PSI - Issue 64

Mercedes Solla et al. / Procedia Structural Integrity 64 (2024) 293–300 M. Solla et al. / Structural Integrity Procedia 00 (2019) 000–000

297

5

To improve real-time visualization of the received signals, the following processing filters were applied on-the-fly while collecting data (although data is stored as raw, no filters): Time-zero correction, gain function (linear and exponential), and background removal. Dewow filtering is not applicable because SFCW are band-limited systems. 4. Results and discussion Figure 5 presents the results produced by the pulsed radar GPR system, while Figure 6 shows the results obtained with the SFCW GPR system. Observing the results from Specimen I, both systems were able to detect all the bars, including the smaller one with a diameter of 8 mm (R6 in Figures 5-a and 6-a). This finding suggests that both GPR systems demonstrated high sensitivity and resolution in detecting individual reinforcement bars, even those with relatively small diameters. In the results considering different cover depths (Specimen II in Figures 5-b and 6-b), both systems successfully detected all the bars. However, it was noted here that the reflection produced by the bar R9 was not completed in the case of the pulse radar system (Figure 5-b). The rough or uneven surface of the specimen most certainly have contributed to interruptions in the movement of the survey wheel, leading to inconsistencies in the recorded distances or positions. This may result in distortions in the recorded data and loss of information, affecting the accuracy of the radar signatures and, subsequently, reliability in detection. Finally, Specimen III (Figures 5-c and 6-c), considering vertical and horizontal overlaps, produced worse results. Overlapping bars can create complex interference patterns in the radar data, especially when the spacing between bars is small, resulting in unique reflection signatures. This effect is observed with both systems for bars R4 to R6, with spacings less than 20 mm. Moreover, detection was also limited when attempting to distinguish closely spaced bars in the vertical direction. In this study, simulating vertical spacings of 4 mm (R13-R14), 7 mm (R7-R8, R11-R12, R15-R16) and 17 mm (R9-R10), the hyperbolic reflections due to the lower bars were not distinguished by neither of the two GPR systems. As additional results, both time domain and frequency domain systems have demonstrated their capabilities to detect damage in concrete. Observing Figure 7, the presence of a superficial crack in Specimen III was identified in the radar data as signal attenuation or loss of strength of the reflected signals (amplitude).

Fig. 5. Results of the pulsed radar GPR, showing the interpretation of the rebars in red color: a) Specimen I, b) Specimen II, and c) Specimen III.