Scholarlink

<p>The inherent spatial heterogeneity of complex geological formations results in large differences in the ultimate bearing capacity of individual piles, and it is difficult to reliably quantify and assess the bearing capacity of individual piles. In this paper, based on the results of on-site static load test and sensitivity analysis method, eight sensitive factors are screened out, and the extreme gradient boosting algorithm (XGBoost) is used to predict the ultimate load carrying capacity of individual piles, however, the computed coefficient of determination is less than 0.9, and the prediction effect needs to be strengthened. On this basis, three kinds of swarm intelligent optimization algorithms are introduced to adaptively match the XGBoost hyper-parameters, and the results show that the HGS-XGBoost hybrid prediction model can more accurately calculate the ultimate bearing capacity of a single pile under the composite strata, and the prediction effect can satisfy the engineering requirements when using the HGS-XGBoost prediction model for the actual project.</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

Prediction of ultimate bearing capacity of single pile in composite formation based on HGS-XGBoost algorithm

  • 期刊论文

  • 发表日期：

    2025年4月3日

查看全文
  • Zong-jun Sun,
  • Ying-feng Han,
  • Feng Jiang,
  • Gang Wang,
  • Wen Zheng,
  • Fei Liu,
  • Yue Wu

摘要

The inherent spatial heterogeneity of complex geological formations results in large differences in the ultimate bearing capacity of individual piles, and it is difficult to reliably quantify and assess the bearing capacity of individual piles. In this paper, based on the results of on-site static load test and sensitivity analysis method, eight sensitive factors are screened out, and the extreme gradient boosting algorithm (XGBoost) is used to predict the ultimate load carrying capacity of individual piles, however, the computed coefficient of determination is less than 0.9, and the prediction effect needs to be strengthened. On this basis, three kinds of swarm intelligent optimization algorithms are introduced to adaptively match the XGBoost hyper-parameters, and the results show that the HGS-XGBoost hybrid prediction model can more accurately calculate the ultimate bearing capacity of a single pile under the composite strata, and the prediction effect can satisfy the engineering requirements when using the HGS-XGBoost prediction model for the actual project.