PSI - Issue 62

Fabio Micozzi et al. / Procedia Structural Integrity 62 (2024) 848–855 Author name / Structural Integrity Procedia 00 (2024) 000 – 000

849

2

1. Introduction Experimental measure of the dynamic response of bridges based on displacement tracking from video processing gained much interest in the past decade, with many applications, mostly involving steel bridges and footbridges, as reported in recent state-of-the-art reviews, e.g., Feng and Feng (2018), Dong and Catbas (2021), Zona (2021), Mousa et al. (2021), Luo et al. (2023). Indeed, there are very interesting features of vision-based dynamic monitoring when bridges are considered, above all the possibility to measure displacements of the deck at various points where conventional displacement transducers could not be applicable due to the absence of stationary points, as well as the possibility that two different methodologies can coexist in the same experimental campaign: a) real time processing of images for displacement extraction of selected points; b) post-processing of the video footage for the extraction of displacements without necessarily pre-defining the specific tracked points. This contribution analyses the performance of an affordable vision-based structural monitoring system applied to an existing medium-span simply supported concrete highway bridge under regular vehicular traffic. The reasons for the interests in such bridge typology is related to the fact that few field studies using vision-based monitoring are available for post-tensioned concrete bridges, the most widespread solution in Italy and other countries. Being the deflections expected in service condition generally lower and their frequency contents generally higher than those in steel bridges and footbridges, concrete bridges are expected to be a more demanding tests of the capacity of vision based monitoring. In addition, it is remarked that many post-tensioned concrete bridges built in the second half of the 20th Century are at the end of their service life; accordingly, a cost-effective solution for monitoring their dynamic response has a strategic importance for the security of our road infrastructures, e.g., Rainieri et al. (2020), Ercoless i et al. (2021), D’Alessandro et al. (2022), D’Angelo et al. (2022), Nicoletti et al. (2023), Kim et al. (2023), Natali et al. (2023). Moreover, a vision-based monitoring system might also integrate additional features such as surveillance, traffic counting, inspection, damage recognition, e.g., Aliansyah et al. (2021), Yu and Nishio (2022), Cardellicchio et al. (2023), further increasing the interest in its potentialities. The experimental results presented in this article were obtained in a field measurement campaign described in Micozzi et al. (2023) where two video cameras with different settings (one also equipped with accelerometers) were simultaneously adopted and compared to contact sensors taken as reference. Attention is here given to the performance of the video camera equipped with accelerometers as compared to the measurements obtained using the contact sensors, integrating the results already presented in Micozzi et al. (2023). 2. Experimental campaign 2.1. Main features of the monitored bridge A three-span concrete bridge located in Central Italy, part of the national highway road network, was made available as testbed structure. Each span (length 32 m) has a single deck made by a concrete slab on six simply supported post-tensioned concrete beams connected transverse beams at the ends and at about 1/3 and 2/3 of the span. A picture taken during the experimental campaign and showing the aerial view of the bridge is reported in Figure 1a. The West first span of this bridge was selected for its reduced distance from the ground, making possible a simple installation of a displacement transducer. The measurements of the bridge response were conducted on July 20th, 2023, while the bridge was normally operational with regular vehicular weekday traffic, mostly cars and a limited number of heavier tracks. 2.2. Vision-based monitoring system The vision-based monitoring system consisted of an industrial computer-vision video camera (Teledyne FLIR BLACKFLY S BFS-U3-23S3M-C) with interchangeable C-mount lens (Kowa LM100JC1MS), installed on a tripod (Manfrotto aluminium video tripod with three-way head), and connected to a laptop through USB3.0. This system is basically a more recent version of the one adopted by Feng and Feng (2021). The video camera was placed under the bridge deck, near the West abutment, as depicted in Figure 1b. Such vantage point was selected as protected from rain and direct sunlight, potentially functional in future permanent monitoring applications. A 100 mm × 100