PSI - Issue 78

Michele Angiolilli et al. / Procedia Structural Integrity 78 (2026) 1807–1814

1813

Fig. 3. a) False floor components; b) Temporary metallic framework supporting the clean room showing the phase in which the dome vessel with the attached TPC below is sitting on a support system; c) Detailed FEM of the clamped solution at the cryostat beam–tertiary membrane interface; d) Buckling FEA of a single foot on the false floor; e) Total displacements under LAr pressure (left) Vs detector assembly load case (right).

Then, a further global FEM was developed to check the floor overall behavior. Its design included the use of: 1D beams for steel footings and IPE100 beams; 2D isotropic shells for the SS plates; and 2D equivalent orthotropic shells for the grating. This FEM, illustrated in Fig. 4a, was first investigated separately through linear and nonlinear quasi static analyses. See results in terms of Von Mises stress on the SS plate in Fig. 4b. Then, it was introduced inside the cryostat (see Fig. 4c) to investigate the pounding forces generated on the cryostat insulation system during nonlinear time history analyses, the detailed aspects of which will be addressed in a dedicated study. A result of the pounding e ff ect from a detailed solid FEM using the estimated pounding force to simulate the lateral device is shown in Fig. 4d.

5.3. Clean room metallic structure

The design of the Clean room metallic exoskeleton required an accurate modelling of the cryostat to determine the stresses transmitted to it. In particular, the cryostat beams were modelled using 1D elements perfectly bonded to each other, with rigid links introduced to preserve the true geometric spacing and ensure that the e ff ective distance between the centroids of the I-beams matches the actual structural configuration. The tertiary membrane was represented by 2D shell thin elements, with appropriate and carefully matched mesh compatibility between the ribs and vertical panels. To simulate the clamped connection between these components, which was fundamental to avoid altering the seismic behavior of the overall structure (see § 5.1), compression-only links were implemented. All elements of the metallic exoskeleton were modeled as 1D beam elements with appropriate rigid links. It is also important to highlight that the crane loads were simulated through influence lines. FEA were crucial for optimizing the structural system, limiting deformations through a network of bracing and horizontal connecting members that ensured a uniform distribution of the significant stresses arising from the crane operational envelope, seismic actions, and combined floor loads, including accidental and dead loads. A flexural moment distribution corresponding to this load envelope is illustrated