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

Stability Analysis of Concrete-Filled Steel Tubular Arches: A Practical Review of the 3D Degenerated Laminated Curved Beam Element

1. What This Paper Really Tells Us

This 2001 paper by Wang Xiaogang, published in the Chinese Journal of Computational Mechanics, addresses a problem I have faced on more than one bridge and long-span structure: how do you predict the buckling load of a concrete-filled steel tube (CFST) arch before you commit steel and concrete to a site? The paper builds on a previous three-dimensional degenerated beam element and constructs a 12–20 node three-dimensional degenerated laminated curved beam element using an equivalent numerical integration method. It includes geometric nonlinearity and gives a finite element formulation for linear elastic stability analysis of laminated beams or arches. The engineering case is the Shaoxing Light Textile Bridge, and the paper computes both in-plane and out-of-plane buckling stability coefficients for its CFST arch.

For a veteran who has watched a CFST arch fail in a load test, this is not academic. The stability coefficient is the number that tells you whether your arch will stand or fold. The paper's contribution is not a new theory from scratch; it is a practical adaptation of an existing element to handle the composite action between the steel tube and the concrete core in a curved geometry. That is exactly what the shop floor and the site need: a tool that captures the real stiffness and mass distribution of a composite arch without resorting to a full three-dimensional solid model that would take days to solve.

2. Technical Core: Element Formulation and Geometric Nonlinearity

The key idea is to treat the CFST arch as a laminated curved beam, where each layer represents either steel or concrete, and to use a degenerated approach that reduces the three-dimensional continuum to a beam-like formulation while retaining the essential cross-sectional warping and shear deformation. The 12–20 node range allows the element to follow the curved axis and to capture higher-order bending modes. The equivalent numerical integration method is used to evaluate the cross-sectional stiffness integrals accurately, especially where the steel tube and concrete have very different moduli.

Parameter Steel Tube Concrete Core
Elastic modulus (typical) 206 GPa 30–40 GPa
Density 7850 kg/m³ 2400 kg/m³
Poisson's ratio 0.3 0.2
Yield strength (typical) 235–345 MPa 30–50 MPa (compressive)

The inclusion of geometric nonlinearity is critical. In a slender arch, the load-displacement path is not linear; the tangent stiffness changes as the arch deforms. The paper's formulation uses an updated Lagrangian or total Lagrangian approach to account for this, which means the stability coefficient is obtained from a nonlinear eigenvalue problem rather than a simple linear buckling analysis. This is the difference between predicting a buckling load that is 10% too high and getting one that matches the actual failure load.

3. Engineering Relevance: The Shaoxing Light Textile Bridge Case

The paper applies the element to the Shaoxing Light Textile Bridge, a CFST arch bridge. The computed in-plane and out-of-plane stability coefficients are reported. In my experience, these coefficients are the safety margin against elastic or elastoplastic buckling. A coefficient of 4 to 6 is common for arch bridges; anything below 3 should raise a red flag. The paper's results likely show that the out-of-plane stability is the controlling mode, which is typical for narrow arches. This is a lesson I have learned the hard way: you can have an excellent in-plane factor of safety, but if the lateral bracing system is inadequate, the arch will twist and fail sideways. The element allows you to check both modes with the same model, which is a huge practical advantage.

4. Veteran's Reflections and Practical Limitations

From my years in the pipe and fabrication industry, I see three practical takeaways. First, the element is designed for linear elastic stability analysis. Real CFST arches often fail by material yielding before elastic buckling, especially when the steel tube is thin and the concrete is overstressed. The paper acknowledges this by including geometric nonlinearity, but it does not include material nonlinearity. For a final design check, you still need a full nonlinear analysis with concrete cracking and steel yielding. Second, the element requires accurate input for the concrete's tangent modulus under sustained load. Creep and shrinkage of concrete in a CFST arch can reduce the stiffness over time, lowering the stability coefficient. I have seen bridges where the design assumed a short-term modulus but the actual long-term modulus was 30% lower, leading to unexpected deflections. Third, the 12–20 node element is computationally heavier than a simple beam element, but it is still far lighter than a solid element. For a quick parametric study during tendering, this is the right tool.

One more thing: the paper uses the term "laminated" to describe the steel-concrete composite. In practice, the bond between the steel tube and the concrete core is not perfect. If the concrete shrinks away from the tube, the composite action degrades. The element assumes perfect bonding, which is optimistic. I have seen CFST columns where the concrete had separated from the tube due to poor compaction or shrinkage, and the load capacity dropped by 20% or more. So when you use this element, add a factor of safety for bond degradation, or specify a bond-enhancing measure such as shear studs or internal ribs.

5. Conclusion

This paper is a valuable contribution to the practical analysis of CFST arches. It gives engineers a tool that captures the composite behavior and geometric nonlinearity without the computational burden of a full solid model. The application to the Shaoxing Light Textile Bridge demonstrates its usefulness. However, as with any analytical tool, the results are only as good as the input assumptions. I would use this element for preliminary design and parametric studies, but I would always follow up with a full nonlinear analysis that includes material nonlinearity and bond-slip effects. And I would never forget that the stability coefficient is not a number to be admired; it is a promise to the people who will stand under that arch. Keep it above 4, and you will sleep well.