*3.1. Preliminary Investigations for Concrete Slabs*

The results for the concrete slabs were first discussed as an initial step towards a more detailed analysis of the floor. Figure 11 shows the experimental GPR B-scans for the concrete slabs with different thicknesses of the air gap. The scan length was approximately 0.4 m. The depth axis was determined with the assumption that the electromagnetic wave velocity was equal to 15 cm/ns (based on the "depth to known reflector method" [45]). For the air gap thickness equal to 1 mm (Figure 11a), the half hyperbolas were clearly detected at the depth of 0.1 (orange arrows) and 0.2 m (green arrows), representing the reflections from the bottom faces of the upper and the lower slabs, respectively. This observation indicated the proper assumption of the electromagnetic wave velocity in the analyzed elements. The bottom faces of both slabs were visible as the lines with an intensity different from the adjacent part of the B-scan. The position of the reflection from the upper slab did not change, whereas the lower slab appeared deeper when increasing the air gap thickness (compare Figure 11a–d). For the air gaps with the thicknesses of 4, 20 and 50 mm, the position of the bottom plate was detected at about 20.5, 22 and 24 cm depths, respectively. This change showed the increase in the distance between both slabs and indicated a high compatibility of the obtained numerical results with the real thicknesses of the gaps. The identified gaps thicknesses were slightly different from the actual ones because the depth axis in Figure 11 was calculated for the velocity of the electromagnetic wave in concrete (15 cm/ns). In addition, the line at the depth of 0.1 m transformed into two separate lines, which indicated the opening of the air gap. It should also be mentioned that the line representing the bottom face of the lower slab became less pronounced for the greater distances between the slabs. This remark allowed concluding that the larger the air gap was, the harder it was to detect elements below it. Moreover, other half hyperbolas (denoted by blue arrows) were present at the depth of about 0.05 m, indicating a technological break during the concreting of the upper slab (visible in all B-scans, see Figure 11a–d). The upper slab was prepared in two stages unlike the lower one, in which no additional perturbations could be seen. A full hyperbola was observed in the center of each B-scan at the depth of about 0.05 m (marked by the red arrow), revealing the presence of the table tennis ball, which was placed in the upper slab between both stages of concreting.

**Figure 11.** Experimental GPR B-scans for concrete slabs with air gap of thicknesses (**a**) 1 mm (trace D-1); (**b**) 4 mm (trace D-2); (**c**) 20 mm (trace D-3); and (**d**) 50 mm (trace D-4).

The numerical GPR B-scans corresponding with the above-described experimental results are presented in Figure 12. The half hyperbolas marked by the orange and green arrows clearly show the bottom faces of the upper and the lower slabs, respectively. The line denoting the bottom face of the lower slab became less noticeable and appeared deeper when increasing the distance between both slabs (cf. Figure 12a–d). For air gaps with thicknesses of 4, 20 and 50 mm, the position of the bottom plate is detected at about 19, 21 and 22 cm depths, respectively. The opening of the air gap was observed as the shift of the reflection from the bottom face of the lower slab occurred and also by the gradual separation of the single line into two lines. The hyperbola revealing the presence of the table tennis ball (red arrow) was clearly visible at the depth of 0.05 m. However, the technological break was not detectable (it was not modelled). The high agreement of the experimental and numerical results allowed concluding that the modelling of electromagnetic wave propagation was correct. The possibility of detecting air gaps and concentrated inclusions was confirmed.

**Figure 12.** Numerical GPR B-scans for concrete slabs with air gap of thicknesses (**a**) 1 mm (model #2.1); (**b**) 4 mm (model #2.2); (**c**) 20 mm (model #2.3); and (**d**) 50 mm (model #2.4).

The UT B-scans for the analyzed slabs are presented in Figure 13. The ultrasonic pressure wave velocity determined before the main tests was equal to 2055 m/s. The scan length was reduced to 0.3 m due to the dimensions of the used UT antenna. There were no noticeable differences between all scans (Figure 13a–d), thus it could be concluded that the thickness of the air gap did not affect the results obtained. The lines at the depth of about 0.1 m representing the reflections from the bottom face of the upper slab were seen in all B-scans. Moreover, these lines were repeated regularly along the depth axis with the step of about 0.1 m (the slab thickness). Ultrasonic waves did not penetrate into the lower slab, they were fully reflected from the bottom face of the upper slab. The deeper lines were only the multiple reflections from the faces of the upper slab. The intensity of these lines decreased with the depth because of the damping of the ultrasonic waves. It is worth noting that the deeper lines appeared only near the edges of the slab. The presence of the boundaries strengthened the reflections, thus the signals were damped slower at the sides compared with the center of the slab. An important observation is that the table tennis ball was not observed at all. The reason might be the limitation of the used UT antenna. The Pundit PL-200PE instruction states that inclusions with a diameter of at least 30 mm should be detected, however, this condition deals with cylindrical elements. Being a relatively small spherical (concentrated) inclusion, the ball was not possible to be detected, despite the fact that its diameter was greater than minimum. To sum up, the air voids could be successfully detected using the UT technique, but without estimating the thickness. On the other hand, the content of the element below the air gap could not be imaged because the ultrasonic waves were entirely reflected from the air gap. Additionally, small concentrated inclusions could not be detected.

**Figure 13.** Ultrasonic B-scans for concrete slabs with air gap of thicknesses (**a**) 1 mm (trace D-1); (**b**) 4 mm (trace D-2); (**c**) 20 mm (trace D-3); and (**d**) 50 mm (trace D-4).
