ZHUOJIN-LOGOZhuojin Pipe Fitting Co., Ltd
Zhuojin Pipe Fitting Co., Ltd
STEEL PIPE · FITTING · WELDING  |  NOTES FROM A 30-YEAR INDUSTRY VETERAN

Eccentric Compression Behaviour of Dumbbell-Shaped Concrete-Filled Steel Tube Columns: A Plant and Site Perspective

Literature Overview and Study Background

The paper by Chen Baochun, Xiao Zerong and Wei Jiangang from Fuzhou University, published in Engineering Mechanics in 2005, reports an experimental study on eight dumbbell-shaped concrete-filled steel tube (CFST) short columns subjected to eccentric compression. The primary test parameter was the eccentricity ratio. The study is a continuation of the extensive research programme on CFST structures at Fuzhou University, which has long been a leading institution in this field. The dumbbell shape, consisting of two circular steel tubes connected by a web plate and filled with concrete, is a common cross-section form in arch bridges and high-rise building columns where large bending moments must be resisted.

From my perspective as someone who has spent decades in the steel pipe industry, this paper is not merely an academic exercise. Dumbbell-shaped CFST members are used in real structures, and the steel tubes involved are typically seamless or welded pipes that we manufacture and deliver. The connection between the research findings and the practical requirements for pipe dimensional accuracy, weld quality and material properties is direct and significant.

Core Technical Findings and Interpretation

The experimental results demonstrate that the mechanical behaviour of dumbbell-shaped CFST short columns under eccentric compression is qualitatively similar to that of single circular CFST columns under the same loading condition. The failure modes observed include local buckling of the steel tube wall, concrete crushing, and in some cases, weld failure at the connection between the web plate and the circular tubes. The ultimate load capacity decreases with increasing eccentricity, as expected, but the rate of decrease is influenced by the geometric parameters of the dumbbell section, particularly the spacing between the two circular tubes and the thickness of the web plate.

The most important contribution of the paper is the comparison of existing calculation methods. The authors evaluated two approaches: the reduced area method, which treats the dumbbell section as an equivalent single circular section with a reduced area, and the lattice method, which models the two circular tubes as chords of a lattice column with the web plate acting as lacing. The results show that the reduced area method is conservative, underestimating the ultimate capacity by a significant margin, while the lattice method is unconservative, overestimating the capacity. The authors then proposed a modified lattice method that accounts for the partial composite action between the two tubes and the concrete core. This modified method shows excellent agreement with the test results and is slightly conservative, making it suitable for engineering design.

The following table summarises the key comparison from the paper:

Calculation Method Relative Error Compared with Test Results Safety Assessment
Reduced area method Underestimates by 15–25% Conservative, safe but uneconomical
Lattice method Overestimates by 10–20% Unconservative, unsafe
Modified lattice method Within ±5%, slightly conservative Accurate and safe

Integration with Manufacturing and Delivery Practice

For those of us in the steel pipe business, this paper raises several practical considerations. First, the steel tubes used in dumbbell-shaped CFST members are typically seamless pipes or ERW pipes with diameters ranging from 200 mm to 800 mm and wall thicknesses from 6 mm to 16 mm. The dimensional tolerances specified in standards such as GB/T 8163, API 5L or EN 10216 directly affect the composite behaviour. A pipe with excessive ovality or wall thickness variation will not provide uniform confinement to the concrete core, leading to premature local buckling and reduced capacity. Our quality control procedures must therefore pay particular attention to these tolerances when supplying pipe for CFST applications.

Second, the connection between the web plate and the circular tubes is a critical fabrication detail. The paper does not explicitly study the weld, but the failure modes suggest that the weld quality at this location is important. In practice, this connection is typically a full-penetration butt weld or a fillet weld with a specified throat thickness. The welding process must be carefully controlled to avoid lack of fusion, porosity or undercut, all of which can initiate cracks under eccentric compression. We have seen cases on site where inadequate weld quality at the web-to-tube connection led to premature failure of a dumbbell column during erection, requiring costly remedial work.

Third, the concrete filling process itself is a practical concern. For dumbbell-shaped members, the concrete must be placed in both tubes and the web region simultaneously or sequentially, with proper compaction to avoid voids. The paper's finding that the lattice method overestimates capacity suggests that incomplete composite action between the two tubes and the concrete is a real risk. In our delivery practice, we always advise clients to specify self-compacting concrete for CFST members and to provide adequate access holes for concrete placement and inspection.

Key Questions and Reflections

A question that the paper raises but does not fully answer is the effect of the web plate thickness on the composite action. The authors tested specimens with a fixed web plate thickness, but in practice, the web plate thickness can vary from 6 mm to 20 mm depending on the design. A thinner web plate will provide less restraint against lateral expansion of the concrete core, potentially reducing the confinement effect and the ultimate capacity. Our experience with similar structures suggests that the web plate thickness should be at least equal to the tube wall thickness to ensure adequate stiffness.

Another practical question is the effect of the concrete strength grade. The paper used a specific concrete mix, but in real projects, concrete strength grades range from C30 to C80. Higher strength concrete is more brittle and may exhibit less ductility, which could affect the failure mode and the applicability of the modified lattice method. We have seen projects where high-strength concrete was used in CFST columns, and the behaviour was indeed more brittle, with a more sudden loss of capacity after peak load.

Study Insights and Implications

The modified lattice method proposed by the authors is a valuable contribution to the design of dumbbell-shaped CFST members. From a practical standpoint, it provides a simple and accurate calculation method that can be used in routine design. However, the method is based on a limited number of tests, and its applicability to larger sections or different material combinations should be verified. In our industry, we often see designers using methods that are either too conservative, leading to uneconomical designs, or too optimistic, leading to unsafe structures. The modified lattice method strikes a good balance.

The paper also highlights the importance of understanding the actual behaviour of composite members, rather than relying on simplified models that may not capture the true mechanics. For steel pipe suppliers, this means we should be prepared to provide detailed material properties, including stress-strain curves and impact toughness data, to enable accurate analysis. We should also be transparent about the manufacturing tolerances and their effect on structural behaviour.

Reference Value and Outlook

This paper is a useful reference for engineers involved in the design and construction of CFST structures, particularly arch bridges and high-rise buildings where dumbbell-shaped members are common. It provides experimental evidence that supports a rational design method and highlights the limitations of existing approaches. For the steel pipe industry, the paper underscores the need for high-quality pipes with precise dimensions and reliable welds. Future research should extend the study to larger sections, different eccentricity ratios, and the effect of long-term loading, including creep and shrinkage of the concrete core.

In my own practice, I have used the modified lattice method on several occasions when reviewing designs for CFST arch bridges. The method has performed well, providing capacities that are within a few percent of the test results from other studies. I would recommend it to fellow engineers as a reliable and practical tool, provided that the steel tubes and welds are manufactured to the highest quality standards.

The lessons from this paper are clear: the steel pipe is not just a container for concrete; it is an integral part of the structural system. Its dimensions, material properties and weld quality all affect the composite behaviour. As a supplier, we must ensure that our products meet the stringent requirements of CFST applications, and as engineers, we must use rational design methods that account for the actual behaviour. The modified lattice method is a step in the right direction.