Scholarlink

<p>The undrained shear strength (<i>s</i><sub>u</sub>) is an important parameter for clays. However, current empirical models for predicting <i>s</i><sub>u</sub> are not sufficiently reliable. A novel data-driven ensemble learning method combining the particle swarm optimization (PSO) algorithm and multiple adaptive regression splines (MARS) has been developed to capture the relationships between <i>s</i><sub>u</sub> and piezocone penetration test (CPTU) parameters based on collected datasets. Four combinations of CPTU measurements (cone tip resistance, <i>q</i><sub>t</sub>; sleeve friction resistance, <i>f</i><sub>s</sub>; and excess pore water pressure, Δ<i>u</i>; or normalized net cone resistance, <i>Q</i><sub>t</sub>; normalized friction ratio, <i>F</i><sub>r</sub>; and pore pressure parameter <i>B</i><sub>q</sub>) and in-situ stresses (total vertical stress, σ<sub>vo</sub>, and effective vertical stress, σ´<sub>vo</sub>) are used as input variables to determine the optimal combination of variables. In this novel model, PSO optimizes the hyperparameters in the MARS algorithm to form the PSO-MARS model, which has the advantage of visualizing the model expression. The performance of the PSO-MARS model is specifically compared with existing empirical correlations and other machine learning (ML) models. The robustness of prediction models is analyzed using Monte Carlo simulation. The results show that the PSO-MARS model demonstrates higher accuracy and robustness in <i>s</i><sub><i>u</i></sub> prediction compared with other ML models and existing empirical correlations. Additionally, the PSO-MARS model provides intuitive expressions of the predicted outcomes. Among the four tested groups, the <i>q</i><sub>t</sub>-<i>f</i><sub>s-</sub>Δ<i>u</i>-<i>σ</i><sub>v0</sub>-<i>σ</i>'<sub>v0</sub> combination is identified as the optimal MARS model and is recommended for <i>s</i><sub>u</sub> prediction in engineering practice.</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

Improved prediction of undrained shear strength of clay from CPTU data using an ensemble PSO-MARS model with robustness analysis

  • 期刊论文

  • 发表日期：

    2025年6月28日

查看全文
  • Wei Duan,
  • Tianren Li,
  • Zening Zhao,
  • Guojun Cai,
  • Xinyao Li,
  • Shaoyun Pu,
  • Xiaoqiang Li,
  • Zhangqian Wu,
  • Songyu Liu

摘要

The undrained shear strength (su) is an important parameter for clays. However, current empirical models for predicting su are not sufficiently reliable. A novel data-driven ensemble learning method combining the particle swarm optimization (PSO) algorithm and multiple adaptive regression splines (MARS) has been developed to capture the relationships between su and piezocone penetration test (CPTU) parameters based on collected datasets. Four combinations of CPTU measurements (cone tip resistance, qt; sleeve friction resistance, fs; and excess pore water pressure, Δu; or normalized net cone resistance, Qt; normalized friction ratio, Fr; and pore pressure parameter Bq) and in-situ stresses (total vertical stress, σvo, and effective vertical stress, σ´vo) are used as input variables to determine the optimal combination of variables. In this novel model, PSO optimizes the hyperparameters in the MARS algorithm to form the PSO-MARS model, which has the advantage of visualizing the model expression. The performance of the PSO-MARS model is specifically compared with existing empirical correlations and other machine learning (ML) models. The robustness of prediction models is analyzed using Monte Carlo simulation. The results show that the PSO-MARS model demonstrates higher accuracy and robustness in su prediction compared with other ML models and existing empirical correlations. Additionally, the PSO-MARS model provides intuitive expressions of the predicted outcomes. Among the four tested groups, the qt-fs-Δu-σv0-σ'v0 combination is identified as the optimal MARS model and is recommended for su prediction in engineering practice.