Scholarlink

<p>Monitoring and early warning of brittle rock failure face inherent challenges due to the scarcity of acoustic emission (AE) events and the short duration of precursor phases. This study developed a passive monitoring methodology based solely on AE signals, integrating wave velocity/damage tomography with multiparameter instability warning techniques, and systematically validated its effectiveness. A three-dimensional (3D) wave velocity inversion algorithm was developed using AE arrival-time data, optimized via the Fast Marching Method (FMM) and a quasi-Newton scheme. The reconstructed wave velocity field was further converted into a damage field through a wave velocity–damage constitutive relationship. Validation using AE event density fields and DIC strain fields confirmed that this methodology accurately reconstructed the dynamic evolution of internal wave velocity and damage under all four loading methods of Brazilian tensile tests. Moment tensor inversion revealed the pronounced influences of loading configurations on microscopic fracture mechanisms. Flat platen loading generated the highest proportion of tensile cracks (47.33%), although these cracks were spatially dispersed, whereas cushion strip loading displayed shear-dominated failure characteristics (40.01%). The patterns of released AE energy exhibited strong correspondence with the evolving crack processes. Both flat platen and arc loading showed abrupt energy release near peak stress, while cushion strip loading exhibited a continuous, dispersed emission of low-energy events. These results indicate a failure evolution trend transitioning from dispersed brittle failure toward more localized progressive failure. Critical slowing down (CSD) analysis of features extracted using an integrated discrete wavelet transform (DWT)-1D convolutional neural network (1D-CNN)-channel attention mechanism achieved early identification of precursory signals across various loading scenarios, demonstrating enhanced robustness in precursor recognition. This study provides novel technical approaches for damage evolution characterization and instability prediction of rock failure. Compared with existing monitoring methods that mainly rely on active excitation or conventional AE statistical parameters, this study integrates passive AE-based damage tomography with extended AE-parameter-based CSD precursor identification. This provides a new framework for internal damage visualization and early warning during rock tensile failure.</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

Damage Tomography and Instability Prediction of Rock Under Tensile Loading Based on Passive Acoustic Emission Monitoring

  • 期刊论文

  • 发表日期：

    2026年6月16日

查看全文
  • Cheng Lyu,
  • Kai Liu,
  • Xiongyu Hu,
  • Yuewen Liu,
  • Wuqiang Cai,
  • Xiao Zhang,
  • Kun Huang

摘要

Monitoring and early warning of brittle rock failure face inherent challenges due to the scarcity of acoustic emission (AE) events and the short duration of precursor phases. This study developed a passive monitoring methodology based solely on AE signals, integrating wave velocity/damage tomography with multiparameter instability warning techniques, and systematically validated its effectiveness. A three-dimensional (3D) wave velocity inversion algorithm was developed using AE arrival-time data, optimized via the Fast Marching Method (FMM) and a quasi-Newton scheme. The reconstructed wave velocity field was further converted into a damage field through a wave velocity–damage constitutive relationship. Validation using AE event density fields and DIC strain fields confirmed that this methodology accurately reconstructed the dynamic evolution of internal wave velocity and damage under all four loading methods of Brazilian tensile tests. Moment tensor inversion revealed the pronounced influences of loading configurations on microscopic fracture mechanisms. Flat platen loading generated the highest proportion of tensile cracks (47.33%), although these cracks were spatially dispersed, whereas cushion strip loading displayed shear-dominated failure characteristics (40.01%). The patterns of released AE energy exhibited strong correspondence with the evolving crack processes. Both flat platen and arc loading showed abrupt energy release near peak stress, while cushion strip loading exhibited a continuous, dispersed emission of low-energy events. These results indicate a failure evolution trend transitioning from dispersed brittle failure toward more localized progressive failure. Critical slowing down (CSD) analysis of features extracted using an integrated discrete wavelet transform (DWT)-1D convolutional neural network (1D-CNN)-channel attention mechanism achieved early identification of precursory signals across various loading scenarios, demonstrating enhanced robustness in precursor recognition. This study provides novel technical approaches for damage evolution characterization and instability prediction of rock failure. Compared with existing monitoring methods that mainly rely on active excitation or conventional AE statistical parameters, this study integrates passive AE-based damage tomography with extended AE-parameter-based CSD precursor identification. This provides a new framework for internal damage visualization and early warning during rock tensile failure.