PSI - Issue 64

Emanuele Gandelli et al. / Procedia Structural Integrity 64 (2024) 685–692 Emanuele Gandelli / Structural Integrity Procedia 00 (2019) 000 – 000

689

5

bridges, since FDD could facilitate the restriction of the portion of a bridge that requires visual inspection by engineers to assess the magnitude of damage.

Fig. 3. Dynamic characterizations of beam- 1: (a) accelerometers’ longitudinal arrangement and detected cracks’ pattern; (b) shear cracks at Fv=800 kN detected through the DIC system.

Fig. 4. Results of ambient vibration on beam-1 processed by FDD technique: (a) experimental mode shape in virgin condition (f 1 =31.25 Hz); (b) experimental mode shape in the damaged condition (f 1,D =29.00 Hz).

4. Detection of small flexural cracks This section presents the outcomes of the dynamic tests performed on beam-2 both in the original undamaged state and in the presence of small flexural cracks (damaged state). As in tests on beam-1, the distance between supports was 6.7 m while a vertical load (F v ) was applied at 2.2 m from one support (Fig. 5-a). A peak force of F v =420 kN, corresponding to a bending moment of 595 kNm, was reached. Right after, the prestress load (N P ) was reduced from the original design value (i.e., 1130 kN) to 678 kN (i.e., -40%). This change of boundary conditions caused sub vertical flexural cracks below the load point (see Fig. 5-b obtained through DIC technique). Then, the vertical load was removed and the prestress level was restored to its original value (i.e., 1130 kN). Ambient vibrations were recorded (sampling frequency of 600 Hz) for both undamaged and damaged conditions using five piezoelectric accelerometers (Wilcoxon 731A model) distributed according to the sensors’ layout shown in Fig. 5-a. V ibrations’ records were processed through the FDD technique. The fundamental frequency detected in virgin condition (f 1 = 32.01 Hz) was affected by some uncertainties. Indeed, the beam was equipped with a post-tensioning system that included a hydraulic jack (175 kg), a load cell (30 kg) and strands (105 kg) whose actual contributions to the beam’s participating mass were unclear. As a result, and the modal analysis could only outline a potential frequency range between 30.90 Hz (taking into account 100% of the additional masses) and 32.14 Hz (considering only the mass of the beam). However, there was no ambiguity regarding the mode shape (shown in Fig. 6-b) identified using FDD which corresponded exactly to the numerical calculation performed with Telaio2D software (Gelfi, 2008). In the damaged configuration, besides the flexural cracks, other “ anomal ies” were introduced by the vertical cuts (about 30 cm long and 2.5 cm deep) near the right support (see Fig. 6-c). These cuts were made as part of a different