Matsui 1985, 1986
In these two papers, the behavior of one-bay frame structures composed of concrete-filled and hollow tube beam-columns spanned by steel beams are studied. The first paper presents preliminary results on 4 CFT frame specimens, whereas the second paper presents a more detailed set of results for the full set of 12 specimens. The papers investigate overall frame behavior and the behavior of the connections between the steel beam and the hollow tube or CFT beam-columns; connection design equations are presented based upon the experimental results. A new value for the limiting width-thickness ratio (D/t ratio) of CFTs is also proposed based upon the performance of the frames and an examination of the post-buckling behavior of CFTs using a plastic limit analysis.
The author suggests that the design of beam to column connections is crucial to the ductile behavior of frames which are composed of CFT columns and steel I-beams. The details for the connection of steel beams to CFTs presented are of two general types: outside stiffener and through stiffener. The outside stiffener connection consists of top and bottom continuous plates which are welded to the outside of the CFT and rigidly connect to the top and bottom flanges of the I-beams. The inside stiffeners are continuos plates (generally with holes which allow concrete to flow through the joint) which penetrate the joint. These connections require field welding of adjacent CFTs. Both types of connection allow the designer to control the strength and behavior of the panel zone.
Experimental Study, Results, and Discussion
Twelve frame specimens were tested both monotonically and cyclically. The loading frame consisted of a hydraulic press which applied axial load to the columns through a spreader beam, and a hydraulic jack which applied horizontal loading under deflection control. The horizontal deflection at various points in the column height, as well as strains, were recorded. The behavior of the connections was monitored closely as was local buckling and post-buckling behavior of the tubes.
The D/t ratio of the steel tubes ranged from 33 to 68. The connections were designed such that the frame specimens with beam-columns having a D/t equal to 33 would exhibit a connection failure and frame specimens with a D/t ratio of 68 (weaker beam-columns) would fail at the base of the beam-columns. The frames composed of CFT columns demonstrated a higher strength (in comparison with predicted plastic limit analysis--see below), and, more importantly, much better post-local buckling behavior than the frames composed of HT columns. The CFTs were able to withstand a larger portion of their ultimate strength after local buckling of the steel tube.
The author cited two main reasons for the improved CFT behavior. First, the infilled concrete results in an outward buckling of the steel tube, rather than an inward buckling, as in hollow tubes. The outward buckling collapse mechanism differs from the inward mechanism, requiring more energy to result in a failure (Matsui, 1985). The second reason for the CFTs' exemplary behavior occurs in the post-buckling region. When the steel buckles, the axial load is transferred to the concrete and the section can maintain a significant portion of its pre-buckling capacity. The cyclically-loaded frame specimens maintained their strength for several cycles showing strength deterioration only after cracking and fracturing of the welded portion of the stiffeners in the connection-failing specimens and extensive local buckling in the column-failing structures.
Analytical Study, Results, and Discussion
The frames were analyzed using a plastic limit analysis for comparison to experimental results. For a given value of axial load and lateral load, the moment capacity of the frame was determined by calculating the ultimate moment of the beam-columns. Formulas to compute this moment were given based upon the full plastic moment of the tube under pure bending and the strength of the concrete in compression. As mentioned above, the frames with the CFT beam-columns exceeded the plastic mechanism line, while the frames with the hollow tube beam-columns were unable to reach the mechanism line due to local buckling of the steel. Based upon the results of the experiments, the author proposed a value for the limiting D/t ratio of CFTs equal to 1.5 times that of hollow tubes. The author concluded that frames composed of CFT beam-columns and steel wide-flange beams with connection designed as demonstrated in the paper produce ductile structures having excellent seismic resistance.
Matsui, C. (1985) “Strength and Behavior of Frames with Concrete Filled Square Steel Tubular Columns Under Earthquake Loading” Proceedings of the International Specialty Conference on Concrete Filled Tubular Structures, pp. 104-111.
Matsui, C. (1986). “Strength and Deformation Capacity of Frames Composed of Wide Flange Beams and Concrete Filled Square Steel Tubular Columns,” Proceedings of the Pacific Structural Steel Conference, Auckland, New Zealand, 4-8 August 1986, Vol. 2, pp. 169-181.