Studying of Rock Failure Mechanisms by Dynamic Scratch Test Data

In this paper, the scratch test method is proposed to estimate geomechanical properties of rock materials. The scratch test is realized in the original experimental set-up, which allows recording the dynamics of force and acoustic emission (AE) signals in the process of scratch test. The data processing method is based on the original results, which reveal the patterns of criticality in the stages of transition from disperse to macroscopic fracture based on the adaptive sequential clustering approach. Three classes of AE signals were determined: “noise” signals with maximum dispersion at a high mean value, characterizing the initial stage of damage development, and two classes of high-energy signals with low dispersion, characterizing the quasi-plastic relaxation mechanism in the formation of collective defect modes, and the quasi-brittle scenario of fracture nucleus formation.