Seismic Design of Rectangular Concrete-Filled Steel Tube Columns with Internal Tie Rods
Literature Overview and Core Concept
This study addresses a structural form that has been gaining ground in high-rise construction: a rectangular concrete-filled steel tube (CFST) column reinforced internally with welded tie rods. The concept is straightforward on paper — place a steel tube, weld a series of transverse tie bars across the hollow section at intervals, then fill the tube with concrete. The tie rods restrain the outward bulging of the steel tube wall under axial compression and improve the confinement of the core concrete. What the paper contributes is a systematic seismic design method for this hybrid system, including the derivation of a strength model, the calibration of ductility factors, and a set of design recommendations for the spacing and diameter of the tie rods.
From my years on the shop floor and on delivery sites, I can tell you that this is not a purely academic exercise. We have shipped thousands of tonnes of rectangular and square hollow sections for CFST columns in seismic zones. The recurring problem on site is not the steel — it is the concrete. When the column is subjected to cyclic loading, the concrete core tends to spall and lose confinement, and the steel tube wall can locally buckle between the stiffener points. The tie rod concept is a direct engineering response to that failure mode, and the paper deserves careful reading for that reason alone.
Technical Points and Interpretation
The paper establishes a design framework based on the concept of a "restrained steel tube" — one where the internal tie rods act as discrete stiffeners that divide the tube wall into sub-panels. The key parameters are the tie rod spacing-to-wall-thickness ratio and the tie rod area-to-tube wall area ratio. The authors propose a limit state design equation that combines the axial load capacity of the concrete core, the contribution of the steel tube, and an additional confinement term attributable to the tie rods. The confinement factor is expressed as a function of the tie rod stiffness relative to the concrete core stiffness.
| Parameter | Symbol | Typical Range | Design Significance |
|---|---|---|---|
| Tie rod spacing ratio | s/t | 20 to 60 | Controls local buckling of tube wall |
| Tie rod diameter | d | 12 to 25 mm | Determines confinement stiffness |
| Tie rod area ratio | As/Ac | 0.5% to 2.0% | Governs additional confinement stress |
| Concrete strength | fcu | 30 to 80 MPa | Core strength and ductility |
| Steel tube yield strength | fy | 235 to 420 MPa | Tube wall contribution |
| Axial load ratio | n | 0.3 to 0.8 | Governs seismic performance level |
The seismic design method follows a displacement-based approach. The authors define a target drift ratio for different performance levels — essentially elastic, repairable, and collapse prevention — and provide a set of equations to determine the required tie rod configuration to achieve that drift. The ductility factor is calibrated against a database of cyclic loading tests, and the paper reports that the proposed method predicts the ultimate drift capacity with a coefficient of variation of approximately 12%, which is acceptable for a design-oriented model.
Integration with Manufacturing and Delivery Practice
Now let me connect this to what we actually do in the plant and on site. The tie rods must be welded to the inner walls of the rectangular tube before the tube is closed and before concrete is poured. This creates several practical challenges that the paper does not fully address but which any experienced fabricator will recognise immediately.
- Welding access and distortion. In a rectangular tube of, say, 400 mm by 400 mm, a welder can reach in through one open end to weld the first few tie rods, but the depth of the tube quickly becomes a problem. For tubes longer than about 1.5 metres, you need either a rotating fixture that allows the welder to work from both ends, or you need to weld the tie rods through access holes that are later plugged and welded. Neither option is cheap, and both introduce the risk of weld defects that are extremely difficult to inspect after the concrete is poured.
- Weld quality and inspection. The tie rod-to-tube wall welds are typically fillet welds of 6 to 8 mm leg length. They are in a confined space, often in the downhand position only if the tube is rotated. After welding, you have to perform magnetic particle inspection (MT) or penetrant testing (PT) on welds that are inside a closed box. That is a practical nightmare. In our shop, we have learned to insist on a full-length access opening on one side of the tube, weld all tie rods, perform 100% MT on every weld, then close the access plate with a full-penetration weld and perform UT on that closure weld. It costs money, but it is the only way to guarantee that the tie rods are actually doing their job.
- Concrete filling and compaction. The presence of tie rods obstructs the free flow of concrete. If you are using a high-strength, self-compacting concrete, the tie rods can create a shadow zone where the concrete does not fully surround the rod, leading to voids and a loss of composite action. The paper's design model assumes full bond and full confinement, but in practice we have seen cases where the concrete did not reach the underside of a tie rod, leaving a gap of 10 to 20 mm. That gap completely negates the confinement benefit of that particular tie rod. The solution is to use a concrete mix with a maximum aggregate size no greater than 20 mm, a slump flow of at least 600 mm, and to vibrate the column from the outside with a formwork vibrator while pouring from the top. Even then, we always specify a post-pour ultrasonic test on a sample of columns to verify the concrete fill.
- Erection and alignment. The internal tie rods add weight to the column, but they do not change the external geometry. The real issue is that the tie rods make the column stiffer in the transverse direction, which changes the natural frequency of the member during lifting. We have had columns that were perfectly straight on the ground but developed a permanent bow of 15 mm after being lifted because the lifting points were not positioned to account for the shifted centre of gravity. The paper does not cover this, but the fabricator and the erector must.
Key Questions and Reflections
The paper is a solid piece of research, but I want to raise a few questions that any engineer should ask before applying this method to a real project.
First, what is the long-term behaviour of the tie rod welds under sustained load and cyclic thermal exposure? The welds are inside a sealed steel tube, so they are not exposed to corrosion, but they are subjected to the creep and shrinkage of the concrete core. Over 50 years, the concrete shrinks by roughly 200 to 400 microstrain, which relieves the compressive stress in the concrete and transfers more load to the steel tube and the tie rods. The tie rod welds, which are already at a stress level of perhaps 30% of yield, will see an increase in stress. Fatigue is not a concern for a static structure, but the long-term stress redistribution is. The paper does not address this, and I would want to see a creep and shrinkage analysis before signing off on a design that relies on the tie rods for confinement.
Second, what happens in a fire? The steel tube is exposed on the outside, and in a fire, the tube wall heats up rapidly while the concrete core lags behind. The tie rods, being embedded in the concrete, will heat up more slowly, but the differential thermal expansion between the steel tube and the tie rods will induce additional stresses in the welds. If the tube wall buckles outward due to thermal expansion, the tie rods will be pulled into tension. The paper does not cover fire scenarios, and I would want to see a fire resistance test or at least a validated finite element model before using this system in a building where fire is a credible threat.
Third, what is the seismic performance of the tie rod welds under low-cycle fatigue? The paper reports ductility factors, but it does not report the low-cycle fatigue life of the tie rod welds. In a severe earthquake, the column may undergo 20 to 30 cycles of large inelastic deformation. Each cycle will strain the tie rod welds. If the welds are of poor quality — and I have seen plenty of poor quality fillet welds in confined spaces — they will crack after a few cycles, and the confinement will be lost. The paper's design method assumes the tie rods remain fully effective throughout the seismic event. That is a bold assumption, and I would want to see the weld quality requirements specified in the design document.
Study Insights and Practical Implications
What I take away from this paper is that the concept is sound, the design method is credible, but the practical implementation is where the real engineering challenge lies. The tie rod concept works only if the welds are sound, the concrete is fully compacted, and the tube is manufactured to tight tolerances. In my experience, the failure of such systems is almost never in the design — it is in the execution.
For the shop floor, this means:
- Specify the tie rod welds as full-penetration or at least as double-sided fillet welds with a minimum leg length of 0.7 times the tie rod diameter.
- Require 100% MT on all tie rod welds before the tube is closed.
- Use a concrete mix designed for confined placement, with a maximum aggregate size of 20 mm and a slump flow of at least 600 mm.
- Perform ultrasonic testing on a minimum of 10% of columns after concrete placement to verify the fill.
- Control the lifting points to account for the shifted centre of gravity.
For the design office, this means:
- Include the tie rod confinement effect in the axial load capacity calculation, but apply a reduction factor of 0.8 to account for potential concrete voids.
- Verify the tie rod welds for the full axial load plus the confinement stress, using a stress concentration factor of 2.0 at the weld toe.
- Check the tie rod welds for low-cycle fatigue with a strain range of 0.5% and a design life of 30 cycles.
- Specify a maximum tie rod spacing of 40 times the tube wall thickness to prevent local buckling.
Conclusion
This paper is a valuable contribution to the design of rectangular CFST columns with internal tie rods. The seismic design method is well-calibrated against test data, and the proposed equations are practical for design office use. However, the real value of the paper lies in what it does not say — the manufacturing and construction challenges that determine whether the tie rods actually deliver the promised confinement. I have seen too many projects where the design was excellent and the execution was poor, and the result was a structure that performed far below its intended capacity. The tie rod concept is no different. If you are going to use this system, you must specify the weld quality, the concrete mix, and the inspection regime with the same rigour that the design method demands. Only then will the tie rods do their job.
Zhuojin Pipe Fitting Co., Ltd