Progressive damage and fracture of biaxially-confined anisotropic coal under repeated impact loads

Abstract

Coal masses in underground mines are in multiaxial stress states after excavation, and they are frequently subjected to dynamic loads particularly from continuous mining-induced seismic events. The changes of in-situ stress conditions and external dynamic loads may induce coal bursts, which involves the violent breaking of coal masses. Meanwhile, the coal bursting process may be influenced by bedding planes that are considered as unique structures in coal masses, which raises uncertainties of coal bursting mechanisms. This paper aims to reveal the progressive damage and burst proneness of anisotropic coal under coupled static stress and repeated dynamic loads. The tested coal specimens have five bedding plane orientations (β = 0°, 30°, 45°, 60°, and 90°), where β is the angle between the bedding plane and dynamic loading direction. A triaxial Hopkinson bar (Tri-HB) system is utilised to apply the biaxial static stress on the specimen and then followed by dynamic loading, and the stress and strain can be recorded by the system accordingly. After the loading test, ultrasonic wave velocity is measured to examine the degree of attenuation, and the quantification of microcrack characteristics is conducted by using the synchrotron-based X-ray computed tomography (CT). Then, the specimen is loaded again under the same biaxial static stress and dynamic load, and measured by ultrasonic sensors and X-ray CT. This process will be repeated until the final failure of the specimen. We found that both dynamic peak stress and P-wave velocity show a decreasing trend due to damage accumulation with increasing number of dynamic tests. The bearing capacities of anisotropic coal are mainly governed by bedding angles, for example, the specimen with β = 30° undergoes the lowest impact numbers, but the one with β = 90° can undergo the highest numbers. The crack initiation and propagation are progressively revealed by two-dimensional (2D) X-ray slices and 3D fracture network reconstructions. As impact number increases, both 2D and 3D fractal dimensions of cracks increase exponentially, whereas dynamic peak stress and P-wave velocity are negatively correlated with them.