Scholarlink

<p>This study examines the bearing capacity of conical foundations near slopes with cohesive frictional soil masses. Based on the three-dimensional upper-bound finite element limit analysis and adaptive mesh refinement technique, the correlation between the bearing capacity factor and six input variables is investigated, including the slope angle, the geometry and position of the conical foundation (i.e., the conical apex angle, the embedment ratio, and the setback ratio), and the soil shear strength (i.e., the internal friction angle and the cohesion ratio). The impacts of these factors on the failure mechanism are also discussed and illustrated. Additionally, a novel hybrid machine learning framework is proposed that integrates five recent nature-inspired optimization algorithms, namely the golden jackal optimization (GJO), mountain gazelle optimizer (MGO), sand cat swarm optimization (SCSO), sea horse optimizer (SHO), and secretary bird optimization algorithm (SBOA), into the extreme gradient boosting (XGBoost) method to increase its performance. The evaluation results indicate that SBOA-XGBoost with the latest optimizer outperforms the default and other hybrid XGBoost models in terms of accuracy. On the basis of the optimized surrogate model, a sensitivity analysis is conducted to estimate the contribution of each input parameter to the conical foundation’s bearing capacity.</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

Stability Evaluation of Conical Footings on Slopes via Three-Dimensional Simulation and the Nature-Inspired Exreme Gradient Boosting Method

  • 期刊论文

  • 发表日期：

    2025年8月5日

查看全文
  • Nhat Tan Duong,
  • Duy Tan Tran,
  • Gia Huy Pham,
  • Suraparb Keawsawasvong,
  • Van Qui Lai

摘要

This study examines the bearing capacity of conical foundations near slopes with cohesive frictional soil masses. Based on the three-dimensional upper-bound finite element limit analysis and adaptive mesh refinement technique, the correlation between the bearing capacity factor and six input variables is investigated, including the slope angle, the geometry and position of the conical foundation (i.e., the conical apex angle, the embedment ratio, and the setback ratio), and the soil shear strength (i.e., the internal friction angle and the cohesion ratio). The impacts of these factors on the failure mechanism are also discussed and illustrated. Additionally, a novel hybrid machine learning framework is proposed that integrates five recent nature-inspired optimization algorithms, namely the golden jackal optimization (GJO), mountain gazelle optimizer (MGO), sand cat swarm optimization (SCSO), sea horse optimizer (SHO), and secretary bird optimization algorithm (SBOA), into the extreme gradient boosting (XGBoost) method to increase its performance. The evaluation results indicate that SBOA-XGBoost with the latest optimizer outperforms the default and other hybrid XGBoost models in terms of accuracy. On the basis of the optimized surrogate model, a sensitivity analysis is conducted to estimate the contribution of each input parameter to the conical foundation’s bearing capacity.