PSI - Issue 78

Andrea Gennaro et al. / Procedia Structural Integrity 78 (2026) 663–670

664

1. Introduction For several decades, Finite Element Analysis (FEA) has been a powerful tool for simulating structural behavior. However, creating an accurate FE model is not an easy task. Numerous modeling strategies and best practice guidelines are available (Rao 2010), c overing aspects such as the selection of element types, degrees of freedom, and appropriate analysis methods. These strategies depend heavily on the analyst's skill and experience, as well as on the intended application of the model, since static and dynamic analyses often require different FE model configurations to achieve the same level of accuracy. With current advances in numerical modeling and computational capabilities, there is a general expectation that an FE model — consistently based on original technical design drawings, on-site geometric surveys, engineering judgment, and assessment processes — should reliably reproduce both the static and dynamic behaviors of a structure. However, acquired experience shows that developing an FE model involves assum ptions and simplifications that may induce considerable errors (Mottershead et al. 2011). Experimental validation is therefore indispensable. Reference data are traditionally obtained from static or quasi-static measurements based on load tests, or dynamic measurements obtained from Ambient Vibrations Test (AVT) (Cunha et al. 2003; Salvalaggio et al. 2024).. In particular, AVT offers several advantages, including the ability to perform measurements without interrupting the normal operation of the structure and to collect data under real operational conditions (Lorenzoni et al. 2019). Finite Element Model Updating (FEMU), also known as model calibration, is a procedure used to determine uncertain parameters in the initial model based on experimental results, aiming to achieve a more accurate updated model of the structure (Friswell and Mottershead 1995). In most FEMU techniques, the stiffness, mass, and damping distributions of a numerical model chosen as a reference configuration are iteratively updated so that the differences between the measured and analytical values of the modal para meters are minimized (Brownjohn et al. 2001). Several works in the dedicated literature report results where modal identification has proven useful for performing model updating of numerical models of existing bridges (De Domenico et al. 2025; Giglioni et al. 2024; Saler et al. 2020).. In this contribution a FEMU based on AVT of an existing Gerber half-joints bridge is presented. The paper is organized as follows: in chapter 2, the description of the case study is presented, including the modal parameters for the model updating, and the preliminary FE model. In chapter 3, results of AVT and FEMU are discussed. 2. Material and Methods 2.1. Description of the case study The case study (Fig. 1) consists of a PRC Gerber half-joints bridge built around 1970 with a Niagara scheme. The bridge is located in the southern Italian region. As the initial project designs were not available, the dimensions of the structural components were determined through an on-site geometric survey. The bridge is composed of a central suspended span (drop-in) and two cantilevered side spans, with a total length of 52.80 m and a total width of 6.50 m. The two side spans are supported by the abutments and the intermediate piers, with a cantilever measuring approximately 2.41 m for the left side and 3.27 m for the right side. The central beams, with an I-section measuring 119 cm in height and 40 cm in width, are supported on the two lateral beams, each with a T-section measuring 119 cm in height and 40 cm in width, by means of half joints. The bearings at the Gerber half-joints are composed of lead plates. On one side of the drop-in, fixed bearings are provided; these consist of lead plates penetrated by appropriately shaped steel bars at the beam locations, designed to resist horizontal forces. On the opposite side, slide bearings are present, consisting of lead plates. It should be noted that the construction details are hypothesized according to Del Giudice (1967) due to the unavailability of the original design documents for the bridge. The longitudinal and transverse view are reported in Fig. 2a and Fig. 2b respectively.