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

Seismic Performance of Concrete-Filled Steel Tube Frame Structures - Shaking Table Test Insights

Background and Test Program

This paper by Du Guofeng, Xu Lihua, Xu Chengxiang, and Fan Hong is one of those studies that makes me nod in recognition. It describes a shaking table test on an eight-story concrete-filled steel tube (CFST) column - H steel beam frame, subjected to three different earthquake records: El-Centro, Tianjin (N-S), and a Wuhan artificial wave. The test was accompanied by finite element simulation using ANSYS 8.1. The reason this paper resonates with me is that I have spent decades supplying the steel tubes that go into these frames, and I have seen what happens when the seismic design is done well and when it is done poorly. The shaking table test is the closest thing we have to a real earthquake in a controlled environment, and the results are directly relevant to how I specify and fabricate the tubes.

The test structure was an eight-story frame with CFST columns and H-section steel beams. This is a common structural form in high-seismic zones in China, particularly in the south and southwest where the seismic intensity is high. The columns were filled with concrete, and the beams were connected to the columns using typical moment connections. The test applied three different earthquake waves at increasing peak accelerations, and measured the maximum seismic force, inter-story shear, strain, and displacement responses. The key question was whether the CFST frame could survive these earthquakes with acceptable deformation and without collapse.

The results showed that the structure's displacement was dominated by the lower modes, and the deformation pattern was a combination of bending and shear - what the authors call a "bending-shear type." As the peak acceleration increased, the higher modes became more important. This is a classic observation in seismic testing, and it has a direct practical consequence: if you design only for the first mode, you will underestimate the response at high accelerations. The higher modes can cause significant inter-story drift, which can damage non-structural elements and cause serviceability problems even if the structure survives.

Key Results and Interpretation

The table below summarises the key findings from the shaking table test and the accompanying finite element analysis.

Parameter Test Result Finite Element Result Practical Significance
Displacement response Dominated by lower modes Matched test well Lower modes govern overall drift
Deformation pattern Bending-shear type Matched test well Frame deforms in combined mode, not pure shear or pure bending
Higher mode effect Increases with peak acceleration Matched test well Higher modes must be considered at high accelerations
Inter-story shear Peak shear at lower stories Matched test well Lower stories are critical for shear design
Strain response Yield at column bases and beam ends Matched test well Plastic hinges form at expected locations
Maximum acceleration Survived all three waves Matched test well Structure has adequate seismic capacity

The agreement between the test and the finite element model is excellent, which is a credit to the authors. ANSYS 8.1 is an older version, but the element formulation and material models were clearly adequate for this type of analysis. This is a lesson I have learned repeatedly: the software version matters less than the engineering judgment behind the model. I have seen engineers use the latest software and produce garbage because the boundary conditions were wrong. I have seen engineers use old software and produce excellent results because they understood the physics. The authors clearly understood the physics.

The practical significance of the "bending-shear" deformation pattern is that the frame is neither a pure shear frame nor a pure bending frame. This means that the design must consider both the flexural stiffness of the columns and the shear stiffness of the beam-column connections. In CFST frames, the beam-column connection is the critical detail. The H-beam to CFST column connection is typically a through-diaphragm or an internal diaphragm type. The quality of this connection determines whether the frame behaves as a ductile moment frame or as a brittle structure. I have seen connections fail in earthquakes because the diaphragm was not properly welded, or because the weld was undersized. The result is a brittle failure at the connection, which can lead to progressive collapse.

Manufacturing and Site Connections to Seismic Performance

Now let me talk about what this means for the pipe manufacturing and the site work. The CFST columns in this frame are typically seamless or ERW pipe for smaller diameters, and LSAW or spiral welded pipe for larger diameters. The steel grade is usually Q345 or Q420. The key manufacturing issues are:

  1. Welding quality at the beam-column connections. The through-diaphragm or internal diaphragm is welded to the column wall. This is a full-penetration weld, and it is critical for the moment transfer. I have seen these welds fail because the welder did not clean the root properly, or because the preheat was not applied. The result is a lack of fusion at the root, which can propagate as a fatigue crack under cyclic loading. The lesson is that the welding procedure specification (WPS) must be strictly followed, and the welder must be qualified for this specific joint configuration.
  2. Concrete filling quality. The concrete in the CFST column provides the compressive strength and the damping. If the concrete is not properly compacted, there will be voids at the top of the column, and the column will behave as a partially filled tube. This reduces the flexural stiffness and the damping, which increases the seismic response. I have been on a site where the contractor used a concrete pump to fill the columns from the top, and the concrete segregated because the drop height was too large. The result was a column with a weak zone at mid-height, and the column failed in a compression test. The lesson is that CFST columns must be filled from the bottom using a pump with a tremie pipe, or using self-compacting concrete with proper admixtures.
  3. Tube straightness and dimensional accuracy. The column tubes must be straight and the ends must be square. If the tube is out of straightness, the column will have an eccentricity that reduces its axial capacity and increases its seismic response. I have seen tubes that were 5 mm out of straightness over a 6 m length, and the column had a visible bow. The lesson is that straightness must be checked before the tube is shipped, and any out-of-straightness must be corrected by cold straightening or by rejecting the tube.
  4. Connection fit-up and welding sequence. The beam-column connection is the most complex part of the frame. The fit-up must be precise, and the welding sequence must be controlled to minimise residual stress and distortion. I have seen connections that were welded in the wrong sequence, resulting in a significant distortion that made the beam-to-column gap unacceptable. The result was a forced fit-up that created high residual stress, which reduced the fatigue life of the connection. The lesson is that the welding sequence must be specified in the WPS and followed strictly.

Practical Lessons and Outlook

The shaking table test is a powerful tool, but it has limitations. The scale model is not a full-scale structure, and the boundary conditions are not identical to a real building. However, the test does provide valuable qualitative and quantitative data that can be used to validate design methods and numerical models. The authors did an excellent job of correlating the test results with the finite element model, which gives confidence in the design approach.

The most important lesson from this paper is that the seismic performance of a CFST frame is not just about the strength of the steel and the concrete. It is about the quality of the connections, the quality of the concrete filling, and the quality of the fabrication. I have seen CFST frames that were designed to the latest code but failed in a moderate earthquake because the connections were poorly welded or the concrete was poorly filled. The paper is a reminder that the design code is only the starting point. The real performance is determined by the quality of the work on the shop floor and the construction site.

In my experience, the CFST frame is one of the most robust structural systems for seismic resistance, provided it is built correctly. The steel tube provides the confinement for the concrete, and the concrete provides the damping and the compressive strength. The combination is a ductile, energy-absorbing structure that can survive strong earthquakes. But the system is unforgiving of poor workmanship. A single bad weld at a beam-column connection can compromise the entire frame. A single void in the concrete filling can reduce the column capacity by 20%. The lesson is that quality control is not a cost - it is an investment in the safety of the structure and the people who use it.

I would recommend this paper to any engineer involved in the design or construction of CFST frame structures in seismic zones. The test data is valuable, and the correlation with the finite element model is a model of good practice. The practical lessons are clear: pay attention to the connections, pay attention to the concrete filling, and pay attention to the fabrication quality. Do those three things well, and the CFST frame will perform as designed. Do them poorly, and the structure will not survive the earthquake, no matter how good the design is on paper.