Scholarlink

<p>In this study, the compressive failure mechanisms of three shotcrete–rock composites (S–G, S–BS, and S–WS) under uniaxial compression were investigated. Non-destructive techniques, including digital image correlation (DIC) and acoustic emission (AE), combined with machine-learning methods, were employed for in-depth insights into deformation and failure behaviors of the composites. Results revealed that strain concentration occurred primarily in the shotcrete, while rock regions exhibited more uniform strain. Displacement field analysis identified shear-dominated failure in the shotcrete, accompanied by tensile splitting in granite (S–G) and black sandstone (S–BS), and shear failure in white sandstone (S–WS), respectively. Strain variations in the <i>x</i>- and <i>y</i>-directions highlighted the significant role of the rock base in shotcrete deformation. Besides, a semi-supervised ternary Naive Bayes (NB) classification model was developed to infer damage sources from unlabeled AE signals. Probabilistic classification outcomes indicated that the average AE parameters for interface damage consistently fell between the values associated with rock damage and shotcrete damage. The number of damage signals in different composites exhibited the greatest increase near 90% of the peak load during the post-peak stage. In S–G, shotcrete damage prevailed during early loading stages, transitioning to interface damage near the peak. S–BS consistently exhibited shotcrete-dominant damage with stable damage proportions, whereas S–WS began with rock-damage dominance, with shotcrete damage becoming more prominent as stress approached the peak. Furthermore, among the composites, S–BS achieved the highest compressive strength, while S–WS exhibited the lowest. These findings offer practical insights for designing and optimizing support systems in complex geotechnical projects.</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

In-Depth Understanding of the Compressive Failure Mechanism in Shotcrete–Rock Composites Using Non-destructive Techniques and Machine-Learning Algorithms

  • 期刊论文

  • 发表日期：

    2026年5月19日

查看全文
  • Dandan Shi,
  • Xudong Chen,
  • Yuzhu Guo,
  • Lu Feng

摘要

In this study, the compressive failure mechanisms of three shotcrete–rock composites (S–G, S–BS, and S–WS) under uniaxial compression were investigated. Non-destructive techniques, including digital image correlation (DIC) and acoustic emission (AE), combined with machine-learning methods, were employed for in-depth insights into deformation and failure behaviors of the composites. Results revealed that strain concentration occurred primarily in the shotcrete, while rock regions exhibited more uniform strain. Displacement field analysis identified shear-dominated failure in the shotcrete, accompanied by tensile splitting in granite (S–G) and black sandstone (S–BS), and shear failure in white sandstone (S–WS), respectively. Strain variations in the x- and y-directions highlighted the significant role of the rock base in shotcrete deformation. Besides, a semi-supervised ternary Naive Bayes (NB) classification model was developed to infer damage sources from unlabeled AE signals. Probabilistic classification outcomes indicated that the average AE parameters for interface damage consistently fell between the values associated with rock damage and shotcrete damage. The number of damage signals in different composites exhibited the greatest increase near 90% of the peak load during the post-peak stage. In S–G, shotcrete damage prevailed during early loading stages, transitioning to interface damage near the peak. S–BS consistently exhibited shotcrete-dominant damage with stable damage proportions, whereas S–WS began with rock-damage dominance, with shotcrete damage becoming more prominent as stress approached the peak. Furthermore, among the composites, S–BS achieved the highest compressive strength, while S–WS exhibited the lowest. These findings offer practical insights for designing and optimizing support systems in complex geotechnical projects.