As per the previous research, the concrete in square CFST is divided into a constraint region and non-constraint regions.
The non-constraint concrete is described with uni-axial stress–strain relation curve of plain concrete referred to GB50010-2002 or author's previous published paper [1].
The constraint concrete employs the modified Mander concrete model (in Fig. 15)proposed by Mander, Priestley and Park in 1988 [16].
The favorable influence of transverse confinement on the axial stress–strain relation is taken into consideration in the constraint concrete.
Related research reveals that, in non-stiffened square CFSTs (S1 and S5), cross-sectional constraint region (shaded area in Fig. 16) distributes near the corners of composite tube and the center of concrete, and carries out transition to the non-constraint regions (blank areas) adjacent to the steel plate center (in Fig. 16(a)).
For these stiffened specimens, the stiffeners restrain steel tube's local buckling at welds, thus the number and location of the welds influence the constraint region.
The boundaries between non-constraint regions and constraint region are described by second order parabola curves.
Fig. 16(b)–(d) displays the constraint region (shaded area) and non-constraint regions (blank areas) in stiffened
cross section of specimens. Fig. 17 shows longitudinal distribution of concrete constraint region (shaded area) and the non-constraint regions (blank areas).
Sections 1–1 is the stiffened cross section in which stiffeners are welded with the steel tube; Sections 2–2 is the weakest cross section between two adjacent stiffened cross sections.
Detailed calculation method can refer to author's previous research [1].