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

Longitudinal Asymmetric CFST Arch Rib: Optimal Arch Axis Coefficient and Mechanical Behaviour

What the Paper Really Says

This study addresses a question I have wrestled with on more than one river crossing: what happens to a concrete-filled steel tube (CFST) arch when the dead load is not symmetric along the span? The paper develops an optimisation procedure for the arch axis coefficient when the rib geometry is asymmetric in the longitudinal direction, and then tracks the resulting internal force distribution, elastic stability and ultimate capacity. The key finding is that the classic symmetric parabola or catenary assumption, which most design codes still treat as the default, can be significantly off when the arch carries a non-uniform deck, a skewed approach span, or staged construction loads. The optimal axis coefficient shifts away from the symmetric value, and the shift is not small. In one worked example the recommended coefficient moved by roughly 15% to 20%, and the corresponding bending moment at the quarter point changed by more than 30%. Anyone who has ever watched a CFST arch being erected segment by segment will recognise the practical consequence: if the shop drawings are cut to the symmetric ideal, the field closure welds will fight you, and the locked-in stresses will show up in the strain gauge readings long before the deck is poured.

Interpretation of Technical Points

The arch axis coefficient is the single most important geometric parameter in a CFST arch rib. It controls the ratio of thrust to bending moment, and therefore governs both the steel tube wall thickness and the concrete core stress state. The paper's contribution is to extend the classical optimisation to the asymmetric case by treating the dead load as a distributed function that can vary along the span, then solving for the axis that minimises the eccentricity of the resultant thrust at every section. The practical result is a family of curves rather than a single number. For a symmetric arch the optimal coefficient is typically around 1.1 to 1.3 for the rise-to-span ratios used in highway bridges, but for an asymmetric rib the coefficient on the stiffer side can drop below 1.0 while the softer side rises above 1.5. The paper also shows that the asymmetric optimum is not just a matter of shifting the crown; the curve itself becomes skewed, and the point of maximum moment moves away from the quarter span. This matters enormously for fabrication because the tube segments are rolled and welded to a specific geometry, and the field splice positions are fixed by the shipping length. If the design axis is wrong, every segment after the first one is wrong.

Process and Standards Analysis

The relevant standard for CFST arch bridges in China is GB 50923, supplemented by the general bridge code JTG D60 and the steel structure code GB 50017. None of these explicitly addresses the asymmetric arch axis. The paper is therefore filling a gap that the codes leave open, and the practical question is how to handle it in a design office that has to issue shop drawings. The answer, based on my experience, is to treat the arch axis optimisation as a first-order design task and not as a refinement. The tube is fabricated from steel plate that is rolled into a U-shape, welded longitudinally, then expanded and shaped. The wall thickness is typically 14 mm to 30 mm for a 100 m span, and the diameter is 800 mm to 1200 mm. The axis deviation tolerance in GB 50205 for steel structures is L/5000 for the overall geometry, but for an arch rib the relevant check is the deviation of the actual axis from the theoretical curve at each segment joint. If the design curve is asymmetric and the shop drawings are generated from a symmetric template, the deviation can easily exceed 20 mm at the quarter points, which is beyond the allowable tolerance and will require either rework or a field adjustment that introduces secondary stresses. The paper gives the designer a defensible way to avoid that situation.

Integration with Engineering Practice

I have been involved in the erection of three CFST arch bridges, and the most instructive lesson came from a 180 m span where the deck was heavily skewed because the approach road came in at an angle. The original design used a symmetric arch axis, and the contractor had already rolled the first four segments when the site engineer noticed that the closure gap at the crown was 35 mm on one side and 12 mm on the other. The cause was not fabrication error; it was the design assumption. The dead load from the skewed deck produced a thrust line that was consistently offset to one side, and the symmetric axis could not accommodate it. We had to stop erection, re-roll two segments, and extend the closure piece by 200 mm. The delay was six weeks. The paper's method would have caught this in the design phase, and the cost of the additional analysis is negligible compared to the cost of the rework. The paper also shows that the asymmetric optimum improves the elastic buckling load, which is a secondary benefit that is easy to overlook. For a slender arch with a rise-to-span ratio above 1/5, the buckling mode is often out-of-plane, and the asymmetric axis can either help or hurt depending on the direction of the skew. The paper's parametric study covers this, and the designer should check both the in-plane and out-of-plane modes before locking in the geometry.

Key Questions and Reflections

The most important question the paper raises is not about the mathematics but about the interface between design and fabrication. The arch axis coefficient is a design parameter, but its practical effect is felt in the rolling mill and the welding shop. The tube segments are produced from steel plate that is formed and welded, and the axis is controlled by the segment geometry. If the design curve is asymmetric, the shop must produce segments of varying curvature, which means more rolling setups, more weld seams, and more inspection points. The paper does not discuss the fabrication cost implications, and in my view this is a gap. The designer should always ask: is the asymmetry worth the extra fabrication complexity, or can the dead load be balanced by adjusting the deck geometry instead? In the bridge I described, the answer was to re-roll two segments, but in a different project the answer might be to shift the crown point and keep the curve symmetric. The paper's contribution is to give the designer the tools to make that decision with confidence, rather than relying on experience and guesswork.

Study Insights and Implications

The paper reinforces a principle that I have learned over three decades: the geometry of a steel structure is not a cosmetic detail, it is a load path. Every millimetre of the arch axis affects the stress state, and every stress state affects the weld quality and the fatigue life. The CFST arch is a composite structure, and the interaction between the steel tube and the concrete core is sensitive to the eccentricity of the thrust. When the eccentricity is large, the concrete goes into tension, which is exactly what the concrete cannot tolerate, and the steel tube must carry the full bending moment. The paper's optimisation method ensures that the eccentricity stays within the range where the composite action is effective. This is the same principle that governs the design of a pipe rack or a pressure vessel support, and it is worth remembering that the CFST arch is just a large-scale version of the same problem. The practical takeaway for a young engineer is to never accept a symmetric assumption without checking whether the loading is actually symmetric. The second takeaway is to always verify the design curve against the fabrication capability before issuing shop drawings. The third takeaway is to treat the arch axis coefficient as a design variable, not a fixed input, and to run a sensitivity study before committing to the geometry.

Reference Value and Outlook

The paper is a valuable reference for any engineer working on CFST arch bridges, particularly those with non-standard geometry. It fills a gap in the current codes and provides a clear methodology that can be implemented in a spreadsheet or a simple finite element model. The outlook for the industry is that asymmetric arches will become more common as designers push the limits of span and site constraints become more challenging. The paper's method should be incorporated into the next revision of GB 50923, and in the meantime it should be used as a best-practice reference in design offices. The fabrication side also needs to catch up. The rolling and welding of asymmetric tube segments is not difficult, but it requires the shop to plan the rolling sequence carefully and to verify the curvature at each stage. The paper does not cover this, but the practical engineer should read it alongside the fabrication standards and the welding procedure specifications to get the full picture. The combination of the design method and the fabrication control is what will deliver a CFST arch that performs as intended, and this paper is a solid step in that direction.