Journal of Constructional Steel Research 67 (2011) 790–799 Contents lists available at ScienceDirect Journal of Constructional Steel Research journal homepage: www.elsevier.com/locate/jcsr Nonlinear analysis of composite beams subjected to combined flexure and torsion E.L. Tan a, , B. Uy b,1 a School of Engineering, University of Western Sydney, Locked Bag 1797 Penrith South DC, NSW 1797, Australia b School of Engineering, Civionics Research Centre, University of Western Sydney, Locked Bag 1797 Penrith South DC, NSW 1797, Australia article info Article history: Received 7 September 2010 Accepted 24 December 2010 Keywords: Composite steel–concrete beams Partial shear connection Flexure Torsion Curved beams Finite element method abstract There are situations in which a composite steel–concrete beam is subjected to torsion, such as members that are curved in plan or straight edge beams in buildings or bridges. The composite action of the steel beam and concrete slab in torsion is usually ignored in design codes of practice. Therefore, a three- dimensional (3D) finite element model is introduced in this paper to simulate composite steel–concrete beams subjected to combined flexure and torsion with the influence of partial shear connection using a commercial software ABAQUS. Brick and truss elements were used with the incorporation of nonlinear material characteristics and geometric behaviour in the model. This is coupled with an extensive parametric study using the validated finite element model using different parameters such as the span length and the level of shear connection. From the analytical study, a new phenomenon has been uncovered, which was validated by the test observation. This phenomenon called torsion induced vertical slip is an important issue, which would make the assumption plane sections remain plane invalid. In addition, difference in span length greatly affected the flexure–torsion interaction relationship of the composite steel–concrete beams, whilst the partial shear connection did not affect the relationship. Design models for readers to take away at the end of this paper are also proposed. © 2010 Elsevier Ltd. All rights reserved. 1. Introduction Composite steel–concrete construction has been widely used around the world. It is commonly used in modern buildings and highways due to their advantages over traditional reinforced con- crete construction. However, due to the complexity of the geom- etry of modern structures, some of the supporting beams such as edge or curved in plan members are subjected to combined loading. One such combination of loading is the application of combined flexure and torsion. These effects of combined flexure and torsion are not currently addressed in the Australian Stan- dards AS 2327.1 [1] or other international standards on compos- ite steel–concrete construction such as the Eurocode 4 [2] or the American Institute of Steel Construction [3]. Moreover, the prob- lem becomes increasingly complex when partial shear connection is introduced to the design. To avoid the huge cost and significant testing time required for conducting full-scale experiments, researchers have often involved Corresponding author. Tel.: +61 2 4736 0403; fax: +61 2 4736 0137. E-mail addresses: e.tan@uws.edu.au, taneeloon@gmail.com (E.L. Tan), b.uy@uws.edu.au (B. Uy). 1 Tel.: +61 2 4736 0228; fax: +61 2 4736 0137. finite element or numerical modelling. Yam and Chapman [4] con- ducted a series of numerical analyses to investigate the inelastic behaviour of composite steel–concrete beams. They produced a predictor–corrector method of systematic numerical integration. However, several assumptions were needed such as the strain distribution have to be linear over the depth of the compos- ite steel–concrete beams, the stress–strain curves for steel were the same for both tensile and compressive regions, the concrete and steel have equal curvatures at all points along the compos- ite steel–concrete beams and there was no separation between the concrete and the steel. Using the elemental formulation from the empirical equation of [4,5] modelled their composite steel–concrete beams using ABAQUS with each shear connector as a 2D truss element with two end nodes and three translational degrees of freedom at both sides. Shell elements were used for the concrete slab and steel beam. Later, Thevendran et al. [6] also used ABAQUS to develop a 3D finite element model to predict the behaviour of curved in plan composite steel–concrete beams. The concrete slab and steel beam were modelled by thin shell elements, whilst the shear connectors were modelled by rigid beam elements. However, full composite action at the concrete–steel interface was assumed. The most recent research was carried out by Erkmen and Bradford [7] whom further extended a 3D elastic total Lagrangian 0143-974X/$ – see front matter © 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcsr.2010.12.015