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

298

6

Fig. 6. Results of the SFCW GPR, showing the interpretation of the rebars in red color: a) Specimen I, b) Specimen II, and c) Specimen III.

Fig. 7. SFCW and pulsed radar GPR results showing signal attenuation on rebar R7 due to a superficial crack (highlighted into dotted red lines).

Figure 8 shows preliminary results of using GPR Insights AI-assisted engine trained to automatically detect hyperbolas from the first level of rebars in reinforced concrete bridge decks. Specimen III was selected as the most challenging to test the capabilities of this AI tool which learning model was originally trained with data from these SFCW GPR devices. Therefore, the highest detection rate has been achieved on this type of data so far. As observed, detection is consistent with the interpretational results of our analysis (see interpretation in Figure 6-c).