PSI - Issue 64

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

690

6

research project aiming to evaluate a new diagnostic technique to measure variations of the prestress level in bridge girders (Mantelli, 2023). As illustrated in Fig. 6-a, the FDD failed to provide a convincing identification of the fundamental frequency of the damaged beam since the peak (around f 1,D =31.91 Hz) was unexpectedly lower than that of the test frame (at 22.25 Hz) and was associated to irregular surrounding frequencies. However, FDD was again capable to localize the two minor damages that indeed corresponded to deviations in “damaged mode shape” (see Fig. 6-c) with respect to the “undamaged mode shape” (see Fig. 6-b). To overcome the uncertainty related to the fundamental frequency in damaged conditions, a hammer test was performed to introduce higher excitations along the beam. A Fourier spectrum with a clear peak at f 1,D =31.88 Hz was obtained by processing the corresponding acceleration s’ record through the “Fast Fourier Transform” (FFT). It is noteworthy that both the identified "damaged frequencies" (i.e., f 1,D =31.91 Hz through FDD of ambient vibrations and f 1,D =31.88 Hz through FFT of hammer test accelerations) were remarkably close to that of the undamaged beam (i.e., f 1 =32.01 Hz). This could be related to two factors: (1) the flexural cracks affected a small portion of the beam and the high prestress level (1130 kN) restored its original elastic flexural stiffness; (2) all the cuts of the "stress-release" tests (Mantelli, 2023) were located in the web without significantly reducing inertia moment of the beam. The lack of influence of these minor damages on f 1 was also confirmed by additional hammer tests carried out with the purpose of eliminating the uncertainty of the participating mass. This was achieved by re-setting the two supports at beam's ends resulting in a span of 9.70 m (see Fig. 7-a). With this simple variation of set-up, the uncertain additional masses (i.e., hydraulic jack, load cell, and plates) insisted directly on the two supports without contributing to the dynamics of the beam. Nine hammer tests were conducted with N P varying from 0 kN to 1130 kN. In the range of prestress loads from 284 to 1126 kN (i.e., from 25% to 100% N P ), all "damaged fundamental frequencies" (f 1,D ) remained almost unchanged and identical (see Fig. 7-b) to that of the virgin beam (i.e., f 1 =17.40 Hz that agrees with Eq. (1)). Otherwise, for prestress loads below 284 kN (i.e., below 25% of N P – see Fig. 7-c,d) flexural cracks became wider resulting in partialized cross-sections with reduced natural frequencies (e.g., f 1,D =15.70 Hz for N P =0 kN). All Fourier spectra of the nine hammer tests are reported in (Gandelli et al., 2024) with further considerations. However, one can conclude that the FDD technique is not suitable to detect variations, if any, of the beam fundamental frequency potentially associated to reductions of the strands’ prestressing load in the most common range (up to -30%).

Fig. 5 Dynamic characterizations of beam- 2: (a) accelerometers’ longitudinal arrangement and detected cracks’ pattern; (b) flexural cracks at F v =420 kN detected through the DIC system; (c) cuts of the “stress - release” tests.