To explore the potential precursors of rock instability, it is necessary to clarify the mechanism of micro-crack from fracturing to failure, which involves the evolution of fracture size, orientation, source model, and their relationships to the loading. The waveforms of acoustic emission (AE) recorded by the sensor network attached rock sample during laboratory tests provide a data basis for solving these problems, since these observations are directly related to the characteristics of the fracturing sources. Firstly, we investigated the source mechanism, looking at the rise angle and the average frequency (RA-AF) trends during five loading stages in a uniaxial compression test. Results show that the proportion of shear events significantly increases when approaching instability. Secondly, we calculated the moment tensor for each event, considering the uncertainties of P-wave polarity, azimuth, and the takeoff angles of the rays. Moment tensor solutions suggest that there are obviously more crack events than shear events in all loading stages. Moment tensor evolutions confirmed that the decreasing of isotropic component and the increment of double-couple can be used as precursors of rock fracturing development. Considering the limitations of these two methods, it is suggested that we should be concerned more about the proportions of individual failure components and their evolutions over time, instead of absolutely classifying the events into a certain source type.
Keywords: RA-AF; acoustic emission; micro-failure precursor; moment tensor; sensor; source type.