PSI - Issue 78

Shahin Sayyad et al. / Procedia Structural Integrity 78 (2026) 277–284

278

Keywords: Finite Element Model Calibration; Genetic Algorithm; Particle Swarm Optimization; Masonry Tower Bell; OpenSeesPy

1. Introduction Historical masonry bell towers, which are prone to vibrations from human activities and seismic excitations, undergo changes, transformations, and restorations over time, leaving permanent marks on their structure and frequently altering their local or even global structural behavior. Hence, dynamic identification along with a calibrated numerical model is a major issue as a starting point for seismic vulnerability evaluation and damage prediction of such structures. This study focuses on improving the accuracy of finite element models (FEM) for these masonry structures using advanced updating/calibrating techniques relying on operational modal analysis (OMA). The purpose is to determine the optimal unknown parameters of the model, such as elastic moduli, mass densities , and Poisson’s ratios , to minimize the objective function that quantifies the discrepancy between experimental and numerical modal properties, such as natural frequencies and mode shapes. FEM calibration methods are generally categorized as manual or automated, and further as direct (non-iterative) or indirect (iterative). Manual methods rely on trial and error to calibrate structural parameters and are adopted when the model is simple, with a limited number of candidate parameters (Altunişik et al., 2019) . However, they may not provide a reasonable physical explanation for changes in results, whereas automated methods are recommended for models with many parameters and can help reduce idealization errors. Direct calibration methods, which have become less common, directly update the stiffness and mass matrices of the FE model without iterative procedures, allowing for computational efficiency. Conversely, iterative methods, which are more prevalent in structural engineering, are further divided into deterministic (maximum likelihood) and stochastic approaches. The finite element model calibration problem is converted into an optimization problem using the Iterative Deterministic Maximum Likelihood Methods (DMLM) (Nozari et al., 2017). In the current study, manual FEM calibration of the bell tower of San Giuseppe in Aci Castello is replaced by GA and PSO as one of the most promising DMLM methods in civil engineering applications, to reinforce the accuracy and efficiency. This study also highlights the interchangeability and synergy between Midas FEA NX and OpenSeesPy (Zhu et al., 2018) for FEM calibration purposes. Midas FEA NX, a commercial finite element software used for simulating and analyzing complex structures with detailed geometry, was replaced by OpenSeesPy due to its lack of flexibility and openness, which are required for automation, scripting, and integration with optimization libraries, limitations shared by many commercial software packages. To address this, OpenSeesPy, the Python interface of the OpenSees framework, was adopted. In the following sections, the research methodology is presented, including the case study description, operational modal analysis, finite element modeling, and model calibration using GA and PSO, followed by the results and discussion. 2. Methodology 2.1. Description of the Case Study The case study for this research is the bell tower of the Church of San Giuseppe, located in the southeast of Aci Castello, a municipality in the Metropolitan City of Catania, Sicily, Italy. The town is located along the eastern side of Mount Etna, one of the world’s most active and historically important volcanoes, recognized as a UNESCO World Heritage Site due to its ongoing volcanic activity that has lasted over 2,700 years. Historical records suggest that the church was constructed around 1748 on the si te of an earlier religious structure dedicated to Sant’Agat a. Architectural features of the bell tower indicate that it may predate the existing church and was possibly part of an earlier construction phase. This hypothesis is supported by the apparent lack of structural connection between the tower and the adjacent church, justifying the tower's treatment as an independent structure in this study. Thus, the bell tower was studied as an isolated structure. Fig. 1 illustrates the side view of the church, along with its adjacent bell tower and its ground floor plan. The tower has a total height of 15.4 m (excluding the tiled roof), with a rectangular cross-section