Experiments were performed to determine the accuracy of proposed theoretical estimates of the ultimate strength of beam-columns. CFT behavior was discussed in detail for concentric and eccentric loading situations. The author began with a discussion of some of the advantages and disadvantages of using CFTs. One advantage is the optimization of the steel strength and stiffness due to its location at the periphery of the cross section. A second advantage is that thinner steel sections may be used because the concrete core forces all local buckling modes outward, ensuring that the tube reaches its longitudinal yield strength before buckling. Finally, two economic advantages arise by using CFTs. Concrete has a lower cost to strength ratio and the tubes act as formwork which decreases construction costs. The author cited two main disadvantages: the questionable strength retention of the steel tubes in fires and the lack of experimental evidence regarding connections and beam-columns. The prime objective of the paper was to provide experimental beam-column data and examine this type of element.
Experimental Study, Discussion, and Results
The tests were performed by applying a constant concentric axial load and increasing the moment to failure. Several conclusions were presented.
Creep. The author found that creep had an influential effect on the specimen behavior. For slower loading rates (i.e., 10 to 15 minutes at a given displacement), creep produced a load reduction of as high as 15%. Conversely, the author suggested that if tests are conducted continuously without pausing for measurements, an additional 10% gain in strength could be achieved.
Concrete-Steel Interaction. The author concluded that the two materials behave independently of one another. At strains below 0.001, the Poisson's ratio of concrete was one-half to two-thirds that of steel. This difference in lateral expansion tended to separate the materials. As strains increased, the concrete expanded laterally at a greater rate than the steel. Above 0.001, the concrete Poisson ratio began to approach that of steel. It was assumed that the steel had attained a confining state when the ratio of the measured circumferential and longitudinal strains in the steel sharply increased. This suggested an outward pressure from the concrete core. Not until about 90% of the load had been added did the ratio between circumferential and longitudinal strains divert substantially from the Poisson's ratio measured from the steel tubes alone in compression. This suggested that little benefit to the concrete arose from confinement activity of the steel encasement. The concrete core did, however, stabilize the steel wall of the tubes, preventing premature local buckling and allowing the tube to attain its full yield capacity.
Stiffness of Axially Loaded CFT Columns. The author suggested that both the steel and the concrete in columns under pure compressive loads should undergo the same amount of longitudinal strain. The modulus of elasticity, or stiffness, of the concrete tends to decrease around a strain of 0.0008. Steel, on the other hand, has a constant stiffness up to a strain of about 0.0010 to 0.0012. Its stiffness also decreases at a lower rate than concrete's stiffness beyond 0.0010. The steel will therefore begin to carry a higher and higher proportion of the load as the corresponding strains increase beyond 0.0008.
Strength of Axially Loaded CFT Columns. A lower limit for the strength of the cross section may be established by summing the force necessary to yield the steel in the longitudinal direction plus the amount of force carried by the concrete at the yield strain. Should the steel exert any confinement on the concrete, the amount of force the concrete carries before the steel yields will increase. The concrete will reach its ultimate strength at a strain value of about 0.002 and may begin to crush. To prevent the onset of crushing before steel yielding, steel with a yield stress above about 60 ksi should be avoided. Furlong recommended a maximum value of steel yield strength of 50 ksi to ensure the steel yields before the concrete crushes. The author also found that confinement of the concrete is much more likely in round sections because the steel may develop an effective hoop tension, whereas the flat sides of a rectangular tube are not effective in resisting perpendicular pressure.
Strength of CFT Beam-Columns. The lower limit to the pure bending capacity of CFTs is the plastic moment on the steel alone. The concrete core may increase the section's resistance to bending, but since the tensile resistance depends on the steel alone, in order for the ultimate moment to increase, the presence of the concrete must cause the neutral axis of the cross section to move toward the compression face. This would increase the tensile moment arm about the geometric centroid of the section, allowing additional capacity. A way to accomplish this is to use a thinner tube and/or higher strength concrete. The author formulated a load-moment interaction diagram in an attempt to predict the capacity of a CFT under combined loading. He plotted Pu/Po versus Mu/Mo, where Pu and Mu are the measured ultimate loads and moments, respectively. Po is the lower limit of axial load estimated by summing the yield strength of the steel and the load on the concrete at the time the steel yields (based on the strain at the yield load). Mo, the lower limit value for moments, is simply the plastic moment capacity of the steel tube alone. An estimate of the lower bound of the strength of a column is given by the equation:
Comparison of Results
The above interaction equation was very conservative for a number of experimental values, but it represents a reasonable lower bound in the absence of computers or a more rigorous approach. The author arrived at better results by considering the composite sections to be reinforced concrete sections and analyzing them as such using the capacity interaction equations specified in the ACI Building Code. The results were illustrated graphically in the paper. Although this method produced better results, the values were quite tedious to compute.
Furlong, R. W. (1967). “Strength of Steel-Encased Concrete Beam Columns,” Journal of the Structural Division, ASCE, Vol. 93, No. ST5, pp. 113-124.