| During the long service period,a structural building is susceptible to local damage induced by some extreme events,for example,explosion,impact,fire and so on.If the structure lacks sufficient robustness,the local damage would spread widely,finally leading to a complete collapse or a large-scale failure disproportionate to the initial local damage,which is defined as the progressive collapse of structural buildings.Despite the low possibility of such events,once happened,it would bring about heavy personnel casualties,substantial loss of property and very negative social influence.With the occurrence of three well-known collapse events,including the one happened to Ronan Point apartment in London in 1968,the one happened to Alfred P.Murrah office building in Oklahoma in 1995 and the one happened to World Trade Center in New York in 2001,scholars worldwide gradually realized the significance of robustness of structural buildings against progressive collapse.Consequencely,more and more studies were devoted into this field.However,it should be noted that,these works,though provided valuable outcomes,are found to focus more on beam-column connections and two-dimensional(2D)plain structures.Relative studies on three-dimensional(3D)structures are fairly limited,especially high-quality experimental tests on 3D composite floor systems.This means the impact of 3D effect on the structural performances of buildings against progressive collapse could not be well investigated.The main work of the current thesis includes:(1)Two large-scale two-bay by two-bay 3D steel frames were performed under central column removal scenarios when subjected to concentrated and uniformly distributed loadings.The latter was approximately simulated by a specially designed 12-point loading system.Based on the test results,the influence of loading methods on the structural responses of 3D steel frames was clarified.Meanwhile,special attention was paid to the contributions of different load-resisting mechanisms,including flexural and catenary action,in resisting progressive collapse.(2)A 3D steel frame-composite floor specimen having the same layout and joints was tested under central column removal scenarios.The load-displacement responses,failure modes as well as stress development among structural components were obtained and discussed in detail.The formation and development of different load-resisting mechanisms as well as the vertical load redistribution after the failure of the central column were also compared and discussed in great detail.(3)Reduced finite models of the collapse tests of 3D steel frame and steel frame-composite floor system were performed by using ABAQUS/Static,in which the steel beams,columns and loading beams were modelled by beam elements,while connections were simulated by several nonlinear discrete spring elements,which were identified based on component-based models.The composite slab was simulated by composite shell elements,which were partitioned into several strong and weak strips in the decking rib directions to consider the nonisotropic responses in parallel and perpendicular directions of the decking ribs.Besides,the composite effect between the steel frame and slab was introduced by defining a series of "weld" connectors at the locations of shear stubs in the real test.The reduced finite model was firstly verified by comparing against the collapse tests of 3D steel frame and composite floor specimen in this thesis under central column removal scenarios,and then against previous tests conducted by the research group under external column and corner column removal scenarios.The verified finite model was used for an extended parametric study,including steel decking thickness,concrete slab thickness,reinforcement spacing along the primary and secondary beams,aspect ratio and boundary condition.(4)Based on the "strip model" of composite slab and energy conservation of isolated structure system,a simplified approach was proposed in this study for collapse assessment of 3D composite floor systems after the failure of an corner column.The main advantages of the newly proposed approach remain that,it takes full consideration of the unceasingly changing rotation center of beam-column connections as well as the stress concentration of slab around them.The accuracy of the proposed approach was firstly validated against the previous collapse test of 3D composite floor system under corner column removal scenario,and then against corresponding numerical results presented in this thesis.(5)Taking Q345 structural steel as the test material,a series of fracture tests covering a variety of strss states were conducted quasi-statically,including tensile tests of smooth and notched round bars,flat grooved plates,flat pure shear plates,flat tension-shear plates and flat plates with a hole.Afterwards,corresponding finite element models were performed with the help of an enhanced Mises yield criterion,based on which the fracture parameters(fracture strain εf-p,stress triaxiality η and Lode parameter Lp)of the tested specimens were successfully extracted.Seven popular fracture models were selected to reconstruct the fracture locus of Q345 structural steel,and their accuracy was compared and discussed in detail in the 3D space of(η,Lp,εf-p).Finally,a special steel plate with three holes,which was designed as a reduced version of bolted end plate connection,was used to further check the predictive capacity of these fracture models at structural level. |