A Numerical Study on the Influence of Debonding in Concrete-Filled Steel Tube Columns on Structural Dynamic Characteristics

Abstract

The influence of debonding in concrete-filled steel tube (CFST) columns on the dynamic characteristics of super high-rise buildings is a common concern that remains insufficiently understood. The abnormal vibration incident of the SEG Plaza on 18 May 2021, also known as the 5·18 incident, serves as a typical case highlighting this issue. After two decades of service, the first-order bending frequency of the building decreased by approximately 6.1%, and extensive CFST column debonding was observed, with the maximum debonding rate reaching up to 97% on certain middle floors. To investigate the influence of CFST column debonding on structural dynamic characteristics, this study first derives a theoretical relationship between debonding parameters, namely angle and distance, and the equivalent bending stiffness of CFST columns. This analytical formulation is then implemented and validated through finite element simulations at multiple scales, including planar frame analysis in ABAQUS, a thin-interlayer simulation method in ANSYS, and full-building modeling in ETABS. Results show that for a planar frame, when a CFST column debonds at 270°, the structural natural frequency decreases by 0.984%; when the debonding angle is 180° with a 2 mm gap, the first-order frequency decreases by 0.141%. Numerical simulation of the SEG Plaza structural model predicts a reduction in the first-order frequency of 0.987% under the observed debonding conditions, confirming that debonding impairs force transmission, reduces structural stiffness, and alters natural frequencies. This study provides a mechanistic basis for evaluating stiffness degradation in long-service super high-rise buildings.

1. Introduction

Concrete-filled steel tube (CFST) columns with high load-bearing capacity, fire resistance, and excellent seismic performance have been widely used in high-rise buildings [1,2,3,4,5,6]. CFST columns are critical components in the design of high-rise building structures [7], with the corresponding analysis theories and design methods being increasingly mature. However, CFST columns can deteriorate over the long-term operation. In practical structures, debonding defects can occur in CFST columns, leading to degraded stiffness and dynamic performance of structures [8].

The abnormal vibration incident of the SEG Plaza in Shenzhen on 18 May 2021, i.e., the 5·18 incident, brought this issue to the forefront. In this 72-story superhigh-rise building, investigations revealed that the direct external cause was vortex-induced resonance of the mast, whose dominant frequency of 2.12 Hz coincided with a modal frequency of the main structure. More importantly, after 20 years of service, the structural dynamic characteristics had significantly changed compared to its completion in 2000: the first-order bending frequency decreased by approximately 6.1%, and the damping ratio dropped by more than half. On-site inspections further found extensive CFST column debonding, with debonding rates reaching as high as 99% on some middle floors. These findings strongly suggest that debonding-induced stiffness degradation contributes to the alteration of structural dynamic properties, thereby affecting the occurrence of resonance. Motivated by this incident, it is essential to investigate how CFST column debonding influences the dynamic characteristics of structural systems.

Extensive research has been conducted on the effect of debonding on the load-bearing capacity of CFST columns. Han et al. [9] studied the mechanical performance of debonded CFST columns under eccentric loads and suggested setting the gap ratio in actual structures to 0.05%. Xue et al. [10] investigated the effects of debonding defects on short CFST columns and found that local buckling was more severe in debonded specimens. Schnabl et al. [11] proposed a mathematical model to derive the changes in buckling load due to annular and partial debonding in slender CFST columns. Liao et al. [12,13,14] conducted axial compression, pure bending, and torsion tests on debonded CFST columns. Ahmad et al. [15] studied the effect of debonding distance on column performance. Manigandan [16], Yu [17], and Wang et al. [18] examined the mechanical performance of short CFST columns with different initial defects. Xue et al. [19] derived new formulas for predicting ultimate load-bearing capacity, reduction coefficients, ductility, and stiffness using evolutionary polynomial regression. Zhang [20] and Liao et al. [21] studied the effect of debonding on CFST members under combined compression–bending–torsion. Han [22] and Zhang et al. [23] examined the hysteretic performance of debonded CFST columns.

In recent years, special-shaped CFST columns (e.g., L-shaped, T-shaped, and cruciform sections) have attracted increasing attention. Hassam et al. [24] developed a finite element model for cross-shaped CFST stub columns and proposed a design-oriented formula for predicting their compressive strength. Li et al. [25] experimentally and numerically investigated the axial compression behavior of cruciform CFST columns stiffened by steel angles. These studies demonstrate that special-shaped CFST columns exhibit mechanical properties comparable to conventional sections. However, the influence of debonding on the global dynamic characteristics of complete building systems that employ either conventional or special-shaped CFST columns remains largely unexplored.

Regarding dynamic characteristics, Chen [26] found that debonding reduces the fundamental frequency of a CFST column. Hu [27] conducted excitation tests on CFST cantilever columns and observed that debonding weakens their vibration characteristics. Guo et al. [28,29] analyzed the causes of debonding and studied the effect of composite defects on bridge supports. In the field of bridge engineering, Lu et al. [30] reported that the natural frequency of arch bridges gradually decreases with increasing annular debonding rate and distance. Lu et al. [31] used finite element software to analyze the effect of debonding in CFST arch ribs on dynamic performance, also noting a reduction in frequency.

In summary, existing research has mainly focused on the static load-bearing capacity of debonded CFST columns or on the dynamic characteristics of debonded CFST arch bridges. Studies on the dynamic performance degradation of building structures due to CFST column debonding are still insufficient. For the safe operation of high-rise buildings, the adverse effect of CFST column debonding on structural dynamic characteristics must be further investigated.

Therefore, this study proposes a mechanism revealing the adverse effect of the CFST column debonding on structural dynamic characteristics. The mapping relationship between the debonding angle and the distance of the CFST columns and structural frequency is established. The structural performance degradation can be evaluated for the operation and maintenance phases of superhigh-rise buildings.

Figure 1 illustrates the methodological framework of the present study. The paper is organized as follows. Section 2 introduces the mechanism by which CFST column debonding affects structural stiffness and frequency. Section 3 uses finite element software to simulate the effects of debonding parameters on structural frequency and proposes a “thin interlayer” simulation method for CFST column debonding. Section 4 studies the effect of CFST column debonding on an actual superhigh-rise structure.

2. Mechanism of the Effect of Debonding in CFST Columns on Structural Dynamic Characteristics

2.1. Effect of Debonding in CFST Columns on the Force Transmission of a Structural System

From the perspective of structural systems, the steel beams and steel–concrete columns are rigidly connected in a healthy state. Under horizontal loads, the shear force at the beam end is transmitted to the column in the form of axial force, while the bending moment at the beam end is transmitted to the column in the form of column bending moment and shear force, creating a structural lateral force resistance system (Figure 2a). In accordance with the joint between the beam and the steel–concrete column (Figure 2b), the shear force and the bending moment at the end of the beam are transmitted to the steel–concrete column through the outer steel pipe of the column. If the steel pipe and column are not debonded, the steel pipe and concrete core column in the CFST column can work together to bear the axial force, shear force, and bending moment transmitted by the beam. However, if debonding occurs, the transmission path of the axial force, shear force, and bending moment borne by the steel pipe to the concrete core column is weakened.

The deformation of the section of the CFST column satisfies the assumption of a flat section when the CFST column is not debonded. The steel tube and concrete core column in the CFST column can synergistically deform to jointly bear the axial force, shear force, and bending moment transmitted at the beam end (Figure 3a). In this case, the structure is a composite structure, and its lateral stiffness is consistent with the design documents.

The debonding of the steel-reinforced concrete column makes the steel tube and the concrete core work independently. The deformation of the cross-section of the CFST column does not meet the assumption of a flat section. The axial force, shear force, and bending moment transmitted from the beam end to the column are borne by the steel tube (Figure 3b). The bearing capacity of the CFST column mainly depends on the steel tube. Accordingly, the structure performance degrades as a steel structure whose lateral stiffness and damping ratio are smaller than those of the composite structure.

Partial debonding of the steel tube and concrete core column exists in the CFST column of most long-term-service structures. However, owing to the presence of stiffeners and other structural measures in the CFST column, the steel tube and concrete column can still work together, but cannot meet the assumption of collaborative work in the planar section (Figure 3c). The structural performance is in the middle state between the above two extreme situations. In this case, the efficiency of the concrete core column in the CFST column is discounted, reducing the bearing capacity of the CFST column and the lateral stiffness of the structure.

2.2. Influence of Debonding in CFST Columns on the Stiffness of the Column Section

The stress on CFST columns after debonding is relatively complex. The following assumptions are made for the derivation process. (1) The stress on the concrete core column and the steel tube linearly changes as the CFST column hollow develops. (2) Given the small steel pipe wall thickness, the influence of steel pipe thickness is ignored in the equilibrium equation section. (3) The debonding distance is significantly smaller than the diameter of the steel pipe, that is, $r-\delta \approx r$ and $d-\delta \approx \mathrm{d}$, where δ is the clearance distance, and $r$ and $d$ are the radius and diameter of the CFST column, respectively. (4) The CFST column does not experience significant buckling. (5) The radial stress distribution at the interface between steel pipes and concrete is uniform.

When the CFST column experiences debonding, the force-transmission mechanism changes. The longitudinal stress diagram of the CFST column [32] is shown in Figure 4a. In the derivation process, the CFST column with debonding defects is divided into a concrete core column and a steel tube. The equilibrium equation is established on the basis of the stress diagram (Figure 4b,c).

Combining the internal force balance equation, deformation coordination equation, and stress distribution assumption of the debonding part of the steel–concrete column, the theoretical expression of the equivalent section stiffness of the steel–concrete column $B$ considering the debonding angle $\theta$ and debonding spacing $\delta$ can be established as shown in Equation (1). The detailed theoretical derivation can be found in Appendix A.

$$B={\displaystyle \frac{\eta e_0{d}_c{E}_s\{f_{c}r[\pi r-\delta (\theta -\frac{1}{2}\sin 2\theta )]+{\sigma }_{c}r[\pi r+\delta (\theta -\frac{1}{2}\sin 2\theta )]+2tr\pi {\sigma }_{s1}(1 + k)\}}{\frac{4}{\pi }[(\frac{4tr{\sigma }_{s1}(1 + k)}{D^2 - d^2}+\frac{64t{r}^2{\sigma }_{s1}(1 + k)}{\pi (D^3 - d^3)}]-2[(\frac{4tr{\sigma }_{s1}(1 + k)}{D^2 - d^2}-\frac{64t{r}^2{\sigma }_{s1}(1 + k)}{\pi (D^3 - d^3)}]}}$$

(1)

where $\delta$ is the debonding distance, $\theta$ is the debonding angle, $r$ is the radius of the steel-reinforced concrete column, $d$ is the inner diameter of the steel-reinforced concrete column, $D$ is the outer diameter of the steel-reinforced concrete column, $t$ is the thickness of the steel pipe, ${\mu }_c$ is the Poisson’s ratio of the concrete material, ${\mu }_s$ is the Poisson’s ratio of the steel pipe material, $E_c$ is the elastic modulus of the concrete, and $E_s$ is the elastic modulus of the steel pipe, $f_c$ is the design value of compressive strength of concrete, $\eta$ is the structural importance coefficient, $e_0$ is the eccentricity of the section axial force, $k$ is the relationship between the steel pipes ${\sigma }_{s1}$ and ${\sigma }_{s2}$, ${\sigma }_c$ is the longitudinal stress of concrete, and ${\sigma }_{s1}$ is the longitudinal stress on the debonding side of the steel pipe concrete.

The CSD10 CFST column in the SEG Building in Shenzhen is selected as an example for section stiffness calculation. The relevant parameter values of CSD10 steel-reinforced concrete columns are shown in

Table 1. Parameter values for CSD10 steel-reinforced concrete columns.

Parameter	Parameter Value	Parameter	Parameter Value
$D$	900 mm	$f_c$	27.5 $\mathrm{M}\mathrm{P}\mathrm{a}$
$d$	872 mm	${\mu }_c$	0.2
$t$	14 mm	$E_s$	20,600 $\mathrm{M}\mathrm{P}\mathrm{a}$
$E_c$	32,500 $\mathrm{M}\mathrm{P}\mathrm{a}$	${\mu }_s$	0.3
$e_0$	0.75	$\eta$	1.1

.

Table 2. Debonding detection results of core concrete in the CFST columns of the building in Shenzhen.

Column Location (Floor)	Concrete Column Debonding Rate Within the Floor (Perimeter Ratio %)	Debonding Distance (mm)
71	66.29	0.05
70	66.08	—
63	97.91	0.05
49	99.03	0.1
34	95.83	0.05–0.2
19	95.45	—
−2	21.04	0.1

shows the results of debonding testing on steel-reinforced concrete columns. The debonding situation in the middle floor of the building is relatively serious, with a debonding rate of over 95% from the 19th floor to the 63rd floor. However, the debonding distance of the CFST columns is relatively small, ranging from 0.05 mm to 0.2 mm.

When the debonding distance $\delta$ remains constant at 2 mm, the parameters of the CSD10 column in Table 1 are substituted into Formula (1), and the equivalent stiffness $B$ of the cross-section varies with the debonding angle $\theta$ as shown in Figure 5a. As shown in Figure 5a, when the debonding angle $\theta$ changes from 0 to 360°, the equivalent stiffness $B$ of the cross-section decreases from 1.806 × 10¹⁵ ${\mathrm{N}·\mathrm{m}\mathrm{m}}^2$ to around 1.6 × 10¹⁵ ${\mathrm{N}·\mathrm{m}\mathrm{m}}^2$, with a decrease of about 11%. The result indicates that as the debonding angle increases, the equivalent stiffness of the cross-section gradually decreases.

Figure 5b shows the relationship between the calculated section stiffness $B$ of the steel–concrete column and the debonding distance $\delta$ when the debonding angle $\theta$ is 180°. As shown in the figure, when the debonding angle $\theta$ is constant, the debonding distance $\delta$ changes from 0.1 mm to 5 mm, and the section stiffness $B$ decreases from 1.806 × 10¹⁵ ${\mathrm{N}·\mathrm{m}\mathrm{m}}^2$ to 1.8035 × 10¹⁵ ${\mathrm{N}·\mathrm{m}\mathrm{m}}^2$, with a variation amplitude of 0.138%. This indicates that when the debonding angle is constant, the effect of debonding distance on section stiffness is relatively small.

The value of stiffness for CFST columns, as composite structures, has always been somewhat uncertain. Significant differences in the calculation methods for structural stiffness for CFST are observed in various countries. For example, in China’s Technical Code for CFST Structures (GB 50936-2014) [33], the axial compression stiffness $EA$, bending stiffness $EI$, and shear stiffness $GA$ of CFST are calculated by adding the stiffness of steel tubes and core concrete, as shown as follows:

$$EA=E_s{A}_s+E_c{A}_c$$

(2)

$$EI=E_s{I}_s+E_c{I}_c$$

(3)

$$GA=G_s{A}_s+G_c{A}_c$$

(4)

where G_s and G_c are the shear moduli of steel pipes and concrete, respectively; A_s and A_c are the cross-sectional areas of steel pipes and concrete, respectively.

After debonding occurs in the CFST column, the stress situation of the column changes. In the previous calculation, the ratio of the bending moment axial force borne by the column component is used to treat the section bending moment. When the bending moment M increases, the eccentricity e will also increase. However, given the different cross-sectional dimensions of the CFST column, the eccentricity (denoted as φ) is proposed to be used for comparative calculation between the section stiffness and bending stiffness, as shown in Figure 6, where the eccentricity is the ratio of the eccentricity to the section radius.

When debonding occurs in CFST, the bending stiffness can be calculated using the standard formula, namely, Equation (3), and the stiffness of the section without debonding defects can be used for verification and correction. According to Equation (1), when the eccentricity is 0.75, the calculation results of equivalent bending stiffness B and the calculation results EI according to the specifications are shown in

Table 3. Comparison of equivalent section stiffness and bending stiffness of CFST.

CFST Number	Bending Stiffness EI (N·mm2)	Equivalent Sectional Stiffness B (N·mm2)	Relative Error (%)
CSD10	1.80962 × 1015	1.80659 × 1015	−0.167
CSD11	1.88449 × 1016	1.89314 × 1016	0.459
CSD12	1.17781 × 1015	1.18321 × 1015	0.459
CSD13	4.26979 × 1015	4.29792 × 1015	0.659
CSD14	1.83516 × 1016	1.83637 × 1016	0.066
CSD15	1.40996 × 1016	1.40972 × 1016	−0.017
CSD16	4.58425 × 1015	4.65649 × 1015	1.576
CSD17	1.33582 × 1016	1.33094 × 1016	−0.366
CSD18	4.0155 × 1015	4.02414 × 1015	0.215
CSD19	1.15166 × 1015	1.15817 × 1015	0.566
CSD20	1.03838 × 1016	1.03839 × 1016	0.001

. When the debonding angle and distance are both zero, other CFST columns are substituted into the trial calculation. The calculated section stiffness B is in good agreement with the bending stiffness EI calculated using the standard, and the error of both is less than 2% for all columns.

2.3. Effect of Debonding on Moment Transfer in Steel Pipe–Concrete Columns

For determining the effect of CFST column debonding on the force transmission of a structural system, a frame is taken as an example to study the bending moment changes in a planar frame under lateral force in different working conditions (Figure 7). Four working conditions are set here: In working condition 1, the CFST columns on both sides of the structure are not debonded, simulating the healthy working state of the structure. Working conditions 2 and 3 have no debonding on one side but have a debonding distance of 2 mm on the other side, with debonding angles of 90° and 270°, respectively. Equation (1) is used to calculate the stiffness reduction in the corresponding steel–concrete columns to simulate the deterioration of the components in the structure, that is, partial debonding in the steel–concrete columns. The debonding side of working condition 4 is an empty steel pipe to simulate extreme debonding situations.

The same lateral force is applied to study the effect of debonding on structural moment transfer. Point a is the top of the left side non-debonded CFST lower column, point b is the bottom of the left side non-debonded CFST lower column, point c is the top of the right CFST lower column, and point d is the bottom of the right CFST lower column. The calculation results of the planar frame bending moment diagram are shown in Figure 8, and the bending moment changes at each point are listed in Figure 9. The calculation results demonstrate that after CFST debonding on one side of the planar frame, the bending moment borne by the CFST column on the non-debonding side gradually increases with an increase in debonding angle. When the debonding side is an empty steel tube, the bending moments borne by the top (point a) and bottom (point b) of the CFST column on the non-debonding side increase by 24.23% and 23.86%, respectively. As the debonding angle increases, the bending moment borne by the CFST column on the debonding side gradually decreases. When the debonding side is an empty steel tube, the bending moments at the top (point c) and bottom (point d) of the steel tube column decrease by 11.75% and 31.41%, respectively. Debonding changes the force transmission of the structure. As the debonding deterioration of the CFST column intensifies, the force on the debonding column gradually decreases, whereas the force on the non-debonding side gradually increases.

3. Finite Element Simulation of the Effect of CFST Column Debonding on Structural Dynamic Characteristics

3.1. Finite Element Simulation Method Using ABAQUS

ABAQUS is used to simulate the effect of debonding in CFST columns on structural dynamic characteristics, and solid elements (C3D8R) are employed for steel tubes, concrete core columns, steel beams, and floor slabs. On the basis of existing research, a secondary plastic flow model for steel, Han Linhai’s constitutive model for concrete [34], and Tie (binding connection) for beam–column connections are used to determine the constitutive relationship of materials.

The 69th and 70th floors of the building in Shenzhen are taken as examples. The frame is composed of CFST columns and steel beams. The outer radius of the steel tube in CFST is 650 mm, the wall thickness of the steel tube is 18 mm, the steel tube has a yield strength of 345 MPa, and the elastic modulus is 2.06 × 10⁵ $\mathrm{M}\mathrm{P}\mathrm{a}$. The Poisson’s ratio is 0.3. The radius of the concrete core column is 632 mm, with a concrete cubic compressive strength of 40 MPa, and the elastic modulus is 3.25 × 10⁴ $\mathrm{M}\mathrm{P}\mathrm{a}$. The Poisson’s ratio is 0.2, the heights of the 69th and 70th floors are 8324 and 3600 mm, respectively, and the sizes of the steel beams are 700 mm × 260 mm × 14 mm × 10 mm and 600 mm × 250 mm × 10 mm × 5 mm. The finite element model for the frame position and parameters is shown in Figure 10.

3.2. Finite Element Simulation of the Influences of Debonding Parameters on Structural Dynamic Characteristics

Keeping the debonding distance of 2 mm constant for steel-reinforced concrete columns, the effects of debonding angles of 0°, 90°, 180°, and 270° on the first three frequencies of structural dynamic characteristics were studied. The schematic diagram is shown in Figure 11, and the simulation results are shown in Figure 12. As shown in Figure 12a–c, the frequencies of the first three orders decrease with the increase in debonding angle, with the first order frequency decreasing from 5.3167 Hz to 5.2644 Hz, the second order frequency decreasing from 5.4508 Hz to 5.4200 Hz, and the third order frequency decreasing from 5.8586 Hz to 5.7667 Hz. As shown in Figure 12d, when the debonding angle is 270°, the change rates of the first three orders are 0.984%, 0.565%, and 1.569%, respectively.

Keeping the debonding angle constant at 180°, as shown in Figure 11c, the effect of varying debonding distance from 0.2 mm to 2 mm on the first three frequencies of the structure was studied. The simulation results are shown in Figure 13. From Figure 13, as the debonding distance of the steel–concrete column increases, the structural frequency gradually decreases. However, the influence of the debonding distance on the structural frequency is relatively small. Compared with the values when the debonding distance is 2 mm, the first-order frequency of the planar frame decreases by 0.141%, the second-order frequency decreases by 0.134%, and the third-order frequency decreases by 0.225%.

It is noted that the parametric study in Section 3 aims to investigate the general sensitivity of structural dynamic characteristics to debonding parameters (angle and distance) using a planar frame model, rather than to replicate the exact debonding conditions of the SEG Building. For this purpose, a conservative debonding distance of 2 mm is adopted to ensure that even small effects are captured. In contrast, the full-building simulation in Section 4 uses the actual debonding distances measured in the SEG Building, which range from 0.05 mm to 0.2 mm (Table 2). Using the analytical formula derived in Section 2.2 (Equation (1)) with a representative actual gap of 0.1 mm and a debonding angle of 180°, the predicted reduction in the equivalent section stiffness is approximately 0.007%, leading to a first-order frequency decrease of less than 0.01% at the frame level. This indicates that the direct effect of the debonding distance alone is negligible for the actual measured gaps. However, the debonding angle effect (e.g., 270° debonding over a significant angular extent) remains the dominant factor, and its influence is adequately captured by the simulations with the conservative 2 mm gap. Thus, the parametric results in Section 3 provide a clear understanding of the relative importance of debonding parameters, while the realistic assessment of the SEG Building is reserved for Section 4.

3.3. Simulation Method for Debonding “Thin Interlayers” in CFST Columns

A superhigh-rise structural system is relatively complex, with a large number of types of components. If solid or refined elements are used for finite element modeling, the time cost for modeling and calculation is relatively high. For saving computational time and cost, on the basis of the custom composite section function of Ansys finite element software, as shown in Figure 14a, Beam 188 or 189 elements are used to simulate the debonding distance through the thickness of the thin interlayer, and the debonding state of CFST is simulated through the properties of the thin interlayer material. When the material properties of the thin interlayer are completely consistent with the concrete, it is considered to be in a fully bonded state, as shown in Figure 14b. When the material properties of the thin interlayer are very small and close to 0, it can be equivalent to a completely annular debonding state, as depicted in Figure 14d. When the material properties of the thin interlayer fall between the above two situations, it can be equivalent to a semi-debonding state, as shown in Figure 14c.

A single column on the mast side of a tower on the 69th floor of the building in Shenzhen (Figure 15a) is extracted for analysis to verify the effectiveness of the above simulation methods and compare the changes in structural dynamic characteristics under different debonding states. The height of the column is 3600 mm, the outer radius of the steel pipe is 650 mm, the inner radius is 632 mm, the steel pipe has a yield strength of 345 MPa, the elastic modulus is 2.06 × 10⁵ $\mathrm{M}\mathrm{P}\mathrm{a}$, the density is 7850 $\mathrm{k}\mathrm{g}/{\mathrm{m}}^3$, and the Poisson’s ratio is 0.3. The core concrete has a cubic compressive strength of 40 MPa, the elastic modulus is 3.25 × 10⁴ $\mathrm{M}\mathrm{P}\mathrm{a}$, the density is 2500 $\mathrm{k}\mathrm{g}/{\mathrm{m}}^3$, and the Poisson’s ratio is 0.2. We plan to analyze three working conditions. Working condition 1 refers to the fully bonded working state of CFST columns. The thickness of the thin interlayer in working condition 2 is 2 mm, and the material properties are consistent with those of the concrete core column. The thickness of the thin interlayer in working condition 3 is 2 mm, and the material properties of the thin interlayer are assigned to simulate the debonding situation of the CFST columns. The parameter settings and thin interlayer materials for each working condition are shown in

Table 4. Column simulation working condition materials and dimensional properties.

Working Condition	Outer Steel Tube		Thin Sandwich		Core Concrete
	Dimension (mm × mm)	Material	Thickness (mm)	Material	Radius (mm)	Material
Condition 1	D1300 × 18	Q345	——	——	632	C40
Condition 2	D1300 × 18	Q345	2	C40	632	C40
Condition 3	D1300 × 18	Q345	2	E = 3 × 10−20 MPa μ = 0.3 × 10−20 ρ = 7.85 × 10−20 kg/m3	630	C40

. Given that the complete debonding of the ring affects the structural force transmission, the strength of the core concrete is reduced by 5% for the simulation. The feasibility of the simulation method is verified through different working conditions. The finite element model is shown in Figure 15b.

The frequencies of the first six orders of the CFST column components are shown in

Table 5. Comparison of frequency calculation results.

Rank Condition	Condition 1 (Hz)	Condition 2 (Hz)	Rate of Change (%)	Condition 3 (Hz)	Rate of Change (%)
1	50.680	50.373	−0.606	49.317	−2.689
2	50.680	50.373	−0.606	49.317	−2.689
3	143.64	142.73	−0.634	144.03	0.272
4	233.30	231.90	−0.600	211.98	−9.138
5	233.30	231.90	−0.600	211.98	−9.138
6	263.05	261.45	−0.608	263.17	0.046

. The comparison of the analysis results shows that the calculated results of condition 2 are similar to those of condition 1, with a maximum error of 0.64% due to the small-element-size error in the section grid division. This similarity demonstrates the feasibility of dividing thin interlayers. Compared with condition 1, condition 3 shows a decrease in the frequency of the components after complete detachment, and the amplitude of frequency changes in the bending and torsion modes increases with an increase in order.

To verify the influence of debonding on the dynamic characteristics of a structural system, this study extracts a frame on the mast side of the 69th and 70th floors of the tower in the SEG Building for analysis. Two working conditions are proposed, with case 1 being a fully bonded frame and case 2 being a frame presenting circular debonding with thin interlayers. A 5% reduction in the stiffness of the core concrete is considered for simulation. The frequency calculation results of the first six orders of the framework are shown in

Table 6. Comparison of the calculation results of the spatial frame frequency.

Frequency Working Condition	Condition 1 (Hz)	Condition 2 (Hz)	Rate of Change (%)	Vibration Mode
1	4.7371	4.6704	−1.408	X-direction first bending
2	5.0269	4.9517	−1.496	Y-direction first bending
3	6.9924	6.8824	−1.573	First torsion
4	22.221	21.439	−3.519	Y-direction second bending
5	23.382	22.524	−3.669	X-direction second bending
6	29.204	28.209	−3.407	Second torsion

, and the vibration mode results are illustrated in Figure 16. The comparison of the analysis results shows that after the CFST column is debonded, the dynamic characteristics of the structural system change, and the frequency decreases. As the frequency order increases, the amplitude decreases gradually. After the CFST column is fully debonded, the core concrete and the outer steel tube do not work together, and the three-dimensional confining pressure state disappears, resulting in decreases in stiffness and structural frequency.

4. Effect of CFST Column Debonding on Real Superhigh-Rise Structures

4.1. Effect of Debonding Formation in CFST Columns on Structural Dynamic Characteristics

A certain building in Shenzhen, the SEG Building, located at the intersection of Shennan East Road and Huaqiang North Road, as shown in Figure 17a,b, is considered for example. It has 72 floors above ground level with a height of 291.6 m, and four floors below ground level with a depth of 21.15 m. The tower of the building adopts a 43.2 m × 43.2 m square octagonal plan, which is a landmark building in Shenzhen that spans a century. The ETABS model of the building is shown in Figure 17c. The structure of the building adopts a CFST column steel beam frame–core tube structure system, and the concrete shear wall of the core tube is a frame tube composed of steel beams and CFST columns [35].

Based on the ETABS software, two methods were used to study the effects of simulated debonding of steel–concrete columns on building dynamic characteristics and engineering parameters. Method 1 is based on the theory derived in Section 2.2 to reduce the equivalent stiffness of steel–concrete columns. On the basis of the detection data in Table 1, theoretical calculations are conducted on the CFST columns of the building. Given that the debonding area of the CFST columns on the middle floor is about 95%, a debonding rate of 95% is considered for theoretical calculation. The calculation results show that the bending stiffness of the CFST columns decreases by about 13.2%. In this part of the simulation, 15% is taken as the reduction value of the bending stiffness of the CFST columns, which tends to be safe. Method 2 uses the thin interlayer theory in Section 3.3 to model and assign material parameters to the debonding layer. Because of the weakening of the collaborative work ability between the steel pipe and the concrete core due to debonding, the stiffness of the core column is reduced by 5% [19]. The thin interlayer is shown in Figure 18.

The effect of CFST column debonding on the first 11 frequencies of the SEG Building is shown in

Table 7. Comparison of the effect of debonding in CFST on the frequency of the building.

Rank	Intact Design	Stiffness Reduction	Reduction in Frequency	Thin Sandwich	Reduction in Frequency
	(Hz)	(Hz)	(%)	(Hz)	(%)
1	0.1723	0.1706	0.987	0.1708	0.871
2	0.1811	0.1792	1.049	0.1795	0.883
3	0.3807	0.3798	0.236	0.3805	0.053
4	0.6727	0.6682	0.669	0.6692	0.520
5	0.7578	0.7526	0.686	0.7538	0.528
6	1.0396	1.0372	0.231	1.0389	0.067
7	1.3397	1.3331	0.493	1.3353	0.328
8	1.4718	1.4670	0.326	1.4693	0.170
9	1.5802	1.5711	0.576	1.5737	0.411
10	1.9829	1.9747	0.414	1.9789	0.202
11	2.1456	2.1350	0.494	2.1389	0.312

. According to Table 7, after the CFST column is debonded, the effect on the structural frequency is relatively small, and the first-order frequency is reduced by less than 1%. The fourth-order bending-torsion coupling in the Y-direction is reduced by less than 0.5%, and the results of the two simulation methods are relatively close, which verifies the validity of the simulation methods.

It should be noted that the measured first-order frequency reduction in the SEG Plaza after 20 years of service is approximately 6.1%, whereas the numerical simulation considering only CFST column debonding predicts a reduction of about 0.987% (Table 7). This discrepancy indicates that debonding alone does not fully account for the observed frequency degradation. Other factors accumulated during long-term service may also contribute, including in-plane stiffness degradation of composite floor slabs, joint connection damage, concrete creep and shrinkage, and variations in live load.

4.2. Effect of Debonding Formation in CFST Columns Under Response Spectrum on Structural Engineering Design Parameters

The changes in the design parameters of the real superhigh-rise structure before and after the debonding of CFST columns are compared by conducting a response spectrum analysis on the structure. The debonding of CFST columns leads to a decrease in the stiffness of the structure, and the changes in the stiffness–weight ratio before and after detachment are shown in

Table 8. Comparison of the influence of CFST debonding on the stiffness–weight ratio of the building.

	EJ (×1011 kN·m2)		G (×104 kN)	Stiffness–Weight Ratio		Rate of Change (%)
	No Debonding	Debonding		No Debonding	Debonding
X-direction	3.73	3.64	262.46	1.78	1.74	2.24
Y-direction	3.42	3.35		1.64	1.60	2.08

. According to Table 8, after the CFST column is debonded, the structural stiffness–weight ratio decreases. The decreases in X-direction and Y-direction are about 2%, both exceeding the specification limit of 1.4, but both are less than 2.7. Thus, second-order effects need to be considered. According to Clause 5.4.4 of the Technical Code for Concrete Structures of Tall Buildings (JGJ 3-2010) [36], a stiffness–weight ratio greater than 1.4 ensures structural stability against gravity second-order effects; Clause 5.4.1 of the same code specifies that when the ratio is less than 2.7, second-order effects should be considered in structural analysis.

Interstory displacement angle is mainly used to limit the horizontal displacement of a structure under normal use to meet the requirements of the structural bearing capacity and stability. The comparison of interstory displacement angles before and after the debonding of CFST is shown in

Table 9. Comparison of the influence of CFST debonding on the interlayer displacement angle of the building.

Story	X-Direction (×10−4)			Y-Direction (×10−4)
	No Debonding	Debonding	Rate of Change (%)	No Debonding	Debonding	Rate of Change (%)
71	7.92	8.08	2.02	7.81	7.98	2.18
70	8.06	8.22	1.99	8.23	8.38	1.82
60	7.89	8.03	1.77	8.24	8.38	1.70
50	7.69	7.82	1.69	7.97	8.10	1.63
40	7.89	8.00	1.39	8.48	8.59	1.30
30	7.36	7.45	1.22	8.07	8.16	1.12
20	5.82	5.89	1.20	6.29	6.35	0.95
10	3.76	3.79	0.80	4.10	4.13	0.73
1	0.63	0.63	0.00	0.77	0.77	0.00

and Figure 19. According to Table 9, after debonding of CFST, the interstory displacement angle of the structure increases; the higher the floor, the greater the change in interstory displacement angle. Specifically, the displacement-angle-change rate at the 71st floor in the Y-direction is highest, reaching 2.18%. However, after debonding, the interstory displacement angle between the X- and Y-direction floors is considerably less than the limit value of 0.002, which meets the requirements of structural safety and use.

The torsional displacement ratio of structures in engineering design is the main basis for determining whether torsional irregularities exist in the structure and for taking targeted measures. The torsional displacement of the building in Shenzhen before and after the detachment of the CFST is shown in

Table 10. Changes in torsional displacement ratio before and after CFST debonding, considering accidental eccentricity in the building in Shenzhen.

Story	X-Direction			Y-Direction
	No Debonding	Debonding	Rate of Change (%)	No Debonding	Debonding	Rate of Change (%)
70	1.0378	1.0372	−0.0611	1.0335	1.0330	−0.0536
60	1.0363	1.0357	−0.0643	1.0358	1.0352	−0.0569
50	1.0395	1.0388	−0.0671	1.0388	1.0382	−0.0592
40	1.0435	1.0427	−0.0690	1.0430	1.0423	−0.0599
30	1.0483	1.0475	−0.0696	1.0482	1.0476	−0.0593
20	1.0531	1.0524	−0.0669	1.0560	1.0553	−0.0588
10	1.1225	1.1212	−0.1101	1.1390	1.1383	−0.0607
9	1.1268	1.1255	−0.1122	1.1358	1.1352	−0.0536
8	1.1311	1.1298	−0.1148	1.1331	1.1325	−0.0491
7	1.1358	1.1344	−0.1174	1.1297	1.1292	−0.0445
6	1.1408	1.1394	−0.1199	1.1258	1.1253	−0.0401
5	1.1463	1.1449	−0.1223	1.1213	1.1209	−0.0360
4	1.1525	1.1510	−0.1245	1.1163	1.1159	−0.0321
3	1.1595	1.1580	−0.1260	1.0984	1.0981	−0.0258
2	1.1679	1.1664	−0.1243	1.0591	1.0589	−0.0138
1	1.1716	1.1704	−0.1028	1.1029	1.1024	−0.0471

. Table 10 demonstrates that the torsional displacement ratio of the structure decreases after the CFST column is detached. Nonetheless, the decrease is insignificant, and the change rate is within 0.15%, which meets the specification requirements.

The SEG Building in Shenzhen has a dual-bias situation of tower and mast. The reason for the decrease in the torsional displacement ratio of the structure is that the stiffness of the structure decreases after the debonding of CFST, resulting in a slight change in the central position of the structure’s stiffness. The relative distance between the center of mass and the center of stiffness decreases, and the torsional effect of the overall structure weakens, leading to a decrease in the torsional displacement ratio of the structure. The changes in the center of mass and stiffness position of the structure before and after the CFST column is debonded are shown in

Table 11. Comparison of the changes in the positions and relative distances of the rigid and mass centers of the building in Shenzhen before and after debonding.

Story	Relative Position of the No Debonding Column (m)			Relative Position of the Debonding Column (m)			Relative Distance Change Rate (%)
	Rigid Center	Mass Center	Relative Distances	Rigid Center	Mass Center	Relative Distances
1	231.080	243.091	−12.011	231.080	243.046	−11.966	−0.380
2	242.556	249.938	−7.381	242.556	249.932	−7.376	−0.069
3	239.421	250.068	−10.647	239.421	250.060	−10.639	−0.073
4	237.520	250.092	−12.572	237.520	250.083	−12.563	−0.072
5	237.520	250.082	−12.562	237.520	250.073	−12.553	−0.077
6	237.520	250.071	−12.551	237.520	250.061	−12.541	−0.078
7	237.571	250.064	−12.494	237.571	250.055	−12.484	−0.077
8	237.567	250.064	−12.497	237.567	250.055	−12.488	−0.073
9	237.768	250.073	−12.305	237.768	250.065	−12.297	−0.070
10	237.521	250.099	−12.578	237.521	250.091	−12.571	−0.061

.

Taking the X-direction as an example, this study investigates the effect of CFST column debonding on the shear–weight ratio of real superhigh-rise structures. The calculation results are shown in

Table 12. Comparison of the influence of CFST debonding on the X-direction shear–weight ratio of the building.

Story	Gravity	No Debonding Model		Debonding Model		Rate of Change (%)
	(kN)	Shear (kN)	Shear–Weight Ratio	Shear (kN)	Shear–Weight Ratio
71	32,034.869	2669.278	0.0833	2663.812	0.0830	−0.2046
70	59,121.137	4536.282	0.0767	4523.412	0.0765	−0.2837
60	255,087.011	8265.374	0.0324	8213.023	0.0322	−0.6334
50	483,053.975	10,834.072	0.0224	10,765.489	0.0223	−0.6331
40	708,010.076	13,119.416	0.0185	13,028.384	0.0184	−0.6939
30	939,075.800	14,932.350	0.0159	14,830.143	0.0158	−0.6845
20	1,168,069.277	16,558.437	0.0142	16,446.681	0.0141	−0.6749
10	1,454,072.497	18,814.933	0.0129	18,694.551	0.0129	−0.6398
1	2,008,761.936	25,115.373	0.0125	24,973.088	0.0124	−0.5665

. According to Table 12, after the CFST is debonded, the shear–weight ratio of the structure slightly decreases, but the change is insignificant. The change rate of each floor is within 0.7%, which is greater than the limit value of 0.012 required by the specification, meeting the requirements.

4.3. Effect of CFST Column Debonding on Seismic Design Parameters Under Frequent Earthquakes

For superhigh-rise building structures, in addition to response spectrum analysis, the mechanical properties and engineering design parameters under seismic action should also be verified. In this section, with the X-direction as an example, EI Centro waves are selected, Taft waves and Joshua waves are subjected to time-history analysis, and the seismic waves are shown in Figure 20. In accordance with Article 5.1.2 of China’s “Code for Seismic Design of Buildings” (GB50011-2010) [37], the maximum value of seismic acceleration time history is selected for time-history analysis, as shown in

Table 13. Maximum seismic acceleration time history used in the time-history analysis (cm/s²).

Seismic Effect	Degree 6	Degree 7	Degree 8	Degree 9
Frequent earthquake	18	35 (55)	70 (110)	140
Rare earthquake	125	220 (310)	400 (510)	620

. A degree 7 frequent earthquake (35 cm/s²) is used for value analysis, and the seismic waves are amplitude-modulated in accordance with the requirements of the code.

The simulation method for stiffness reduction in Section 4.1 is used to simulate the detachment of CFST in ETABS for time-history analysis. The changes in displacement, base shear force, and overturning moment of the real superhigh-rise structure before and after debonding of CFST columns are obtained and compared.

The vertex displacement time history is an important indicator for controlling the overall displacement of building structures, reflecting the horizontal displacement of the structures under earthquake action. In the process of analyzing the vertex displacement time history, the maximum displacement of floors under different seismic waves is first analyzed, and then the vertex displacement time history of the building in the SEG Building is studied. The original model is compared with the debonding model, and the analysis results are shown in Figure 21 and

Table 14. Comparison of the maximum vertex displacement change in the building in Shenzhen before and after CFST debonding.

Seismic Wave Working Condition		EI Centro Wave		Taft Wave		Joshua Wave
		Time (s)	Displacement	Time (s)	Displacement	Time (s)	Displacement
Maximum vertex displacement (mm)	No debonding	53.1	69.90	49.5	129.34	47.7	95.2
		44.5	−63.57	46	−132.84	44.4	−102.13
		Amplitude	69.90	Amplitude	132.84	Amplitude	102.13
	Debonding	48.2	71.26	49.1	132.83	36.8	97.98
		44.7	−66.26	46.3	−134.07	39	−106.20
		Amplitude	71.26	Amplitude	134.07	Amplitude	106.20
	Rate of change (%)	Positive	1.95%	Positive	2.70%	Positive	2.92%
		Negative	4.23%	Negative	0.93%	Negative	3.99%
		Amplitude	1.95%	Amplitude	0.93%	Amplitude	3.99%

.

According to Table 14, the maximum displacement of the structural vertices of the building under seismic waves increases after the CFST column is debonded. Under the action of Taft waves, the moment of maximum displacement of the structural vertex before and after the debonding of the steel–concrete column is relatively close, and the amplitude of the maximum displacement of the vertex increases by 0.93%. Under the action of Joshua waves, the time of maximum displacement of the vertex changes, and the amplitude increases by 3.99%. Under the action of EI Centro waves, the maximum displacement of the unidirectional vertex increases by 4.23%, and the amplitude increases by 1.95%. Under the action of seismic waves, the debonding of steel–concrete columns leads to an increase in vertex displacement of the structure, with increments all within 5%.

The base shear force under X-direction seismic waves can be obtained, as shown in Figure 22 and

Table 15. Comparison of base shear force change in the building in Shenzhen before and after CFST debonding.

Seismic Wave Working Condition		EI Centro Wave		Taft Wave		Joshua Wave
		Time (s)	Force	Time (s)	Force	Time (s)	Force
Base shear force (kN) ×103	No debonding	41.4	71.74	45.3	95.13	30.1	75.86
		35.3	−65.64	10.4	−85.92	36.3	−86.87
		Amplitude	71.74	Amplitude	95.13	Amplitude	86.87
	debonding	44.1	77.43	50.8	91.42	44.8	92.84
		54	−81.98	48.5	−105.75	48.5	−91.18
		Amplitude	81.98	Amplitude	105.75	Amplitude	92.84
	Rate of change (%)	Positive	7.35	Positive	−3.90	Positive	22.38
		Negative	24.89	Negative	23.08	Negative	4.96
		Amplitude	14.27	Amplitude	10.98	Amplitude	6.87

. The calculation indicates that the base shear force of the building in Shenzhen increases under the action of seismic waves after the CFST column is debonded. Under the action of EI Centro and Joshua waves, the amplitude of the base shear force before and after the debonding (the absolute value of the base shear force under the action of seismic waves) increases. Under the action of EI Centro waves, the maximum change rate of the forward base shear force of the structure before and after debonding is 24.89%, and the maximum change rate of amplitude is 14.27%. The debonding increases the base shear force of the structure under seismic action.

Under the Taft wave, the positive base shear force decreases by 3.9% after debonding, while the negative base shear force increases by 23.08%, and the absolute amplitude increases by 10.98%. This is attributed to the asymmetric stiffness degradation induced by debonding, which makes the positive and negative responses no longer symmetric. The increase in absolute amplitude indicates an overall rise in seismic demand, consistent with the reduction in structural stiffness.

Overturning moment is an important factor that must be considered in the design of high-rise structures. Reasonable design and measures can effectively resist overturning moment, ensuring the safety and stability of buildings. The overturning moment of the building in Shenzhen under X-direction seismic waves can be obtained and compared through time-history analysis, as shown in Figure 23 and

Table 16. Comparison of base overturning moment changes in the building in Shenzhen before and after CFST debonding.

Seismic Wave Working Condition		EI Centro Wave		Taft Wave		Joshua Wave
		Time (s)	Moment	Time (s)	Moment	Time (s)	Moment
Base overturning moment (kN·m) ×103	No debonding	51.2	331.27	54.1	412.49	46.9	380.05
		46.4	−292.54	57.2	−418.47	43.8	−378.88
		Amplitude	331.27	Amplitude	418.47	Amplitude	380.05
	Debonding	51.4	307.21	54.9	514.82	34.8	466.13
		39	−340.02	51.3	−535.54	44.8	−439.51
		Amplitude	340.02	Amplitude	535.54	Amplitude	466.13
	Rate of change (%)	Positive	−7.26	Positive	24.81	Positive	22.65
		Negative	16.23	Negative	27.98	Negative	16.00
		Amplitude	2.64	Amplitude	27.98	Amplitude	22.65

. Based on the calculation, the overall amplitude of the overturning moment of the SEG Building under seismic wave action increases after the CFST column is debonded. Under the action of Taft and Joshua waves, the overturning moment before and after the debonding increases. The overturning moment of the structure under the action of Taft waves varies greatly, with a maximum difference of 27.98% before and after debonding. Debonding increases the overturning moment of the structure under seismic action.

5. Conclusions

This study provides a mechanistic understanding of how localized debonding in CFST columns affects the global dynamic characteristics of super high-rise buildings. The main novelties include the analytical stiffness formula that explicitly incorporates the debonding angle and distance, and the thin-interlayer simulation method that enables efficient full-building dynamic analysis. Although the research is motivated by the 5·18 incident of the SEG Plaza, the proposed theoretical framework and numerical methodology are general in nature and can be applied to evaluate the dynamic characteristic degradation of other high-rise buildings employing CFST columns.

(1)

The effect of the debonding angle and the debonding distance on the bending stiffness of steel–concrete columns is theoretically derived. The structural stiffness gradually decreases as debonding defects develop in steel–concrete columns, affecting the structural force transmission and decreasing the distributed bending moment of the debonding column. The structural frequency steadily decreases with an increase in the debonding angle of the CFST column. The increase in the column debonding distance slightly decreases structural frequency.

(2)

The influences of debonding parameters on structural dynamic characteristics are studied through finite element simulation. The debonding angle of CFST has a significant effect on structural frequency. When the CFST column in a planar frame is debonded at 270°, the structural frequency decreases by 0.984%. When the debonding angle is fixed at 180°, the first-order frequency of the planar frame containing a 2 mm debonding distance CFST column with a clearance decreases by 0.141%. When the clearance distance of the CFST column is small, the influence of the clearance distance on the structural frequency is relatively small. Ansys is used for finite element simulation. The feasibility of the simulation method is demonstrated through the simulation of a single column and a spatial frame.

(3)

The effect of CFST column debonding on a real high-rise structure in Shenzhen is studied. After the CFST column debonding, the first-order frequency of the structure decreases by 0.987%, which verifies the effectiveness of the “thin interlayer” simulation method for CFST column debonding. The debonding defects increase the vertex displacement, base reaction force, overturning moment, and floor shear force of the structure, indicating that the stiffness of the steel tube concrete structure decreases after debonding. The maximum difference in the overturning moments before and after the debonding of the structure under the Taft wave reaches 27.98%, proving that the debonding of steel-reinforced concrete is detrimental to the seismic resistance of structures.