The yield strength, tensile strength, and elongation of the   Grade-

800 steel were fy ¼ 746 MPa, fu ¼ 835 MPa, and 21。6%, respec- tively。 The properties of the Grade-400 steel were fy ¼ 301 MPa, fu ¼ 466 MPa, and 33。8%, respectively。 The yield strength of the Grade-800 steel was defined using the 0。2% offset method。 As shown in Fig。 2, the high-strength steel exhibits early strain- hardening behavior without an apparent yield  plateau。

Fabrication and Test  Setup

Built-up box sections fabricated by complete-joint-penetration (CJP) groove welding were used for all specimens。 For Grade- 800 high-strength steel, flux cored wire of nominal tensile strength FEXX ¼ 860 MPa [classified as E121T-1 according to AWS A5。29 (2010)] was used and 100% CO2 shielding gas was used during welding operation。 For Grade-400 mild steel, flux cored wire of nominal tensile strength 580 MPa [classified as E71T-1 according to AWS A5。20 (2005)] was used。 When the mild-steel vertical stiff- eners were welded to the high-strength tube plates, double-fillet welding using the lower-strength weld material for  mild  steel was used。 In all specimens, the net column  length  excluding rigid ends was Lc ¼ 700 mm。 In order to relieve the stress concen- tration at the top and  bottom  of the columns  and to prevent   the

。 Stress-strain relationships of tensile  coupons

corresponding premature failure, exterior stiffeners of 100 mm depth were welded at the column  ends。

Fig。 3 shows the test setup for eccentric axial loading。 An axial force was applied at the top of the column through knife-edges, which were used to realize the hinge conditions。 The column height between the hinges was Le ¼ 1,720 mm。 The loading was con- trolled by the vertical displacement of 0。005 mm=s。 Two linear var- iable displacement transducers (LVDTs) were used to measure the vertical displacements in the compression and tension sides, and four LVDTs were used to measure the horizontal  displacements of the deflected column (Fig。 3)。 Strain gauges were used to mea- sure the axial and lateral strains of the tube   plates。

In actual construction of CFT columns, nonshrinkage concrete is used to avoid the shrinkage of the core concrete。 In the test spec- imens, however, to avoid complexity of concrete mix design, ordi- nary concrete mix design was used。 For this reason, shrinkage of the concrete occurred during curing。 Therefore, in order to apply the uniform axial load to the steel tube and concrete, high-strength epoxy was grouted to the casting hole (one ł150 hole in square specimens and two ł100 holes in rectangular specimens) for the surface treatment 1 week before  testing。

Test Results

Load-Displacement  Relationship

Figs。 4 and 5 show the axial load-lateral deflection relationships of the specimens。 The lateral deflection was measured at the mid- height of the columns。 Loading was terminated when the load- carrying capacity decreased to 70% of the peak load。 The test results are summarized in Table  2。

In square-column Specimens E1, E3, and E4, the peak strength was the greatest in E4 (high-strength steel plus stiffeners), and the lowest in E3 (mild steel)。 In particular, when compared to Specimens E1 (high-strength steel) and E4, Specimen E3 showed significant early stiffness degradation due to local buckling and early yielding of the tube flange, although the section was classified as a compact section。 In the rectangular specimens, the load- carrying capacity of E5 (high-strength steel plus stiffeners) was greater than that of E2 (high-strength steel)。 The test result indicates that the vertical stiffeners increased the load-carrying capacity by restraining the plate local buckling and providing confinement to the filled concrete。