PSI - Issue 24

Alessandro Castriota et al. / Procedia Structural Integrity 24 (2019) 279–288 A. Castriota et al. / Structural Integrity Procedia 00 (2019) 000 – 000

281

3

b

a

Fig. 1. Panel reinforced with 5 stringers and 2 ribs: (a) axonometric view of the panel with saw-cut; (b) the saw-cut in the central part

3. Numerical model

Two numerical FEM models were created using the open-source Code_Aster. The models represent in one case the undamaged panel and in the other case the panel with the presence of the cut. The numerical models allowed evaluating the deformation field and calculating the first local instability load of the panel in the two simulated conditions. From the two models, the relative displacement load curves were also calculated to evaluate the stiffness of the component in the considered structural conditions. In both cases, the modeling of the various parts was carried out using 4-node shell elements (DKT), while the riveted joints were made using beam elements (POU_D_E). The bonding of the traverse skin was simulated with a rigid coupling of the corresponding nodes (LIASON_COQUE) belonging to crosspieces and skin. In order to obtain two comparable meshes, the damaged panel, which geometry is complicated by the presence of the cut, was modeled firstly. A sensitivity analysis carried out on a linear elastic analysis allowed determining the best compromise for the mesh generation, which counts 17136 elements. Subsequently, the model of the intact panel was obtained by filling the cut with appropriate elements restoring the original structural integrity of the panel. The stress/strain field of the damaged panel was used to determine the location of the strain gauges applied to the real panel, in order to verify the correct application of loads and constraints during the experimental test. Furthermore, the corresponding nodes on the numerical model were present in correspondence with the strain gauges, in order to carry out the appropriate comparisons. The application of the constraints respected the constraints that would exist on the real component, trying to simulate the continuity of the material that the panel would have in its original structure. In detail, one end was totally fixed, while the axial displacement of the other loading end was the only admitted. On the lateral edges, the movements outside the plane and the rotations were blocked to simulate the presence of the surrounding material (Fig. 2). In the experimental test this function has been achieved using anti-buckling guides. The load was applied using a load distributed over the length of the cross section, according to the effective and experimental load condition. The deformation field, the load displacement curve and the stiffness of the panel, in the two cases, was evaluated applying a static load equal to P, corresponding to the maximum static service load for the case of damaged component. 3.1. Constraints and load