ABSTRACT
Classification of acoustic emission (AE) signals emitted during failure processes relative to the cracking mode enables assessment of the condition of a concrete structure.The reason is the sequence of fracture modes starts with concrete cracking due to moderate tensile stresses and leads to shearing phenomena before final failure.Therefore, reliable characterization ofAE sources, may lead to warning before fracture reaches its final stage. The technique of AE uses piezoelectric sensors on the surface of the material to record the stress waves produced by the motion of the crack tips and transform them into electric waveforms, (Grosse & Ohtsu 2008). Basic waveform characteristics are; the maximum voltage of the waveform (amplitude, A in dB), the number of threshold crossings (counts), the duration (time between the first and last threshold crossings) and the rise time (RT inµs) which is the time between the first threshold crossing and the waveform peak amplitude, see Figure 1. A frequency indicator is the average frequency (AF in kHz), which is defined as the number of counts over the duration, while the central frequency, CF (centroid of the FFT spectrum in kHz) is also used. Additionally the “RA” value is the ratio of RT over A and is measured in µs/V. The number of recorded signals (hits) yields information on the active fracturing points and contributes to the monitoring of processes like fracture (Carpinteri et al. 2008, Karihaloo et al. 2013), creep (Verstrynge et al. 2009),
corrosion (Uddin et al. 2006) and self-healing (Van Tittelboom et al. 2012). Localization of the sources is also possible in three dimensions if multiple sensors are used (Luo et al. 2006). In order to obtain qualitative information on the cracking mode several waveform shape parameters have been examined. In recent studies it is shown that frequency characteristics measured by AF, decrease for the shear mode of cracking relatively to the tensile. Additionally, RA increases substantially due to the different wave modes emitted. These trends have been exploited in laboratory conditions with quite successful characterization of the AE sources as to their mode (Aggelis 2011, Shahidan et al. 2013, Farhidzadeh et al. 2013).The classification is based on the AF-RA axes in which the tensile and shear cracking AE signal populations exhibit limited overlap in many laboratory cases so that they can be separated into clusters by a single line.
