Scholarlink

<p>This study assesses the combined effect of geogrid reinforcement and lime stabilization on the load-bearing and settlement characteristics of silty sand supporting a rectangular footing. A series of laboratory model load–settlement tests was conducted, incorporating variations in geogrid placement depth and lime content to evaluate the influence of these parameters on ultimate bearing capacity (UBC) and settlement behavior. To complement and extend these experimental findings, advanced computational approaches, random forest (RF), extreme gradient boosting (XGB), categorical boosting (CAT), and adaptive boosting (ADA), were employed to develop predictive models for UBC estimation. Model accuracy was assessed using widely accepted statistical indices, with results validated against experimental data to ensure reliability. Among the proposed algorithms, the CAT model achieved the highest predictive performance with training R<sup>2</sup> = 1.00 and testing R<sup>2</sup> = 0.951, followed by XGB, ADA, and RF models. From the results obtained, it can be concluded that the developed models effectively capture complex nonlinear interactions among soil properties, reinforcement parameters, and footing behavior. Integrating experiments with machine learning provides a faster, more cost-effective tool for geotechnical design, enabling optimized reinforcement use and efficient footing design.</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

Performance Analysis of Lime-Treated Geogrid-Reinforced Silty Sand for Rectangular Footings Using Experimental and Computational Approaches

  • 期刊论文

  • 发表日期：

    2026年5月15日

查看全文
  • Syed Md Yousuf,
  • Furquan Ahmad,
  • Sufi Md Gulzar,
  • Divesh Ranjan Kumar,
  • Warit Wipulanusat

摘要

This study assesses the combined effect of geogrid reinforcement and lime stabilization on the load-bearing and settlement characteristics of silty sand supporting a rectangular footing. A series of laboratory model load–settlement tests was conducted, incorporating variations in geogrid placement depth and lime content to evaluate the influence of these parameters on ultimate bearing capacity (UBC) and settlement behavior. To complement and extend these experimental findings, advanced computational approaches, random forest (RF), extreme gradient boosting (XGB), categorical boosting (CAT), and adaptive boosting (ADA), were employed to develop predictive models for UBC estimation. Model accuracy was assessed using widely accepted statistical indices, with results validated against experimental data to ensure reliability. Among the proposed algorithms, the CAT model achieved the highest predictive performance with training R2 = 1.00 and testing R2 = 0.951, followed by XGB, ADA, and RF models. From the results obtained, it can be concluded that the developed models effectively capture complex nonlinear interactions among soil properties, reinforcement parameters, and footing behavior. Integrating experiments with machine learning provides a faster, more cost-effective tool for geotechnical design, enabling optimized reinforcement use and efficient footing design.