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

Initial Stress in CFST Steel Tubes: A Fabrication Perspective

The Core Problem and Why It Matters

Han and Yao's paper from 2003 is a classic in the CFST field, and it addresses a problem that I have seen cause real trouble on real projects: what happens to the bearing capacity of a concrete-filled steel tube beam-column when the steel tube already carries a significant axial stress before the concrete is poured? In my world, this is not a hypothetical scenario. When we fabricate a CFST column, we often have to lift it, transport it, and erect it before the concrete is placed. During that period, the tube is carrying its own weight, plus the weight of any temporary bracing, plus wind loads and erection loads. If the tube is part of a larger frame, it may also be carrying load from adjacent members that have already been erected. All of that creates an initial stress in the steel tube, and when the concrete is poured later, the tube is already in a partially stressed state.

The paper does two things. First, it presents experimental results for square CFST beam-columns with initial stress. Second, it develops a theoretical model for both circular and square sections, accounting for the initial stress coefficient, the member slenderness, the steel ratio, the load eccentricity, and the strengths of the steel and concrete. The model is validated against the experiments, and then the paper explores the parameter space to identify which factors dominate the reduction in capacity.

The key finding, as I read it, is that the initial stress reduces the ultimate bearing capacity of the composite member, and the reduction is more pronounced for members with higher slenderness and higher load eccentricity. That makes physical sense: if the tube is already stressed, it has less reserve capacity to resist the additional load from the concrete and the applied loads. But the paper goes further and quantifies the effect, which is what the industry needs.

What the Shop Floor and the Delivery Site Actually Care About

From a fabrication standpoint, this paper is a wake-up call about the sequencing of operations. I have seen contractors pour concrete into a CFST column that was already supporting a significant portion of the bridge deck, because they wanted to save time by pouring the concrete before the deck was fully erected. That is a dangerous practice. The paper shows that the initial stress can reduce the capacity by 10% or more in some cases, and that reduction is not covered by the standard safety factors in most design codes. The codes assume that the concrete is poured when the tube is stress-free, which is rarely true in practice.

The practical implication is that you need to control the erection sequence so that the tube does not carry excessive load before the concrete reaches its design strength. That means you may need to add temporary supports, or you may need to delay the placement of adjacent members until the concrete has cured. It also means you need to measure the actual stress in the tube before pouring, not just assume it is zero. We have done this on several projects by installing strain gauges on the tube at the fabrication shop and monitoring them during erection. The gauges tell us the real initial stress, and we can then decide whether the pour sequence is safe.

Another practical issue is the effect of initial stress on the tube's weld quality. If the tube is welded while it is under stress, the weld will experience a different stress state than if it were stress-free. This can affect the weld's residual stress distribution and its susceptibility to cracking, especially in high-strength steel. I have seen a case where a tube was welded while carrying a significant bending stress, and the weld developed a transverse crack during the concrete pour because the combined stress exceeded the yield strength in the heat-affected zone. The paper does not address this directly, but it is a real concern that any fabricator should be aware of.

Practical Lessons and My Own Judgment

The paper's theoretical model is impressive, but I would caution against applying it blindly to real structures. The model assumes a perfect bond between the steel and the concrete, and it assumes that the concrete is fully confined by the tube. In practice, the bond is rarely perfect, and the confinement is often reduced by the presence of internal stiffeners, shear connectors, or even just the surface roughness of the tube. I have seen CFST columns where the concrete did not fill the entire cross-section, leaving a gap at the top, and that gap completely changed the load path. The initial stress model would not capture that, and the capacity would be overestimated.

Another point I would raise is the interaction between initial stress and the concrete's shrinkage and creep. When the concrete shrinks, it pulls away from the tube, reducing the bond and the confinement. When it creeps under sustained load, the stress redistributes from the concrete to the steel tube, increasing the steel stress. If the tube already has initial stress, the creep-induced redistribution can push the steel stress above its yield point, causing premature yielding and a loss of stiffness. This is a long-term effect that the paper does not address, but it is critical for structures that are designed for a 100-year service life.

In my own practice, I have started specifying that any CFST column must be designed with an initial stress coefficient of at least 0.2, even if the construction sequence suggests that the initial stress will be lower. This gives a margin of safety that covers the uncertainties in the erection sequence, the concrete placement, and the long-term creep and shrinkage effects. It is a conservative approach, but it has saved us from problems on more than one occasion.

Study Insights and Implications for Practice

The paper's most valuable contribution is the parametric study that identifies the relative influence of each factor. It shows that the initial stress coefficient is not the dominant factor in all cases; for stocky members with low eccentricity, the effect is small, but for slender members with high eccentricity, it can be significant. This means that the design approach should be tailored to the specific geometry and loading conditions, rather than applying a blanket reduction factor.

For the steel pipe industry, this paper reinforces the need for accurate measurement of the tube's stress state during fabrication and erection. We have invested in strain gauge technology and data logging systems that allow us to monitor the tube's stress in real time, from the moment it is fabricated until the concrete reaches its design strength. This is not cheap, but it is a fraction of the cost of a structural failure.

I would also note that the paper's findings have implications for the way we specify the tube's material properties. If the tube is going to carry initial stress, it should have a higher yield strength than would otherwise be required, because the initial stress reduces the available margin. We have started specifying a minimum yield strength of 355 MPa for CFST tubes that will be erected before the concrete is poured, even if the design only requires 275 MPa. The extra cost is small, and the safety margin is worth it.

Reference Value and Outlook

This paper is a must-read for anyone involved in the design or construction of CFST structures. It addresses a real problem that is often ignored in design codes, and it provides a framework for quantifying the effect. The experimental data is limited to square sections, but the theoretical model covers both circular and square, and the parametric study gives useful guidance for design.

The industry needs to update its design codes to account for initial stress, and it needs to develop standard procedures for measuring and controlling the initial stress during construction. Until that happens, engineers should use the paper's model as a guide, and they should be conservative in their assumptions. The cost of being conservative is small; the cost of a failure is enormous.

My final thought is this: the paper is from 2003, and it is still relevant today. That tells you how slow the industry has been to adopt these findings. I have seen CFST columns erected with no consideration of initial stress, and I have seen them perform poorly under load. The paper gives you the tools to avoid that, but only if you use them. The next time you specify a CFST column, ask yourself: what is the initial stress in the tube, and how am I going to measure it? If you cannot answer that question, you are not ready to build.