Scholarlink

<p>Precise calculation of soil bearing capacity is critical in geotechnical engineering to ensure ground stability and structural safety. Traditional evaluation techniques such as the Standard Penetration Test (SPT), Cone Penetration Test (CPT), and Plate Load Test (PLT) are well-established but often time-consuming, labor-intensive, and spatially constrained. This study presents a semi-automated, real-time method for estimating soil bearing capacity by integrating torque, force, and rotational speed sensors into conventional drilling equipment. Unlike prior Measuring-While-Drilling (MWD) approaches, which have largely focused on granular formations and deeper borehole profiling, this work introduces a custom-built torque measurement system specifically designed for shallow-depth, low-permeability clayey soils. The system offers improved sensitivity to subtle resistance changes encountered during cohesive soil penetration, thereby enhancing prediction accuracy in scenarios where conventional MWD systems typically underperform. Laboratory and field tests were performed on four clayey soil types (CH, MH, SC, CL), and the collected drilling parameter data were analyzed using Multiple Linear Regression (MLR) and Response Surface Methodology (RSM). The MLR model explained 95.6% of the variability in soil bearing capacity (R<sup>2</sup> = 0.956, MAPE = 7.87%), although it was limited in capturing non-linear interactions. In contrast, the RSM model accounted for 99.7% of the variability (R<sup>2</sup> = 0.997, MAPE = 0.72%) and more effectively modeled the complex relationships among drilling parameters. Among all inputs, torque emerged as the most significant predictor of bearing capacity. The developed framework enables faster, more cost-effective, and sensor-integrated evaluation of soil strength, especially for cohesive soils offering a practical alternative to conventional testing. Future work will extend this approach to mixed and granular soils, deeper borehole conditions, and adaptive, ML-driven real-time control systems to enhance field-scale geotechnical applications.</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

Real-time prediction of soil bearing capacity in clayey soils using drilling parameters and statistical modeling

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年10月29日

查看全文
  • Prashant Pande,
  • Jayant Giri,
  • Jayant Raut,
  • Hassen Sabeur,
  • Rajesh Bhagat,
  • Muzhda Azizi

摘要

Precise calculation of soil bearing capacity is critical in geotechnical engineering to ensure ground stability and structural safety. Traditional evaluation techniques such as the Standard Penetration Test (SPT), Cone Penetration Test (CPT), and Plate Load Test (PLT) are well-established but often time-consuming, labor-intensive, and spatially constrained. This study presents a semi-automated, real-time method for estimating soil bearing capacity by integrating torque, force, and rotational speed sensors into conventional drilling equipment. Unlike prior Measuring-While-Drilling (MWD) approaches, which have largely focused on granular formations and deeper borehole profiling, this work introduces a custom-built torque measurement system specifically designed for shallow-depth, low-permeability clayey soils. The system offers improved sensitivity to subtle resistance changes encountered during cohesive soil penetration, thereby enhancing prediction accuracy in scenarios where conventional MWD systems typically underperform. Laboratory and field tests were performed on four clayey soil types (CH, MH, SC, CL), and the collected drilling parameter data were analyzed using Multiple Linear Regression (MLR) and Response Surface Methodology (RSM). The MLR model explained 95.6% of the variability in soil bearing capacity (R2 = 0.956, MAPE = 7.87%), although it was limited in capturing non-linear interactions. In contrast, the RSM model accounted for 99.7% of the variability (R2 = 0.997, MAPE = 0.72%) and more effectively modeled the complex relationships among drilling parameters. Among all inputs, torque emerged as the most significant predictor of bearing capacity. The developed framework enables faster, more cost-effective, and sensor-integrated evaluation of soil strength, especially for cohesive soils offering a practical alternative to conventional testing. Future work will extend this approach to mixed and granular soils, deeper borehole conditions, and adaptive, ML-driven real-time control systems to enhance field-scale geotechnical applications.