PSI - Issue 64

Dario De Domenico et al. / Procedia Structural Integrity 64 (2024) 784–790 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

788

5

the behavior of Mesnager hinges, pinned restraints at the intermediate piers and moment release at the half-joint to simulate the Gerber scheme. For the sake of brevity, in this section, preliminary results for some of the overpasses with the most recurrent configuration (deck width of 10 m with 5 girders) are discussed. Specifically, Table 1 shows a comparison between numerical and estimated modal frequencies for some of the investigated overpasses. It is noted that negligible discrepancies are present among the experimental frequencies of each structure as well as a good agreement with values estimated via the FE model with relative errors lower than 10%. Similarly, Table 2 summarizes the estimated modal damping ratios of the deck, showing a reasonable consistency among the estimates of each road overpass for the first modes. Further, the identified damping ratios are in the order of magnitude of 1 – 2%, which well agrees with typical values in serviceability conditions observed in PC structures with similar span lengths.

Table 1. Comparison between numerical and identified modal frequencies for the investigated overpasses. Modal frequency [Hz] Mode number FE model Overpass A Overpass B Overpass C Overpass D 1 2.66 2.64 2.83 2.84 2.79 2 2.85 3.05 3.41 3.11 3.12 3 4.89 4.56 4.83 4.87 4.77

Table 2. Identified modal frequencies for the investigated overpasses. Modal damping ratio [%] Mode number Overpass A Overpass B Overpass C

Overpass D

1 2 3

0.90 1.88 2.40

0.68 1.31 1.63

0.57 0.70 1.09

0.70 0.68 0.95

Finally, Fig. 3 shows the identified mode shapes and their numerical counterparts of a representative road overpass. As can be observed, the first two mode shapes are representative of the vertical flexural and torsional behavior of the central suspended span, respectively, whereas the third one reflects the flexural behavior of the cantilever spans with the rigid rotation of the central one influenced by the displacements at the half-joints. A general good agreement is achieved between numerical and experimental counterparts as shown in Fig. 3. A quantitative assessment of the degree of similarity between the two types of mode shape estimations has been also carried out via the Modal Assurance Criterion (MAC), which resulted in values over 98% for all the three modes. It is worth noting that “ minor ” damages (that are not located at supports or joints) would tend to manifest deviations from the ideal numerical model at higher order identified frequencies and mode shapes, not just the first three modal shapes. Consequently, dynamic measurements along, in this case, are not sufficient to conclude that the bridge deck is not undergoing any damage. Based on the above critical remark, in addition to dynamic field tests, also static load tests were performed on the investigated road overpass decks under traffic load models compliant with load model 1 of the current Eurocode 1 - Part 2, and experimental results in terms of deflections were compared with theoretical estimates obtained by FE models equipped with ideal material properties. In all cases, the experimental deflections were consistent with numerical estimates and this result suggests that the stiffness of the bridge deck is currently not far from that evaluated in the undamaged scenario. In other words, the bridge deck remains uncracked under the maximum serviceability load prescribed by the Eurocode 1. The critical interpretation of the dynamic test results along with the static load tests has made it possible to conclude that no significant damage process is taking place in the analyzed overpass deck under serviceability conditions.