Scholarlink

<p>This paper proposed a coupling framework based on finite element limit analysis (FELA) and artificial neural network (ANN) to investigate the bearing capacity of ring foundation on slope crest. Six design parameters are considered, namely (i) geometry of slope: slope angle (<i>β</i>), setback ratio (<i>s/B</i>); (ii) geometry of foundation: radius ratio (<i>r</i><sub><i>i</i></sub><i>/r</i><sub><i>o</i></sub>), embedded depth ratio (<i>D/r</i><sub><i>o</i></sub>); (iii) strength of soil: internal friction angle (<i>φ</i>), strength ratio (<i>c/γB</i>). The parametric study based on FELA results shows that the increase in <i>D/r</i><sub><i>o</i></sub>, <i>c/γB</i>, <i>s/B</i>, and <i>φ</i> leads to an increase in the bearing capacity factor (<i>p/γB</i>). Meanwhile, the opposite trend is observed in the correlation between <i>p/γB</i> and the variables <i>β</i> and <i>r</i><sub><i>i</i></sub><i>/r</i><sub><i>o</i></sub>. The failure mechanism is also discussed in detail, considering each design variable’s impact on the development of shear bands. Finally, various ANN models with different structures and activation functions were trained and evaluated using the adaptive moment estimation (Adam) algorithm. The results indicate that the ANN model using the structure 6–27-28–1 and the rectified linear unit function makes the most accurate predictions, with <i>R</i><sup>2</sup> = 0.9894.</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 Analysis of Ring Foundations on Slope Crest: 3D FELA and ANN

  • 期刊论文

  • 发表日期：

    2025年1月6日

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

摘要

This paper proposed a coupling framework based on finite element limit analysis (FELA) and artificial neural network (ANN) to investigate the bearing capacity of ring foundation on slope crest. Six design parameters are considered, namely (i) geometry of slope: slope angle (β), setback ratio (s/B); (ii) geometry of foundation: radius ratio (ri/ro), embedded depth ratio (D/ro); (iii) strength of soil: internal friction angle (φ), strength ratio (c/γB). The parametric study based on FELA results shows that the increase in D/ro, c/γB, s/B, and φ leads to an increase in the bearing capacity factor (p/γB). Meanwhile, the opposite trend is observed in the correlation between p/γB and the variables β and ri/ro. The failure mechanism is also discussed in detail, considering each design variable’s impact on the development of shear bands. Finally, various ANN models with different structures and activation functions were trained and evaluated using the adaptive moment estimation (Adam) algorithm. The results indicate that the ANN model using the structure 6–27-28–1 and the rectified linear unit function makes the most accurate predictions, with R2 = 0.9894.