PSI - Issue 5

Rachel Martini et al. / Procedia Structural Integrity 5 (2017) 1108–1115 Martini et al/ Structural Integrity Procedia 00 (2017) 000 – 000 These equations represent the velocities of both waves in homogeneous, elastic and semi-infinite media. These equations are also applied to a stone masonry, as there are no other analytical relations available. Although not considered a homogeneous material, this value will represent the general mechanical properties of masonry (Luis Miranda, Lorenzo Cantini, João Guedes 2016). For the direct test it is assumed that the propagation of the wave ray is along the thickness of the wall (55 cm). It is then considered a representative P-wave velocity characteristic of the two sheets of masonry. While for the ISIM, the propagation of the wave is assumed to follow wall surface. Thus, the distances used for the calculation of velocities were 20 cm, 55 cm and 90 cm, for the 1st, 2nd and 3rd accelerometer, respectively. 1111 4

(b)

(a)

Fig. 1: (a) GPR and (b) sonic tests.

The treatment of the results obtained by the indirect test was performed using linear regression models fitted to the distance vs. time graphs, whose slope is assumed to represent the wave velocity. This methodology, used in seismic refraction exploration, should be used with caution due to possible propagation wave ray paths distinct from the ones assumed liner along the surface. Then, it is often ambiguous the type of the first arrival waves.

3. Discussion

3.1. GPR

Tareco et al., (2009) carried out experiments to characterize the internal structure of a damaged adobe wall having used this research methodology in masonry with high resolution GPR (1.6 GHz). Figueiredo et al., (2013) used the GPR to analyze the damage caused by earthquakes in buildings with adobe masonry. The use of other techniques in line with the GPR method is widely used in order to address limitations of the GPR. An example of this is the use of GPR with sonic methods to analyze interventions made from damaged walls (Anzani et al., 2006). The GPR tests comprises 462 radargrams with the 1.6 GHz antenna (MALA) and 126 radargrams with 900 MHz antenna (GSSI). All these radargrams were processed using the software ReflexWin, according to the following sequence of operations: set time zero, dewow filtering, background removal and Butterworth bandpass filtering. Fig. 2 (a) shows a processed radargram from the stone masonry wall PP1, relative to a vertical parallel linear radargrams, spaced 5cm and performed with GPR/GSSI, 900MHz antenna. The highlighted in blue area indicates a change in material. Since the wall consists of two sheets, this area corresponds to the meeting of the two sheets, which interface may contain air, a greater amount of mortar and change material (stone, sand, stone), which introduces disturbances in wave propagation and causes strong reflections. However, the size of this area is larger than the actual thickness between two sheets, because the radargram includes areas with heterogeneous materials, then with different velocities of wave propagation (material characteristics). The time-depth conversion was performed using the known