Predicting fracture precursors in cementitious materials using natural time analysis coupled with non-extensive statistical mechanics

Abstract

In recent years, there has been growing interest in using acoustic emission (AE) time series data to develop fracture precursors for predicting imminent failure in cementitious materials. In this context, natural time (NT) analysis used in seismology has proven to be a useful tool, offering insights into the critical stage, namely the region of criticality, that precedes the mainshock event. Therefore, in the present study, the occurrence of impending macroscopic fracture in cementitious composites is predicted using the NT analysis of acoustic emission. To achieve this, the parameters of natural time, namely the variance, \(\kappa _{1}\), the change in entropy, \(\Delta S\), and the complexity measure, \(\Lambda _{i}\), were utilized as precursors. Furthermore, the Tsallis q-index, obtained from the non-extensive statistical mechanics framework, was used in conjunction with the NT parameters. It was observed that the NT parameters, \(\kappa _{1}\) reached a critical value of 0.07, \(\Delta S\) attained a global minimum, and \(\Lambda _i\) exhibited an abrupt increase before the mainshock event, similar to that observed in seismicity prior to major earthquakes. Furthermore, a sudden drop in Tsallis q-index was observed before the mainshock event. In addition, the behavior of the cementitious composite material closely resembled that of the Olami–Feder–Christensen earthquake model, as evidenced by the significant increase in cumulative AE energy observed after the region of criticality. Therefore, \(\kappa _{1}\), \(\Delta S\), and \(\Lambda _i\) in combination with the q-index could be utilized as precursors for detecting impending macroscopic fracture in cementitious composites.