Identifying impending failure in heterogeneous materials: A study on acoustic emission time series

Abstract

In the present paper, precursors to identify impending failure in heterogeneous materials are investigated by analyzing Acoustic Emission (AE) time series. Three main approaches are considered: (1) long-range correlations using the Hurst (H) analysis, (2) Natural Time (NT) analysis, and (3) a proposed Method of Critical Fluctuations-Based (MCF-B) approach. The above approaches are employed to data collected from an experimental AE test on a concrete cubic specimen subjected to an axial compression loading. The results indicate that: persistent long-range correlations tend towards near-random behavior prior to imminent failure; critical regions, characterized by the convergence of natural time parameters, show a direct correlation with imminent increments in the acoustic emission activity and the structure's collapse; the proposed MCF-B approach as precursor, is capable to identify critical states, being the results in agreement with those from NT and Hurst analysis.