PSI - Issue 84

Vincenzo Gattulli et al. / Procedia Structural Integrity 84 (2026) 41–48

47

operational conditions, making it particularly suitable for underground environments such as LNGS, where controlled excitation is not feasible. Applied to COSINUS acceleration data, SSI identifies multiple vibration modes, including roto-translational motion in the principal directions and rotational behavior, providing a detailed characterization of the structural dynamic response. As shown in Fig. 4, the stability diagram highlights consistent modal frequencies across model orders, confirming the robustness of the identified parameters. These frequencies, reported in Table 2, are used as experimental references for comparison with the numerical model.

Fig. 4. SSI stability diagram showing consistent modal frequencies across model orders, COSINUS experiment.

Table 2. Modal frequencies identified from SSI analysis for different sensor layouts and processing (SSI reference-based)

Layout 1 (SSI_REF_1)

Layout 2 (SSI_REF_2)

Mode 1: 4.93 Hz Mode 2: 6.58 Hz Mode 3: 8.98 Hz

Mode 1: 5.05 Hz Mode 2: 6.71 Hz Mode 3: 8.69 Hz

Order: 63

Order: 75

5.3. Finite Element Modelling, Numerical Modal Analysis, and Model Validation and Interpretation To support the interpretation of the experimentally identified modal properties, a finite element (FE) model of the structural system associated with the LUNA experimental area within Hall B is developed as a numerical reference for interpreting the experimental results. The model is implemented using Midas Gen software and includes the main structural components of the LUNA experimental area The modal properties of the structure are obtained by solving the eigenvalue problem in Eq. (2): (K − 2 M) = 0 (2) where K and M represent the stiffness and mass matrices, respectively. The first natural frequencies obtained from the numerical model are f 1 =12.11 Hz, f 2 =12.11 Hz, and f 3 =20.95 Hz. These values represent the expected dynamic behaviour of the numerical model and are used as a reference for a preliminary comparison with experimentally identified modal properties. A subsequent model updating procedure will be applied in future work to reduce the differences between experimental and numerical frequencies. A comparison between experimental measurements and numerical predictions is performed by considering modal frequencies obtained from SSI/PSD analysis and finite element modelling. The difference between experimental and numerical frequencies is expressed as reported in Eq. (3): (%)= | − | × 100 (3) The values of Δ f provide an indication of the difference between experimental observations and numerical predictions. In the present study, discrepancies are expected because the experimental data refer to the COSINUS setup, whereas the numerical model represents a different structural configuration associated with the LUNA experimental area. Therefore, the comparison should be interpreted as a preliminary and qualitative assessment rather than a direct validation. It is worth noting that discrepancies between experimental and numerical modal frequencies may arise from modelling simplifications, uncertainty in boundary conditions, and the influence of non-structural components that are difficult to represent in the numerical model. In this context, the comparison between measured and simulated