PSI - Issue 64

Margherita Autiero et al. / Procedia Structural Integrity 64 (2024) 1798–1805 Author name / Structural Integrity Procedia 00 (2019) 000–000

1804

7

section has reached a temperature equal to 600°C more or less and so the steel has lost about half of its strength. The load decreases until 7.8 minutes with an inward deflection for a tension load, with a little catenary effect for the beam. By assessing that the global behaviour of the structure without one element is not changed, the load and the displacements are the same in shape and entities until the last step of the previous analysis. For these reasons, another structural model was built by removing both the 2 nd and 3 rd horizontal beams of the central shoulder and applying the internal forces as times change, that the previous thermos-mechanical analysis had provided. This procedure was continued until to reach a structural model without the beam at the 2 nd , 3 rd , and 4 th load levels and the truss’ diagonal. The following figure shows the results of this analysis without these elements, in particular Fig. 4b shows the deformed shapes of the structure at last step 742.5 seconds (i.e., 12.4 minutes). From the deformed shape it is possible to see that at that time the critical element was the transversal beam of the central shoulder at the 5 th load level, which is a coupled C section 150x50x15mm and 2mm thick-ness, which shows a buckling shape. Fig.4c shows the comparison between the axial force provided by SAFIR in the beam element and the buckling resistances calculated with both capacity methods. The first thing that it is possible to appreciate is the fact that the last step of the structural analysis provided by SAFIR time increased from 10.2 to 12.4 minutes, moreover the comparison confirms the critical element is the selected transversal beam of the 5 th level, which has continued its heating corresponding to an increase of the horizontal displacements, and the axial load in compression, until 9 minutes when the load starts to decrease. The beam starts to deflect inward, at this point the section has reached a temperature equal to 600°C and so the steel has lost about half of its strength. The load decreases until 12.4 minutes with an inward deflection for a tension load, with a little catenary effect for the beam. Currently, it is possible to also see that the uprights belonging to the central shoulder without the horizontal beams that have collapsed, show a global buckling. (a) (c) Fig. 4. (a) comparison between stress and resistance (b); deformed shape at 12.4 minutes in the Diamond environment of the structure without the beam at the 2nd, 3rd, and 4th load levels and the truss’ diagonal; (c) comparison between stress and resistance. Finally, the shown proposed procedure has confirmed that to correctly estimate the collapse times and the shape of the global mechanism, it is necessary to go beyond the last step of the structural analysis provided by an implicit analysis, by manually eliminating the elements that collapse. Indeed, since SAFIR implements only implicit analysis, this type of analysis stops when convergence problems are reached that usually correspond to the collapse of the most stressed structural elements. 5. Conclusions and future developments This work presents the study of a typological ARSW, from the definition of the fire modelling to the study of collapse mechanism. A fire model that allows vertical and horizontal propagation, starting from a localized fire, evaluated, and validated against experimental results available in the literature, is proposed. It was found that to obtain different temperature distributions within the structure it was necessary to divide the whole geometry of the fire model into different cells as the load level of the structure, i.e., where the combustion material is placed. In this way, A CFAST model of the ARSW structure characterized by cells communicating among themselves through horizontal openings (ceiling/floor vent) was obtained in the examined case study. Several thermo-mechanical analyses were performed by using the SAFIR software, which allows modeling Class 4 steel elements like beam elements but considering the local instabilities that can occur in these slender sections. These analyses were carried out by considering different fire models for different structural elements (LOCAFI and zone models’ results). The results (b) -200 -150 -100 -50 0 50 100 150 200 0 1 2 3 4 5 6 7 8 9 101112131415 time [min] NEd,fi Nb,Rd,fi - Actual EC Nb,Rd,fi - New EC N Rd,fi , N Ed,fi [KN] t collapse = 12.4 min -200 -150 -100 -50 0 50 100 150 200 0 1 2 3 4 5 6 7 8 9 101112131415 time [min] NEd,fi Nb,Rd,fi - Actual EC Nb,Rd,fi - New EC N Rd,fi , N Ed,fi [KN] t collapse = 7.8 min