Journal of Building Engineering 45 (2022) 103554

Contents lists available at ScienceDirect

Journal of Building Engineering

Dynamic response of RC beam-slab substructures following instantaneous

removal of columns
Kai Qian a, b, Xin-Yi Chen a, Ting Huang b, ∗
a College of Civil Engineering and Architecture, Guangxi University, Nanning, 530000, China
b College of Civil Engineering and Architecture, Guilin University of Technology, Guilin, 541004, China

ARTICLE INFO ABSTRACT

Keywords: High fidelity solid-element-based numerical modeling is employed in this paper to study the dynamic response of
Progressive collapse reinforced concrete (RC) beam-slab substructures following instantaneous removal of columns. After validation
RC beam-Slab substructures of the numerical model against experimental results, incremental dynamic analysis is conducted to obtain the dy-
Sudden column removal
namic structural resistance to progressive collapse. The numerical results for development of internal forces and
Dynamic response
distribution of stress and strain are used to illustrate the load redistribution and load transfer mechanisms of RC
Multi-column removal
Numerical simulation beam-slab substructures subjected to sudden loss of columns. The numerical model is further employed to ex-
plore parameters affecting the dynamic response and structural resistance of beam-slab substructures under sud-
den column removal scenario. The parameters studied include damping ratio, slab thickness, location of re-
moved columns, multi-column removal sequence and interval. The numerical result shows that the dynamic pro-
gressive collapse resistance of the beam-slab substructure is much smaller than the quasi-static test result due to
a dynamic increase factor (DIF) of 1.14. Moreover, a minimum slab thickness of 1/45 of span length is found vi-
tal for the improvement of progressive collapse resistance of RC beam-slab substructures through developing
compressive membrane action (CMA) and tensile membrane action (TMA) as well as enhancing the development
of compressive arch action (CAA) and tensile centenary action (TCA) of beams. When the damping ratio is
smaller than 5%, the increase of damping ratio can significantly benefit the structural resistance against progres-
sive collapse by producing smaller deflection and milder vibration. The additional loss of the corner column be-
sides the penultimate perimeter column is found to form the most hazardous situation among the double-column
removal scenarios. In contrast to the location of removed columns, the column removal sequence and interval
tend to little affect the maximum deflection developed and consequently the progressive collapse resistance of
substructures.

1. Introduction progressive collapse and special design guidelines [8,9] have also been
proposed. So far, the most recommended method to analyze the resis-
The failure of a few columns due to accidental loading such as blast tance of structures against progressive collapse is the alternate load
and impact may result in progressive collapse. The progressive collapse path analysis (ALPA) due to its straightforwardness and effectiveness.
refers to “the spread of an initial local failure from element to element, The ALPA essentially evaluates the competence of the remaining build-
eventually resulting in the collapse of an entire structure or a dispropor- ing in providing sufficient alternate load paths for load redistribution
tionately large part of it” [1,2]. Some well-known incidents of progres- after notional loss of one or several vertical load bearing elements, such
sive collapse include the partial collapse of Ronan Point apartment in as columns. In practice, the ALPA can be implemented either experi-
London [3], the collapse of Alfred P. Murrah Federal Building in Okla- mentally or numerically.
homa City [4] and twin towers of the World Trade Centre in New York Extensive experimental studies have been conducted to investigate
City [5]. In the event of progressive collapse, the collapse extent and se- the structural performance of reinforced concrete (RC) structures
vere casualties far exceed the expected consequences of local failures. against progressive collapse based on the ALPA. Especially, many fo-
Thus, over the past few decades, existing building design codes [6,7] cused on the progressive collapse resistance of RC substructures under
have been increasingly supplemented with provisions of design against various column removal scenarios based on the quasi-static tests

∗ Corresponding author.
E-mail address: [email protected] (T. Huang).

https://doi.org/10.1016/j.jobe.2021.103554
Received 3 August 2021; Received in revised form 20 October 2021; Accepted 30 October 2021
Available online 3 November 2021
2352-7102/© 2021 Elsevier Ltd. All rights reserved.
K. Qian et al. Journal of Building Engineering 45 (2022) 103554

[10–16] and pushdown analysis [15,17–19]. In contrast to beam- ing corner, perimeter and/or interior column, and the effects of column
column substructures reliant solely on compressive arch action and ten- removal sequence and interval on the dynamic response.
sile catenary action (CAA and TCA) [20–23], the presence of slabs not
only significantly enhances the progressive collapse resistance of beam- 2. Numerical model of RC beam-slab substructures
slab substructures through the additional development of compressive
and tensile membrane action (CMA and TMA) [24–29], but also makes 2.1. Overview of the reference test
the structural resistance to progressive collapse dependent on loading
scheme, i.e., uniformly distributed load (UDL) or concentrated load The numerical study is carried out on the basis of the tests con-
(CL). The results show that, although CL is feasible to demonstrate the ducted by Qian and Li [37] on RC beam-slab substructures subjected to
essential load transfer mechanisms against progressive collapse, UDL is instantaneous removal of the penultimate perimeter column. They
a more representative loading scheme to explore the structural resis- tested two 1/4 scaled RC beam-slab substructures, namely D-0.91 and
tance of beam-slab substructures [30,31]. D-1.16, subjected to sudden column removal by using a specially de-
Since accidental loss of columns usually occurs in an instant and signed Instantaneous Column Removal Device (ICRD). It was shown
progressive collapse is in nature a dynamic response, the quasi-static that the duration of column removal was fast enough to simulate acci-
approaches may result in inappropriate design since the damping and dental impact or terrorist attack with the ICRD. The two specimens, the
inertial effects caused by the dynamic response are not taken into ac- prototype of which was a four-story RC building designed in accor-
count. Previous studies indicate that the failure patterns in the quasi- dance with Singapore Code CP65 [46], had identical dimensions, rein-
static tests are similar to those in dynamic tests, however, it is necessary forcement details and boundary conditions, except that different UDL of
to relate the static response to dynamic response by using the dynamic 10.5 kPa and 11.6 kPa had been applied on the slab before column re-
increase factor (DIF) approach [32,33], energy-based approaches moval which amounted to 0.91 and 1.16 times the design load
[34,35] or equivalent single-degree-of-freedom model approaches (DL + LL), respectively. Each specimen consisted of two bays in the
[36,37]. To calibrate the DIF for RC structures, a few dynamic tests longitudinal direction and one bay in the transverse direction, and the
have been carried out on RC substructures under sudden column re- beam span was 2250 mm in both directions as shown in Fig. 1. The
moval caused by contact detonation [32] or rapid release of the support beams and slabs were extended by 563 mm at two adjacent sides and
force via various devices [38–43]. While an DIF less than 1.4 has been imposed with distributed design load (see cross-hatch area in Fig. 1) to
generally reported for RC beam-column substructures [38,39,42], a account for the rotational constraints from interior bays. The cross-
much larger DIF of 1.86 was determined when introducing initial dam- sectional dimensions and reinforcement details of the beams and slabs
age to the remaining structural members by the explosion [32]. To date, are also summarized in Table 1. The concrete cover was 10 mm and
few studies have been reported about the dynamic response of beam- 7 mm, respectively, for the beams and slabs. While the compressive
slab structures. While Adam et al. [44] tested a full-scale two-story RC strength of concrete was measured as 24.4 MPa and 25.1 MPa for D-
flat-slab building structure with loads reflecting design conditions sub- 0.91 and D-1.16, respectively, the properties of steel reinforcement are
jected to sudden removal of a corner column and found that the slab listed in Table 2. The test set-up is shown in Fig. 2. The specimen was
membrane action was not a significant load redistributing mechanism supported by five remaining columns with cross-sectional area of
for the case they investigated, Qian and Li [37] showed that the TCA 170 mm × 170 mm which were bolted to steel supports (Item 7 in Fig.
and TMA, since they were slightly activated by the dynamic response to 2), while the column stub of the notionally removed column with cross-
sudden removal of a column and thus little weakened by the associated sectional area of 125 mm × 125 mm was supported by the ICRD (Item
damage, made a major contribution to the residual strength of the 5 in Fig. 2) which would be knocked down later to simulate dynamic
beam-slab substructures as compared to the CAA and CMA. Nyunn et al. column removal. Note that the cross-sectional area of five remaining
[45] deduced that the DIF values for a multi-story RC beam-slab build- columns were enlarged to prevent them from being crushed upon re-
ing subjected to the loss of an edge column or a corner column were moval of the target column. The loading scheme of UDL was repro-
close to 1.6, regardless of infill walls, based on structural analysis using duced by placing weights (Item 1 in Fig. 2) on top of the slab. The total
SAP2000. However, the remaining structures in these studies have re- mass of the weights on D-0.91 and D-1.16 were 9000 kg and 12,000 kg,
sponded almost within the elastic range or limited to early plastic stage, respectively. Both specimens could survive after the sudden removal of
providing little demonstration on the dynamic load-carrying capacity the penultimate perimeter column. More details about the specimen de-
and value of DIF for beam-slab structures under large deformation sign and test setup of the reference experiment can refer to the paper
stage. [37].
In short, the study of RC beam-column structures against progres-
sive collapse is still limited in the realm of dynamic response into plas- 2.2. Specifications of the numerical model
tic range. As an alternative to expensive and technically demanding dy-
namic tests, this paper aims to investigate the robustness of RC beam- To further study the dynamic response of RC beam-slab substruc-
slab substructures subjected to sudden column removal under incre- tures following instantaneous removal of a column, the numerical
mental load and provide insights into the load transfer mechanisms model built using commercial software LS-DYNA [47] as shown in Fig.
against progressive collapse by adopting high-fidelity finite element 3 was employed. The numerical model was validated by the abovemen-
(FE) numerical modeling. After validation of the numerical modeling tioned tests on D-0.91 and D-1.16.
against the experimental dynamic response of beam-slab substructures
following sudden removal of the penultimate perimeter column, the FE 2.2.1. Element types
model is further used to explore the dynamic load redistribution behav- The concrete was modelled by 8-node solid elements with reduced
ior, load transfer mechanisms, progressive collapse resistance and the integration scheme. As each solid element adopts only one integration
value of DIF for beam-slab substructures. The FE model is also used to point, this scheme can decrease the computational time effectively with
illustrate the contributions of beams and slabs to the progressive col- reasonable calculation accuracy. All the reinforcement was simulated
lapse resistance of beam-slab substructures, the effects of slab thickness by 2-node Hughes-Liu beam elements with 2 × 2 Gauss quadrature in-
and damping ratio on the dynamic response of the substructure follow- tegration at the cross section, which can model the behavior of tension,
ing sudden removal of columns. Finally, the numerical modeling is em- transverse shear and bi-axial bending well [48]. Besides, the beam for-
ployed to explore the robustness of RC beam-slab substructures sub- mulation is based on the degenerated 8-node solid element so that the
jected to removal of multiple columns in case of extreme events, includ- reinforcement elements can be fully compatible with the concrete ele-

2
K. Qian et al. Journal of Building Engineering 45 (2022) 103554

Fig. 1. Dimensions and reinforcement details of the specimen (unit: mm) [30].

Table 1
The dimensions and reinforcement details for the beams and slabs [30].
Section dimensions (mm) Top reinforcement Bottom
reinforcement

Width Depth Thickness End Middle

Interior Beam 80 138 N.A. 2T13+1T10 3R6 2T10

Perimeter 80 112 NA 2T13+1T10 3R6 2R8
Beam
Slab N.A. N.A. 64 R3@80 R3@50 R3@50

Note: R and T represent plain round bar and deformed bar, respectively.

Table 2
The properties of steel reinforcement [30].
Reinforcement Diameter Yield Ultimate Yield Ratio of
(mm) strength strength (MPa) strain elongation
(MPa) (10−6)

R3 3 358 439 2014 10.9%

R6 6 445 561 2190 17.5%
R8 8 450 556 2474 16.1%
Fig. 2. Test setup of the dynamic test [30].
T10 10 503 635 2609 15.1%
T13 13 473 592 2343 16.9%
control meshing and actual element sizes may be adjusted depending
on the dimensions.
ments. Based on the mesh sensitivity analysis as shown in Fig. 4, the el-
The perfect bond and bond-slip behavior between the reinforcement
ement size was decided as 35 mm for both concrete and reinforcement,
and surrounding concrete were both considered in this numerical study.
since a finer mesh could not provide a more accurate solution but sig-
The perfect bond was defined by using the keyword *CONSTRAINED_L
nificantly increased the computational time. It should be noted that LS-
AGRANGE_IN_SOLID. In a more refined modeling, the bond-slip behav-
DYNA uses the defined element size as the maximum element size to
ior was considered in the beam-column connections where severe

3
K. Qian et al. Journal of Building Engineering 45 (2022) 103554

Fig. 3. Numerical model of test specimen.

Fig. 3. The five intact columns bolted to the strong steel supports in the
test were connected to the steel supports in the numerical model by us-
ing the keyword *CONTACT_TIED_NODES_TO_SURFACE, while the no-
tionally removed column was connected to the dummy ICRD with con-
tact surface defined by keyword ∗CONTACT_AUTOMATIC_SURFACE_T
O_SURFACE. To completely reproduce the tests, the weights in the test,
as shown in Fig. 2, were modelled. They were modelled as rigid bodies
placed on the top surface of slabs with contact surfaces defined by the
keyword ∗CONTACT_AUTOMATIC_SURFACE_TO_SURFACE. Besides,
adequate static and dynamic coefficients of friction were provided at
the contact surfaces to prevent any sliding of the weights. The gravity
load generated by the weight of the specimen itself and the additional
weights was applied by using keyword *LOAD_BODY and special care
was taken to eliminate unintended dynamic vibration caused by the ap-
plication of the gravity load with keyword *CONTROL_DYNAMIC_RE-
LAXATION. The weights placed on the slab extensions to account for ro-
tational constraints from adjacent slab panels was simulated using uni-
Fig. 4. Mesh sensitivity analysis. form pressure.
During the test, the ICRD was knocked over with a high speed. The
cracking occurred, and the following bond stress-slip relationship as measurement of the vertical force in the ICRD indicated that the dura-
recommended by the FIB Model Code 2010 [49] for deformed bar in tion of column removal in D-0.91 and D-1.16 were 0.012 s and 0.008 s,
concrete was defined by using the keyword *DEFINE_FUCTION once respectively, less than 1/10 of the natural period of the free vibration of
the reinforcement elements had been connected to the concrete ele- the specimens, which is fast enough to simulate instantaneous column
ments via the keyword *CONSTRAINED_BEAM_IN_SOLID. removal. In the numerical modeling, the instantaneous column removal
was achieved by defining the erosion of the dummy ICRD using key-
word *MAT_ADD_EROSION. Fig. 5 compared the variation of the verti-
cal reaction force in ICRD for both Specimen D-0.91 and D-1.16, con-
(1) firming that the erosion of the dummy ICRD can simulate instantaneous
column removal accurately. Finally, for the dynamic modeling, the
damping ratio of 5% was defined.
where τbmax is the maximum bond stress with fc indicating compres-
2.2.3. Material models and strain rate effects
sive strength of concrete; τbf is the residual bond stress; s1 is the slip cor-
For the simulation of RC structures subjected to dynamic effect, the
responding to maximum bond stress; s2 is the slip corresponding to pull-
three material models in LS-DYNA, including Karagozian and Case Con-
out failure; and s3 equals to half the bar rib spacing. Based on the sug-
crete (K&C) model, continuous surface cap model (CSCM) and Riedel-
gestions from FIB Model Code 2010 [49], the values of parameters α, s1,
Hiermaier-Thoma (RHT) model, are frequently used [47,48]. All of
s2 and s3 are selected as 0.4, 1 mm, 2 mm, and 10 mm, respectively.
these three models can model post-damage softening, shear dilation,
confinement effect as well as strain rate effects for normal strength con-
2.2.2. External load, boundary conditions and column removal
crete with unconfined compressive strength ranging from 28 MPa to
All nodes at the base of the steel supports and ICRD were restrained
58 MPa [47]. In this study, the simplified version of CSCM with the key-
with no translational and rotational degree of freedoms, as indicated in

4
K. Qian et al. Journal of Building Engineering 45 (2022) 103554

2.3. Validation of the numerical model

Before validating against the reference dynamic tests, the static

structural resistance of the studied beam-slab substructure missing a
column and the failure pattens at the ultimate state were predicted us-
ing the model with quasi-static uniformly distributed loading. The re-
sults were compared to the experimental data, aiming to validate the
model in terms of full-range structural behavior and damage patterns at
large deformation stage. Fig. 6 compares the load-displacement curves
of FEM and experimental results. It can be seen that they are in good
agreement in terms of first peak and ultimate resistance, as well as de-
velopment trend during the entire loading stage. Moreover, a compari-
son of FEM results with and without consideration of bond-slip behav-
ior suggested that the bond-slip effect was insignificant and the model
with perfect bond agreed even better with the experimental result. Sim-
Fig. 5. Comparison of numerical and experimental support force underneath ilar findings were also reported by other numerical studies on various
the stub of the notionally removed column. RC structures resisting progressive collapse [54–56]. Fig. 7 compares
the failure modes at the beam ends. The numerical model can effec-
word *MAT_CSCM_CONCRETE was chosen for the concrete material tively show the severe concrete spalling and rebar fracture with the ero-
due to its numerical stability. The constitutive model consists of a linear sion of concrete elements and reinforcement elements. Particularly, the
pre-peak part and a softening post-peak part. This simplified version numerical prediction of concrete separation agreed reasonably well
only requires input parameters of unconfined compressive strength fc′, with the actual damage pattern, confirming the validity of the selected
maximum aggregate size Ag, erosion criterion and units [48]. The other erosion limit for the concrete element. Again, the model with and with-
default model parameters will be generated automatically as functions out consideration of bond-slip behavior developed similar damage pat-
of fc′ and Ag through the use of the FIB Model Code 2010 [49]. In the terns. Therefore, the model with perfect bond was adopted for the dy-
tests, fc′ was measured as 24.4 MPa and 25.1 MPa for D-0.91 and D- namic simulation to reduce the computational time.
1.16, respectively. Ag was taken as 19 mm considering the thickness of For dynamic behavior, the column stub of the notionally removed
the slab was 64 mm. The element erosion was defined for concrete ele- column will move downward due to the gravity load after the column
ment for the sake of graphical representation of separation of concrete removal. The numerical and experimental results of the vertical dis-
caused by severe cracking and spalling. The maximum principal strain placement time-history are compared in Fig. 8. It is seen that the verti-
has been shown to be a proper erosion algorithm for concrete material cal displacement response including the first peak and corresponding
[50,51]. By this criterion, the concrete element will be deleted if the time as well as the permanent displacement after stabilization can be
maximum principal strain εmax exceeds the principal strain at failure εf. closely predicted for both D-0.91 and D-1.16. Fig. 9 compares the devel-
Since the erosion limits are dependent on mesh size, the appropriate opment of reinforcement strain at various points. Although the peak
value of εf was carefully selected based on the specific mesh size by trial strains tended to be underestimated, the permanent strains were rea-
and error. The value of 0.1 was found to give the best fit with test re- sonably predicted. In particular, the yield strains were reached at all the
sults in terms of concrete cracking and spalling. designated points in the tests with the exception of G3 in D-0.91, which
To account for the dynamic effect due to the instantaneous removal was well reproduced in the simulations.
of column, the rate effect option of the CSCM_CONCRETE model is The crack pattern and damage mode of D-0.91 and D-1.16 as shown
turned on by setting the model parameter “IRATE” to 1 [48]. By this op- by effective plastic strains were also compared with the experimental
tion, the CSCM model becomes a rate-dependent model since the ones in Fig. 10. It is seen that the area with the distribution of extensive
strength of the material increases with an increasing rate for compli-
ance with published experimental results [52]. This strength enhance-
ment due to high strain rate is usually expressed by a strength enhance-
ment factor which is the ratio of dynamic strength to static strength
with regard to a specific strain rate. The rate effect formulations imple-
mented in the CSCM_CONCRETE model is based on Simo et al. [53] ex-
tension of the commonly used Duvaut-Lions formulation to cap models
[48]. These default formulations were also applied in this study to con-
sider the effect of strain rate on concrete material.
A symmetric piecewise elastic-plastic model was used for steel rein-
forcement by using keyword *MAT_PIECEWISE_LINEAR_PLASTIC. This
model is featured by multi-linear relationship after yielding. For this
model, the mass density, Poisson's ratio, elastic modulus, yield strength
and ultimate strength should be provided. Besides, the erosion of rein-
forcement element was also applied to represent fracture of rebars
when the strain exceeds the ultimate elongation ratio of steel. The de-
tailed properties of reinforcing bars with different diameters can refer
to Table 2. For reinforcing steel, the rate effect formulations proposed
by Cowper and Symonds [47] is implemented and the strength en-
hancement is applied to both yield strength and ultimate strength [47].

Fig. 6. Comparison of quasi-static load-displacement curves between FEM and

experiment.

5
K. Qian et al. Journal of Building Engineering 45 (2022) 103554

Fig. 7. Comparison of damage patterns at failure under quasi-static loading.

of the numerical model. The model was then utilized to investigate the
load transfer paths and progressive collapse resistance of beam-slab
substructures following sudden removal of a penultimate perimeter col-
umn. Moreover, the effects of pertinent parameters, such as damping
ratio, slab thickness and column removal scenarios, were evaluated sys-
tematically.

3. Dynamic structural behavior against progressive collapse

The investigation on the variation of axial forces of columns as well

as the development and distribution of internal forces or stresses in the
slabs and beams helps shed light on the structural behavior of beam-
slab substructures against progressive collapse. Moreover, the dynamic
progressive collapse resistance of beam-slab substructures is evaluated
by using incremental dynamic analysis, the result of which is employed
to assess the reliability of the dynamic increase factor method and en-
ergy-based method for estimating the structural resistance under sud-
Fig. 8. Comparison of numerical and experimental vertical displacement. den column removal scenarios.

cracks and the direction of cracks agree well with the experimental 3.1. Gravity load redistribution
ones. Basically, wide flexural cracks concentrated around the perimeter
of the slab near the interior beams, and concrete spalling and crushing Fig. 11 (a) and (b) plot the numerical results of axial forces of
associated with formation of plastic hinges were observed at the longi- columns before and after the sudden loss of the penultimate column,
tudinal and transverse beam ends as well as along the slab diagonals. i.e., B4, for Specimen D-0.91 and D-1.16, respectively. It is obvious
Furthermore, the edge transverse beam A3-A4 was subjected to com- that the gravity load originally carried by B4 should be redistributed to
bined bending and torsion originated from the downward deformation the adjacent columns. As can be seen from Fig. 11, both D-0.91 and D-
of the slab. Obvious spiral cracks developed in the edge beam of Speci- 1.16 experience increases in the axial compressive forces of columns
men D-1.16 due to the heavier load (Fig. 10(c)). This is also well indi- A4, C4 and B3 but slight decreases in the axial compressive forces of
cated by the strain distribution of the numerical result (Fig. 10(d)). columns A3 and C3. In particular, column B3 normally exhibits the
In summary, the good agreement in terms of vertical displacement largest increase in the axial compressive force. For D-0.91, the axial
response, reinforcement strain and crack pattern confirmed the validity compressive force of column B3 increases from 40.2 kN to 61.1 kN at

6
K. Qian et al. Journal of Building Engineering 45 (2022) 103554

Fig. 9. Comparison of development of reinforcement strain at: (a) G1; (b) G2; (b) G3.

7
K. Qian et al. Journal of Building Engineering 45 (2022) 103554

Fig. 10. Comparison of numerical and experimental crack pattern: (a) Test of D-0.91; (b) FEM of D-0.91; (c) Test of D-1.16; (d) FEM of D-1.16.

the instant of collapse of B4 and eventually stabilizes at 56.2 kN at the sistance of RC beam-slab substructures will be discussed in next sec-
end of disturbance, which undertakes up to 44% of the load shifted tion.
from unloaded columns including A3 and C3. At the meanwhile,
columns A4 and C4 take over 26% and 30%, respectively, of the reallo- 3.2. Dynamic progressive collapse resistance
cated load. Similarly, the columns B3, A4 and C4 gain 39%, 29% and
32%, respectively, of the shifted load in the case of Specimen D-1.16. Given the tested beam-slab substructure specimen subjected to sud-
Details about the variation of axial forces in the remaining columns are den removal of the penultimate perimeter column did not fail under the
listed in Table 3. It is clear that the tributary load of B4 has been redis- UDL up to 11.9 kPa [37], the dynamic progressive collapse resistance of
tributed primarily through the longitudinal and transverse beams to the specimen was determined by means of incremental dynamic analy-
the adjacent columns, while columns A3 and C3 with no beams con- sis based on the validated numerical model. The obtained vertical dis-
necting to B4 suffer from uplift action as a result of the deformation of placement time-histories at the column stub of B4 with increasing UDL
the slab. It is also worth noting that the transient increases in the axial are shown in Fig. 12. As the UDL increased from 6.9 kPa to 20.6 kPa,
forces of relevant columns at sudden loss of B4 are much more signifi- i.e., 0.73(DL + LL) to 1.82(DL + LL), the stabilized vertical displace-
cant than the permanent changes due to dynamic effect, reaching ment after removal of B4 increased from 13 mm to 258 mm. The speci-
37.9% and 10.8% more for the D-0.91 and D-1.16, respectively. The men would be unable to survive from the loss of the penultimate
dynamic effect has actually made the substructure more vulnerable to perimeter column once the UDL further increased to 21.5 kPa, since the
progressive collapse. The dynamic effect on the progressive collapse re- substructure could not find a new balanced position with the vertical

8
K. Qian et al. Journal of Building Engineering 45 (2022) 103554

Fig. 11. The change of axial forces in columns:(a) Specimen D-0.91; (b) Specimen D-1.16.

Table 3
The change of axial forces in columns after load distribution of Specimen D-0.91 and D-1.16.
Specimen A3 B3 C3 A4 B4 C4

ΔP(kN) ΔR(kN) ΔP(kN) ΔR(kN) ΔP(kN) ΔR(kN) ΔP(kN) ΔR(kN) ΔP(kN) ΔR(kN) ΔP(kN) ΔR(kN)

D-0.91 −4.2 −3.4 20.9 16.0 −5.2 −4.9 12.8 9.5 −27.7 13.9 10.8
D-1.16 −5.8 −4.3 19.4 16.9 −6.4 −4.9 14.9 12.7 −34.4 16.0 14.1

Note:ΔPandΔRdenote the transient peak and residual axial force relative to the origin value, respectively.

Fig. 12. Vertical displacement time-histories based on dynamic incremental

analysis.

displacement steadily increasing with the running time. The dynamic

progressive collapse resistance of the substructure subjected to sudden
removal of the penultimate perimeter column is thus determined as Fig. 13. Load-displacement capacity curves obtained from different ap-
20.6 kPa. The load-displacement capacity curve in terms of the magni- proaches.
tude of UDL versus the stabilized vertical displacement after sudden
loss of the column based on the incremental dynamic analysis is plotted approaches. The quasi-static load-displacement curve as shown in Fig.
in Fig. 13. The curve exhibits three phases as identified by the structural 13 was obtained from the quasi-static pushdown test on an identical
stiffness. Substantial reduction in the stiffness is found with the dis- specimen under the UDL loading scheme [28]. In this manner, the static
placement between around 50 mm and 120 mm, which is attributable progressive collapse resistance of the specimen is obtained as 23.5 kPa.
to the crushing of concrete destroying the function of CAA and CMA, The dynamic increase factor (DIF), defined as the ratio of the static pro-
whereas the subsequent partial recover of the stiffness is owing to the gressive collapse resistance to dynamic progressive collapse resistance,
development of TCA and TMA. This will be discussed in section 3.3. is thus obtained as 1.14 for the beam-slab substructure. The value of the
For comparison purpose, the progressive collapse resistance of the DIF obtained in this work is smaller than the value of 1.6 reported by
RC beam-slab substructure under penultimate perimeter column re- Nyunn et al. [45] for RC beam-slab building subjected to the loss of an
moval scenario was also evaluated using quasi-static and pseudo-static edge column or a corner column, and is much smaller than the recom-

9
K. Qian et al. Journal of Building Engineering 45 (2022) 103554

mended value of 2.0 according to DoD [8]. The smaller DIF of the load ment and TCA at large vertical displacement. The maximum axial com-
in this work can be attributed to the structural resistance at failure in- pression force is largely dependent on the lateral constraint stiffness at
volving significant development of plasticity, whereas the structural re- the beam ends, while the maximum axial tension force depends on the
sponse covered in the prior study was limited to the elastic range under lateral constraint stiffness as well as reinforcement ratio and reliable
typical design load. This is consistent with the knowledge that the DIF anchoring [31]. This explains the diverse axial forces developed at the
depends on the ductility of structures, with structures with greater duc- cross-sections L1 and R1 of the longitudinal beam near the corner col-
tility having lower DIFs as discussed in DoD [8] Annex C. On the other umn and intermediate perimeter column. This also explains the weak
hand, the ultimate deflection of the specimen under quasi-static loading CAA and TCA of the transverse beam at the cross-section T2 near the
is almost double the maximum dynamic displacement that could be sus- column stub as indicated by the slight internal force, since the rota-
tained before eventual collapse, indicating that the dynamic effect is tional stiffness of the longitudinal beam can provide limited constraint
even more detrimental to the deformation capacity of the structure. for the transverse beam.
For the pseudo-static approach, the pseudo-static response as shown On the other hand, the development of CMA and TMA of slabs is
in Fig. 13 is derived by employing the Izzuddin's energy method [34] as most simply illustrated by the horizontal movement seen at the oppo-
follows site sides of slab as shown in Fig. 15. The movement is first outward
(negative) and then in gradual transition into inward movement (posi-
(2) tive), indicating the development of CMA and TMA of slabs at the small
and large deflection, respectively.
Fig. 16 shows the reinforcement stress of the RC beam-slab sub-
where PPS(ud) denotes the pseudo-static load resisting capacity at
structure at two different deformed states against the progressive col-
the displacement demand ud; and PQS(u) represents the load-
lapse, i.e., at the vertical displacement of 35 mm and 400 mm. At the
displacement capacity curve under quasi-static loading. The progres-
small deflection as shown in Fig. 16(a), the large stress mainly concen-
sive collapse resistance determined in this way is found to be 20.2 kPa,
trates at the beam and slab reinforcement around the column stub and
which is very close to the estimation of 20.6 kPa based on the incre-
the beam ends. Besides, the development of sagging moment can be
mental dynamic analysis. It is thus suggested that the pseudo-static
clearly identified at the column stub region of the longitudinal beam
analysis essentially relying on energy equilibrium is competent to pre-
with the top and bottom reinforcement in compression and tension,
dict the dynamic progressive collapse resistance on the basis of mea-
whereas the hogging moment is identified at both far ends of longitu-
sured quasi-static load-displacement curve. However, the deformation
dinal beam. This confirms the CAA mechanism is dominant in the lon-
capacity of the substructure is still tremendously overestimated with
gitudinal beam at the small deflection. At the large deflection as
this approach as inherited from the quasi-static response.
shown in Fig. 16(b), while near the beam end the tension stress of top
reinforcement has greatly exceeded the compression stress of bottom
3.3. Load transfer mechanisms reinforcement, the bottom reinforcement outside the hogging moment
region of longitudinal beam is largely mobilized into considerable ten-
The development and distribution of internal forces and stresses will sion, confirming the dominant development of TCA. On the contrary,
be explored to help illustrate the load transfer mechanisms of RC beam- the transverse beam shows the dominant development of CAA in the
slab substructures against progressive collapse. Fig. 14 shows the axial overall process. This can be attributed to the composite effect between
forces developed at the beam sections as labeled for model D-1.82, the slab and beam working compatibly as the T-section beam, making
where D-1.82 indicates the application of the UDL equal to 1.82 times the flexural mechanism of the transverse beam prevail even at very
the design load of the prototype frame and corresponds to the case large deflection. This has enhanced the contribution of transverse
where the specimen will eventually collapse after loss of the penulti- beam to the progressive collapse resistance of the specimen at the
mate perimeter column as discussed above. As can be seen, the axial small deflection. However, after quick reduction of axial compression
force development at beam cross-sections L1, R1 and T1 near the re- at the cross-section T1 due to the concrete crushing, the transverse
maining columns show similar characteristics. These axial forces are beam is still slow in evolving into TCA as shown by Fig. 14, resulting
initially compression and then turn into tension, indicating the sequen- in far less contribution at the large deflection. Most slab reinforcement
tial mobilization of CAA of beams at relatively small vertical displace- (both at the top and bottom) except those near the interior longitudi-

Fig. 14. Development of beam axial force in D-1.82 against progressive col-
lapse. Fig. 15. Horizontal movements developed at opposite sides of slab for D-1.82.

10
K. Qian et al. Journal of Building Engineering 45 (2022) 103554

Fig. 16. Reinforcement stress of D-1.82: (a) at deflection of 35 mm; (b) at deflection of 400 mm.

nal beam are in remarkable tension at the large deflection, confirming structure resistance was used for the demonstration. In the B-D model,
the mobilization of TMA of slabs. the UDL was applied in terms of the tributary loads to the respective
beams and attention was paid to ensure the vertical reaction force at the
3.4. Contribution of beams and slabs base of the column stub was the same as that in the beam-slab model.
Similar to the beam-slab substructure, the dynamic progressive collapse
To better demonstrate the relative contribution of the beams and resistance of the bare frame under sudden removal of penultimate
slabs to the dynamic progressive collapse resistance of the substructure, perimeter column was determined by conducting incremental dynamic
the beam-column model (denoted as B-D hereafter) built by removing analysis with the B-D model. The obtained load-displacement capacity
the slab of the beam-slab model was studied and the ratio of the beam- curve of B-D is plotted in Fig. 17, along with that of the beam-slab
column substructure resistance to the corresponding beam-slab sub- model (denoted as BS-D). As can be seen, the dynamic progressive col-

11
K. Qian et al. Journal of Building Engineering 45 (2022) 103554

55 mm to 40 mm, while the vibration lasts for a shorter period as can

be seen in Fig. 18. With the damping ratio further increasing to 8%, no
noticeable changes in the structural response are observed.

4.2. Effect of slab thickness

On the basis of the model D-0.91, the effect of slab thickness on the
dynamic response of RC beam-slab substructures under column re-
moval is investigated by gradually reducing the slab thickness from
64 mm in the reference test to 35 mm while maintaining same rein-
forcement details. Note that the reinforcement ratios of the slab with
different slab thicknesses all satisfied the requirement of the minimum
and maximum reinforcement ratio. Considering the beam depth of
112 mm, the studied slab thickness of 64 mm, 57 mm, 50 mm, 42 mm
and 35 mm correspond to a slab thickness-to-beam depth ratio of 0.57,
0.51, 0.45, 0.38 and 0.31, respectively. The comparison of the dynamic
response as shown in Fig. 19 suggests that the effect becomes very evi-
dent once the slabs are thinner than 50 mm, which is about 1/45 of the
span. While the change of the resulted vertical displacement is within
37% when the slab thickness varies between 64 mm and 50 mm, the
maximum vertical displacement rises to 95.49 mm and 147.01 mm
Fig. 17. Comparison of the load-displacement capacity curve between B-D and when the slab thickness is reduced to 42 mm and 35 mm, respectively,
BS-D. which are 121% and 241% larger than that with slabs of 64 mm thick.
It is apparent that thinner slabs are weaker in developing CMA. How-
lapse resistance of the bare frame was found to be 15.9 kPa, only 77% ever, thin slabs are also unfavorable for development of CAA and TCA
of the beam-slab substructure resistance, whereas the corresponding of beams. Fig. 20 shows that slabs, as the thickness reducing from
vertical displacement significantly increased from 258 mm to 459 mm
as a result of elimination of the slabs. It is obvious that the presence of
slabs contributes to both the structural resistance and stiffness of the
beam-slab substructures. Furthermore, the contribution of slabs is esti-
mated by subtracting the load-displacement curve of B-D from that of
BS-D as shown in Fig. 17. It is seen that the slabs yield at around 7 kPa,
which agrees well with the predicted yield load of the slabs using yield-
line theory [28]. Overall, the slabs are mobilized prior to beams and
make dominant contribution to the structural resistance at deflections
smaller than 200 mm which is approximately three times the slab
depth.

4. Effect of structural parameters and column removal scenarios

4.1. Effect of damping ratio

Fig. 18 shows the structural responses of beam-slab substructures

with different damping ratios subjected to sudden column removal. As
the damping ratio increases from 0% to 5%, smaller deformation can Fig. 19. Effect of slab thickness on the dynamic response of substructures.
develop in the substructure following the sudden loss of the column
since the resulted vertical displacement is found to decrease from

Fig. 18. Effect of damping ratio on structural deformation capacity. Fig. 20. Effect of slab thickness on axial force of A4-B4 beam left end section.

12
K. Qian et al. Journal of Building Engineering 45 (2022) 103554

64 mm to 35 mm, not only leads to 34% reduction in the maximum ax-

ial compression force in the left end of A4-B4 beam, but also make the
beam evolve into axial tension at a smaller deflection and develop 37%
less ultimate axial tension.

4.3. Effect of column removal scenarios

4.3.1. Effect of lost column(s)

Apart from the loss of a single bearing column, the accidental loss of
multiple columns is also likely to take place. The different scenarios of
removing one more column adjacent to B4 are investigated, such as re-
moval of the corner column A4, interior column B3 and another
perimeter column C4. Fig. 21, showing the dynamic displacement re-
sponse of substructure under the different column removal scenarios,
suggests that the progressive collapse resistance of beam-slab substruc-
tures is strongly associated with the lost columns. Obviously, the loss of
more bearing columns would make the substructures more prone to
progressive collapse. It can be seen that the loss of one more adjacent
column besides the penultimate perimeter column would incur at least
3 times larger deflection. Furthermore, among the double-column re-
moval scenarios, the loss of the corner column with the penultimate col-
umn represents the most hazardous situation to the progressive collapse
resistance of the substructure. As is seen, the substructure imposed with
0.91 times the design load will undergo considerable deflection shortly
after the removal of columns A4 and B4, and the deflection keeps in-
creasing slowly afterwards until leading to collapse. This indicates a dy-
namic progressive collapse resistance of 0.91 times the design load,
which is only 50% of the resistance to progressive collapse under re-
moval of only B4. Similar findings were reported in the experimental
study by Qian et al. [28]. The main reason for this is that no reliable
secondary load-carrying mechanisms, including CAA, TCA, CMA and
TMA, can be developed after the loss of both corner and penultimate
columns.

4.3.2. Effect of column removal sequence and interval

Fig. 22 shows the vertical displacement time-histories at the col-
umn stub of A4, assuming columns A4 and B4 are removed simultane-
ously, or one after another with very short, short and long interval to
have increasingly less overlap between dynamic response of substruc-
ture after sudden removal of each column. The results indicated that
the multi-column removal sequence and interval has little effect on the
maximum vertical displacement developed. Besides, the final changes
of the axial forces in the columns are also identical to those in the si-
multaneous removal scenario as shown in Fig. 23, despite the fact that Fig. 22. Vertical displacement of A4 for: (a) sequential removal of A4 and B4;
(b) sequential removal of B4 and A4.
the gravity load is redistributed twice for the sequential removal of the
two columns. However, the various multi-column removal sequences
produce slightly different damage patterns in the substructure as
shown in Fig. 24. The beams and slabs in span B–C exhibit wider
cracks when columns A4 and B4 are removed at the same time,
whereas the extreme damage can spread to span A-B especially at the
top of beam end adjoining B4 when B4 is removed some time after the
removal of A4 and at the bottom of beam ends when B4 is removed
prior to A4.

5. Conclusions

In this paper, a series of high fidelity solid-element-based numerical

models of RC beam-slab substructures are calibrated based on the test
data from Qian and Li [37], and then are utilized to investigate the dy-
namic progressive collapse resistance and load transfer mechanisms un-
der the penultimate perimeter column removal scenario. The pertinent
geometric and structural parameters affecting the progressive collapse
resistance of beam-slab substructures, as well as diverse column re-
moval scenarios, are explored. The main conclusions are drawn as fol-
lows:
Fig. 21. Dynamic response with loss of different column(s).

13
K. Qian et al. Journal of Building Engineering 45 (2022) 103554

Fig. 23. Reaction forces for: (a) simultaneous and (b) sequential removal of A4 and B4.

Fig. 24. Crack pattern for: (a) simultaneous; (b) sequential removal of A4 and B4; (c) sequential removal of B4 and A4.

1. After loss of a bearing column, the gravity load is redistributed 2. Based on the dynamic progressive collapse resistance
primarily through the transverse and longitudinal beams to the determined via the incremental dynamic analysis, a dynamic
adjoining columns, and axial forces of these columns can be increase factor (DIF) of 1.14 is obtained by dividing the
increased by as much as 30%–60%. Moreover, with the column experimental quasi-static progressive collapse resistance by the
suddenly removed, the transient increases in the axial forces of dynamic resistance, which is significantly less than the value of
remaining columns are much more significant than the permanent 2.0 as advised by the DoD design guidelines for design of
changes due to the dynamic effect, making the substructure more buildings to resist progressive collapse. On the other hand, the
vulnerable to progressive collapse. energy equilibrium-based pseudo-static analysis can provide very
precise estimation of the structure's progressive collapse

14
K. Qian et al. Journal of Building Engineering 45 (2022) 103554

resistance, implying a simple alternative tool to the dynamic Facil. 12 (3) (1998) 110–112.
[5] K.A. Seffen, Progressive collapse of the World trade center: simple analysis, J.
analysis. However, both quasi-static and pseudo-static Eng. Mech. 134 (2) (2008) 125–132.
approaches significantly overestimate the deformation capacity [6] American Society of Civil Engineers (ASCE), Minimum Design Loads for
of substructures. Buildings and Other Structures, ASCE7-05, Reston, VA, 2005.
[7] B.S. Institution, National Annex to Eurocode 1: Actions on Structures - Part 1-7,
3. The progressive collapse resistance of the RC beam-slab
Accidental Actions, 2006.
substructures is mainly provided by the compressive arch action [8] Department of Defense (DoD), Design of buildings to resist progressive collapse,
(CAA) of the longitudinal beam and especially the flexural ufc 4-023-03, Change (1-27 January 2010).
mechanism of the transverse beam, owing to the significant [9] General Services Administration (GSA), Progressive collapse analysis and design
guidelines for new federal office buildings and major modernization projects, 2003.
“composite effects” between the slab and beam, at the small [10] J. Yu, K.H. Tan, Structural behavior of RC beam-column subassemblages under a
deflection, and by the tensile membrane action (TMA) of the middle column removal scenario, J. Struct. Eng. 139 (2) (2012) 233–250.
slabs and the catenary action (TCA) of the longitudinal beam at [11] Y. Su, Y. Tian, X. Song, Progressive collapse resistance of axially-restrained frame
beams, ACI Struct. J. 106 (5) (2009) 600–607.
the large deflection. [12] J. Yu, K.H. Tan, Experimental and numerical investigation on progressive
4. The presence of slabs can improve the progressive collapse collapse resistance of reinforced concrete beam column sub-assemblages, Eng.
resistance of the RC beam-column counterpart by as much as Struct. 55 (1) (2013) 90–106.
[13] W.J. Yi, Q.F. He, Y. Xiao, S.K. Kunnath, Experimental study on progressive
30%. Especially, the slabs are mobilized prior to beams and make collapse resistant behavior of reinforced concrete frame structures, ACI Struct. J.
dominant contribution to the structural resistance against 105 (4) (2008) 433–439.
progressive collapse at deflections smaller than three times the [14] F. Sadek, J.A. Main, H.S. Lew, Y. Bao, Testing and analysis of steel and concrete
beam-column assemblies under a column removal scenario, J. Struct. Eng. 137 (9)
slab depth. However, a minimum slab thickness about 1/45 of
(2011) 881–892.
span length is vital to the development of CMA of slabs as well as [15] F.L. Wang, J. Yang, Sandeep. Shah, effect of horizontal restraints on progressive
to improve the development of CAA and TCA of beams. collapse resistance of precast concrete beam-column framed substructures, J. Civ.
5. A large damping ratio helps dissipate energy, enabling smaller Eng. 24 (3) (2020) 879–889.
[16] Y. Zhou, T. Chen, Y. Pei, H.J. Hwang, X. Hu, W. Yi, L. Deng, Static load test on
deflection and milder vibration. But excessive damping with the progressive collapse resistance of fully assembled precast concrete frame structure,
damping ratio greater than 5% hardly bring about further Eng. Struct. 200 (2019) 109719.
improvement. [17] M. Sasani, J. Kropelnicki, Progressive collapse analysis of an RC structure, Struct.
Des. Tall. 17 (4) (2010) 757–771.
6. The loss of one more column dramatically increases the risk of [18] D.C. Feng, S.C. Xie, Y. Li, L. Jin, Time-dependent reliability-based redundancy
progressive collapse. The location of the removed columns is assessment of deteriorated RC structures against progressive collapse considering
found to have greater impact on the structural resistance against corrosion effect, Struct. Saf. 89 (2021), 102061.
[19] D.C. Feng, S.C. Jun, Y. Li, K. Qian, Robustness quantification of reinforced
progressive collapse than the column removal sequence and concrete structures subjected to progressive collapse via the probability density
interval. The loss of the corner column in addition to the evolution method, Eng. Struct. 202 (2020) 109877.
penultimate perimeter column represents the most hazardous [20] A.t. Pham, K.H. Tan, Experimental study on dynamic responses of reinforced
concrete frames under sudden column removal applying concentrated loading,
situation among the double-column removal scenarios, which can
Eng. Struct. 139 (2017) 31–45.
reduce the progressive collapse resistance of the substructure by [21] J. Yu, K.H. Tan, Experimental and numerical investigation on progressive
50%. The maximum deflection developed is hardly affected collapse resistance of reinforced concrete beam column sub-assemblages, Eng.
whether these two columns are removed simultaneously or one Struct. 55 (2013) 90–106.
[22] D.C. Feng, S.C. Jun, W.N. Deng, Z.D. Ding, Probabilistic failure analysis of
after another with very short, short and long interval. reinforced concrete beam-column sub-assemblage under column removal scenario,
Eng. Fail. Anal. 100 (2019) 381–392.
Author statements [23] D.C. Feng, S.C. Xie, C.L. Ning, S.X. Liang, Investigation of modeling strategies for
progressive collapse analysis of RC frame structures, J. Perform. Constr. Facil. 33
(6) (2019) 04019063.
Kai Qian: Conceptualization. Xin-yi Chen: Methodology and Formal [24] P.X. Dat, K.H. Tan, Experimental study of beam-slab substructures subjected to a
Analysis. Ting Huang: Conceptualization and Writing- Original Draft penultimate-internal column loss, Eng. Struct. 55 (2013) 2–15.
[25] K. Qian, B. Li, Slab effects on response of reinforced concrete substructures after
Preparation. loss of corner column, ACI Struct. J. 109 (2012) 845–855.
[26] P.X. Dat, K.H. Tan, Experimental response of beam-slab substructures subject to
Declaration of competing interest penultimate-external column removal, J. Struct. Eng. 141 (2015) 014170.
[27] X.Z. Lu, K.Q. Lin, Y. Li, H. Guan, P.Q. Ren, Y.L. Zhou, Experimental investigation
of RC beam-slab substructures against progressive collapse subject to an edge-
The authors declare that there is no conflict of interest regarding the column-removal scenario, Eng. Struct. 149 (2017) 91–103.
publication of this paper. [28] K. Qian, B. Li, Z.W. Zhang, Influence of multicolumn removal on the behavior of
RC substructures, J. Struct. Eng. 142 (5) (2016) 04016006.
[29] Z. Tan, W.H. Zhong, L.M. Tian, Y.H. Zheng, B. M, S.C. Duan, Numerical study on
Acknowledgements collapse-resistant performance of multi-story composite frames under a column
removal scenario, J. Build. Eng. 44 (2021) 102957.
The authors would like to acknowledge the financial support by the [30] K. Qian, B. Li, Resilience of flat slab structures in different phases of progressive
collapse, ACI Struct. J. 113 (3) (2016) 537–548.
National Natural Science Foundation of China (Nos. 52022024, [31] J. Yu, L.Z. Luo, Y. L, Numerical study of progressive collapse resistance of RC
51778153), the Natural Science Foundation of Guangxi Province beam-slab substructures under perimeter column removal scenarios, Eng. Struct.
(No.2021GXNSFFA196001) and the Research Foundation of Guilin 159 (2018) 14–27.
[32] J. Yu, T. Rinder, A. Stolz, K.H. Tan, W. Riedel, Dynamic progressive collapse of
University of Technology (No. GUTQDJJ 2019124). an RC assemblage induced by contact detonation, J. Struct. Eng. 140 (6) (2014)
04014014.
References [33] M. Liu, A new dynamic increase factor for nonlinear static alternate path analysis
of building frames against progressive collapse, Eng. Struct. 48 (2013) 666–673.
[34] B.A. Izzuddin, A.G. Vlassis, A.Y. Elghazouli, D.A. Nethercot, Progressive collapse
[1] ASCE/SEI 7, Recommendations for Designing Collapse-Resistant Structures,
of multistorey buildings due to sudden column loss-part 1: simplified assessment
Structural Engineering Institute-American Society of Civil Engineers, Reston, VA,
framework, Eng. Struct. 30 (2008) 1308–1318.
2010.
[35] A.T. Pham, K.H. Tan, J. Yu, Numerical investigations on static and dynamic
[2] F. Kiakojouri, V.D. Biagi, B. Chiaia, M.R. Sheidaii, Progressive collapse of framed
responses of reinforced concrete sub-assemblages under progressive collapse, Eng.
building structures: current knowledge and future prospects, Eng. Struct. 206
Struct. 149 (2017) 2–20.
(2020) 110061.
[36] K. Qian, B. Li, Quantification of slab influences on the dynamic performance of
[3] H. Griffiths, A. Pugsley, O.A. Saunders, Report of the Inquiry into the Collapse of
RC frames against progressive collapse, J. Perform. Constr. Facil. 29 (1) (2015)
Flats at Ronan Point, presented to the Minister of Housing and Local Government,
04014029.
HMSO, Canning Town, 1968, pp. 58–66.
[37] K. Qian, B. Li, Dynamic and residual behavior of reinforced concrete
[4] W.G. Corley, P.E.M. Sr, M.A. Sozen, C.H. Thornton, The Oklahoma City bombing:
substructures following instantaneous removal of a column, Eng. Struct. 148
summary and recommendations for multihazard mitigation, J. Perform. Constr.
(2017) 175–184.

15
K. Qian et al. Journal of Building Engineering 45 (2022) 103554

[38] K. Qian, B. Li, Dynamic performance of RC beam-column substructures under the Manual – Volume II, Material Models, California, 2020.
scenario of the loss of a corner column—experimental results, Eng. Struct. 42 [48] Y.D. Murray, User’s Manual for LS-DYNA Concrete Material Model 159, APTEK,
(2012) 154–167. Inc., Colorado, 2007, p. 92.
[39] Y. Tian, Y. Su, Dynamic response of reinforced concrete beams following [49] CEB-FIP Model Code, Design of Concrete Structures, Fédération Internationale
instantaneous removal of a bearing column, Int. J. Concr. Struct. Mater. 5 (1) du Béton fib/International Federation for Structural Concrete, 2010.
(2011) 19–28. [50] J. Yu, L.Z. Luo, Y. Li, Numerical study of progressive collapse resistance of RC
[40] Z.H. Peng, S.L. Orton, J.R. Liu, Y. Tian, Experimental study of dynamic beam-slab substructures under perimeter column removal scenarios, Eng. Struct.
progressive collapse in flat-plate buildings subjected to an interior column removal, 159 (2018) 14–27.
J. Struct. Eng. 144 (8) (2018) 04018094. [51] A.T. Pham, K.H. Tan, J. Yu, Numerical investigations on static and dynamic
[41] Z.H. Peng, S.L. Orton, J.R. Liu, Y. Tian, Experimental study of dynamic responses of reinforced concrete sub-assemblages under progressive collapse, Eng.
progressive collapse in flat-plate buildings subjected to exterior column removal, J. Struct. 149 (2017) 2–20.
Struct. Eng. 143 (9) (2017) 04017125. [52] D.L. Grote, S.W. Park, M. Zhou, Dynamic behavior of concrete at high strain rates
[42] S.L. Orton, J.E. Kirby, Dynamic response of a RC frame under column removal, J. and pressures: I. experimental characterization, Int. J. Impact Eng. 25 (2001)
Perform. Constr. Facil. 28 (9) (2014) 04014010. 869–886.
[43] Y. Zhou, X. Hu, Y. Pei, H.J. Hwang, T. Chen, W.J. Yi, L. Deng, Dynamic load test [53] J.C. Simo, J.G. Kennedy, S. Govindjee, Non-smooth multisurface plasticity and
on progressive collapse resistance of fully assembled precast concrete frame viscoplasticity. Loading/unloading conditions and numerical algorithms, Int. J.
structures, Eng. Struct. 214 (2020) 110675. Numer. Methods Eng. 26 (10) (2010) 2161–2185.
[44] J.M. Adam, M. Buitrago, E. Bertolesi, J. Sagasete, J.J. Moragues, Dynamic [54] J. Yu, Y.P. Gan, J. Wu, H. Wu, Effect of concrete masonry infill walls on
performance of a real-scale reinforced concrete building test under a corner-column progressive collapse performance of reinforced concrete infilled frames, Eng.
failure scenario, Eng. Struct. 210 (2020) 110414. Struct. 191 (2019) 179–193.
[45] S. Nyunn, F. Wang, J. Yang, Q.F. Liu, I. Azim, S. Bhatta, Numerical studies on the [55] Y.H. Weng, K. Qian, F. Fu, Q. Fang, Numerical investigation on load
progressive collapse resistance of multi-story RC buildings with and without redistribution capacity of flat slab substructures to resist progressive collapse, J.
exterior masonry walls, Struct 28 (2020) 1050–1059. Build. Eng. 29 (2020) 101109.
[46] CP 65, Structural Use of Concrete, Part1. Code of Practice for Design and [56] J. Yu, Y.P. Gan, J. Liu, Numerical study of dynamic responses of reinforced
Construction, 1999, Singapore Standard. concrete infilled frames subjected to progressive collapse, Adv. Struct. Eng. 24 (4)
[47] Livermore Software Technology Corporation (LSTC), LSDYNA: Keyword User’s (2020).

16