PSI - Issue 13

Ekaterina Damaskinskaya et al. / Procedia Structural Integrity 13 (2018) 298–303 Author name / Structural Integrity Procedia 00 (2018) 000 – 000

300

3

cylinders 10 mm in diameter and 20 mm high. The experiments were carried out on Westerly granite samples. Westerly granite consists of 28% of quartz, 33% of plagioglase, 33% of K-feldspar, and 5% of mica (3.5% of biotite, 1.9% of muscovite) [Chayes (1950)] and has an average grain size of 0.075 mm [Stesky (1978)]. The spatial resolution of tomographic images for the samples of this geometry was ~3 mkm. The resolution we achieved was the maximum possible one for the samples of a given size [Tóth and Hudák (2013)] if we take into account the physical principles of tomography and structural features of the X-ray tube and tomographic camera. A series of tests was carried out (11 samples). The tomographic survey of the entire series of samples carried out prior to the mechanical tests showed that the initial samples had no defects of the type of cracks and pores with dimensions larger than 3 mkm. This is an important result, since the defects existing before the start of deformation can lead to a concentration of stresses (i.e., to increased stresses as compared with the average local stresses), and, hence, can predetermine the site where a fault will be formed. To verify this result, further studies were carried out. Several samples from the series were used to make polished sections. The studies were carried out with an SEM Vega 3 Tescan electron microscope and a Nikon Eclipse LV 100 pol polarization microscope. The analysis of backscattered electron image and photographs of thin sections did not reveal defects of the crack type. Quasistatic tests of Westerly granite samples were carried out under the conditions of uniaxial compression at a Shimadzu AGX-Plus electromechanical testing machine. To monitor acoustic emission in real time, an Amsy-5 Vallen System was used. Two broad-band piezoelectric transducers of acoustic emission AE105A with a bandwidth of 450-1150 kHz were attached to the ends of the sample, which allowed a linear location of AE signal sources. During the experiment, a database in which parameters of individual AE signals, i.e., the emission time, coordinate of the source hypocenter and energy, were recorded. The accuracy of determining the coordinates of AE source hypocenters was ~2 mm. A granite sample was deformed in steps. At each stage the load was slowly increased (at a rate of ≈ 5 N/s) to a certain magnitude F and then the sample was held under this load F until the activity of the AE signals fell to zero (Fig. 1a). At the next stage the load F was increased by 0.08 Fmax (Fmax is the breakdown force). The experiment was stopped when an avalanche-like increase in the activity of acoustic emission began, which indicated that the destruction of the sample was approaching (Fig. 1b). As a result, the sample preserved its integrity, and a final tomographic survey could be carried out. After each loading the sample was unloaded and placed into the tomographic chamber to perform tomographic survey. In total, 11 stages of loading-unloading and tomographic surveys were carried out.

a)

b)

Fig. 1. Variations in stress (solid line) and energy of individual AE signals (points) at the first (a) and final (b) stages of sample deformation.

3. Results of experiments and discussion

After each tomographic survey of the sample, a set of slices (graphic images) consisting of ~6500 files was formed (the slice is the section perpendicular to the cylindrical sample axis). The distance between the slices corresponded to the spatial resolution (3 mkm). Gradations of gray color in the images corresponded to different material densities. The defects, such as cracks or pores, were of a black color.