Experimental study on identifying catastrophic failure in the brittle fracture process via multi-source acoustic characteristics.

Identifying and predicting the catastrophic failure of brittle rock remains a challenging task, yet it is crucial for developing early warning systems and preventing dynamic rock hazards. In this study, we employed the propagative parameters of ultrasonic waves and information from acoustic emission (AE) events to characterize the brittle failure of a flawed sandstone sample under uniaxial compression. A sliding event window method was developed to obtain the temporal b-value, effectively revealing microcrack growth based on the frequency-magnitude distribution of AE events. The precursory period candidate (PPC) was defined based on the specific change pattern of the temporal b-value. The peak amplitude, velocity, and differential acoustic attenuation term (DAAT) of the transmitted ultrasonic waves were analyzed to reflect the damage state and stress conditions of the sample. Furthermore, a comprehensive analysis was carried out to model the correlation between the ultrasonic propagative parameters and the AE temporal b-value. The results revealed that the DAAT exhibited the most distinct characteristics in two PPCs, suggesting that a global increase in the DAAT, combined with a sudden drop and subsequent slight recovery of the temporal b-value, could be used to identify the impending catastrophic failure of the sandstone sample. These findings from multi-source acoustic characteristics provide new insights for understanding the development of ultimate catastrophic failure in brittle rock and may serve as a theoretical reference for early alerts regarding dynamic rock disasters in practical engineering applications.