Scholarlink

<p>Accurately predicting the load-bearing capacity of cable bolts in underground excavations remains a critical challenge due to the complex interplay of geometric, material, and geological parameters. This study proposes a novel artificial neural network (ANN) model to optimize a cable bolt design by leveraging data from 437 in situ pull-out tests conducted at the Imiter silver mine in Morocco. Nine key parameters—including cable diameter, hole diameter, grout strength, embedment length, water-cement ratio, and rock mass properties—were systematically analyzed. The ANN model demonstrated superior predictive accuracy (<i>R</i><sup>2</sup> = 0.98, RMSE = 17.57) compared to support vector machines (SVM), <i>k</i>-nearest neighbors (KNN), and group method of data handling (GMDH). Parametric analyses revealed the dominance of water-cement ratio, grout strength, and cable geometry in load-bearing behavior, with failure modes predominantly characterized by a grout-cable slippage. Finite element simulations and kinematic analyses further validated the role of rock mass relaxation zones and structural discontinuities in instability. The ANN framework provides a robust, data-driven tool for enhancing ground support reliability, reducing over-conservative designs, and improving safety in underground mining operations. This research advances machine learning applications in geotechnical engineering, offering actionable insights for optimizing cable bolt systems in complex geological settings.</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

Machine Learning Models for Cable Bolt Load-Bearing Capacity in Sublevel Open Stope Mining

  • 期刊论文

  • 发表日期：

    2025年4月11日

查看全文
  • Amine Soufi,
  • Youssef Zerradi,
  • Anas Bahi,
  • Mohamed Souissi,
  • Latifa Ouadif

摘要

Accurately predicting the load-bearing capacity of cable bolts in underground excavations remains a critical challenge due to the complex interplay of geometric, material, and geological parameters. This study proposes a novel artificial neural network (ANN) model to optimize a cable bolt design by leveraging data from 437 in situ pull-out tests conducted at the Imiter silver mine in Morocco. Nine key parameters—including cable diameter, hole diameter, grout strength, embedment length, water-cement ratio, and rock mass properties—were systematically analyzed. The ANN model demonstrated superior predictive accuracy (R2 = 0.98, RMSE = 17.57) compared to support vector machines (SVM), k-nearest neighbors (KNN), and group method of data handling (GMDH). Parametric analyses revealed the dominance of water-cement ratio, grout strength, and cable geometry in load-bearing behavior, with failure modes predominantly characterized by a grout-cable slippage. Finite element simulations and kinematic analyses further validated the role of rock mass relaxation zones and structural discontinuities in instability. The ANN framework provides a robust, data-driven tool for enhancing ground support reliability, reducing over-conservative designs, and improving safety in underground mining operations. This research advances machine learning applications in geotechnical engineering, offering actionable insights for optimizing cable bolt systems in complex geological settings.