Zhaohui Huang

The influence of shear connectors on the behaviour of composite steel-framed buildings in fire

Abstract A three-dimensional non-linear procedure is presented for modelling partial shear connection in composite steel-framed buildings in fire. The model has been incorporated within the computer program VULCAN which has been developed to model such structures. The accuracy and reliability of the model are demonstrated by analyses of four previously tested simply-supported composite beams, of which two were tested at ambient temperature, and a major full-scale fire test conducted in the Cardington composite test frame. The model is clearly capable of predicting the response of composite members and structures in fire with acceptable accuracy, although data is needed on the degradation of shear stud connectors at elevated temperatures. It is shown that, in statically determinate components, the local failure of shear studs can very rapidly alter the behaviour at the failure temperature, although whether this is important depends on the loading level. The inclusion of partial interaction is shown to make a small, but potentially significant, amendment to the analytical results in the high temperature range for the three-dimensional full-scale fire test.

Effective stiffness modelling of composite concrete slabs in fire

Abstract In this paper a modified layered slab element is developed based on the layered procedure previously developed for the modelling of composite slabs in fire. In the development reported here, the ribs forming the lower part of any slab cast onto metal decking are included in the slab modelling. The basic idea is to use the nominal thickness of the composite slab as the thickness of the slab element, and to use an effective-stiffness factor to modify the material stiffness matrices of plain concrete in order to take into account the orthotropic properties of the slab. A maximum-strain failure criterion is used in the modelling of concrete. Three standard fire tests on metal-deck composite slabs and one full-scale natural fire test on the Cardington composite building frame are modelled for validation. The validations indicate that the model proposed is clearly capable of predicting the fire resistance of this type of composite slab in fire with reasonable accuracy. It is evident from this study that the influence of the ribs across the bottom of the slabs is significant and should be accounted for. The calculation of effective stiffness factors, which are based on the theory of elastic beam bending, is adequate and efficient, and the maximum-strain failure criterion is simple and suitable for such problems.

Non-linear structural modelling of a fire test subject to high restraint

Abstract The computer code VULCAN has been developed for the three-dimensional structural analysis of composite and steel-framed buildings in fire. In this paper, the main features of the program are outlined, with particular emphasis on the most recent development to the layered procedure for modelling of concrete floor slabs. This development has introduced geometric non-linearity into the modelling of slabs, whose layer structure already allowed temperature distributions and change of material properties through the thickness, as well as modelling the effect of the ribs at the bottom of composite decking slabs. The capabilities of the model are firstly tested at ambient temperature for a uniformly loaded ribbed reinforced concrete slab with simply supported edges, and this is followed by a very detailed modelling of the Cardington restrained beam fire test. In both cases the development of membrane action is demonstrated and the structural behaviour is compared with the geometrically linear case. A number of studies are carried out to demonstrate the influence of the major floor slab details on the behaviour of the structure in fire conditions. These studies provide evidence that when exposed steel temperatures are less than 400°C the concrete slab has little influence, other than to play a part in generating thermal curvature to composite beams. For temperatures higher than about 500°C the effect of the slab progressively becomes much greater, and it is very important to model concrete slabs correctly. The influence of membrane action cannot be ignored, particularly when the fire compartment is subject to high restraint because it is surrounded by cool, stiff structure. At very high temperatures the floor slab becomes the main load-bearing element and the floor loads above the fire compartment are carried by the membrane forces developed in the slab, with tension being carried mainly by the steel anti-cracking mesh or reinforcing bars. However, the effect of the very high in-plane restraint to thermal expansion in the particular Cardington test considered is to enhance the peripheral zone of compressive membrane force and to reduce the extent of the central area of tensile force compared with more usual cases.

Development and Validation of 3D Composite Structural Elements at Elevated Temperatures

A three-dimensional (3D) eight-noded brick element, which is capable of representing the performance of composite structures subjected to 3D stress conditions at ambient and high temperatures, has been developed and incorporated into a finite-element analysis program Vulcan. In the formulation of this element, geometric nonlinearity, material nonlinearity, material degradation, and thermal expansion at elevated temperatures have been taken into account. The von Mises and Drucker-Prager theories were chosen as the 3D failure criteria for steel and concrete, respectively. In particular, a series of postfailure criteria and corresponding 3D constitutive relations for concrete at high temperatures were defined in this study. The accuracy and efficiency of this newly developed structural element was verified against the results from a number of tests on composite structures subjected to 3D stress conditions at both ambient and elevated temperatures. The proposed 3D structural element can be used to model a large range of composite structures in fire, and perform more detailed studies on them.

Nonlinear analysis of orthotropic composite slabs in fire

In this study an orthotropic slab finite element is developed to model orthotropic slabs in fire, using a layered 9-noded isoparametric slab element and a 3-noded beam element. The element is assembled from a solid slab element which represents the continuous upper portion of the profile, and a special beam element which represents the ribbed lower portion. An equivalent width for the cross-section of this beam element is determined according to the dimensions of the solid slab element and the cross-section of the ribbed profile, and the beam shares the nodes of the solid slab element. The temperature within each layer of the slab element can vary between adjacent Gauss integration points so as to reflect temperature variations in the horizontal plane. Several fire tests on composite slabs have been modelled to validate the approach. Cases of orthotropic slabs with wide range of parameters defining the ribbed profile have been studied, which show that the orthotropic slab model is robust and effective in reflecting the influence of the shape of ribs on the thermal and structural performance of the slabs in fire. The study shows the influence of decking shape on the thermal and structural behaviours of orthotropic slabs. A simple evaluation method for profile selection is proposed.

Behaviour of Frame Columns in Localised Fires

A static/dynamic version of the software Vulcan has recently been developed, in which the numerical singularity of a static analysis, induced by a local instability of a structure, for instance the buckling of a column, can be covered by switching to the explicit dynamic procedure. This version of Vulcan allows the post-buckling behaviour of a member to be traced, finding a re-stabilized state if it exists.In this paper a study of the behaviour of steel columns in localised fires is presented. A simplified model is developed, taking into consideration the axial restraint of the column in fire. An axial elastic stiffness is used to represent restraint from the superstructure, and the full post-buckling behaviour is covered. Results obtained from the simplified model have been validated against previous numerical studies. A full-frame model has been created for comparison. The simplified model has been extended to investigate the effect of beam yielding on the columns restraint conditions. This is modelled...

Analysis of Industrial Steel Portal Frames in Fire

This paper presents a robust numerical model for dealing with temporary instabilities which occur in the numerical analysis of steel structures under fire conditions. The model adopts the combined static-dynamic solution procedure to model ‘snap-through’ behaviour of industrial steel portal frame in fire. This new method allows solution procedure automatically switch between static and dynamic approaches, with the objective mainly for overcoming a transitory stage of instability in structural modelling. The current model is computationally very efficient compared to conducting full dynamic analysis of the structures for the whole duration of fire. The method could easily be applied for modelling composite and reinforced concrete buildings under fire conditions. The snap-through instability of the pitched portal frame has been modelled successfully by the new procedure. The method could provide a very useful modelling tool for the perform-based fire safety design of industrial buildings, as a much more rea...

A Simplified Model for Analysis of End-plate Connections Subjected to Fire

In this paper a robust 2-noded connection element has been presented for modelling the bolted end-plate connections between steel beam and column at elevated temperatures. The connection element allows the element nodes to be placed at the reference plane with offset and the non-uniform temperature distributions within the connection. In this model the connection failure due to bending, axial tension, compression and vertical shear are considered. The influence of the axial tensile force of the connected beam on the connection is also taken into account. This model has the advantages of both the previous simple and component-based models. A series of numerical studies was carried on a 2D steel frame under ISO834 and Natural fires. The results indicated that the deflections of beams are significantly affected by using different types of the connections. However, the axial forces of the connected beam are less significantly affected by different types of the connections. Another finding is that the axial tensile force in the beams generated due to catenary action is relative small compared to the forces caused by the thermal shrinkage of the beam during cooling phase of the real fire.

Use of sub-structuring in modelling of composite building response to compartment fires

Publisher Summary The chapter describes an appropriate procedure that employs a sub-structuring technique together with selective node re-numbering, for the numerical modeling of a building response to fires which are restricted to compartments within the building. In this procedure, user-defined cool regions of the building model are sub-structured and condensed into a linear “super-element,” which has nodes connected to the non-linear sub-frames located within and in the immediate vicinity of the fire zone. The procedure is incorporated into the non-linear program Vulcan, which has been developed to model the structural response of loaded structures to fire attack, and two full-scale fire tests are modeled to examine the computational efficiency of the method. The method makes it much easier to investigate the influence of the surrounding cool structure on the behavior of the elements within the fire compartment, and it is investigated in the two full-scale fire tests described in the chapter.

The Influence of Tensile Membrane Action in Concrete Slabs on the Behaviour of Composite Steel-Framed Buildings in Fire

A computer program VULCAN has been progressively developed for some years at the University of Sheffield, with the objective of enabling three-dimensional modelling of the behaviour of composite buildings in fire. In this paper the theoretical basis of the non-linear layered procedure used to model the reinforced concrete floor slabs, which includes both geometric and material non-linearity, is briefly outlined. Several of the full-scale fire tests carried out in 1995-96 on the composite frame at Cardington, representing cases in which different degrees of in-plane restraint are provided by the adjacent structure, are modelled to evaluate the influence of tensile membrane action in the concrete slabs on the structural behaviour in fire. In order to illustrate the influence of membrane action and its relationship with boundary restraint, all cases have been analysed using both geometrically linear and nonlinear slab elements. A series of parametric studies has been carried out as an initial investigation into the characteristics of steel reinforcement which allow this action to take place. It is clear that the influence of tensile membrane action of concrete slabs on the behaviour of such composite structures in fire is very important, and should be accounted for in later additions and amendments to structural fire engineering design codes. Introduction In 1995-96 six large fire tests were carried out on a full-scale composite multi-storey building at the BRE Fire Research Laboratory at Cardington UK. The test results confirmed that steel members in real multi-storey buildings have significantly greater fire resistance than isolated members in the ISO834 (1985) Standard Fire Test. Although steel temperatures became considerably greater than Eurocode4 (1992) critical temperatures for the loading levels used, no run-away failures were observed. A general observation was that the composite slab appears to play a significant part in preventing structural collapse, and in consequence it is very important to model the action of concrete floor slabs correctly in fire. The numerical software VULCAN (Najjar and Burgess, 1996; Bailey et al., 1996; Huang et al., 1999, 2000a, 2000b) has been developed in recent years at the University of Sheffield for three-dimensional analysis of the structural behaviour of composite and steelframed buildings in fire. The objective of the current phase of the research has been to extend the layered procedures previously developed by the authors (Huang et al., 1999, 2000b) for the modelling in fire of solid reinforced concrete slabs and orthotropic slabs, including their ribbed lower part, to include geometric non-linearity. A total Lagrangian approach is adopted throughout, in which displacements are referred to the original configuration.