Scholarlink

<p>Accurate assessment of rock mass quality is essential for the safe and cost-effective design of deep underground infrastructure, particularly in geologically heterogeneous environments. Among the key geomechanical parameters, uniaxial compressive strength (UCS) plays a fundamental role in rock mass characterization and geotechnical design. Conventional UCS determination relies on borehole sampling and laboratory testing, which are invasive, costly, time-consuming, and spatially limited, especially at great depths. This study presents a non-invasive geophysical framework for predicting UCS distributions to depths exceeding 1&#xa0;km using controlled-source audio-frequency magnetotellurics (CSAMT). A generalized resistivity–UCS relationship was developed from 219 paired resistivity–UCS observations obtained from 35 boreholes across six major underground engineering sites encompassing granite, granite–tuff, tuff–basalt, and tuff–sandstone lithologies. The empirical model was subsequently applied to inverted CSAMT resistivity data to generate continuous 2D and pseudo-3D UCS models for deep subsurface characterization. Independent leave-one-site-out (LOSO) validation demonstrated strong predictive performance, yielding R² = 0.91, MAE = 6.85&#xa0;MPa, and RMSE = 11.83&#xa0;MPa, indicating good transferability across diverse geological settings. The resulting UCS models successfully delineate lithological heterogeneity, weathered and fractured zones, and major fault structures, enabling early identification of mechanically weak regions and supporting geotechnical risk assessment. Unlike previous studies that focused primarily on site-specific correlations, this work demonstrates a validated geophysical–geomechanical framework applicable across multiple geological environments. The proposed methodology provides a scalable and cost-effective tool for deep underground infrastructure planning, geomechanical modeling, and subsurface development in data-scarce regions.</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

Non-invasive geophysical prediction of deep rock mass UCS in complex geological settings

  • 期刊论文

  • 发表日期：

    2026年7月10日

查看全文
  • Muhammad Hasan,
  • Lijun Su,
  • Yanjun Shang

摘要

Accurate assessment of rock mass quality is essential for the safe and cost-effective design of deep underground infrastructure, particularly in geologically heterogeneous environments. Among the key geomechanical parameters, uniaxial compressive strength (UCS) plays a fundamental role in rock mass characterization and geotechnical design. Conventional UCS determination relies on borehole sampling and laboratory testing, which are invasive, costly, time-consuming, and spatially limited, especially at great depths. This study presents a non-invasive geophysical framework for predicting UCS distributions to depths exceeding 1 km using controlled-source audio-frequency magnetotellurics (CSAMT). A generalized resistivity–UCS relationship was developed from 219 paired resistivity–UCS observations obtained from 35 boreholes across six major underground engineering sites encompassing granite, granite–tuff, tuff–basalt, and tuff–sandstone lithologies. The empirical model was subsequently applied to inverted CSAMT resistivity data to generate continuous 2D and pseudo-3D UCS models for deep subsurface characterization. Independent leave-one-site-out (LOSO) validation demonstrated strong predictive performance, yielding R² = 0.91, MAE = 6.85 MPa, and RMSE = 11.83 MPa, indicating good transferability across diverse geological settings. The resulting UCS models successfully delineate lithological heterogeneity, weathered and fractured zones, and major fault structures, enabling early identification of mechanically weak regions and supporting geotechnical risk assessment. Unlike previous studies that focused primarily on site-specific correlations, this work demonstrates a validated geophysical–geomechanical framework applicable across multiple geological environments. The proposed methodology provides a scalable and cost-effective tool for deep underground infrastructure planning, geomechanical modeling, and subsurface development in data-scarce regions.