Structures 65 (2024) 106743

Contents lists available at ScienceDirect

Structures
journal homepage: www.elsevier.com/locate/structures

Self-centering friction beam-column joint: A promising approach to seismic

and progressive collapse resilience
Xiao Lu *, 1 , Hang Xu , Longhe Xu
School of Civil Engineering, Beijing Jiaotong University, Beijing 100044, China

A R T I C L E I N F O A B S T R A C T

Keywords: Multi-hazard resilience is currently at the forefront of research in civil engineering. Progressive collapse and
Self-centering beam-column substructures seismic collapse are two typical failure scenarios that buildings may suffer during their service life. One prom­
Progressive collapse ising approach to enhancing seismic resilience of frames is the use of self-centering beam-column joints.
Mid-column failure
Although these joints offer significant seismic benefits, their resistance mechanism against progressive collapse
Numerical simulation
Parameter analysis
remains unclear. To address this issue, this paper focuses on the high seismic resilience self-centering friction
frame and establishes a numerical analysis model for self-centering friction beam-column substructures (SCFS),
validated by experimental data. Then, the progressive collapse process and resistance mechanism of SCFS under
the scenario of mid-column failure are investigated. Finally, the influence of design parameters on the pro­
gressive collapse resistance of SCFS are also discussed. Results demonstrate that the load-carrying capacity of
SCFS is predominantly achieved through the flexural action and catenary action. When compared to ordinary
reinforced concrete substructures (ORCS) of identical dimensions and reinforcement, the load-carrying capacity
of SCFS at combined flexural action and compressive arch action phase is 13.5 % lower than that of ORCS, while
the load-carrying capacity at catenary action phase is 42.5 % higher than that of ORCS. Additionally, the span-to-
depth ratio of the beam, reinforcement ratio of the intact beam section, and activation force of the self-centering
connector have the most significant impact on the progressive collapse resistance of SCFS.

1. Introduction compressive membrane action of the slab. While under large de­
formations, the tensile membrane action and CA can provide significant
Progressive collapse refers to the propagation of initial localized resistance. Qian et al. [8] emphasized that the progressive collapse
damage caused by unexpected disaster loads in a structural system, mechanism of corner-column frame substructures is predominantly
resulting in disproportionate collapse or even total structural failure [1]. defined by FA. Furthermore, Feng et al. [9] and Qian et al. [10] high­
Once progressive collapse occurs, it poses a serious threat to human life lighted the substantial contributions of floor slabs and infill walls to the
safety and leads to significant property damage. Consequently, miti­ progressive collapse resistance of reinforced concrete frames. Although
gating the risk of progressive collapse in building structures has become static tests can effectively study the resistance evolution mechanism,
a matter of paramount concern. they cannot capture the dynamic response that accompany actual col­
Frames have been extensively examined in progressive collapse lapses. Yang et al. [11], Zhao et al. [12], Qian et al. [13], Pham et al.
research, involving experiments and simulations on various failure [14] and Panahi et al. [15] studied the dynamic response of sub­
scenarios, including mid-column, edge-column, and corner-column structures of reinforced concrete frames during column removal and
failures. For instance, Qian et al. [2], Lin et al. [3] and Huang et al. explored the dynamic amplification factor of components. Building
[4] pointed out that the progressive collapse mechanisms of frames upon these accomplishments, scholars have systematically proposed
include the flexural action (FA), compressive arch action (CAA), and design methods for progressive collapse resistance, such as the tie force
catenary action (CA). Weng et al. [5], Yu et al. [6] and Pang et al. [7] method, alternative path method, and enhanced local resistance
revealed that under small deformations, the slab-column structure method.
mainly relies on the bending capacity of the floor slab and the On the other hand, earthquake-induced collapse during the service

* Corresponding author.
E-mail address: [email protected] (X. Lu).
1
ORCID: 0000-0003-1395-7701

https://doi.org/10.1016/j.istruc.2024.106743
Received 21 November 2023; Received in revised form 21 May 2024; Accepted 9 June 2024
Available online 14 June 2024
2352-0124/© 2024 Institution of Structural Engineers. Published by Elsevier Ltd. All rights are reserved, including those for text and data mining, AI training, and
similar technologies.
X. Lu et al. Structures 65 (2024) 106743

life of buildings constitutes a pivotal factor affecting structural safety. 2. The proposed SCFS
The escalation of socio-economic standards has heightened expectations
for seismic performance, prompting a shift in seismic design objectives 2.1. Construction detail of SCFS
from "collapse prevention " to "rapid post-earthquake recovery" [16].
Consequently, novel structural systems, like self-centering frames, have The construction details of the self-centering friction frame are
been proposed to enhance their seismic resilience [17–20]. For instance, shown in Fig. 1. The end of the frame beam features a reduced section
Mokhtarnejad et al. [17] introduced self-centering capabilities in that provides sufficient shear and limited bending capacity. The SFCs are
beam-column joints through strategic gaps at beam ends, substituting symmetrically installed on the upper and lower surfaces of the beam’s
steel bars with shape memory alloy bars. Li et al. [18], Cai et al. [19], reduced section, providing the main flexural resistance. As the SFC
and Ye et al. [20] utilized prestressed tendons passing through possesses both self-centering and energy dissipation capabilities, the
beam-column joints to assemble prefabricated concrete beams and col­ entire beam-column joint exhibits high self-centering capacity, along
umns into an integral unit. They also added Z-shaped steel plates, angle with stable strength and energy dissipation capabilities. The SFC con­
steel, and replaceable metal damper at the beam ends to improve the sists of a T-shaped central plate, a rectangular central plate, upper and
energy dissipation capacity of the joints, thereby reducing residual lower cover plates, several assembled disc springs, and high-strength
deformation and damage to the joints. However, research on the pro­ bolts. The central and cover plates have complementary grooves and
gressive collapse of these innovative structural systems remains limited. are assembled using high-strength bolts and preloaded disc springs. Due
Furthermore, seismic and progressive collapse designs impose distinct to the presence of slot holes on the central plate, the central and cover
demands on components. Seismic design of frames prioritizes resisting plates can slide within the grooves. The vertical component of the
global horizontal forces, while progressive collapse design ensures the groove motion compresses the disc springs, increasing the contact
structure can maintain stability even after local vertical load-bearing pressure between the plates’ contact surfaces, thereby increasing fric­
elements fail, relying on the remaining elements to redistribute both tion and providing sufficient energy dissipation capacity. During
vertical and horizontal forces. Some scholars even pointed out that the unloading, the internal pressure of the disc springs causes the central
ductility design in seismic design may weaken the progressive collapse plate to slide back to its initial position.
resistance of structures after losing columns [21,22]. In particular, these The SFCs bear approximately 80 % of the total moment at the beam
self-centering joints weaken the overall integrity of the frames, which end. Under design basis earthquakes, the entire frame primarily dissi­
may be detrimental to the structure’s resistance to progressive collapse. pates energy through the SFCs, while under maximum considered
Therefore, it is essential to conduct progressive collapse research on earthquakes (MCE), the damage to the frame beams is concentrated in
these self-centering frames with high seismic resilience. the reduced sections. The self-centering forces provided by the SFCs
In previous research, a self-centering friction beam-column joint was result in minimal post-earthquake residual deformation of the frame,
proposed by the authors [23], which significantly improves seismic and repairing the concrete cracks in the reduced sections can quickly
performance. However, its progressive collapse resistance mechanism accomplish the restoration process. The maximum seismic design
remains unclear and requires further investigation. Therefore, based on deformation of the SFC is established based on the structure’s maximum
the previous experimental work on the seismic performance of allowable story drift ratio under the MCE (adopted as 3.5 % in this
self-centering friction beam-column joints, this study conducts numeri­ study). Under MCE conditions, the SFC is expected to function normally
cal investigation on self-centering friction beam-column substructure without sustaining damage. However, for earthquake events exceeding
(SCFS) under mid-column failure scenarios. Firstly, refined finite MCE intensity, upon reaching the design deformation limit, yielding
element models (FEM) are established for one SCFS and one ordinary may occur in the center or cover plates of the SFC, necessitating post-
reinforced concrete substructure (ORCS). The accuracy of the simulation earthquake replacement.
method is verified using experimental data. A detailed comparative
analysis of the damage evolution process, deformation characteristics, 2.2. Seismic performance of the self-centering friction beam-column joint
and resistance mechanism of SCFS under mid-column failure scenarios is
carried out. Through parameter analysis, the effects of design parame­ To investigate the seismic performance of the self-centering friction
ters such as the activation force, maximum seismic design displacement beam-column joint (SCFJ), cyclic loading tests are conducted on an SCFJ
and post-activation stiffness of self-centering frictional connector (SFC), and an ordinary reinforced concrete joint (ORCJ). The main dimensions
span-to-depth ratio of beams, reinforcement of the beam, and concrete and reinforcement details of the specimens are shown in Fig. 2, and
strength on the progressive collapse resistance of SCFS are studied. additional detailed information can be found in reference [23]. The
frictional force developed within the SFC is directly proportional to the
initial bolt pre-tightening force and the friction coefficient of the con­
tacting surfaces between the center and cover plates. The experimentally
determined friction coefficient is approximately 0.13. Each SFC

Fig. 1. The construction details of the self-centering friction frame.

2
X. Lu et al. Structures 65 (2024) 106743

Fig. 2. The main dimensions and reinforcement details of the specimens.

incorporates 12 bolts pre-tensioned to a force of approximately 30 kN. ratio of the SCFJ is approximately 0.7 %, which is only 34 % of that for
These values are consistently utilized in the benchmark numerical model ORCJ. A negative correlation often exists between the number and width
for SFCs presented in Section 3. The hysteresis curves and damage of cracks in concrete members subjected to similar loading conditions
comparison of ORCJ and SCFJ are depicted in Fig. 3. Compared to the and having the same dimensions. This implies that fewer, wider cracks
ORCJ, the hysteresis curve of the SCFJ exhibits a flag-shaped behavior, may be observed in SCFJ compared to ORCJ. However, due to the
indicating excellent self-centering capability. Several studies confirm reduced section at the beam end of SCFJ, the bending moment it can
lower energy dissipation in self-centering components compared to RC sustain is significantly lower than that of the intact section in ORCJ.
counterparts [24]. This is evident in Fig. 3. Initially, the SCFJ specimen Consequently, the crack width in SCFJ is not larger than that of ORCJ.
outperforms the ORCJ specimen. However, exceeding 1.5 % drift, the The maximum crack widths for SCFJ and ORCJ are presented in Table 1.
ORCJ specimen exhibits higher dissipation due to delayed steel yielding, Crack width measurements are taken when the displacement load at
offering greater plastic dissipation than the frictional mechanism in the each level reached its peak amplitude. These data clearly demonstrate
SCFJ specimen. Existing research demonstrates that increased post-yield that the crack width of SCFJ is smaller than that of ORCJ under equiv­
stiffness of self-centering components or structures can significantly alent displacement conditions. Furthermore, Fig. 3b illustrates that the
reduce the displacement response during an earthquake [25]. This number of cracks in SCFJ is significantly lower compared to ORCJ. More
translates to a potential for lower energy dissipation requirements in detailed comparison of damage can be found in the literature [23].
self-centering structures compared to conventional concrete structures These findings collectively indicate that SCFJ experiences less severe
for mitigating displacement. Therefore, even though the energy dissi­ damage, primarily localized in the designated reduced section of the
pation of the SCFJ specimen in this study at the maximum loading concrete beam.
displacement is 37 % lower than the ORCJ, it can still effectively reduce
the structure’s displacement response. Furthermore, for structures with
Table 1
higher energy dissipation demands, energy-dissipating steel bars [26] or
The maximum crack widths for SCFJ and ORCJ at different drift ratios.
friction dampers [27] can be incorporated at the joints to augment their
Drift ratio θ= 0.25 % θ= 1.5 % θ= 2.5 % θ= 3.5 %
energy dissipation capacity. Under the same applied displacement, the
residual displacements of the SCFJ are consistently smaller than those of Crack width ORCJ start to crack 1 3 5
the ORCJ. At the maximum tested drift ratio of 3.5 %, the residual drift (mm) SCFJ not cracked 0.5 1 4

Fig. 3. The hysteresis curves and damage comparison of ORCJ and SCFJ.

3
X. Lu et al. Structures 65 (2024) 106743

Seismic performance tests have demonstrated that the SCFJ effec­ beam is 3200 mm (250 mm × 400 mm cross-sectional). The reduced
tively mitigates beam-column damage, reduces residual displacements, length of the SCFS at the beam end is 440 mm, and the height of the
and facilitates rapid post-earthquake recovery. Nonetheless, it is crucial reduced section is 200 mm. The longitudinal steel bar diameter is
to acknowledge that these joints introduce weakening effects at the 16 mm, and the stirrup diameter is 8 mm. The stirrup spacing at the
beam ends, and the longitudinal bars experience discontinuity between beam ends is 50 mm, while for all other locations, it is 100 mm. The
the reduced and intact beam sections. The implications of these specific dimensions of the SFC is detailed in Fig. 4b. The yield strength of the
construction details on the progressive collapse resistance of frames steel plate is 345 MPa. The disc springs are combined in series, with each
remain ambiguous, underscoring the immediate need for dedicated group consisting of 7 pieces. The disc spring’s dimensions comprise a
research into the progressive collapse resistance capabilities of SCFS. 45 mm outer diameter, a 21 mm inner diameter, and a 5 mm thickness.
Additionally, the strength of the bolt is 900 MPa. Both specimens share a
3. Numerical modeling method concrete strength of 34.7 MPa, accompanied by a concrete elastic
modulus of 34,839 MPa. The 8 mm and 16 mm diameter steel bars
3.1. Specimen design exhibit yield strengths of 407 MPa and 436 MPa, along with ultimate
strengths of 575 MPa and 537 MPa, respectively. All these material
To investigate the progressive collapse resistance of the self- properties are obtained from the prior experiment [23] and directly
centering friction frame, a SCFS and an ORCS are designed and shown incorporated into the numerical simulation.
in Fig. 4, and only half of the specimens is displayed due to symmetry.
Informed by existing experimental data on progressive collapse, this
study employs a 1:2 scale SCFS specimen for numerical simulation. The 3.2. Element types and material constitutive models
design of the beams and columns adheres to the relevant Chinese
building codes, specifically the "Code for design of concrete structures" Refined FEMs for the SCFS and ORCS specimens are established
and the "Code for seismic design of buildings". To ensure comparability, using MSC.Marc general finite element software and shown in Fig. 5.
identical beam-column dimensions, materials, and longitudinal steel Concrete and steel plates are simulated using hexahedral solid element,
bars are employed, with the design bending moment at the beam ends while rebars and SFCs are simulated using two-node truss element.
essentially equal for both SCFS and ORCS. The edge-column length is Deformation compatibility is achieved through shared nodes among
1800 mm, mid-column length is 975 mm, and both have a concrete, rebars and SFCs. To capture the bond-slip behavior within the
400 mm × 400 mm cross-section. The clear span of the single-span plastic hinge region between rebars and concrete, the elastic modulus of
longitudinal rebars is appropriately adjusted to approximate the

Fig. 4. The dimensions and reinforcement of the specimens.

4
X. Lu et al. Structures 65 (2024) 106743

Fig. 5. The refined FEM of SCFS.

observed behavior [28,29]. Fig. 6. The equivalent tension force-displacement skeleton curve of SFC.
The concrete compressive constitutive relationship is based on Saenz
model [30] as shown in Eq. (1). In the equation, Ec=fc/ε0 is the secant stiffeners are incorporated at the T-shaped extremity of the center plate
elastic modulus at the peak stress, where fc is the compressive strength of (as illustrated in Fig. 4). Consequently, the numerical analysis does not
concrete, ε0 is the corresponding strain at the peak stress, taken as incorporate compressive buckling of the SFC. Hence, given the sym­
0.0018 based on experimental values. E0 denotes the elastic modulus of metrical nature of the tension-compression characteristics of SFC, this
concrete. The tensile behavior of concrete is modeled using the smeared study exclusively employs the tension force-displacement skeleton curve
crack model, with the peak tensile stress of 2.68 MPa, softening modulus as an illustrative example.
set to one-tenth of the elastic modulus. The concrete’s ultimate crushing In the initial phase, the SFC retains its inherent stiffness, designated
strain is taken as 0.0033, and the interface shear transfer coefficient is as ka, and experiences elastic deformation. The initial stiffness ka is
assumed as 0.1. approximately equal to the axial stiffness of the central plate with the
The reinforcement constitutive behavior is modeled using Eq. (2) same length as SFC, that is, ka=EA/l, where E is the elastic modulus of
[31]. In the equation, fy and εy are the yield strength and yield strain of the steel, A is the area of the thinnest part of the central plate, and l is the
the reinforcement, respectively. Es represents the elastic modulus. The total length of SFC. At this stage, there is no relative sliding between the
parameter values k1 , k2 , and k4 are determined from experimental tests cover and the central plate, and the assembled disc springs sustains its
of the tensile behavior of the reinforcement and are taken as k1 = 7, precompressed height until the activation force Fa of the SFC is reached,
k2 = 45, and k4 = 1.32. which can be obtained through Eq. (3).
E0 εc sinα + μcosα
σ= ( )( ) ( )2 (1) Fa = 2nb Fpr ( ) (3)
cosα − μsinα
1 + EE0c − 1 εε0c + εε0c
Where nb is the number of bolts on each splice, Fpr is the pre-load applied
⎧
⎪ Es ε, ε ≤ εy to each bolt, μ is the friction coefficient between steel plates, and α is the
⎪
⎪
⎪
⎪ angle of the grooves.
⎨ fy , εy < ε < k1 εy As the deformation increases, the SFC is activated. At this point, the
σ= (2)
⎪
⎪ Es (1 − k4 ) ( )2 loading stiffness kb can be calculated using Eq. (4). As the deformation
⎩ k4 fy + ε (k − k )2 ε − k2 εy , ε > k1 εy
⎪
⎪
⎪
y 2 1 further increases, the disc springs are fully compressed, and the SFC
reaches its maximum seismic design displacement. The displacement at
The SFC significantly influences the load characteristics of the SFCS, this stage Δ1 can be calculated using Eq. (5). Where k is the stiffness of
and its precise constitutive model directly impacts simulation accuracy. the assembled disc springs on each bolt, nj is the number of center plates
In seismic design, to ensure the functionality of SFC under earthquakes, acting in the series, and δt is the maximum allowable deformation of the
the deformation requirements of SFC are determined based on the assembled disc springs. In the other word, Δ1 corresponds to the
structural deformations expected during MCE. Thus, the operational allowable lateral deformation of the slotted hole. Based on Eq. 5, Δ1 can
phases of SFC mainly include the overall elastic deformation before be determined, which subsequently allows for the design of the slotted
activation and the relative sliding deformation between the central and hole dimensions.
the cover plate after activation. When subjected to earthquakes, the
force-displacement skeleton curve of SFC exhibits a bilinear behavior nb sinα + μcosα
kb = 2k tanα( ) (4)
[23,32]. However, the drift ratio of the frames is approximately 2 % to nj cosα − μsinα
3 % under MCE, while the vertical deformation at the column end
F
during progressive collapse can reach up to 1/5 of the span. At this (δt − kpr )nj
Δ1 = (5)
point, the deformation demand imposed on SFC markedly surpasses the tanα
permissible deformation for seismic resistance, leading to a scenario After the disc springs are fully compressed, the SFC undergoes
where the assembled disc springs in SFC are fully compressed. Subse­ overall deformation again, and at this stage, the stiffness is approxi­
quently, the SFC begins to deform as a whole without relative sliding mately equal to the initial stiffness. With further deformation, the SFC
until the SFC yields and eventually fails. Consequently, the can experience three potential yielding modes: yielding of the steel
force-displacement of SFC during the process of progressive collapse can plate, yielding of the bolts connecting to the column, and yielding of the
be divided into four stages, and its equivalent force-displacement skel­ bolts connecting to the beam. The yielding capacity (Fy) of the SFC is
eton curve is visualized in Fig. 6. To guarantee the functionality of the determined by the minimum value among these three scenarios. Based
SFC, buckling analyses are also conducted on its center plate during the on the dimensions of the SFC in this study, steel plate yielding occurs
design phase. In most cases, the center plate’s thickness is governed by prior to the other two damage modes. After yielding, the stiffness of the
its buckling resistance rather than its yield strength. To further enhance SFC can be estimated using the steel’s hardening modulus, approxi­
stability and prevent buckling induced by excessive end rotation, mated at 0.01 times the initial stiffness [33]. Finally, when the ultimate

5
X. Lu et al. Structures 65 (2024) 106743

elongation of the steel is reached, the SFC experiences failure. Consid­ of the specimens and simulate the fracture failure of the steel rein­
ering that the SFC is modeled using truss elements for an equivalent forcement. The numerical method proposed in Section 3 can be
representation, the corresponding stress-strain relationship of the truss employed to analyze the progressive collapse behavior of the SCFS
elements can be rapidly calculated based on the force-displacement under the condition of middle column failure.
relationship shown in Fig. 6. The key parameter values of the SFC’s
force-displacement curve are listed in Table 2. 4. Discussion on progressive collapse behavior

3.3. Boundary and loading conditions 4.1. Damage progression

To simplify the boundary conditions of the beam-column substruc­ When loaded to the maximum vertical displacement of 640 mm, a
ture, some experiments have assumed that the deformation of the sub­ comparison of deformation and damage patterns between the SCFS and
structure can be effectively constrained by adjacent beams and columns. ORCS is shown in Fig. 11. It’s observed that the SCFS exhibits significant
Therefore, it is assumed that both ends of the substructure are fully fixed damage at the beam-column interfaces and the intact beam sections near
[9]. Within the scope of the numerical simulations conducted within this the SFC, while the ORCS specimen’s damage is concentrated at the
study, the same assumption is adopted, meaning that all the degrees of beam-column interfaces. To better illustrate the entire failure process of
freedom of the edge columns are constrained. To prevent local failure the specimens, four critical observation sections are selected in the SCFS
and ensure a more uniform load distribution at the top of the middle specimens: the reduced beam section near the edge column (A), the
column, a loading steel plate is introduced at the top of the middle intact beam sections near the SFC (B and C), and the reduced beam
column. Displacement load is applied to the middle column until the section near the middle column (D). The damage evolution process of
deformation of the specimen exceeds 640 mm (one-fifth of the net span the specimens is described through the development of concrete strains
length). (Fig. 12a) and steel strains (Fig. 12b and c) at these sections, as well as
the axial force development of the SFC (Fig. 12d).
3.4. FEM validation In the case of the SCFS specimen, specific milestones are observed as
the middle joint displacement (MJD) progresses. At an MJD of 13 mm,
In order to ensure the rationality of the material parameters in the tension activation begins in the SFCs at the upper part of section A and
SCFS model, the seismic performance tests of the SCFS are first used for the lower part of section D (refer to Fig. 12d). Upon reaching an MJD of
validation. Utilizing the aforementioned modeling approach, the FEM of 37 mm, the steel bars in the upper part of section A and the lower part of
the SCFJ presented in Section 2.2 is established (as shown in Fig. 7a). section D undergo yielding under tension (see Fig. 12b and c). Further
The comparison of hysteresis curves between simulation and experi­ progression to an MJD of 42 mm leads to the attainment of the
mental results is shown in Fig. 7b. As the drift ratio reaches 2 %, the compressive peak strain in the concrete at the lower part of section A and
relative strength error between simulation and experiment is generally the upper part of section D (depicted in Fig. 12a). Subsequently, around
less than 13 %. Similarly, at a drift ratio of 3.5 %, the maximum relative an MJD of 49 mm, the steel bars at the lower part of section A and the
error in cumulative energy dissipation between simulation and experi­ upper part of section D yield under compression (as shown in Fig. 12b
ment is 10.5 %. This alignment between the simulated hysteresis curve and c). As the MJD reaches 66 mm, the concrete strain at the lower part
and the experimental counterpart highlights the reasonableness of the of section A and the upper part of section D reaches the ultimate
numerical simulation methodology employed. compressive strain, resulting in substantial damage to the concrete
The investigation into the progressive collapse mechanism primarily within sections A and D, as depicted in Fig. 12a.
revolves around pushdown analysis. In order to verify the efficacy of the Continuing the progression, an MJD of 267 mm prompts tension
suggested numerical analysis approach, simulations are conducted on yielding in the steel bars at the upper part of section B and the lower part
two specimens: the S4 specimen by Yu et al. [4] and the B2 specimen by of section C (Fig. 12b and c). At an MJD of 272 mm, the SFCs at the
Ren et al. [9]. The dimensions, reinforcement details, and material upper part of section A and the lower part of section D attain the
properties of both specimens are shown in Fig. 8. The relationship be­ maximum seismic design displacement, leading to an abrupt increase in
tween the simulated vertical displacement of the middle column and the SFC stiffness, subsequently intensifying the damage within sections B
applied load is shown in Figs. 9 and 10. and C. Consequently, rapid growth of steel strain is observed. Upon
Regarding the S4 specimen, there is a 10 % disparity between the reaching an MJD of 325 mm, the concrete at the lower part of section B
peak load capacity simulated during the CAA phase and the experi­ and the upper part of section C reaches the peak compressive strain
mental outcome. Similarly, in the CA phase, the variation in peak load (Fig. 12a).
capacity between simulation and experiment is 14 %. It is important to Further progression to an MJD of 411 mm triggers yielding in the
note that the stiffness of the horizontal pins is not specified in the test. SFCs at the upper part of section A and the lower part of section D
Consequently, the notable inconsistencies mentioned earlier can be (Fig. 12d). As the MJD increases to 464 mm, the concrete strain at the
attributed significantly to the uncertainty arising from the selected lower part of section B and the upper part of section C reaches the ul­
stiffness values in the simulation. In contrast, the B2 specimen possesses timate compressive strain, leading to significant concrete cracking near
more clearly defined boundary conditions, resulting in a closer align­ sections B and C, as shown in Fig. 12a. Notably, when the MJD reaches
ment between the simulated and experimental curves at various phases. 594 mm, an observation in Fig. 12b and c reveals the fracture of a steel
Furthermore, Fig. 9b compares the horizontal reaction force - middle bar at the upper part of section B. This fracture leads to an instantaneous
column displacement curve of the S4 specimen recorded in the experi­ reduction in axial forces within the SFCs upon steel fracture. The final
ment with the curve obtained from finite element analysis, demon­ stage occurs as the MJD reaches 642 mm, resulting in the fracture of one
strating a substantial agreement between the finite element analysis steel bar at the lower part of section D. At this moment, the loading
results and the experimental outcomes. Therefore, the proposed nu­ displacement has surpassed 1/5 of the net span of the beam, resulting in
merical simulation method can reasonably predict the FA, CAA, and CA the finalization of the loading process.

Table 2 4.2. Load-resisting mechanisms

The key parameters values of the SFC.
Parameter Fa ka kb Δ1 Fy 4.2.1. Load-resisting mechanisms of ORCS
Fig. 13 shows two curves representing the simulated applied load-
Values 181 kN 593 kN/mm 12.09 kN/mm 21.3 mm 972.9 kN
MJD and horizontal force-MJD of the ORCS specimen (depicted by the

6
X. Lu et al. Structures 65 (2024) 106743

Fig. 7. The FEM and comparison of experimental and simulation results of SCFJ.

Fig. 8. The dimensions and reinforcement details of S4 and B2.

black dotted line). Consistent with prior test findings [4,9], the applied phase. Specifically, when the MJD reaches 640 mm, the peak bearing
load-MJD curve follows a distinct two-stage pattern during loading. capacity of the CA mechanism is 383 kN.
Initially, a combination of FA and CAA (FA-CAA) governs the behavior,
with bending dominating at small MJD values. With increasing MJD, the 4.2.2. Load-resisting mechanisms of SFCS
beam experiences a rapid increase in axial compression, transitioning The simulated applied load-MJD curve and horizontal force-MJD
into the CAA mechanism. The specimen’s bearing capacity steadily in­ curve of the SFCS specimen are shown in Fig. 13 (The red solid line in
creases, reaching a peak of 340 kN at an MJD of 78 mm, coinciding with the figure), revealing distinct disparities from the ORCS findings. In the
the beam’s axial pressure nearing its maximum value. At an MJD of initial loading phase, the structural resistance is primarily provided by
480 mm, the beam’s axial force shifts to tension, marking the start of the the bending moment carried by sections A and D and the force couple
CA mechanism. Notably, the bearing capacity of the specimen continues sustained by the SFCs, indicating the FA mechanism. Both SCFS and
to increase with MJD, surpassing the peak observed during the CAA ORCS substructures exhibit similar trends in horizontal force versus

7
X. Lu et al. Structures 65 (2024) 106743

Fig. 9. The applied load–displacement and horizontal force-displacement curves of S4.

formation is hindered, resulting in a negligible contribution of CAA to

the specimen’s load-bearing capacity. Consequently, the load-bearing
capacity mainly relies on the flexural resistance of the reduced beam
section and the force couple facilitated by the SFCs. Since the SFC
maintains stable stiffness after activation, the force couple from the SFCs
continues to grow steadily, leading to a stable rise in the load-
displacement curve of the SCFS specimen.
As MJD continues to increase, the horizontal force in SCFS transi­
tions to tension earlier and exhibits a steady increase. Specifically, upon
reaching an MJD of 243 mm, the axial force within the SFCJ beam
transitions back to tension, yielding a load-bearing capacity of 279 kN,
marking the onset of the CA mechanism. At an MJD of 272 mm, the SFC
reaches its maximum seismic design displacement, restoring its initial
stiffness, and damage rapidly transfers to sections B and C. When the
MJD is 411 mm, the yielding of the SFC causes another reduction in SFC
stiffness, resulting in a slower increase in the lateral force-displacement
curve. At an MJD of 593 mm, the steel bars at the top of section B and
bottom of section C fracture, leading to a rapid decline in the load-
Fig. 10. The applied load–displacement curves of B2. displacement curve. Subsequently, the tension in steel bars at the bot­
tom of section B and top of section C continues to contribute to the load-
MJD. In the initial stage, the horizontal force is compressive and in­ bearing capacity in the CA phase, causing the load to continue
creases with increasing MJD. However, the peak horizontal force increasing. The load-bearing capacity of the CA mechanism, defined as
attained by SCFS is significantly lower (~47 %) compared to ORCS. corresponding load applied to the middle column when the first fracture
Fig. 14 vividly illustrates the principal compressive stress distributions of longitudinal bar in the beam, reaches 666 kN. Conversely, the hori­
of the SCFS and ORCS specimens at their respective maximum horizonal zontal force in ORCS experiences a later transition to tension, but the
forces. It can be observed that the compressive stress zone of the SCFS subsequent increase in tension force is more rapid. At a displacement of
specimen’s beam appears in the middle intact beam section. Due to the 593 mm, the difference in horizontal force between the two systems
presence of the reduced section at the beam ends, a complete arch becomes insignificant, with SCFS reaching roughly 92 % of the ORCS’s

Fig. 11. Damage pattern comparisons.

8
X. Lu et al. Structures 65 (2024) 106743

Fig. 12. The damage evolution process.

horizontal force.
Despite the comparable horizontal forces at larger MJD, SCFS dem­
onstrates a superior vertical load-bearing capacity. This is attributed to
the contrasting behavior of the steel reinforcement within the beams
during the tension phase. In ORCS, all steel bars experience tension, and
the load-bearing capacity is solely dependent on the CA mechanism.
Conversely, within the SCFS beam, the steel bars exhibit a combined
tension and compression profile, with the SFC also experiencing this
behavior. This implies that the beam ends continue to resist a significant
bending moment through the FA mechanism. Consequently, the overall
load-bearing capacity of SCFS is a combined effect of the FA and CA
mechanisms. Additionally, the larger end rotation observed in the SCFS
beam translates to a greater contribution from vertical components of
internal tension within the beam, further enhancing its vertical load
resistance.
Furthermore, the influence of the CA mechanism on the SCFS’s
resistance to progressive collapse evolves with increasing MJD, as
illustrated in Fig. 15. It’s observed that the CA mechanism becomes
active at an MJD of approximately 243 mm. Subsequently, its contri­
bution to the bearing capacity rises rapidly with increasing displace­
ment. At an MJD of 465 mm, the CA mechanism surpasses the FA
mechanism in terms of its contribution to the bearing capacity. Finally,
when the MJD reaches 593 mm and steel bar fracture occurs, the cate­
nary mechanism contributes up to 58 % of the total bearing capacity.
From the aforementioned analysis, noticeable distinctions emerge in
the load-bearing mechanisms between ORCS and SCFS specimens. In the
Fig. 13. The simulated force-displacement curves of the ORCS and SCFS. initial phases, ORCS primarily achieves load-bearing through FA-CAA
mechanism, with CAA predominantly influential. In contrast, due to
the presence of reduced beam sections in SCFS, the prominence of the

9
X. Lu et al. Structures 65 (2024) 106743

Fig. 14. The principal compressive stress distributions of the SCFS and ORCS.

Fig. 15. The contribution of CA mechanism to SCFS’s bearing capacity at

different MJD.

CAA mechanism diminishes, leading the load-bearing capacity to pre­

dominantly hinge on the FA mechanism. Remarkably, the FA mecha­
nism load-bearing capacity of SCFS surpasses that of ORCS by 22 %.
Nevertheless, within the entire early phase of the FA-CAA mechanism,
the peak load-carrying capacity of SCFS is 13.5 % lower than that of
ORCS. In the subsequent phase, ORCS relies on the CA mechanism to
provide load-carrying capacity, while SCFS relies on FA-CA mechanism
to provide load-carrying capacity. When MJD reaches 640 mm, the CA Fig. 16. The vertical deflection distribution of the specimens’ beams.
peak load-carrying capacity of SCFS is 42.5 % higher than that of ORCS
specimens. This observation highlights SCFS’s superior capability in ORCS by 0.063 rad.
achieving great resistance to progressive collapse during the CA phase. Generally, in the CA phase, the beam deflection curve of SCFS
demonstrates clear inflections at sections B and C, creating a three-
segment curve, while the deflection curve of ORCS is roughly linear.
4.3. Deformation characteristics This indicates that under the same applied displacement during the CA
phase, the SCFS specimen experiences a more pronounced rotation at
The vertical deflection distribution of the specimens’ beams is shown the mid-span section, leading to a larger vertical component of the
in Fig. 16. Utilizing differences in vertical displacements between tensile force of the longitudinal reinforcement in the beam. Conse­
neighboring measurement points and their horizontal spacing allows quently, the SCFS specimen exhibits a higher CA resistance.
computation of each beam section’s local rotation [4]. Additionally, the
local rotations of critical sections at MJD of 50 mm, 200 mm, 400 mm, 5. Parameter analysis
and 600 mm are listed in Fig. 16. As both specimens exhibit symmetric
vertical deflections, the deflection of just one half of the specimen is To further reveal the influence of key design parameters on the
depicted. progressive collapse resistance of SCFS, an extensive parametric analysis
At MJD 50 mm, both SCFJ and ORCJ specimens show similar is conducted. Prior research has established that influential factors for
deformation, with equal local rotations per section. At 200 mm MJD, the progressive collapse of ORCS encompass beam span-to-depth ratio,
SFCS has a higher rotation at section A, but smaller rotations at other concrete strength, and reinforcement ratio [9]. Drawing from the con­
sections compared to ORCS. By MJD 400 mm, ORCS’s section A rotation struction specifics of SCFS and the aforementioned analysis outcomes,
surpasses SCFS; however, SCFS’s section B and mid-span rotations potential design parameters influencing the progressive collapse resis­
already exceed ORCS. At 600 mm MJD, ORCS’s section A rotation ex­ tance of SCFS encompass SFC’s activation force, SFC’s post-activation
ceeds SCFS by 0.035 rad, while SCFS’s mid-span rotation surpasses

10
X. Lu et al. Structures 65 (2024) 106743

stiffness, SFC’s maximum seismic design displacement, beam 5.2. Effect of SFC’s maximum seismic design displacement
span-to-depth ratio, reinforcement ratio of the intact beam section,
concrete strength, reinforcement ratio in the reduced section, and the The load-displacement curves of the specimens with varying
depth of the reduced section. Given the prerequisites of self-centering maximum seismic design displacements (Δ1 ) are depicted in Fig. 18a. Up
behavior and shear resistance within seismic design [23], the range of to MJD 140 mm, the curves largely overlap, but differences become
variation in the height and reinforcement ratio of the reduced section noticeable after that. For the specimens with Δ1 = 10 mm, 15 mm,
remains relatively constrained. Preliminary calculations have indicated 20 mm, and 30 mm, the respective MJD values at which SFC reaches
that such variations produce a negligible impact on the progressive their maximum seismic design displacements are 140 mm, 198 mm,
collapse load-bearing capacity. Therefore, the influence of the rein­ 275 mm, and 421 mm. At these specific points, the curves exhibit in­
forcement ratio and depth of the reduced section is excluded from this flection points with an instantaneous increase in load-bearing capacity.
analysis. The final selected parameters and their corresponding ranges However, for the specimen with Δ1 = 40 mm, SFC never attains its
are detailed in Table 3, with the values of the baseline model indicated maximum seismic design displacement, consequently the curve lacks an
in parentheses. inflection point. The CA peak load capacities of the five specimens are
650 kN, 660 kN, 666 kN, 678 kN, and 724 kN, respectively.
5.1. Effect of SFC’s activation force A positive correlation exists between the peak load capacity and the
maximum seismic design displacement. This can be attributed to the
As evident from Eq. 3, the bolt pre-tightening force directly in­ following mechanism. When the SFC reaches its maximum seismic
fluences the activation force of the SFC. Consequently, the bolt pre- design displacement, the damage of the specimen will quickly transfer to
tightening force is adjusted to achieve the desired activation force. sections B and C. Therefore, a larger maximum seismic design
The dimensionless parameter ζ is introduced to express the relative displacement value can delay the damage of sections B and C and reduce
magnitude of the SFC’s activation force and its calculation is derived the plastic rotation of these sections, thereby effectively protecting the
from Eq. (6). Where Fa is the activation force of the SFC, d is the distance longitudinal bars in the beam from premature fracture and consequently
between the upper and lower SFCs, Mb is the yield moment of the intact enhancing the load-bearing capacity.
concrete beam section, and Mrb is the yield moment of the reduced
section. 5.3. Effect of SFC’s post-activation stiffness
Fa ⋅d
ζ= (6) Fig. 19a illustrates the load-displacement curves of specimens with
Mb − Mrb
varying post-activation stiffness of SFC. The post-activation stiffness of
To ensure that the beam end remains a vulnerable section in seismic SFC minimally affects the initial phase of the load-displacement curve,
design, the value of ζ should not exceed 1. The load-displacement curves predominantly influencing the latter portion of the FA mechanism. A
obtained under varying activation forces are shown in Fig. 17a. It be­ heightened post-activation stiffness leads to an increased peak load ca­
comes evident that the activation force of SFC notably influences the FA, pacity of the FA mechanism. Furthermore, the post-activation stiffness
with an increasing in ζ resulting in an almost linear rise in the peak load has a negligible impact on the load-bearing capacity of the CA mecha­
of FA, as depicted in Fig. 17b. This phenomenon arises due to the nism. Fig. 19b offers a detailed depiction of the impact of post-activation
symmetrical arrangement of SFCs at both the upper and lower sections, stiffness on the peak load capacity of the FA mechanism. With an
which collectively bear a substantial portion of the entire beam’s elevation in post-activation stiffness, the peak load capacity of the FA
bending moment during FA. Consequently, the load-bearing capacity of mechanism exhibits a linear growth trend. This is attributed to the fact
SCFS during FA hinges on the bending resistance of the beam end. that a higher post-activation stiffness accelerates the augmentation of
Hence, a greater activation force corresponds to an elevated peak load- SFC’s load-bearing capacity. And during the FA stage, the load-bearing
bearing capacity during FA. capacity of the SCFS specimen predominantly relies on SFC.
The impact of the SCF’s activation force on the initial stage of the CA
is marginal; however, it does have effect on the peak load-carrying ca­ 5.4. Effect of the beam span-to-depth ratio
pacity. As the activation force increases, the peak load-carrying capacity
of CA experiences a slight decline. This is because a larger activation In this study, while maintaining a constant beam height of 400 mm,
force leads to a diminished disparity in the strength between sections A the influence of varying the span-to-depth ratio by changing the span is
and B. As a result, section B enters the plastic range earlier and experi­ explored. Fig. 20a demonstrates the influence of span-to-height ratios on
ences more severe damage. the progressive collapse performance, while Fig. 20b provides a detailed
As shown in Fig. 17b, a higher activation force results in a larger representation of how the peak load during the FA and CA phases change
rotation at section B, indicating more severe damage and larger strain in with the span-to-depth ratios. To ensure consistent horizontal axis
the steel reinforcement. This accelerates the occurrence of bar fracture, scaling, owing to varying vertical displacement across different spans,
consequently contributing to a reduction in the peak load-carrying ca­ the horizontal coordinate has been adjusted to represent the ratio of the
pacity of CA. MJD to the span length (L). Evidently, the span-to-depth ratio exerts a
significant influence on the load-displacement curves of the SCFS
specimens. With an increase in the span-to-depth ratio, the initial stiff­
ness and FA peak load-bearing capacity gradually decrease. This is
Table 3
because the bending resistance capacity at the beam end remains rela­
The range of values for the design variables of SFC.
tively constant, and as the span-to-depth ratio increases, the lever arm
Design variables Value range Design variables Value becomes longer, resulting in reduced load applied at the middle column.
range
The CA mechanism’s peak load capacity is the highest when the span-to-
Activation force (kN) ζ = 0.24–0.98 Reinforcement ratio of 0.5–1.69 depth ratio is 9. However, beyond this ratio, the CA mechanism’s peak
(0.98) the intact beam section (0.89)
load-bearing capacity experiences a swift decline, stabilizing at
(%)
Maximum seismic 10–40 Span-to-depth ratio 9–18 approximately 420 kN thereafter.
design displacement (20) (9)
(mm) 5.5. Effect of beam reinforcement ratio
Post-activation 5–15 Concrete strength (MPa) 24–54
stiffness (kN/mm) (12) (34)
In this study, the reinforcement ratio of both the upper and lower of

11
X. Lu et al. Structures 65 (2024) 106743

Fig. 17. The Effect of SFC’s activation force on progressive collapse performance.

Fig. 18. The Effect of SFC’s maximum seismic design displacement on progressive collapse performance.

Fig. 19. The Effect of SFC’s post-activation stiffness on progressive collapse performance.

the intact beam sections is varied between 0.5 % and 1.69 %. As shown reinforcement ratio increases, the inflection point of the load-
in Fig. 21, alterations in the reinforcement ratio have minimal effect on displacement curve becomes more pronounced when the SFC attained
the load-bearing capacity prior to an MJD of 270 mm, indicating its maximum seismic design displacement, typically occurring around
negligible influence on the FA mechanism. However, an increased an MJD of 270 mm, resulting in a higher peak bearing capacity during
reinforcement ratio results in an enhanced load-bearing capacity of the the CA phase. Furthermore, the influence of both the top and bottom
CA, thereby mitigating the risk of progressive collapse. Moreover, as the reinforcement ratios on the SCFS’s progressive collapse exhibited

12
X. Lu et al. Structures 65 (2024) 106743

Fig. 20. The Effect of beam span-to-depth ratio on progressive collapse performance.

Fig. 21. Load-Displacement curves of specimens with different reinforcement ratios.

remarkable consistency. When the top and bottom reinforcement ratios 5.6. Effect of the concrete strength
are increased from 0.5 % to 1.69 %, the CA peak load-bearing capacity
increased by 159 kN and 198 kN, respectively. The impact of concrete strength on the progressive collapse perfor­
mance of SCFS is explored, considering span-to-depth ratios of 9 and 18.
The resulting load-displacement curves are presented in Fig. 22. When

Fig. 22. Load-Displacement curves of specimens with different concrete strengths.

13
X. Lu et al. Structures 65 (2024) 106743

the span-to-depth ratio is 9, an elevation in concrete strength leads to a CRediT authorship contribution statement
slight increase in the load-bearing capacity of specimens within the
20 mm to 180 mm MJD range. This can be attributed to the fact that at a Xiao Lu: Writing – review & editing, Writing – original draft, Su­
smaller span-to-depth ratio, higher concrete strength amplifies the CAA, pervision, Methodology, Investigation, Conceptualization. Longhe Xu:
consequently enhancing axial compression within the beam (When the Supervision, Methodology. Hang Xu: Writing – original draft, Visuali­
concrete strength increases from 24 MPa to 54 MPa, the concrete pres­ zation, Resources, Investigation.
sure increases by 22.5 kN) and thereby augmenting its bending resis­
tance. Additionally, specimens with lower concrete strength Declaration of Competing Interest
experienced more severe damage at sections B and C, leading to larger
section rotations. When the concrete strength increased from 24 MPa to We wish to confirm that there are no known conflicts of interest
54 MPa, the section rotations at sections B of the four specimens are associated with this publication and there has been no significant
0.253 rad, 0.236 rad, 0.236 rad, and 0.232 rad, respectively. Lesser ro­ financial support for this work that could have influenced its outcome.
tations delay the failure of steel reinforcement. Consequently, with an We confirm that the manuscript has been read and approved by all
increase in concrete strength, the CA load capacity of SCFS experiences a named authors and that there are no other persons who satisfied the
minor increment. criteria for authorship but are not listed. We further confirm that the
However, for a span-to-depth ratio of 18, concrete strength exerts order of authors listed in the manuscript has been approved by all of us.
neglectable influence on the progressive collapse performance. This is We confirm that we have given due consideration to the protection of
primarily attributed to the heightened span-to-depth ratio substantially intellectual property associated with this work and that there are no
diminishing the CAA and promoting a more even distribution of section impediments to publication, including the timing of publication, with
rotations along the beam. Consequently, the influence of concrete is far respect to intellectual property. In so doing we confirm that we have
less significant compared to the effects of steel bars and SFC. followed the regulations of our institutions concerning intellectual
property.
6. Conclusions
Acknowledgments
To investigate the progressive collapse performance of the SFCS with
high seismic resilience, three-dimensional nonlinear FEMs are estab­
The authors extend their sincere appreciation for the financial sup­
lished. The study explored the progressive collapse damage evolution
port received from the Fundamental Research Funds for the Central
process of SFCS and revealed its progressive collapse mechanism.
Universities of China (Grant No. 2022JBMC051) and National Natural
Finally, the influence of design parameters on SFCS’s progressive
Science Foundation of China (Grant No. 52125804).
collapse behavior is assessed. The main conclusions are summarized as
follows:
References
(1) In the scenario of middle column failure, SFCS primarily resists
progressive collapse through the early-phase mechanism of FA, suc­ [1] Kiakojouri F, Sheidaii MR, De Biagi V, Chiaia B. Progressive collapse of structures:
ceeded by the later-phase mechanism of CA and FA. However, due to the a discussion on annotated nomenclature. Struct 2021;29:1417–23.
presence of reduced sections at the beam ends, SFCS fails to establish a [2] Qian K, Weng YH, Fu F, Deng XF. Numerical evaluation of the reliability of using
single-story substructures to study progressive collapse behaviour of multi-story RC
complete arch, resulting in a less pronounced manifestation of CAA. frames. J Build Eng 2021;33:101636.
Conversely, ORCS predominantly relies on CAA mechanism during the [3] Lin K, Chen Z, Li Y, Lu X. Uncertainty analysis on progressive collapse of RC frame
initial phase to mitigate progressive collapse. As the CA phase unfolds, structures under dynamic column removal scenarios. J Build Eng 2022;46:103811.
[4] Huang Y, Tao Y, Yi W, Zhou Y, Deng L. Numerical investigation on compressive
SCFS displays a three-segment deflection curve, while the deflection
arch action of prestressed concrete beam-column assemblies against progressive
curve of ORCS demonstrates an approximate linearity. collapse. J Build Eng 2021;44:102991.
(2) For the SFCS and ORCS specimens of equal size and reinforce­ [5] Weng YH, Qian K, Fu F, Fang Q. Numerical investigation on load redistribution
capacity of flat slab substructures to resist progressive collapse. J Build Eng 2020;
ment in this study, SFCS exhibits a more pronounced FA mechanism
29:101109.
compared to ORCS. However, ORCS shows a more significant CAA [6] Yu J, Luo LZ, Fang Q. Structure behavior of reinforced concrete beam-slab
mechanism. This ultimately results in SCFS having a 13.5 % lower peak assemblies subjected to perimeter middle column removal scenario. Eng Struct
load in the early phase of FA-CAA mechanism compared to ORCS. In the 2020;208:110336.
[7] Pang B, Wang F, Yang J, Nyunn S, Azim I. Performance of slabs in reinforced
later phase, SFCS’s peak load-bearing capacity in the CA mechanism is concrete structures to resist progressive collapse. Structures 2021;33:4843–56.
42.5 % higher than that of ORCS, indicating greater resistance to pro­ [8] Qian K, Cheng JF, Weng YH, Fu F. Effect of loading methods on progressive
gressive collapse under large deformations. collapse behavior of RC beam-slab substructures under corner column removal
scenario. J Build Eng 2021 2021;44:103258.
(3) The activation force and post-activation stiffness of SFC have a [9] Feng DC, Zhang MX, Brunesi E, Parisi F, Yu J, Zhou Z. Investigation of 3D effects on
significant impact on the FA mechanism, while the reinforcement ratio dynamic progressive collapse resistance of RC structures considering slabs and
of the intact beam section and maximum seismic design displacement of infill walls. J Build Eng 2022 2022;54:104421.
[10] Qian K, Lan DQ, Li SK, Fu F. Effects of infill walls on load resistance of multi-story
SFC have a substantial influence on the CA mechanism. The beam’s RC frames to mitigate progressive collapse. Structures 2021;33:2534–45.
span-to-height ratio has a noticeable effect on both the FA and CA [11] Yang Z, Li Y, Guan H, Diao M, Gilbert BP, Sun H, et al. Dynamic response and
mechanisms. However, the effect of concrete strength on the progressive collapse resistance of RC flat plate structures subjected to instantaneous removal of
an interior column. Eng Struct 2022;264:114469.
collapse performance of SCFS remains marginal. Furthermore, the peak
[12] Zhao Z, Liu Y, Li Y, Guan H, Yang Z, Ren P, et al. Experimental and numerical
load associated with the CA mechanism is closely related to the rotation investigation of dynamic progressive collapse of reinforced concrete beam-column
of sections B and C under large deformations. assemblies under a middle-column removal scenario. Struct 2022;38:979–92.
[13] Qian K, Chen XY, Huang T. Dynamic response of RC beam-slab substructures
While the present study employed numerical simulations to provide
following instantaneous removal of columns. J Build Eng 2022;45:103554.
initial evidence of the SCFS’s promising performance in terms of both [14] Pham AT, Tan KH. Experimental study on dynamic responses of reinforced
seismic resistance and progressive collapse resistance, the investigation concrete frames under sudden column removal applying concentrated loading. Eng
into progressive collapse is limited to the mid-column removal scenario. Struct 2017;139:31–45.
[15] Panahi S, Zahrai SM. Performance of typical plan concrete buildings under
Future research should encompass the removal of corner columns, side progressive collapse. Struct 2021;31:1163–72.
columns, and other potential failure scenarios to achieve a more [16] Castillo JGS, Bruneau M, Elhami-Khorasani N. Seismic resilience of building
comprehensive understanding of the entire self-centering frame struc­ inventory towards resilient cities. Resilient Cities Struct 2022;1(1):1–12.
[17] Mokhtarnejad N, Garmeh V, Shariatmadar H, Askariani SS. Numerical study of a
ture’s progressive collapse resistance. novel SMA-based self-centering beam-to-column connection. Structures 2023
2023, January;47:1033–49.

14
X. Lu et al. Structures 65 (2024) 106743

[18] Li Y, Geng F, Ding Y, Wang L. Experimental and numerical study of low-damage [25] Fan X, Xu L, Li Z. Seismic performance evaluation of steel frames with pre-pressed
self-centering precast concrete frame connections with replaceable dampers. Eng spring self-centering braces. J Constr Steel Res 2019;162:105761.
Struct 2020;220:111011. [26] Wang HS, Marino EM, Pan P, Liu H, Nie X. Experimental study of a novel precast
[19] Cai X, Pan Z, Zhu Y, Gong N, Wang Y. Experimental and numerical investigations prestressed reinforced concrete beam-to-column joint. Eng Struct 2018;156:68–81.
of self-centering post-tensioned precast beam-to-column connections with steel top [27] Song LL, Guo T, Yu G, Cao ZL. Experimental study of a self-centering prestressed
and seat angles. Eng Struct 2021;226:111397. concrete frame subassembly. Eng Struct 2015;88:176–88.
[20] Ye J, Yan G, Liu R, Xue P, Wang D. Hysteretic behavior of replaceable steel plate [28] Sezen H, Setzler EJ. Reinforcement slip in reinforced concrete columns. Acids
damper for prefabricated joint with earthquake resilience. J Build Eng 2022;46: Struct J 2008;105(3):280.
103714. [29] Pan WH, Tao MX, Nie JG. Fiber beam–column element model considering
[21] Lin K, Li Y, Lu X, Guan H. Effects of seismic and progressive collapse designs on the reinforcement anchorage slip in the footing. B Earthq Eng 2017;15:991–1018.
vulnerability of RC frame structures. J Perform Constr Fac 2017;31(1):04016079. [30] Saenz LP. Discussion of ‘Equation for the stress-strain curve of concrete’ by Desayi
[22] Li M, Sasani M. Integrity and progressive collapse resistance of RC structures with P and Krishnan S. ACI J 1964;61:1229–35.
ordinary and special moment frames. Eng Struct 2015;95:71–9. [31] Lu X, Lu X, Guan H, Ye L. Collapse simulation of reinforced concrete high-rise
[23] Lu X, Xu H, Zhang X, Xie L. Experimental investigation on seismic performance of building induced by extreme earthquakes. Earthq Eng Struct D 2013;42(5):705–23.
self-centering frictional cast-in-situ beam-column joints. Eng Struct 2023;285: [32] Hashemi A, Zarnani P, Masoudnia R, Quenneville P. Seismic resistant rocking
116062. coupled walls with innovative Resilient Slip Friction (RSF) joints. J Constr Steel Res
[24] Zhang Y, Xu LH. Experimental investigation of a new self-centering shear wall with 2017;129:215–26.
resilient hinge devices. Eng Struct 2022;266:114657. [33] Cao X, Wu L, Li Z. Behaviour of steel-reinforced concrete columns under combined
torsion based on ABAQUS FEA. Eng Struct 2020;209:109980.

15