An analysis of the meso-structural damage evolution of coal using X-ray CT and a gray-scale level co-occurrence matrix method

• Published in: International journal of rock mechanics and mining sciences (Oxford, England : 1997) Vol. 152; p. 105062
• Main Authors: Wu, Yan, Wang, Dengke, Wang, Lei, Shang, Zhengjie, Zhu, Chuanqi, Wei, Jianping, Yuan, Anying, Zhang, Hongtu, Zeng, Fanchao
• Format: Journal Article
• Language: English
• Published: Berlin Elsevier Ltd 01.04.2022; Elsevier BV
• Subjects: Algorithms; Cluster analysis; Clustering; Coal; Coal damage; Compression; Computed tomography; Control systems; Crack initiation; Damage detection; Damage evolution; Deformation; Evolution; Fractures; GLCM method; Homogeneity; Image acquisition; Image contrast; Matrix methods; Normal distribution; Scanning; Statistical analysis; Statistical features; Statistics; Structural damage; Vector quantization; X-ray CT scan; GLCM method; Damage evolution; Statistical features; Coal damage; X-ray CT scan
• ISSN: 1365-1609; 1873-4545
• DOI: 10.1016/j.ijrmms.2022.105062

To study the meso-damage evolution in coal under loading conditions, the industrial CT scanning tests under the uniaxial compression condition were carried out using an industrial CT scanner equipped with a loading control system. The gray level co-occurrence matrix (GLCM) theory was applied to quantitatively analyze the meso-damage evolution and the fracturing characteristics using the acquired CT images at each scanning stage. Four statistical features (i.e., contrast, energy, correlation, and homogeneity) of the CT images were extracted to evaluate the internal damage evolution in coal. The results show that the contrast generally took a changing trend of first decreasing and then increasing with the closure, initiation, and expansion of the fractures, that is, it decreased slowly at the compaction stage, slightly increased at the elasticity stage, and drastically increased at the post-peak stage. On the contrary, the energy, correlation, and homogeneity showed a changing trend opposite to that of the contrast. The changing tendency of mean values of the four statistical features could be depicted by the Boltzmann function, and in addition, the statistical features followed Gaussian distribution during the complete stress-strain process. The maximum damage cross-section of the coal sample could be transferred with the increase of deformation in the coal sample, indicating that the final failure location may not be the maximum initial damage location. The fracture ratio measured by the binarization and K-Means clustering algorithm methods typical showed three changing stages (i.e., a slow reduction process, a slight increase process, and a dramatic increase process) and could be used to reflect the meso-damage development process in coal at different deformation stages.