PSI - Issue 3

D. Triantis et al. / Procedia Structural Integrity 3 (2017) 346–353 2 D. Triantis, E.D. Pasiou, I. Stavrakas, I. Dakanali and S.K. Kourkoulis / Structural Integrity Procedia 00 (2017) 000–000

347

1996; Vidya Sagar & Raghu Prasad 2012). Proper analysis of the amplitudes’ distribution of these events, known as either b-value (Colombo et al. 2003; Main 1989) or improved b-value (Ib) analysis (Shiotani et al. 2001), is nowadays widely used for the study of fracture process in rock-like and cement-based materials. Identification and evaluation of various stages of the fracture process is, also, implemented based on a series of acoustic parameters, like for example, the acoustic event (or hit) rate, the released energy rate, the cumulative energy and the ring down count. From another point of view, valuable information about the damage process in quasi-brittle materials is provided, also, by the detection of low-level electric signals. Indeed, the formation of micro-cracks in a quasi-brittle non metallic material produces electric charges, i.e. a complicated charged system of electric dipoles (Vallianatos et al. 2004; Va rotsos et al. 2002). These electric dipoles produce an electric potential along the crack resulting to flow of current. Detection of such electric currents is considered as precursor of large-scale fractures and this is already verified at both laboratory (Frid et al. 2009; Vallianatos et al. 2004) and geodynamic scale (Varotsos 2005). These weak electrical current emissions are known as Pressure Stimulated Currents (PSC) (Stavrakas et al. 2004). In the laboratory scale they are usually detected by a pair of electrodes, properly attached on the tested specimen, and they are recorded with the aid of a very sensitive electrometer. The PSC technique was initially applied on marble (Stavrakas et al. 2003, 2004; Triantis et al. 2006) and amphibolite specimens (Triantis et al. 2007), subjected to uniaxial compression. Later on, the technique was successfully applied on cement-based specimens (Kyriazopoulos et al. 2011; Triantis et al. 2012). Nowadays, the PSC technique is widely adopted worldwide, either in its original form (Cartwright-Taylor et al. 2014; Li et al. 2015; Sun et al. 2002) or properly modified for specific applications (Archer et al. 2016; Dann et al. 2014). In the present paper, both AE and PSC techniques are used to study cement-mortar specimens under three-point bending (3PB) at various loading rates in order to detect possible correlation of the respective experimental data. 2. Materials and experimental arrangement The cement-mortar specimens used were made of ordinary Portland cement, sand and water at a ratio (per weight) 1:3:0.5, respectively. Fifteen prismatic specimens (five groups of three specimens each) of dimensions 50 mm x 50 mm x 250 mm were prepared and they were properly cured for 90 days in order to attain 95% of their strength. All tests were carried out using an Instron DX-300 electromechanical loading frame (of 300kN capacity), under load-control mode at loading rates varying from 28 N/s to 135 N/s for each group of specimens. The specimens were supported by two metallic rods, each one positioned at a distance of about 100 mm from the specimen’s central cross section. The load was applied at the central cross-section by means of a third metallic rod. Thin teflon sheets (of thick ness equal to 2 mm) were placed between the specimen and the three rods to electrically isolate the specimen (Fig.1). A pair of electrodes was attached at the lower surface of each specimen (Fig.1) on either side of the central cross section (at a distance l≈a/5 from each other, where α is the specimen’s span) to detect the electric signal. The specific arrangement was chosen since it was proved that strong PSC recordings are ensured and the influence of the electrical noise is minimized. An electrometer of very high sensitivity (Keithley, model 6517) was used to record the electric signal, the data were recorded in real time and they were stored on a hard disk through a GPIB interface. The applied mechanical load was recorded using an analog-to-digital (A/D DAQ) data acquisition device (Keithley model KUSB 3108). The whole experimental set-up was enclosed in a Faraday shield to blockade any external electrical noise. To detect the acoustic emissions, two R15a sensors were coupled (using proper silicone) one at the mid-span of the front surface and one at the right side of the specimen (Fig.1). The 2-channel PCI-2 AE acquisition system by the Physical Acoustics Corp was used, with two preamplifiers, the gain of which was 40dB. The signals were band-pass filtered between 20-400 KHz. The threshold value was set at 40 dB after a series of pencil lead breaks (5mm, HB).

3. Results and discussion 3.1. Mechanical behaviour

All specimens failed at a load-level within the 3.0-3.5 kN range, independently of the loading rate used. The average fracture load for each group of specimens is shown in Table 1. It is clear that the range of the loading rates used for the five groups of specimens is not large enough to affect the fracture load. The fracture plane of all specimens was almost normal to the axis of the specimens (Fig.1c). Slight inclinations observed are well attributed to small asymme tries of the set-up. Since the deviations are within acceptable limits the data of all specimens were taken into account.