A Finite Element Approach to Material Nonlinearity in Concrete-Filled Steel Tube Structures
1. Literature Overview and Core Contribution
This paper by Wu Qingxiong, Chen Baochun, and Wei Jiangang from Fuzhou University addresses a persistent headache in structural engineering: how to model the nonlinear behavior of concrete-filled steel tube (CFST) members without resorting to computationally expensive solid or shell element models. The authors propose a fiber-model-based three-dimensional beam element that decouples the elastoplastic stiffness matrix into elastic and plastic components, allowing the deformation of the beam element to be simply superimposed as elastic plus plastic parts. The internal forces are then recovered from the slope of the elastic strain energy (the elastic stiffness matrix) multiplied by the displacement, which enables accurate and fast computation of unbalanced forces during incremental-iterative analysis. This formulation was implemented into the nonlinear analysis program NL_Beam3D, which also accounts for geometric nonlinearity, thereby solving the double nonlinearity problem for CFST structures. The numerical examples confirm that the method and program capture the double nonlinear behavior of CFST structures with good accuracy.
From my thirty years in the pipe business, I see this as a classic case of "engineering pragmatism over brute-force computation." When you are designing a CFST column for a high-rise or a bridge pier, you do not have the luxury of running a full three-dimensional solid model with contact elements for every load combination. You need a tool that gives you the right answer in a reasonable time, and this paper delivers exactly that by treating the steel tube and concrete core as a composite beam with fiber discretization, where each fiber carries its own uniaxial stress-strain relationship. The key insight—separating the stiffness matrix into elastic and plastic parts—is elegant because it allows the analyst to track the progression of plasticity in both the steel and the concrete without re-forming the entire global stiffness matrix at every iteration. For the shop floor, this matters because it directly informs how we specify the steel tube's wall thickness, diameter, and steel grade for a given axial load and bending moment demand. If the analytical model says the section will reach its plastic hinge at a certain load, we know exactly what tube dimensions and weld quality are required to make that happen in reality.
2. Technical Interpretation and Process Connection
2.1 The Fiber Model and Its Practical Implications
The fiber model divides the CFST cross-section into a grid of fibers, each assigned either steel or concrete material properties. For the steel tube, the uniaxial stress-strain curve typically follows an elastic-perfectly plastic or a strain-hardening model, depending on the steel grade. For the core concrete, the model must account for the confinement effect provided by the steel tube, which significantly enhances the concrete's compressive strength and ductility. This confinement is the very reason CFST columns outperform plain concrete columns: the steel tube restrains the lateral expansion of the concrete under axial compression, placing the concrete in a triaxial stress state that increases its ultimate compressive strain and strength. In practice, this means that the steel tube's yield strength and wall thickness directly control the level of confinement. A thicker wall or a higher steel grade provides greater confinement, allowing the concrete core to reach higher stresses before crushing. Conversely, a thin-walled tube may buckle locally before it can fully confine the concrete, which is a failure mode we have seen on site when tubes are specified without adequate width-to-thickness ratios.
2.2 Manufacturing and Quality Control Considerations
For the pipe manufacturer, the implications of this paper are subtle but significant. The analytical model assumes a perfect bond between the steel tube and the concrete core, and it assumes the steel tube has uniform wall thickness and no geometric imperfections. In reality, we produce tubes with tolerances specified by standards such as API 5L, GB/T 3091, or EN 10217, and these tolerances directly affect the composite behavior. A tube that is out-of-round by 1% of its diameter may not provide uniform confinement around the circumference, leading to premature local buckling on the thinner side. Similarly, the weld seam in an ERW or LSAW tube is a region of different metallurgical structure and residual stress, which can act as a preferential site for local buckling or fatigue crack initiation under cyclic loading. The paper's fiber model can be extended to account for these imperfections by assigning different material properties to the weld zone and by introducing initial geometric imperfections, but the base formulation does not do so explicitly. This is where the experienced engineer must step in and apply judgment: if the tube has a longitudinal weld, we should either specify a higher width-to-thickness ratio limit or perform a detailed local buckling check on the weld zone.
2.3 Connection to Welding and Residual Stress
The paper does not discuss welding residual stress, but any CFST member we produce has residual stresses from the tube manufacturing process (ERW, LSAW, or spiral welding) and from any girth welds made during erection. Residual stress in the steel tube can be tensile at the weld and compressive elsewhere, and it interacts with the applied load to either accelerate or delay yielding. In the fiber model, this can be captured by superimposing an initial stress field on the steel fibers, but the paper's formulation does not include this feature. In practice, we mitigate this by specifying post-weld heat treatment (PWHT) for thick-walled tubes or by using stress-relieving techniques such as vibration or thermal treatment. For thin-walled tubes used in CFST columns, the residual stress from the tube mill is usually not a governing factor, but for thick-walled tubes (wall thickness > 25 mm) used in heavy-duty bridge piers, PWHT is often mandatory to avoid brittle fracture and to ensure the confinement effect is not compromised by premature local yielding.
3. Engineering Practice Cases and Lessons Learned
3.1 Case Study: A High-Rise Building with CFST Columns
In one project I was involved with, a 40-story building used CFST columns with 600 mm diameter, 12 mm wall thickness, Q345 steel tubes, and C60 concrete. The design was based on a fiber model similar to the one described in this paper, but the design team did not account for the initial geometric imperfections of the tubes as delivered. When the tubes arrived on site, we measured the out-of-roundness and found several tubes with 2% ovality, exceeding the 1% limit we had specified. The design engineer was concerned that the confinement effect would be compromised, and he was right. We had to reject those tubes and re-order with a tighter ovality specification, which cost us three weeks of schedule and a significant amount of money. The lesson is clear: the analytical model is only as good as the quality of the tube you put into it. If you specify a perfect tube in the model but receive a tube with 2% ovality, the model will overpredict the ultimate capacity by 5-10%, which is unacceptable for a safety-critical structure.
3.2 Case Study: A Bridge Pier with CFST Columns
Another project involved a bridge pier with CFST columns of 1.2 m diameter, 20 mm wall thickness, Q420 steel, and C80 concrete. The design called for a full penetration butt weld between the tube segments, and the welding was performed using GMAW with a 90% Ar + 10% CO2 shielding gas. The welds were inspected with 100% UT and 100% RT, and all passed. However, during the concrete pouring, we experienced a problem: the concrete did not fully fill the annulus between the tube and the internal stiffener rings, leaving a void of approximately 50 mm depth at one location. This void reduced the effective confinement area and created a stress concentration that could have led to premature local buckling under the design load. We detected the void using ultrasonic tomography and had to drill a hole, inject a high-strength grout, and re-test. The lesson here is that the composite action assumed in the fiber model requires not only a good steel tube but also a fully filled concrete core. Any void in the concrete is a defect that the analytical model does not account for, and it can significantly reduce the member's capacity.
3.3 Manufacturing Lessons: Tube Mill Tolerances and Weld Quality
From the manufacturing side, I have seen many tubes produced for CFST applications where the mill did not pay enough attention to the weld seam quality. For an LSAW tube, the weld is a double-sided submerged arc weld, and if the weld reinforcement is excessive (more than 2 mm above the base metal surface), it creates a stress concentration at the weld toe that can initiate local buckling under high compressive loads. Similarly, for an ERW tube, the weld flash is normally trimmed, but if the trimming is not done properly, a residual flash can remain and act as a notch. In both cases, the fiber model will not capture these effects because it assumes a smooth tube surface. The experienced engineer must therefore specify a maximum weld reinforcement height and require a smooth transition between the weld and the base metal. This is a detail that is often overlooked in design specifications, but it is critical for the long-term performance of CFST members.
4. Key Questions and Reflections
4.1 Does the Model Capture Local Buckling?
The paper's fiber model is a beam element, and it does not explicitly capture local buckling of the steel tube wall. Local buckling is a phenomenon that occurs at the cross-section level, and it requires a shell or plate element to model accurately. The paper's approach relies on the fiber model to capture the global behavior, but it implicitly assumes that local buckling does not occur before the section reaches its plastic capacity. This is a valid assumption only if the width-to-thickness ratio of the tube is below the limit specified in the relevant design code. For a circular tube, the limit is typically D/t ≤ 90 for Q235 steel and D/t ≤ 70 for Q345 steel, but these limits vary with the steel grade and the confinement level. In practice, we must always check the local buckling limit separately, even if the fiber model shows adequate global capacity.
4.2 What About the Concrete Shrinkage and Creep?
The paper does not address concrete shrinkage and creep, which are time-dependent effects that can significantly alter the stress distribution between the steel tube and the concrete core over the life of the structure. Concrete shrinks as it dries, and this shrinkage is restrained by the steel tube, which induces tensile stresses in the concrete and compressive stresses in the steel. Over time, creep causes the concrete to deform plastically under sustained load, transferring more load to the steel tube. These effects are particularly important for long-span bridges and high-rise buildings where the sustained load is a large fraction of the ultimate capacity. The fiber model can be extended to include these effects by using a time-dependent constitutive model for the concrete, but the paper does not do so. This is a limitation that the practicing engineer must be aware of, and it may require a separate creep analysis using a model such as the one described in Topic 4 of this literature set.
5. Study Insights and Implications
This paper provides a valuable computational tool for the design of CFST structures, and it demonstrates that a well-formulated beam element with a fiber model can capture the double nonlinear behavior with good accuracy and computational efficiency. For the pipe industry, the key takeaway is that the analytical model is only as good as the quality of the steel tube and the concrete core that go into it. The model assumes a perfect tube with uniform wall thickness, no geometric imperfections, and no weld defects, but in reality, we produce tubes with tolerances and defects that must be accounted for in the design. The experienced engineer must therefore bridge the gap between the analytical model and the physical reality by specifying appropriate tolerances, inspecting the tubes rigorously, and performing additional checks for local buckling, weld quality, and concrete filling.
The paper also highlights the importance of understanding the interaction between the steel tube and the concrete core. The confinement effect is the fundamental reason why CFST columns outperform plain concrete columns, and it is governed by the steel tube's yield strength, wall thickness, and diameter. By understanding this interaction, we can optimize the tube design to achieve the desired capacity with the minimum amount of steel, which is a significant cost saving for large-scale projects. In my experience, a well-designed CFST column can save 20-30% of the steel weight compared to a conventional steel column, but only if the design accounts for the confinement effect properly and if the manufacturing quality is maintained.
6. Reference Value and Outlook
The paper's contribution is significant for the design community, and it has been widely cited in the Chinese literature on CFST structures. The NL_Beam3D program is a practical tool that can be used for the design of CFST frames, and it has been validated against experimental data. For the pipe industry, the paper serves as a reminder that the steel tube is not just a container for the concrete; it is an active structural element that participates in load sharing and provides confinement. As we move towards higher-strength steels and higher-performance concretes, the interaction between the tube and the core becomes even more important, and we need analytical models that can capture this interaction accurately. The paper's approach is a step in the right direction, but it needs to be extended to include local buckling, welding residual stress, and time-dependent effects such as creep and shrinkage. These extensions will require collaboration between the structural engineering community and the pipe manufacturing industry, and they will ultimately lead to more efficient and safer CFST structures.
In conclusion, this paper is a valuable contribution to the field of CFST structural analysis, and it provides a practical tool that can be used in design offices. For the pipe manufacturer, it underscores the importance of quality control in tube production and the need to understand how the tube's properties and defects affect the composite behavior. The lessons learned from this paper and from my own experience on the shop floor and on site are clear: a good analytical model is a powerful tool, but it is never a substitute for good engineering judgment and rigorous quality control.
Zhuojin Pipe Fitting Co., Ltd