PSI - Issue 64

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

787

4

being S i,k the i th area defined by the k th IMF. Repeating the procedure for all the extracted IMFs, the modal damping ratios identification is achieved. Finally, the mode shapes are evaluated based on the phase shifts of the various synchronous sensor recordings. Specifically considering a network of Q sensors, assuming t p as the time instant corresponding to a local peak value of the k th IMF, the k th mode shape vector can be assembled as follows:

T

(1) v t

( ) q v t

( ) Q v t

( ) q v t

1 /max q Q

k 

(5)

k

p

k

p

k

p

k

p

The procedure is therefore repeated for each set of IMFs corresponding to the same mode and recorded at different locations for the estimation of all the K mode shape vectors. 4. Field tests and discussion of results An extensive monitoring campaign has been performed on the above-mentioned family of road overpasses, including static and dynamic load tests as well as non-destructive tests and concrete coring for material characterization. Specifically, for the aim of this study, the attention is focused on dynamic load tests. In this regard, free vibration tests were performed on each overpass for the modal characterization. The selected excitation source consisted of a heavy truck with a gross weight of 400 kN crossing the overpass at an average speed of 50 km/h and free vibrations were recorded after its transit. To this aim, a network of 8 accelerometric sensors (S1 – S8) with a sensitivity of 1000 V/g was adopted: 6 sensors were rigidly anchored to the deck at the central span in a symmetric fashion and the remaining ones were mounted favoring the east lane, one for each span as sketched in Fig. 2. This sensor layout was designed and implemented by taking into consideration the expected modal shapes of the Niagara type Gerber static bridge scheme, which comprise vertical vibrations in both the central (suspended) span and in the lateral (cantilever) spans. In particular, sensors labeled S2, S7, S3 and S8 were placed besides the half-joint to capture the influence of this constraint between the central and lateral spans, whereas sensors S1 and S6 were useful to investigate the mid-span vertical vibrations in the two outermost lanes. Sensors placed along two opposite outermost lanes of the bridge deck were aimed to capture torsional modal shapes and represents a typical scheme adopted in other dynamic field tests of bridge decks (De Domenico et al. 2022). The signals were recorded with a sampling frequency of 1 kHz.

Fig. 2. Layout of the sensors (S1-S8) on the overpass deck during the dynamic load tests (measures in [m]).

The acquired time series were pre-processed via Butterworth filter considering as cut-off frequencies 0.5 Hz and 15 Hz, respectively. Adopting the modal identification procedure described in Section 3, it was possible to identify the modal parameters of the investigated overpasses for the first three modes. Further, the estimated frequencies and mode shapes have been compared with the previsions of a FE model made in SAP2000 (Computers and Structures Inc. 2016). Specifically, girders and transverse diaphragm were modeled via beam elements (six degrees of freedom per node) whereas 2D shell elements were adopted for the RC slab and the connection between elements was achieved via rigid links and joint offset between the centroid of the girder and the mid-plane of the slab. The material properties were set based on the information obtained from the material characterization tests. The selected boundary conditions were clamping restraints at the abutment to better simulate