Scholarlink

<p>Accurate prediction of the drilling rate index (DRI) is critical for efficient excavation and quantifying risk in underground engineering. However, the inherent randomness and nonlinear interdependence structure of rock mass mechanical parameters restrict the accuracy of DRI prediction. To surmount the limitations, a correlation-embedded intelligent probabilistic evaluation framework is proposed in this study. A kriging surrogate model is first employed to establish the nonlinear relationship linking DRI to three input parameters: uniaxial compressive strength (UCS), Brazilian tensile strength (BTS), and point load strength (Is(50)). Subsequently, the C-vine copula is introduced to capture the nonlinear dependence structure of the input parameters, thus overcoming the limitations of the "variable independence" assumption prevalent in traditional analyses. Leveraging this foundation, statistically sufficient random input samples meticulously preserving the original correlation characteristics can be generated via Monte Carlo simulation (MCS). Validation using three test samples demonstrates that the correlation-embedded model effectively constrains the simulated sample space, and the distinguishability of the predicted DRI grades is superior to that of the independent model. Additionally, sensitivity analysis based on case data reveals that Is(50) is the dominant factor influencing DRI variability. The research findings provide a statistically robust and physically credible theoretical paradigm for DRI assessment and drilling parameter optimization.</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

Quantifying Uncertainty in Rock Mass Drillability: A Probabilistic Framework with Embedded Parameter Dependence

  • 期刊论文

  • 发表日期：

    2026年4月6日

查看全文
  • Guangkun Li,
  • Yiguo Xue,
  • Yi Han,
  • Xiying Cheng,
  • Youquan Liu,
  • Zhiwen Tong

摘要

Accurate prediction of the drilling rate index (DRI) is critical for efficient excavation and quantifying risk in underground engineering. However, the inherent randomness and nonlinear interdependence structure of rock mass mechanical parameters restrict the accuracy of DRI prediction. To surmount the limitations, a correlation-embedded intelligent probabilistic evaluation framework is proposed in this study. A kriging surrogate model is first employed to establish the nonlinear relationship linking DRI to three input parameters: uniaxial compressive strength (UCS), Brazilian tensile strength (BTS), and point load strength (Is(50)). Subsequently, the C-vine copula is introduced to capture the nonlinear dependence structure of the input parameters, thus overcoming the limitations of the "variable independence" assumption prevalent in traditional analyses. Leveraging this foundation, statistically sufficient random input samples meticulously preserving the original correlation characteristics can be generated via Monte Carlo simulation (MCS). Validation using three test samples demonstrates that the correlation-embedded model effectively constrains the simulated sample space, and the distinguishability of the predicted DRI grades is superior to that of the independent model. Additionally, sensitivity analysis based on case data reveals that Is(50) is the dominant factor influencing DRI variability. The research findings provide a statistically robust and physically credible theoretical paradigm for DRI assessment and drilling parameter optimization.