PSI - Issue 44

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

2104

7

Fig. 7. Results of the monitoring system: (a) optimized sensors’ layout; (b) identified resonance frequencies before and after the optimization of the sensors ’ position; (c) trends of internal and external temperatures.

As mentioned above, the OSP analysis was performed using the EI method; an iterative method in which the set of candidate positions is rapidly reduced to the number of available sensors. The mathematical framework of the method can be found in the work of Kammer (1991). EI has proven to be a robust and simple method to apply. Furthermore, unlike genetic algorithms, the EI method does not rely on computationally intensive search techniques. A very useful numerical example for the implementation of the EI algorithm is provided in the work by Jiang et al. (2017). Given the complexity of the structure, the method was applied to the modal displacements of the calibrated model to overcome any spatial aliasing issues. Selecting a set of 4 uniaxial sensors, the optimised positions are shown on the right-hand side of Figure 7a. At the time of the analysis, the 3 sensors on the façade had already been installed and Figure 7b shows the improvement in the effectiveness of the monitoring system in correctly tracking all the 10 preliminarily identified modes. Figure 7c shows the temperature values recorded on the securing steel system and inside the church. Possible correlations between frequency and temperature will be investigated in the future. 5. CONCLUSIONS The design of a dynamic monitoring system requires preliminary detailed dynamic identification tests from which the optimal position of sensors can be studied. In this paper, the preliminary results obtained from the design and installation of a dynamic monitoring system for the church of Santa Maria in Via in Camerino have been shown. Finite element modelling can contribute substantially to support the interpretation of experimental dynamics, and to the optimal sensor placement development, especially in case of complex structures, where uncertainties on the material characteristics and boundary conditions can significantly affect the system dynamics.