A discrimination method for rock damage mechanisms: insights from AE waveform complexity and the morphological parameter

Accurate identification of rock damage mechanisms is crucial for the safe exploitation of deep energy resources and the long-term stability of underground storage. This study performs full-process acoustic emission (AE) monitoring during uniaxial compression and Brazilian splitting tests on granite, sandstone and marble, and systematically investigates the relationships between AE waveform complexity (Kt), information entropy density (IED) and rock damage mechanisms. A ternary classification framework is developed to distinguish tensile, shear and mixed-mode failure using AE waveform parameters. Validation against moment tensor inversion and the conventional AF–RA technique demonstrates the reliability of the method, with a maximum deviation of only 5.15% from moment tensor results. The median IED of tensile signals is consistently lower, whereas their median Kt is higher than those of shear signals, indicating that these two parameters are effective indicators of damage mechanisms. Compared with AF–RA, the framework enables three-mode classification and provides clearer physical interpretation and quantitative criteria, while, unlike moment tensor inversion, it does not rely on precise arrival-time picking or phase identification. The results reveal that rock damage evolves from an initial tensile stage to mixed-mode and finally shear-dominated failure.