi-component admixture system were analyzed by scanning electron microscopy (SEM) and energy dispersive spectroscopy (EDS). The results showed that the dosage of silica fume and slag powder had the most significant effect on the strength and workability of SSRPC. In contrast, the interaction of the two similarly had the most significant effect on the performance of SSRPC. When the SF: SG: FA dosage ratio was 20:8:16, the hydrated calcium aluminosilicate hydrate (C-(A)-S–H) and ettringite (AFt) in the matrix intertwined with each other, which contributed to a significant increase in the strength. The relative error accuracies in the optimized matrix ratios for the RSM and the ANN-GA were 4.73% and 2.07%, respectively. Compared with the RSM model, the ANN-GA model has more accurate prediction performance with mean relative error (MRE), root mean square error (RMSE), and goodness of fitting R<sup>2</sup> of 0.003, 0.41, and 99.2%, respectively, and therefore, the application of this model can achieve high-precision fit ratio optimization for SSRPC.</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

Study on performance and mix proportion optimization of silica fume-slag powder-fly ash ternary mineral admixture system for sea sand reactive powder concrete

  • 期刊论文

  • 发表日期：

    2025年12月8日

查看全文
  • Lincai Ge,
  • Haitao Li,
  • Shuai Liu,
  • Zixian Feng,
  • Mahdi Hosseini

摘要

In order to study the influence of large volume of mineral admixtures on the compressive strength and working performance of sea sand reactive powder concrete (SSRPC), Box-Behnken design response surface method (RSM) was used to optimize the experimental design. Based on the response surface test results, a high-precision prediction model of genetic algorithm optimized artificial neural network (ANN-GA) was constructed. Finally, the microscopic morphology and hydration products of the multi-component admixture system were analyzed by scanning electron microscopy (SEM) and energy dispersive spectroscopy (EDS). The results showed that the dosage of silica fume and slag powder had the most significant effect on the strength and workability of SSRPC. In contrast, the interaction of the two similarly had the most significant effect on the performance of SSRPC. When the SF: SG: FA dosage ratio was 20:8:16, the hydrated calcium aluminosilicate hydrate (C-(A)-S–H) and ettringite (AFt) in the matrix intertwined with each other, which contributed to a significant increase in the strength. The relative error accuracies in the optimized matrix ratios for the RSM and the ANN-GA were 4.73% and 2.07%, respectively. Compared with the RSM model, the ANN-GA model has more accurate prediction performance with mean relative error (MRE), root mean square error (RMSE), and goodness of fitting R2 of 0.003, 0.41, and 99.2%, respectively, and therefore, the application of this model can achieve high-precision fit ratio optimization for SSRPC.