PSI - Issue 78

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

1812

tightly coupled to both the mechanical performance limitations of the cryostat and the complex logistics of the detector installation. A 3D model of this structure - without the view of the lateral and top panels -is illustrated in Fig. 3b, and it shows a specific installation phase in which the vessel dome, resting on temporary supporting beams bolted to the cryostat flange, has the TPC suspended beneath it.

5. FEA: Modeling approach and result summary

This section focuses on detailed aspects of the cryostat and its ancillary structures, modeled through dedicated Finite Element Models (FEMs) to accurately capture boundary conditions, installation phases, and site constraints. These models were continuously validated against the maximum stresses approved in the cryostat executive design. The Finite Element Analyses (FEA) were performed using SAP2000 and MIDAS FEA NX. 5.1. Cryostat beams -tertiary membrane connection e ff ect In the design stage, welded connections were initially planned - and numerically assumed - between the IPE600V beam flanges and the tertiary membrane. However, as detailed in Section 2.1, clamped connections were adopted during installation. To realistically capture geometry and local deformations from these clamps, shell elements were used for the beams and tertiary plate, while the clamps (present only in the as-built model) were modeled with solid (brick) elements. In the welded model (WM), the inner beam flanges and tertiary membrane were connected via “welding” con tacts (perfect adhesion), whereas Clamped Model (CM) employed “general contact” definitions for all interacting parts (clamps–tertiary membrane, clamps–beam flanges, tertiary membrane–metal sheet, metal sheet–beam flanges), including normal and tangential sti ff ness, friction coe ffi cient, and shear / normal breaking strengths for clamp bolts. See a FEM detail in Fig.3 c. Due to this contact formulation, nonlinear sequential analyses (dead and assembly loads followed by operational and pressure loads) were performed for both models to allow direct comparison. Results in terms of global displacements show that the CM exhibits maximum displacements by over 70% higher than WM. This is primarily due to the limited capacity of the clamps to restrain rotations of the tertiary plate relative to the IPE600V beams. Consequently, the lack of continuous restraint also allows the IPE600V beams to deform more under the thrust actions. In contrast, in WM, the beams and tertiary plate mutually constrain each other, resulting in smaller displacements. Nonetheless, the maximum displacement observed in CM still meets the deformability requirements set by NTC18 (2018), so the design change was authorized to be implemented. 5.2. False floor Several FEMs were developed to investigate this critical composite structure. Initially, analytical calculations were complemented by a dedicated model of a portion of the structure, employing solid brick elements and general sliding contacts. The objective was to ensure adequate sti ff ness and strength to minimize stress concentrations on the finish ing SS plate, while e ffi ciently transferring loads through the steel feet to the underlying insulation system. This was achieved by optimizing the thickness of the SS plate, the type of grating (aimed at drastically reducing the SS thick ness), the beam sections, and the number of support feet. This was a lengthy iterative process, the details of which are not reported here for the sake of brevity; a comprehensive description of the final structure is instead provided in § 4. Please refer to Fig. 3d for the buckling analyses, which illustrate the transition from Linear Buckling Analyses (LBA) performed on both 1D and 3D elements to more advanced evaluations including Material Nonlinear Analysis with Imperfections (MNIA), Geometrically Nonlinear Analysis with Imperfections (GNIA), and Geometrically and Mate rially Nonlinear Analysis with Imperfections (GMNIA). These progressively refined analyses highlight the drastically reduced capacity of the elements compared to the classic Euler closed-form solution. The same FEM model was further refined by explicitly incorporating all the insulation layers beneath the footprints of the support feet to evaluate their deformability under concentrated loads. The displacement induced to the primary membrane by the uniform LAr under operational conditions was also assessed and compared to the most critical case, namely the fully assembled TPC enclosed within the SS vessel (total load of approximately 30 tons) resting on its eight support feet (each with a 250 mm × 200 mm footprint) on the false floor. The resulting displacement fields at the SS membrane from these FEA are shown in Fig. 3e.