Damage classification in concrete, mortar and cement-paste beams under bending by Acoustic Emission Technique (AET)

Abstract

In concrete, AET has been used to characterize the different fracture modes in flex-ural beams by RA method (described in RILEM TC 212 ACD recommendation). In this study, three-point-bending tests have been conducted on cement-based specimens (concrete, mortar and cement-paste) to promote tensile mode and/or shear mode in these specimens. Acoustic Emission (AE) raw data were filtered to remove noise from signals. Then, k-means algorithm, as a Pattern Recognition Analysis (PRA) method, was used to identify the similarities within the recorded signals and to classify the filtered-signals into groups. Clustering results by k-means according to a proposed process have proven efficiency in providing information on RA, AF, Rise-time, Duration and Absolute Energy (ABEN) related to mode I and mode II. Then, two damage classification techniques were used, in which RA method was applied to filtered signals. By these techniques, the proportion of mode I and II was specified and the location of demarcation line also fixed.