Predicting rock failure with the critical slowing down theory

Abstract Early warning signals of critical transitions in rock before its sudden failure are key for forecasting geological disasters in rock engineering. Here we examined the critical slowing down characteristics of acoustic emission (AE) and precursory signals preceding catastrophic rock failure. We conducted laboratory experiments on different types of rock specimens subjected to tension and compression, instrumented by real-time AE monitoring. Based on the critical slowing down theory, AE signals monitored over the deformation and failure of the rock specimens were parametrized statistically. The results show that the critical slowing down characteristics of AE signals exist for rocks under various experimental conditions. The sudden and significant increase in the variance and the autocorrelation coefficient of AE can be used as precursors for rock failure. The time intervals between the precursory signals and rock rupture for brittle rocks under tension are smaller than those under compression. These early warning time intervals are longer and the ratios of the early warning to peak states are lower during Brazilian tests than during direct tensile tests. Furthermore, relationships between multiple influencing factors and precursory signals were discussed. We found that the precursory signals are strongly related to the intrinsic properties of rock and the external experimental conditions. Finally, the applicability and reliability of precursory signals through the critical slowing down theory were verified for rock failure by statistically analysing the laboratory data from a wide range of rocks.