Correlational fractal analysis of the Kaiser effect in rock: implications for salt-cavern engineering under triaxial cyclic loading and unloading

The Kaiser effect exhibits stress-induced changes in the acoustic properties of rock, so accurately identifying the Kaiser point of rock under complex conditions is essential for understanding the damage and failure mechanisms. In this paper, triaxial loading and unloading tests were conducted on salt rock using MTS815 material testing machine and PCI-II acoustic emission (AE) system. By introducing the Kaiser point calculation approach involving AE signal optimisation, Grassberger-Procaccia (GP) algorithm, K-means clustering, and random sample consensus (RANSAC) algorithm, we achieved enhanced accuracy in analysing the Kaiser effect for salt rock. The cumulative AE energy exhibited a distinct surge point during the cycle number (N) was less than 21, closely aligned with the optimised key AE energy point. For N > 21, the cumulative energy displayed stable changes, rendering conventional method ineffective. By utilising screened AE signals (8796 samples) within the GP algorithm, the surge point of cumulative energy consistently aligned with the sudden drop point of the correlation dimension when N<21, with a maximum error of 6 s. Remarkably, the GP algorithm identified the prominent Kaiser point even when N > 21, suggesting its applicability throughout the loading cycle.