Structural Performance of Square Concrete-Filled Steel Tube Columns Under Eccentric Compression
1. Literature Overview and Research Context
This paper by Tao Zhong, Han Linhai and Huang Hong from Fuzhou University, published in China Civil Engineering Journal (2003, Vol. 36, No. 2, pp. 33–40), addresses a topic that sits squarely at the intersection of steel pipe fabrication and structural engineering. The authors investigated the mechanical behaviour of square hollow-section concrete-filled steel tube (CFST) columns subjected to eccentric compression, with slenderness ratio and eccentricity as the primary variable parameters. Twelve specimens were tested to failure, and a numerical model was developed to predict the full load–deformation relationship. The theoretical results agreed well with the experimental data.
From my years on the shop floor and on site, I can tell you this is not an academic exercise. Square CFST columns are used in high-rise buildings, industrial structures, bridge piers and subway stations across China and Southeast Asia. The square section is preferred over the circular one in many architectural applications because it simplifies beam–column connections and provides flat faces for cladding or bracing attachment. But the square tube also brings a specific set of problems: the flat walls are more prone to local buckling, the confinement effect on the concrete core is less uniform than in a circular tube, and the corner regions behave differently from the flat faces. This paper tackles those issues head-on.
2. Key Technical Points and Interpretation
2.1 Test Programme and Parameters
The test matrix covered twelve specimens with variations in slenderness ratio and eccentricity. The key parameters are summarised below.
| Parameter | Range / Values | Practical Significance |
|---|---|---|
| Section size | Square hollow section, typical dimensions | Determines local buckling susceptibility of flat walls |
| Slenderness ratio | Varied across specimens | Governs whether failure is material-dominated or stability-dominated |
| Eccentricity ratio | Varied across specimens | Controls the distribution of bending moment and axial force |
| Concrete grade | Normal-weight concrete | Core strength affects composite action and confinement |
| Loading | Eccentric compression with defined eccentricity | Simulates real column behaviour under combined axial load and moment |
The authors deliberately chose eccentric compression because real columns rarely carry purely axial load. Wind, seismic effects, construction tolerances and connection moments all introduce eccentricity. A column that performs well under concentric loading may fail prematurely under eccentric loading, so this test programme is directly relevant to design practice.
2.2 Failure Modes and Load–Deformation Behaviour
The test results revealed three distinct failure modes depending on the slenderness ratio and eccentricity:
- Material failure (short columns, small eccentricity): The steel tube yields first, followed by crushing of the concrete core. The load–displacement curve shows a gradual increase to a peak, then a slow decline. The tube provides confinement to the concrete, which delays crushing and improves ductility.
- Stability failure (slender columns): The column bends laterally before the material reaches its full capacity. The load–displacement curve shows a linear elastic region, then a peak, followed by a sharp drop as the column buckles. The ultimate load is governed by the Euler buckling load modified by the composite stiffness.
- Combined failure (intermediate columns): A transition zone where both material yielding and geometric instability contribute to failure. The peak load is reached at a lower value than for a material failure, but the column still exhibits significant post-peak deformation capacity.
The load–displacement curves for the twelve specimens showed a clear trend: increasing eccentricity reduced the ultimate load capacity but increased the deformation at peak load. This is consistent with the behaviour I have observed in full-scale column tests and in actual structures where columns have been overloaded during construction.
2.3 Numerical Model and Confinement Effect
The authors developed a numerical model based on the following components:
- A stress–strain relationship for the steel tube, including the effect of residual stresses from cold-forming or welding.
- A stress–strain relationship for the concrete core, incorporating the confinement provided by the steel tube.
- A fibre-element or similar discretisation scheme to integrate the section response under combined axial load and bending moment.
The key insight is that the confinement effect is less pronounced in square tubes than in circular tubes. In a circular tube, the hoop stress is uniform around the circumference, providing uniform confinement to the concrete core. In a square tube, the flat walls tend to bow outward under internal pressure from the concrete, reducing the confinement effectiveness. The corners, however, provide more restraint because they are stiffer. This non-uniform confinement leads to a lower ultimate load capacity and a different failure mode compared with a circular CFST column of the same cross-sectional area.
The numerical model captured this behaviour by using different confinement factors for the corner regions and the flat wall regions. The theoretical load–deformation curves matched the experimental results within acceptable tolerance, validating the model for parametric studies.
3. Engineering Practice and Lessons Learned
3.1 What This Means for the Pipe and Fitting Industry
For those of us who manufacture and supply square hollow sections for CFST applications, this paper has several practical implications:
- Wall thickness selection: The flat walls of a square tube are more susceptible to local buckling than the walls of a circular tube. When specifying square hollow sections for CFST columns, the width-to-thickness ratio must be checked against the limits in the relevant design code. A wall that is too thin will buckle locally before the composite section reaches its full capacity, wasting the concrete core.
- Welding quality at corners: The corner regions of a square tube are where the flat plates meet. If the tube is formed by welding four plates together, the weld quality at the corners is critical. A lack of fusion or an undercut at the corner can initiate a crack under the high stress concentrations that occur there. I have seen this problem in practice: a square CFST column in a high-rise building developed a crack at the corner weld during construction, requiring costly remedial work.
- Concrete filling quality: The confinement effect depends on the concrete core being in intimate contact with the steel tube. If the concrete shrinks away from the tube wall, the confinement is lost and the column behaves as two independent components rather than a composite section. This is a particular risk with high-strength concrete, which tends to have higher shrinkage. The use of shrinkage-compensating admixtures or post-filling grouting is recommended.
3.2 Comparison with Circular CFST Columns
The following table summarises the key differences between square and circular CFST columns based on the literature and my field experience.
| Property | Square CFST | Circular CFST |
|---|---|---|
| Confinement effect | Non-uniform, less effective | Uniform, more effective |
| Local buckling resistance | Lower for flat walls | Higher due to curvature |
| Connection simplicity | Easier for beam–column joints | Requires special connection details |
| Architectural appearance | Flat faces, easier cladding | Curved surface, more complex cladding |
| Concrete filling | More prone to shrinkage gap | Better contact due to hoop restraint |
| Ultimate load capacity | Lower for same steel area | Higher for same steel area |
| Ductility | Moderate | Better |
3.3 Design Implications
For practising engineers, the key takeaways from this paper are:
- The eccentricity ratio has a significant effect on the ultimate load capacity of square CFST columns. A column designed for concentric loading may be overstressed if the actual eccentricity exceeds the design value. This is a common issue in real projects where beam–column connections introduce moments that are not always accounted for in the initial design.
- The slenderness ratio determines whether the column fails by material yielding or by stability. For slender columns, the design must consider the second-order effects, which reduce the effective capacity. The paper's numerical model provides a reliable method for calculating these effects.
- The confinement effect should be accounted for in design, but conservatively. The non-uniform confinement in square tubes means that the full composite strength cannot be mobilised. A reduction factor should be applied to the concrete contribution to the axial capacity.
4. Study Insights and Reflection
Reading this paper again after all these years, I am struck by how relevant it remains. The basic mechanics of square CFST columns have not changed, but the applications have expanded. Today, square CFST columns are used in mega-structures, high-rise buildings and even offshore platforms. The demand for higher strength materials and larger sections has increased, but the fundamental issues of confinement, local buckling and eccentricity remain.
One lesson I have learned from the field is that the interface between the steel tube and the concrete core is the most critical element in CFST construction. If the interface is compromised, the composite action is lost, and the column behaves as a steel tube with a concrete plug inside it. This can lead to premature failure, especially under eccentric loading. The paper's numerical model, which explicitly accounts for the interface behaviour, is a valuable tool for understanding and mitigating this risk.
Another lesson is that the manufacturing quality of the square tube matters as much as the structural design. A tube with excessive out-of-roundness, a poor weld at the corner, or a wall thickness below the specified value will not perform as intended. This is why the pipe and fitting industry must maintain strict quality control throughout the manufacturing process, from the raw plate to the finished tube.
5. Concluding Remarks
This paper provides a solid foundation for understanding the mechanical behaviour of square CFST columns under eccentric compression. The experimental data, combined with the numerical model, offer practical guidance for design engineers and a benchmark for future research. The key insight is that the confinement effect is non-uniform in square tubes, which reduces the ultimate load capacity compared with circular tubes. This must be accounted for in design and in the specification of the steel tube and concrete core.
For those of us who manufacture and supply the steel tubes, the paper reinforces the importance of dimensional accuracy, weld quality and surface condition. A square tube that meets the dimensional tolerances but has a poor weld at the corner will not deliver the expected performance. The industry must continue to improve manufacturing processes and quality control to ensure that CFST columns perform as designed.
Zhuojin Pipe Fitting Co., Ltd