PSI - Issue 84

Prajwal Giri et al. / Procedia Structural Integrity 84 (2026) 1119–1126

1122

The bridge configuration, including the longitudinal layout and cross-sectional geometry, is illustrated in Figures 2(a–c), while the detailed girder profile and tendon arrangement are presented in Figure 2(d). The structure is simply supported single-span bridge with a total length of 20 m, representative of a typical medium-span bridge. Transverse sti ff ness and load distribution are enhanced by four transverse beams, two located 0.55 m from the supports and two positioned at 6.3 m from each support. Additionally, a 3.1 m-long ramp extension is provided on one side of the bridge to facilitate vehicle access and load application during static testing, thereby simulating realistic operational conditions. Each girder is reinforced with five prestressing strands and one post-tensioned tendon, each strand having a cross-sectional area of 1.39 cm². The post-tensioned tendon, composed of 12 strands, is placed approximately 12.5 cm above the bottom axis to counteract tensile stresses under service loads. The prestressing strands are arranged in two layers: one strand at 30 cm from the bottom of the beam and four strands at 6 cm from the bottom, horizontally distributed at distances of 7 cm, 13 cm, 37 cm, and 43 cm from the external edge. 4. Model Identification and Finite Element Modelling To establish a baseline for subsequent monitoring and damage assessment, ambient vibration tests (AVTs) were conducted on the full-scale prestressed concrete bridge deck in its undamaged state. The setup employed ten triaxial DEWESOFT MEMS accelerometers (IOLITE ® 3xMEMS-ACC-INC, outdoor version), each with a noise density of 25 µ/ √ and a dynamic range of ± 2 g . The sensors were symmetrically arranged along the girder, with five sensors on each side (Figure 3). Fig. 3. Longitudinal profile of the deck showing the position of the accelerometers . Measurements were collected under natural excitation on October 17, 2025, with a sampling frequency of 200 Hz and a total duration of 30 min. The recorded responses were processed using the MOVA / MOSS operational modal analysis software (García-Macías and Ubertini, 2020) based on the covariance-driven stochastic subspace identification (CoV-SSI) method, enabling reliable extraction of natural frequencies, mode shapes, and damping ratios while e ff ectively distinguishing physical modes from numerical artifacts. Based on the identified dynamic characteristics, a FEM of the bridge deck was developed in SAP2000 (Figure 4a). The girders and transverse beams were modelled using frame elements with lumped torsional masses at the joints, while the concrete slab was represented by shell elements. Simply supported boundary conditions were assumed, with a pinned support at one end and a roller support at the other, both located 0.30 m from the deck extremities. Material properties were defined according to design specifications, with elastic moduli of 35.54 KN / mm 2 for the girders and 32.59 KN / mm 2 for the slab, a unit weight of 25 KN / m 3 , ’ 0. . w y x displacements at sensor locations. Prestressing and post-tensioning e ff ects were incorporated through equivalent loads. An initial prestressing stress of 1250 MPa (approximately 67.2% of = 1860 MPa) was assumed for both tendon systems. After accounting for friction, anchorage set, relaxation, shrinkage, and creep, e ff ective stresses of 966.85 MPa for the pre-tensioned tendons and 897.08 MPa for the post-tensioned tendons were adopted. These values were subsequently used for damage scenario simulations.

Table 1. Comparison of experimentally identified modal frequencies with FEM predictions before and after calibration. Mode No. Mode type [Hz] Uncalibrated model [ z] . . [%] AC

Calibrated model

.

. [%]

AC

6.6 5

5.95

- 0.

0.98 0.97 0.86

6. 9

- .

0.98 0.97 0.86

/

.5

.977

- .97

.6 9

0.78

8.0

6.98

-5.8

8. 98

. 7