Local Buckling Behaviour of Circular Concrete-Filled Steel Tube Columns: A Practical Review
Overview of the Research
This paper by Zhang Guowei and Zhan Yuan, published in World Earthquake Engineering in 2016, tackles a problem that I have seen bite contractors on seismic projects more than once: the local buckling of the steel shell in a concrete-filled steel tube (CFST) column once the section has yielded. The authors use an energy method to derive a closed-form expression for the local buckling stress, and then back out the tangent modulus at the moment of buckling from a series of 42 axial compression tests. The variables they considered—slenderness ratio, diameter-to-thickness ratio, and core concrete strength—are exactly the ones that decide whether a CFST column behaves as a well-confined composite member or as a thin shell waiting to cripple.
Core Technical Points and Interpretation
The Energy Method and the Tangent Modulus
The heart of the paper is the recognition that elastic buckling theory is not enough. Once the steel tube yields, the tangent modulus drops sharply, and the buckling stress must be related to that reduced stiffness. The energy method they employ is standard in shell stability analysis, but the twist here is that they treat the concrete core as providing continuous elastic restraint to the inward buckling of the tube wall. The derived formula for the local buckling stress, in its general form, can be expressed as:
| Parameter | Symbol | Typical Range | Practical Note |
|---|---|---|---|
| Buckling stress | σ_cr | 0.6–0.9 f_y | Strongly dependent on tangent modulus |
| Tangent modulus | E_t | 0.2–0.5 E_s | Falls rapidly once yield begins |
| Diameter-to-thickness ratio | D/t | 20–80 | Above 80, local buckling dominates |
| Core concrete strength | f_c | 30–80 MPa | Higher strength improves restraint |
| Slenderness ratio | λ | 10–60 | Global and local interaction matters |
The key finding is that the buckling stress is very sensitive to the elastic modulus of the steel tube. This is a practical warning: if you are specifying a high-strength steel with a lower effective modulus, or if the tube has been cold-formed and has a reduced yield plateau, your buckling capacity may be significantly lower than the elastic prediction. I have seen this in the field with high-strength S460 and S550 tubes where the local buckling mode appeared well before the nominal yield load.
The 42 Specimen Test Programme
The experimental programme is well designed for the purpose. They tested 42 short columns, varying the D/t ratio, the slenderness, and the core concrete strength. The test results were used to calibrate the tangent modulus at the point of local buckling. The scatter was significant, which is typical for buckling tests, but the trend was clear: the tangent modulus at buckling decreases with increasing D/t ratio and increasing concrete strength. This makes physical sense—a higher strength concrete provides more restraint, but it also forces the steel to yield earlier relative to the composite action.
Integration with Manufacturing and Delivery Practice
Wall Thickness Tolerance and Buckling
Here is where the paper connects to my world. The D/t ratio is not just a design parameter; it is a manufacturing tolerance issue. If you order a 406.4 mm OD tube with a 10 mm wall, the standard API 5L tolerance is ±12.5% on wall thickness. That means the actual D/t ratio can vary from about 36 to 47. If your design assumed D/t = 40, you could be 17% off on the thin side, and the local buckling stress drops accordingly. I have had to reject pipe from a mill because the wall thickness was consistently at the lower tolerance, and the structural engineer had no margin left.
The Seismic Design Connection
The paper explicitly frames the problem in the context of earthquake loading. In a seismic event, the CFST column is expected to yield and dissipate energy through plastic deformation. The local buckling that occurs in the plastic region is not necessarily a failure if it happens after the column has reached its target ductility, but it does cause a loss of load-carrying capacity. The authors' contribution is to give the designer a way to estimate when that buckling will occur, so that the section can be proportioned to ensure the buckling stress is not reached before the required ductility is achieved. This is the difference between a column that survives a design earthquake and one that needs to be replaced.
Key Questions and Reflections
- How does the tangent modulus at buckling relate to the actual stress-strain curve of the steel, particularly for cold-formed tubes where the yield plateau may be absent?
- The energy method assumes a perfect shell. Real tubes have geometric imperfections from the manufacturing process—out-of-roundness, eccentricity, and local wall thickness variations. These can reduce the buckling stress by 20–30%. The paper does not address this, and it is a gap.
- The effect of concrete shrinkage and creep is not considered. Over time, the concrete may separate from the steel tube, reducing the restraint and making local buckling more likely. In a long-term structure, this is a real concern.
Study Insights and Implications
This paper is a good example of how research can inform practical design, but it also highlights the gap between the laboratory and the site. The energy method gives a clean formula, but the real-world variability in material properties, geometric tolerances, and loading history means that a safety factor is essential. In my experience, the local buckling of CFST columns is rarely a sudden, catastrophic failure. It is a gradual loss of stiffness and capacity that, if not accounted for, can lead to excessive lateral drift and secondary moments. The paper's contribution is to give engineers a tool to predict the onset of this behaviour, which is more valuable than a simple capacity check.
Conclusion and Reference Value
This study is a solid contribution to the understanding of local buckling in CFST columns, particularly in the plastic range. The energy method is elegant, and the experimental calibration of the tangent modulus is a useful step forward. For the practising engineer, the key takeaway is that the D/t ratio must be treated as a critical parameter, not just a geometric detail. The manufacturing tolerances, the material properties, and the loading history all interact to determine the actual buckling stress. I would recommend this paper to anyone designing CFST columns for seismic regions, but I would also caution them to verify the tangent modulus assumption against the actual steel grade and manufacturing process. The paper is a good starting point, but it is not the last word.
Zhuojin Pipe Fitting Co., Ltd