An experimental study of the fire performance of non-sway loaded concrete-filled steel tubular column assemblies with extended end plate connections

Abstract This paper presents the results of an experimental study of the behaviour of concrete filled steel tubular column assemblies using extended end plate connections. The objectives of this study are to investigate the effects of rotational restraint on column bending moments and column effective lengths. Two series of column assemblies have been fire tested at the University of Manchester’s fire testing laboratory. In each test, a hot-finished rectangular steel tube of grade S275 and dimension of either RHS 200×100×5mm or RHS 200×100×12.5mm, filled with nominally C30 concrete, was connected to a pair of steel beams at one end using extended end plate connections. Each series of tests consist of four combinations of two levels of total axial load in the column, with either equal or unequal loads on the connected beams. Design calculations are performed by using a finite element heat transfer computer program (ABAQUS) in combination with the ambient temperature design method for composite columns in Eurocode 4 Part 1.1, but with modified composite cross-sectional capacities at elevated temperatures. Due to different thicknesses of the steel tubes, the two series of fire tests show different behaviour. For tests using the thinner tube, local buckling was observed. From the results of design calculations, it is found that the position of local buckling of the steel tube has direct influence on the effective length of a concrete filled column. The effective length of a concrete filled column with a pin end may be taken as the distance from the largest local buckle of the steel tube to the pin end. For tests using the thicker tube, no local buckling was observed. The results of design calculations suggest that the effective length of these columns can be determined according to the method in Eurocode 4 Part 1.2. For these test columns, this means that the column effective length should be taken as 0.7 times the column height. The design column bending moment may be taken as the unbalanced beam load acting eccentrically from the column centreline.