Abstract The understanding of failure mechanisms of rocks containing multiple cavities is important for stability assessment of rock excavations in rock engineering applications. However, published research is still limited on the failure mechanism between cavities and quantitative assessment of the crack development characteristics for different cavity configurations. In this research, specimens containing two cavities were investigated for their displacement characteristics and failure mechanisms under uniaxial compression. Digital image correlation (DIC) was used to measure the displacements of the specimens in experiments, while the failure process was modelled using the Particle Flow Code (PFC) software. A density-based clustering method was used to detect micro-crack clusters and the well-known statistical tool of Principal Component Analysis (PCA) was applied innovatively in this work to define the macro-cracks. The clustering method was further developed in this work for vector analysis so that the displacement vectors around macro-cracks could be classified into different types based on the failure mechanism to understand the crack development characteristics. The crack development between two cavities and failure mechanism for different geometrical configurations were subsequently studied in detail using experimental and numerical results.

中文翻译：

---

类岩石材料单轴压缩两腔间裂纹扩展特性的试验与数值研究

摘要 理解多腔体岩石破坏机理对于岩石工程应用中岩石开挖稳定性评估具有重要意义。然而，已发表的研究仍然局限于腔之间的失效机制和不同腔配置的裂纹发展特征的定量评估。在这项研究中，研究了包含两个腔体的试样在单轴压缩下的位移特性和破坏机制。实验中使用数字图像相关（DIC）来测量试样的位移，而使用粒子流代码（PFC）软件对破坏过程进行建模。使用基于密度的聚类方法检测微裂纹簇，并在这项工作中创新地应用了著名的主成分分析（PCA）统计工具来定义宏观裂纹。在这项工作中进一步发展了聚类方法进行矢量分析，以便根据破坏机制将宏观裂纹周围的位移矢量分为不同类型，以了解裂纹发展特征。随后使用实验和数值结果详细研究了两个腔之间的裂纹发展和不同几何配置的失效机制。在这项工作中进一步发展了聚类方法进行矢量分析，以便根据破坏机制将宏观裂纹周围的位移矢量分为不同类型，以了解裂纹发展特征。随后使用实验和数值结果详细研究了两个腔之间的裂纹发展和不同几何配置的失效机制。在这项工作中进一步发展了聚类方法进行矢量分析，以便根据破坏机制将宏观裂纹周围的位移矢量分为不同类型，以了解裂纹发展特征。随后使用实验和数值结果详细研究了两个腔之间的裂纹发展和不同几何配置的失效机制。