PSI - Issue 44

Davide Arezzo et al. / Procedia Structural Integrity 44 (2023) 2098–2105 D. Arezzo et al./ Structural Integrity Procedia 00 (2022) 000 – 000

2099

2

1. Introduction In recent years, Vibration-Based Structural Health Monitoring (VB-SHM) systems are becoming increasingly popular, to such an extent that many seismic risk mitigation frameworks are based on dynamic monitoring systems (Costanzo et al 2021). In order to make the monitoring economically feasible, the optimization of the number and position of sensors is critical in the design and implementation of an effective VB-SHM (Li et al 2012). To this purpose, the minimum number of degrees of freedom (DOF) of the dynamic system to be monitored must be identified, in order to obtain sufficient information to describe the dynamic behavior of the structure with an adequate level of accuracy (Meo et al 2005). This type of optimization is called Optimal Sensors Placement (OSP) and one of the first applications, namely the Effective Independence (EI) method, was presented in the work of Kammer (1991). Subsequently, OSP received considerable attention and nowadays several applications can be found in the literature (Heo et al 1997, Guo et al 2004, Flynn et al 2010, Worden et al 2001, Jiang et al 2017, Lenticchia et al 2018, Civiera et al 2021). The OSP procedure requires an in-depth knowledge of the dynamics of the structure to be monitored; generally, dynamic identification tests are performed with a limited number of sensors and the identified modal shapes, especially those at higher frequencies, can be affected by spatial aliasing. Therefore, the availability of a carefully calibrated Finite Element Model (FEM) contributes substantially to the OSP phase. This is especially true for complex structures, where uncertainties in material characteristics and boundary conditions can be significant. This scenario is typical for historical masonry structures. Model Updating (MU) of FEM constitutes essentially in minimizing an error function that measures the difference between the dynamic behavior of the model and that of the real construction. As mentioned, the model of a historical masonry structure can reach a high degree of complexity, with a high number of parameters to be calibrated. In such cases, a family of artificial intelligence algorithms, Swarm Intelligence (SI), can provide substantial support. SI is a family of population-based meta-heuristic algorithms inspired by the collective behavior of insects, such as ants, termites, bees, wasps and other animals capable of performing certain intrinsic social actions (Blum et al 2008). Within SI, one of the most widely applied algorithms is the Particle Swarm Optimization (PSO) algorithm (Kennedy et al 1995), and applications of PSO for MU of civil engineering FEM can be found in many works (Saada et al 2013, Tran Ngoc et al 2018, Cancelli et al 2020, Zacharakis and Giagopoulos 2022). This paper presents some results obtained from the optimization of the monitoring system of the historic Santa Maria in Via church in Camerino, which was severely damaged by the 2016 Central Italy seismic sequence. Firstly, a description of the church and of its dynamics, identified through environmental vibration tests, is provided. Then, a refined FEM of the structure and the results obtained with the PSO-assisted MU procedure are presented. Finally, the performance of the optimized monitoring system is shown. 2. The Santa Maria in Via Church The church of Santa Maria in Via (Figure 1) is located in Camerino, Central Italy, and consists of many architectural elements typical for this type of building: the tiburium, which rises 8 m above the internal structure, the bell tower, with an octagonal plan and a rather slender belfry, and a façade consisting of two levels, a lower one incorporated into the main body of the church, and an upper one rising next to the tiburium to a height of around 23 m above the ground level. In 2016, the Central Italy earthquake severely struck the church of Santa Maria in Via, together with many other churches of the Marche Region (Morici et al 2020, Vitale et al 2022), causing damage that were successively aggravated by the adverse environmental conditions of the winter months of the year 2017. In detail, the church suffered the partial collapse of the bell tower and of the wooden roof, together with part of the fake dome, and many diffused cracks on the interior and exterior walls of the façade (Figure 2a,b). One of the most serious damages relates to the collapse of the rear part of the tiburium (Figure 2a), which collapsed mainly inside the nave. Both the interior and exterior walls of the nave and apse were not affected by significant collapses or cracks, with the exception of the two side altars adjacent to the façade section, which suffered significant damage to the arches and vaulting. The lower level of the façade suffered an out-of-plane rotation, while the upper level suffered a rigid translational movement that was probably responsible for the collapse of part of the tiburium adjacent to the façade.