Composite Columns of Hollow Steel Sections Filled with High Strength Concrete
Doctoral thesis, 1993
In the design of buildings today, great effort is made to increase the structural flexibility. This has resulted in a demand for columns with reduced cross sections. To achieve a high load bearing capacity with a small cross sectional area, it is worthwhile to study the possibility of utilizing high strength concrete and composite effects.
The aim of the present study was to evaluate possible advantages of high strength concrete, possible confining effects of composite sections and the effect of shear transfer at the interface.
In the present study, composite columns consisting of hollow steel sections filled with concrete have been studied. The 39 tested columns were hinged at both ends and loaded with a compressive axial force normally applied with an initial end eccentricity of 20 mm. The basic parameters were the concrete grade, the steel yield strength and the steel section type. In complementary tests, the effect of debonded interface, combined axial and lateral loads and the way of load introduction at the ends were examined. For reference, the squash load was evaluated from stub tests, one for each column. The squash load can be regarded as an upper limit for the ultimate capacity of the corresponding column.
The principal assumptions in the theoretical analysis were complete interaction between the steel section and the concrete core, linear strain distribution, bilinear stress-strain relationship for the steel and parabolic stress-strain relationship for the concrete.
For the stub tests, the load bearing capacity was slightly higher than the sum of the nominal ultimate capacities for the steel and concrete sections. This finding shows that if there were any confinement effects, the reduced strength of the steel section was compensated by the increased strength of the concrete core. The higher compressive strength of the concrete increased the maximum load of the stub specimen.
For all of the composite columns except those with special connection details at the supports, the maximum capacity was determined by overall buckling with no sign of local buckling at the critical cross section. High strength concrete increased the load bearing capacity, especially in cases of small bending moments.
To predict the maximum load and deflection for the columns, a model that uses the moment-curvature relationship for the critical cross section has been used. The model predicts the maximum load with an accuracy of 10%. The parabolic stress-strain relationship for the concrete seemed to be an appropriate measure to consider both normal and high strength concrete in the calculation model. It was found not to be important to consider confinement effects for the eccentrically loaded composite columns.
The debonded interface seemed not to be of importance when the load was applied on the total cross section. However, when the steel-concrete interface was debonded and the load was applied on the steel section only, the concrete core did not contribute to the behavior and the column behaved as an unfilled steel tube.