Defence Technology 25 (2023) 174e191

Contents lists available at ScienceDirect

Defence Technology
journal homepage: www.keaipublishing.com/en/journals/defence-technology

Effects of connection types and elevated temperature on the impact

behaviour of restrained beam in portal steel frame
Yu-Xu Guo a, c, Feng Xi a, c, *, Ying-Hua Tan a, c, Feng Liu b, Ya-Chao Hu a, b
a
School of Civil Engineering, Shandong Jianzhu University, Jinan, 250101, PR China
b
Shandong Provincial Key Laboratory of Civil Engineering Disaster Prevention and Mitigation, Shandong University of Science and Technology, Qingdao,
266590, PR China
c
Key Laboratory of Building Structural Retrofitting and Underground Space Engineering, Shandong Jianzhu University, Jinan, 250101, PR China

a r t i c l e i n f o a b s t r a c t

Article history: Based on the background of structural protection and Disaster Reduction Engineering, the dynamic
Received 10 February 2022 behaviour and failure mechanism of restrained beams in portal steel frames in localised fire are inves-
Received in revised form tigated via experimental measurement and numerical simulation techniques. Comprehensive parametric
21 March 2022
studies are carried out to discuss the influence of end connection types, temperature, impact velocity,
Accepted 26 April 2022
Available online 4 May 2022
impact mass and span-to-depth ratio (SDR) on the dynamic response of the beams. The characteristics of
deformation, internal force and energy distribution about the restrained beams and its joints are
investigated. A temperature dependent criterion for evaluating the frame joint performance is proposed
Keywords:
Restrained beam
to measure the degree of performance degradation and impact resistance of the joint. The dynamic
Connection displacement amplification factor in different temperature environments are proposed for the different
Temperature beam end constraint types and SDRs. Results of the experimental and numerical analysis show that the
Impact welded connection (WC) of three typical joint types is the strongest, and the extended endplate
Failure modes connection (EEC) is the weakest in terms of the impact resistance performance. With regard to the failure
Internal force mechanism mode of the joint, the failure positions of the WC and the welded-bolted connection are located in the
inner web of the column. Meanwhile, the EEC is located in the connection position between the beam
and the endplate. Three different internal force stages and two obvious critical temperature boundaries
of the restrained beams emerge with the increase in temperature, and they have significant character-
istics in terms of deformation trend, internal force transfer and energy distribution. During the impact, a
phenomenon known as “compression arch action” develops into “catenary action” with the increase in
deflection in the frame beam mechanism.
© 2022 China Ordnance Society. Publishing services by Elsevier B.V. on behalf of KeAi Communications
Co. Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/
licenses/by-nc-nd/4.0/).

1. Introduction behaviour of structures under extreme loads, such as fire, explosion

and impact, are essential in the study of the performance of steel
Concerns regarding the design of structures to withstand fire structures, which can be seriously affected by these extreme loads.
and blast have intensified after the “9$11” event in the USA. The Recently, the study of steel structures subjected to distinct
Federal Emergency Management Agency, in conjunction with the extreme loads (e.g. fire and impact) has attracted considerable
American Society of Civil Engineers, highlighted in the accident attention. Studies on the fire [2e4] and impact [5e7] resistance of
investigation report [1] that ‘the performance design of steel single components, such as steel beam [8,9], column and plate
frames under the combined effects of fire, explosion and impact [10,11], have been extensively investigated. A number of scholars
should be undertaken’. The dynamic response and failure worldwide have discussed the fire resistance of steel frame struc-
tures through various methods, such as experiments and numerical
simulations. Zhao and Kruppa [12] investigated the overall stability
* Corresponding author. School of Civil Engineering, Shandong Jianzhu University of steel frames subjected to fire via several fire tests. The relevant
Jinan, 250101, PR China. parameters were analysed by numerical simulation. Chen and
E-mail address: [email protected] (F. Xi). Wang [13] proposed the three numerical simulation modelling
Peer review under responsibility of China Ordnance Society

https://doi.org/10.1016/j.dt.2022.04.017
2214-9147/© 2022 China Ordnance Society. Publishing services by Elsevier B.V. on behalf of KeAi Communications Co. Ltd. This is an open access article under the CC BY-NC-
ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Y.-X. Guo, F. Xi, Y.-H. Tan et al. Defence Technology 25 (2023) 174e191

methods for studying steel frames exposed to fire, and their explored through finite element (FE) analysis to better understand
strengths and weaknesses were also discussed. The importance of the response behaviour and connection performance of restrained
joints for steel frame resistant fire performance cannot be exten- beams subjected to impact loading during a fire. The values of the
sively emphasised. A large number of numerical analytical models dynamic displacement amplification factor in different tempera-
on steel frame bolt nodes are used to investigate the fire resistance ture environment are suggested for the impact protection design of
of structures under the influence of different bolt diameters, bolt steel structures with different beam end constraint types and SDRs.
strength ratings, endplate dimensions and strengths [14]. Lindberg Moreover, the structural dynamic response characteristics of
and Pedersen [15] focused on examining the plastic dynamic restrained beams in fire are discussed and compared in detail to
response of portal frame through drop hammer experiment and ambient temperature. The significance of the above work was to
numerical simulation in the study of the impact resistance of steel integrate structural behaviour in engineering with theoretical
frame structures. The influence of certain key parameters, such as studies for future research on structural behaviour theory and
connection type, impact location, flange thickness and span-to- refinement of safety and security design guidelines.
depth ratio (SDR), on the load carrying mechanism, load carrying
capacity, damage mode and energy absorption of steel frames has
been highlighted [16e18], suggesting that the SDR of the beam has 2. Experiment presentation
a great influence on the impact resistance of steel frames.
At present, research on the structural behaviour under the 2.1. Structural model of the impact experiment at an elevated
combined action of extreme loads, such as fire, impact and explo- temperature
sion, is still in the development stage. Accordingly, only a small
number of relevant studies have been published. Liew et al. [19e22] The present test design is inspired by the pioneering work on
conducted a study associated with the behaviour of a multi-storey the steel frame by B. Lindberg [15], in which a one-story steel frame
3D steel frame structure subjected to fire following an explosion. with rectangular cross-section members are used and only impact
They also analysed the interaction effects of explosion and fire in loading is considered. The present study aims to investigate the
steel structures via numerical simulation. Liew and Chen [23] characteristic response of the steel frame when subjected to com-
suggested a hybrid element method for the analysis of steel frame bination of fire and impact loading. Therefore, considering the
structures subjected to fire following a local explosion to investi- feasibility of the experiment and avoiding the influence of more
gate the ultimate load carrying capacity of multi-storey steel frames factors, the single-story structure is the optimal choice in the initial
and steel columns. Sun et al. [24] investigated the blast resistance stage of exploring effects of such complex loads. In spite of this, to
and local damage analysis of steel tubes after being exposed to fire closer to the practical situation, I-shaped section of members and
by experimental methods. Forni et al. [25,26] investigated the three types of connections wield used in current practical engi-
strain rate effect of S355 steel at high temperatures through SHTB neering are adopted in present test model.
experiments, identified the parameters associated with the JeC Referring to the impact experiment of the portal frame model
constitutive model and examined the blast response of steel col- discussed in Ref. [15], a steel frame structure (Fig. 1) is made up of a
umns in fire. The structural response to fire and impulsive load can beam and two columns and their connections. The three different
be catalogued by the following two cases: (a) impulsive load fol- typical connections shown in Fig. 2 are considered. Rigid connec-
lowed by a fire (IFF) and (b) impulsive load during a fire (IDF) [27]. tions (RC) include welded connection (WC) and welded-bolted
In the IFF case, the duration of action of the explosive or impact load connection (WBC), which are the welding technique and the
is much shorter than that of the fire, and no significant coupling combination of bolting and welding technique, respectively. The
effect was observed between the two loads. Scholars worldwide extended endplate connection (EEC) is regarded as a semi-rigid
can possibly investigate this case by decoupling the thermal and connection (SRC) that uses a combination of bolting and welding
dynamic responses. The IDF case showed a relatively complex techniques. These three types of connections are widely used in
coupling effect compared with the IFF case. The strain rate effect at engineering practice. These three connections with different rota-
high temperatures is one of the most significant challenges, and the tional stiffness mean different end constraints. Therefore, they can
effect of thermal expansion on the mechanical properties of the be applied to reflect the end restraint effect of adjacent structures
material must be taken into account. Relevant research needs to be
supplemented and improved.
However, experimental research and in-depth parameter anal-
ysis on the dynamic behaviour of structures under the combined
action of fire and impact loading are scarce, which is the motivation
for this study. Currently, the main criterion for evaluating the joint
performance is the rotational angle under static load (FEMA-355D;
AISC-360). To the author's knowledge, there is no report on the
relevant methods to evaluate the joint connection performance
under impact load and the influence of temperature effect.
In this study, experiments and numerical simulation had been
carried out to investigate the failure modes and impact resistance
of three common joint types of restrained beams subjected to
impact following a localised fire. A novel temperature-dependent
criterion for effectively evaluating the degree of degradation of
joint performance was proposed. The deformation of restrained
beams and their characteristics during deformation, such as inter-
nal forces and energy, had been analysed in detail to examine the
influence of joint types, temperatures and SDRs, impact velocity
and impact mass on the dynamic response. The failure processes
and mechanisms of the restrained beams had been extensively Fig. 1. Simple portal frame subjected to fire and impact loading.

175
Y.-X. Guo, F. Xi, Y.-H. Tan et al. Defence Technology 25 (2023) 174e191

Fig. 2. Details of the beam-to-column connections: (a) WC; (b) WBC and (c) EEC.

on the frame beam. The portal frame in Fig. 1 is a sub-structure. On heating system, a drop hammer impact machine and a record-
the one hand, the columns and connections in the structure reflect acquisition system. Three sets of components, namely, seven
the translational and rotational restraint effects on the beam ends. specimens, were employed in the experimental design scheme. The
On the other hand, the frame beam is a more general end elastic drop hammer has a mass 137 kg and fell from 7.2 m. The impact face
restrained beam. is semi-circular with a radius of 12.5 mm and 80 mm width
The two steel columns and the connecting joints are assumed to orthogonal to the beam axis. The overall height of the column ho
be protected from fire, so only the beam heats up in the event of a and the span length of the specimens between two columns lo were
fire in the room. When the frame beam reaches a certain temper- 0.9 m and 1.6 m, respectively. During the experiment, the 10# I-
ature T, it is subjected to impact by the upper falling layer. Mass M is beam was applied to the steel beams and columns, which were
assumed to impact the central section of the beam with velocity vo. made of Q235B. Grade 8.8 M12 high strength bolts were used for
The temperature is assumed to be constant because the impact connection in the specimen. The details of the connections and
process is short (in ms). The portal fame is characterised by the dimensions are shown in Fig. 2. During the impact test, stiff re-
overall height ho and length lo, and the cross-sectional dimensions straint beams are placed in front of and behind the upper section of
are shown in Fig. 1. the column to prevent the out-of-plane deformation of the frame
column as shown in Fig. 3. The temperature is monitored by ther-
mocouple to reach the established temperature.
2.2. Experimental program The experiment was carried out according to the following
steps:
The dynamic behaviour of portal steel frame structures in fire
subjected to impact loading was investigated at elevated temper- (a) The steel frame was fixed to the support base of the exper-
atures via drop hammer impact experiments (Fig. 3). The device iment bench by bolts.
used in the experiments is made up of several systems, including a

Fig. 3. Experimental set-up: 1 traction chains, 2 counterweight blocks, 3 hammerhead, 4 directional track, 5 RHC, 6 moving track, 7 support base, 8 track-supporter, 9 bolts for fixing
the base, 10 steel frame specimen, 11 heating system and 12 stiff restraint beams.

176
Y.-X. Guo, F. Xi, Y.-H. Tan et al. Defence Technology 25 (2023) 174e191

(b) The frame beam was placed in the resistive heating chamber flexural buckling of the frame beam occurred when the tempera-
(RHC), and asbestos was used to wrap the inner wall of the ture increased. Moreover, only the slightest bulge occurred to the
RHC for insulation. The heating process follows the ISO-834 endplates of the ECC steel frame at ambient temperature. However,
standard fire curve. the crack between one and two rows of bolts at the endplate
(c) Thermocouple temperature sensors were used to monitor exhibited a typical characteristic of shear fracture at 400
C. A slight
the temperature of the steel beams in real time. Once the local buckling of the impact response was observed in the ambient
temperature reached the specified value (20
C, 400
C and temperature environment. The different degrees of
800
C), the RHC was unlocked and removed along the torsionaleflexural buckling of the beam at elevated temperature,
moving track as soon as possible. including distortional buckling, are caused by the temperature ef-
(d) The drop hammer was released by the loosening control fect. This phenomenon is uncommon in the ambient temperature
clamps after the anti-disengagement safety device had been environment. Furthermore, this phenomenon is not limited to
suspended. The beam was subjected to impact of the drop plastic deformation at the joints, but develops towards fracture.
hammer at the middle span. The measured results of the frame subjected to impact in
localised fire, including the vertical and lateral displacement at the
During the experiment, several calibrated high-speed cameras mid-span of the beam and axial displacement at the column top,
were employed to capture the process of the impact. After the are shown in Table 1. The aforementioned table illustrates minimal
experimental steps were completed, the final deformation of the differences in the impact resistance of the three types of joints at
steel frame structure was determined by traditional measurement ambient temperature. The deflection was greatest for the EEC steel
methods to support the results of the high-speed camera. frames and least for the WC steel frames when the temperature
increased to 400
C. The above-mentioned phenomena suggested
that better impact resistance was provided by the WC compared
2.3. Experimental results
with the WBC, and EEC was the worst at elevated temperatures.
This phenomenon was significantly intensified with the increase in
The experimental stages of portal steel frame structure sub-
temperature.
jected to extreme load mainly consists of two loading process
phases, namely, static heating stage and impact response stage
when reaching a certain temperature. It should be noted that the 3. FE model
temperature is basically unchanged due to the ephemeral impact
phase. The physical state of the steel frame structure is shown in The experimental behaviour of the above-mentioned structure
Fig. 4(a) at the end of the first stage, i.e. when a certain temperature can be reproduced to test the validity of the numerical results using
has been reached. It can be seen that at this time, the temperature ABAQUS for numerical simulation. Further parameter analysis can
of the frame beam has increased to the set value, while no signif- be carried out to comprehensively and extensively investigate the
icant temperature increase occurred in the frame columns and structural response and connection performance.
connections due to temperature isolation measures. In the second
stage, i.e. after being impacted by free falling hammer, the struc- 3.1. Structural models and analytical steps
tural deformation state is shown in Fig. 4(b). It can be observed that
the frame beam has exhibited significant in-plane bending defor- Given the structural symmetry, semi-structured methods can be
mation, local deformation in the mid-span and a certain degree of employed to effectively improve computational efficiency. In the
out-of-plane displacement. numerical simulation, the drop hammer was replaced by a Mass.
The diagrams of the final deformation and permanent dis- The Mass was 68.5 kg, and its velocity is 11.87 m/s before the
placements can be created using photographs and data, and part of impact. In this study, the SDR was in the range of 8e32, and the
the experimental result figures is shown here. The overall defor- temperature was in the range of 20e800
C to investigate the ef-
mation of the steel frames subjected to impact by a drop hammer at fects of SDR and temperature on the impact response of the
400
C is shown in Fig. 5. The mid-span displacements of the beam structure. The data of the main models are shown in Table 2.
of the three types of joints gradually increased with the increase in During the impact, different parts of the frame components
temperature. Only a slight local deformation of the impacted area came into contact with each other. The different types of contact
was observed at ambient temperature. An exacerbation of local must be reasonably simplified and simulated. “Tie” contact was
deformation in the span and significant out-of-plane torsional- used to set the position connected together by welding [18].

Fig. 4. Experimental results of two stages: (a) Static heating stage; (b) Impact response stage.

177
Y.-X. Guo, F. Xi, Y.-H. Tan et al. Defence Technology 25 (2023) 174e191

Fig. 5. Experimental results of the three types of joints: 400
C and 11.87 m/s: (a) WC; (b) WBC; (c) EEC.

Table 1
Measurement results of the experiment.

Type of connection T/
C Displacement at the mid-span/cm Axial displacement at the column top/cm

Vertical Lateral

WC 20 11.4 2.0 0
400 16.6 3.0 2.5
800 28.0 4.5 5.7
WBC 20 12.0 2.4 0
400 21.0 3.5 3
EEC 20 12.2 2.3 0
400 27.7 4.6 4.3

Table 2 surface of the frame beam is simulated by radiation and convection,

Part of the model conditions. and the temperature field distribution in the steel beam is deter-
Type of joint l/h T/
C mined by heat conduction, as shown in Fig. 7. The radiation factor,
WC, WBC, EEC 8, 12, 16, 20, 24, 28 and 32 20e800
convection coefficient, and thermophysical parameters should be
defined. According to the EC1 [28], the convection coefficient of the
directly heated surface for the frame beams was given as 9 W/m2

C, whilst the other areas were taken as 4 W/m2
C. The radiation
Surface-to-surface contact was applied in the other parts of the
factor of the directly heated surface for the frame beams was set to
specimen to ensure the correctness of the contact and continuation
1.0, whilst the other areas were taken as 0.8.
of the subsequent calculations. The master and slave faces of each
The coupled temperatureedisplacement static and dynamic
interface pairs must be manually searched to guarantee that no
analysis modules were adopted, and the calculation process was
penetration will occur between the meshes. The tangential and
carried out as follows:
normal behaviours of contact were accounted for in the contact
properties. Penalty friction formulation was selected for the
(a) Buckling analysis was conducted to obtain the buckling
tangential behaviours of contact, and its value was set to 0.3 [14,16].
mode and the initial geometric imperfection of the beam to
The nonlinear penalty method was used in Fig. 6 to improve the
predict the possible out of plane behaviour of the structure
convergence of the calculation and solver stability for the normal
According to EC3 [29], the amplitude of the geometric
contact behaviours. The value of Ke was the representative under-
imperfection was 1/150 of the deformation value in the first
lying element stiffness. The heating effect of the heating box on the
mode.

178
Y.-X. Guo, F. Xi, Y.-H. Tan et al. Defence Technology 25 (2023) 174e191

an overall view of the 3D detailed FE model of an assembly by using

the WBC and EEC.

3.3. Constitutive model with temperature and strain rate effect

The full process of the material from elasticity, plasticity to

damage failure, as well as the corresponding strain rate and tem-
perature effects, needs to be accounted for to achieve refinement in
the numerical simulation. The material model is shown in Fig. 9.
The engineering stressestrain curve can be obtained through the
uniaxial tensile test of the material. The classical metal plasticity
model in ABAQUS was used to simulate the nonlinear behaviour of
the materials. The material properties in the model should be
defined by the true stressestrain relationship instead of the engi-
neering stressestrain relationship. The main indexes of members
other than bolts are as follows: Young's modulus is 206 GPa, the
yield stress is 290 MPa, and the ultimate tensile stress is 430 MPa.
Meanwhile, the Young's modulus and yield stress of the bolts are
Fig. 6. Nonlinear penalty pressureeoverclosure relationship curve.
210 GPa and 640 MPa.
The structural steel materials have two damage modes, namely,
ductile failure and shear failure (DeS). The initial equivalent plastic
strain is related to stress triaxiality. Relevant established studies
have concluded that the smaller two initial equivalent plastic
strains can be applied [17] and calculated by using Eq. (1) and Eq.
(2):
8
>
> 1
> 1:13
> h
>
> 3
>
εD <
p
1 10
¼ 0:04 þ 0:86expð  0:7hÞ   h  (1)
εu >> 3 3
>
>
>
>
>
: 0:12 10
h
3
8
> 5
>
> 0:43 h
>
> 3
p
εS <
¼ 5 (2)
εu > > 0:38 þ 0:40exp½6:69ðh  2Þ   h 2
>
> 3
>
:
0:78 2h

p p
where εD is the initial equivalent plastic strain for ductile failure; εS
is the initial equivalent plastic strain for shear failure; εu is the
tensile fracture plastic strain of steel; and h is stress triaxiality.
DUCTCRT and SHRCRT have been defined as the ductile and
shear damage initiation criteria, respectively. The corresponding
Fig. 7. Schematic of the method of thermal transfer.
damage factors according to Eq. (3) and Eq. (4) [30] are employed to
characterise the development degree of the material from the yield
(b) Static analysis was carried out to introduce gravity on the to the damage initial point, as shown in the (b) and (c) stages of
whole structure and application of the bolt load [30,31]. Fig. 9.
(c) Heat transfer analysis of the steel frame was conducted, ð
dεp
which was raised to the specified temperature in accordance uD ¼ p (3)
εD ðh; ε_ Þ
p
with the ISO-834 standard fire curve.
(d) Impact dynamic analyses were performed.
ð
dεp
uS ¼ p (4)
εS ðh; ε_ Þ
p

3.2. Elemental types The energy method is used to describe the progression of
damage. Fracture energy Gf can be calculated according to Eq. (5)
Components, such as beams, columns and endplates, including
[30], where εpl
0 is the equivalent plastic strain at the onset of
bolts, were modelled using solid elements (C3D8RT that considers
the temperature degrees of freedom). Mass adopts discrete rigid damage, εpl
f
is the equivalent plastic strain at failure, and L is the
shell element (R3D4). All components have been meshed utilised characteristic length of the element. The process must be idealised
the structured meshing technique, except for bolts and mass block, because the curve of the damage evolution stage measured by the
which are meshed using the free meshing technique. Fig. 8 shows unidirectional tensile test is actually in a complex stress state. At
179
Y.-X. Guo, F. Xi, Y.-H. Tan et al. Defence Technology 25 (2023) 174e191

Fig. 8. FE models of WBC and EEC.

Fig. 10. Strain rate and temperature effects on εpl

cr .

Fig. 9. Stressestrain curve.

The reduction in material stiffness after the onset of damage was

ambient temperature, the peak tensile stress and failure strain εu characterised using the stiffness reduction rate SDEG (i.e. “D” in
obtained from uniaxial tensile test is 430 MPa and 0.287, respec- Fig. 9). When SDEG >0 was satisfied, material damage started to
occur, and when SDEG ¼ 1 was satisfied, damage to the material
tively. Eq. (1) can be used to calculate εpD (i.e., point c in Fig. 9). The
completely and macroscopically manifests as fracture failure and
fracture energy related to the element characteristic length L can be
ultimate strength, as recommended by EC3 (Table 3).
calculated base on the approach [29] of using a monotonically
decreasing linear function to describe the damage evolution pro-
cess and Eq. (5). The fracture energy Gf at ambient temperature is
εpf
calculated to be 4.176  104 L N/m2. The calculation method of ð
fracture energy at elevated temperatures are consistent with those Gf ¼ Lsy dεp (5)
at ambient temperatures. The influence of temperature effect on εpD εp0
and εu can be considered through ICR in Fig. 10 and EC3 [29].

180
Y.-X. Guo, F. Xi, Y.-H. Tan et al. Defence Technology 25 (2023) 174e191

Table 3
Reduction of the material properties by the effect of temperature.

T/
C E/GPa fy;q /MPa fu;q /MPa ICT

20 206 290.0 430.0 1.00

200 185 290.0 362.5 1.00
400 144 290.0 290.0 1.03
600 61.8 133.4 133.4 1.32
800 16.5 34.8 34.8 2.03

8
<0 for T < Tt
T * ¼ ðT  Tt Þ=ðTm  Tt Þ for Tt < T < Tm (6)
:
1 for Tm < T

where T is the current temperature, Tm is the melting temperature,

and Tt is the transition temperature defined as the one at or below
which there is no temperature dependence of the yield stress.
With regard to the temperature effect, the reduction factors of
Young's modulus and yield strength was adopted. The Fig. 11. Strain energy and deformation curves with different element sizes.
CowpereSymonds equation was used to account for the strain rate
effect, where the relevant parameters are taken as follows: C ¼ 40.4
small. Hence, the effectiveness of the calculation model is verified.
and P ¼ 5 at T ¼ 20
C [32]; C ¼ 400 and P ¼ 1 at T ¼ 1000
C [23,33].
The typical results of deformation comparison between the
However, accurate data about temperature-related parameters C
experiment and the FE are shown in Fig. 13. The basic trend and size
and P are lacking in the literature, and linear interpolation method
of deformation are consistent. In the WC-16-800, the buckling
has been widely accepted. In this study, the values are taken by
produced by the web at the impacted position of the beam is
linear interpolation when the temperature is between 20
C and
consistent with the FE results. In the EEC-16-400, a significant shear
1000
C [27]. The effect of temperature and strain rate on the
fracture occurred between the first and the second rows of bolts in
damage initiation equivalent plastic strain of Q235B has been
the endplate during the experiment. Fig. 13(c)eFig. 13(e) show a
investigated in previous studies [34]. The initiation equivalent
comparison of the structural transient deformation captured with a
plastic strain was amplified or reduced by introducing influencing
high-speed camera with the numerical simulation results. In terms
factors related to the strain rate and temperature, defined as ICSR
of both time course and deformation trends, both results for the
and ICT, respectively (Fig. 10). Parameter T * in Fig. 10 is related to
several typical joints are basically consistent. The comparison re-
the material properties and temperature, and it is calculated by
sults show that the same torsionaleflexural buckling characteris-
using Eq. (6). The effect of temperature and strain rate on the
tics (including direction, amplitude and moment) are developed at
density r and Poisson's ratio n of the materials was neglected,
the mid-span. In summary, the experimental behaviour can be
where r is taken as 7850 kg/m3 and n as 0.3.
accurately simulated by using a material constitutive model
The recommended values of expansion coefficient, conductivity
considering damage evolution.
and specific heat capacity with temperature in EC3 were adopted
for the thermophysical parameters [29].
4.2. Connection failure modes
3.4. Mesh sensitivity study
In this section, ductile damage and shear damage are used to
The influence of element size on hourglass energy and defor- analyse the failure mode of joints. Fig. 14 depicts the contours of
mation mode should be considered to improve the accuracy of the damage for three joint types at 400
C.
FE results. Fig. 11 presents the curves of the element sensitivity The damage areas of WC and WBC are mainly located in the web
results, and Fig. 12 shows the mid-span deformation of the beam, of the column. The ductile damage and shear damage are distrib-
where ALLAE is the total artificial strain energy; and ALLIE is the uted. The damage area of EEC is mainly concentrated in the
total strain energy. Deformation and energy stability were used as connection between the endplate and the beam, including the
criteria to assess the accuracy of the model. No local buckling second row of bolt holes, and the damage type is dominated by
occurred at the mid-span when the element size is less than shear damage. It is worth pointing out that the heights of the three
3.5 mm. Based on energy and deformation stability analysis, the joint damage areas are situated in the tension zone above the
element size of beam and joints is 2.5 mm, and that of column is neutral axis of the beam. The direction of the displacement field in
3.5 mm. Therefore, the number of elements with different joint the plane is plotted in Fig. 15. The aforementioned figure shows that
types and SDR for the FE models can be determined (Table 4). EEC presents a state of unidirectional tension, whilst the WC and
WBC demonstrate the state of upper tension and lower compres-
4. Numerical results sion. In terms of failure modes, there is a kind of cooperative failure
mode among beams, columns and joints occurring in WC and WBC,
4.1. Comparison with the experimental results defined as connection failure. The synergistic failure of the beam
and endplate, defined as the end of beam failure, occurs at the EEC.
The mid-span displacement of the frame beam obtained by This notion means that a certain load carrying capacity is still
experimental and numerical simulation is shown in Table 5. The available after the failure, and the strength of the endplate plays an
comparative analysis shows that the error between the results is important role in the performance of the EEC.

181
Y.-X. Guo, F. Xi, Y.-H. Tan et al. Defence Technology 25 (2023) 174e191

Fig. 12. Deformation comparison with different element sizes.

Table 4
The number of element.

Connection type WC WBC EEC

SDR SDR ¼ 8 211,590 211,926 271,370

SDR ¼ 32 544,230 531,126 604,010

Table 5 column. There was a novel approach that the degree of reduction in
Comparison of the experimental (Exp) and FE results of deformation. impact resistance of the three joints steel frame in fire was assessed
Item # dV/cm Error/% dL/cm Error/%) dA/cm Error/% by comparing the internal forces of the three connections with the
Exp FE Exp FE Exp FE
TRC in this section. The schematic of the TRC is shown in Fig. 16.
The resisting bending moment at the connections was an
WC-16-20 11.4 12.1 6.14 2.0 2.2 10.00 0 z0 z0
important index of frame performance of the steel frames. Quan-
WC-16-400 16.6 15.2 8.43 3.0 2.9 3.33 2.5 2.4 4.00
WC-16-800 28.0 26.1 6.78 4.5 4.2 6.67 5.7 5.3 7.02 titative analyses of joint performance degradation were conducted
WBC-16-20 12.0 11.3 5.83 2.4 2.2 8.33 0 z0 z0 by comparing the bending moments of the column section at the
WBC-16-400 21.0 19.6 6.67 3.5 3.2 8.57 3 2.7 10.00 connection and bending moment degradation coefficient b defined
EEC-16-20 12.2 11.9 7.37 2.3 2.1 8.70 0 z0 z0 as follows:
EEC-16-400 27.7 26.3 5.03 4.6 4.4 4.35 4.3 4.1 4.65

Note: WC-16-20 represents the WC frame with 16 of the SDR of the beam at 20
C. X
MHm
dV, dL and dA represent vertical displacement at the mid-span, lateral displacement at b¼ X
(7)
the mid-span and axial displacement at the column top, respectively. MRm

X
where MHm is the maximum bending moment of the column
4.3. Mechanism analysis and parametric study X
section at the WC, WBC and EEC with the SDR of X, and MRm is for
the TRC steel frame.
4.3.1. Internal forces in plane and deformations
During the impact, the out-of-plane internal forces were far less
The crossover point of the neutral axis extensions at the beam
than the in-plane internal forces even though the out-of-plane
and column sections was chosen as the coupling intersection. The
displacement occurred in the frame beam. Thus, the major focus
beam and columsn sections were linked at the crossover point to
in this section was on in-plane forces (i.e. axial force and bending
simulate the theoretical rigid connection (TRC) of the beam and
moment).
182
Y.-X. Guo, F. Xi, Y.-H. Tan et al. Defence Technology 25 (2023) 174e191

The history curves of the axial force and bending moment of the
beam sections at the connection are shown in Fig. 17. The afore-
mentioned figure demonstrates that the axial forces in the beam at
ambient temperature were expressed as pressure throughout the
impact. The axial force exhibited pressure first with the increase in
temperature. Then, the axial force gradually decreased to transform
into tension during the impact. It was worth noting that this phe-
nomenon in which the direction of the axial force changed inten-
sified with the increase in temperature. Meanwhile, the peak and
duration of pressure tended to decrease.
In addition, the temperature significantly affected the bending
moment at the connection of beam section, and a negative corre-
lation between temperature and bending moment was shown. The
“compression arch action (beam action)” was confirmed by the
phenomenon of axial force in the beam. An increasing axial force
was observed during the impact, which was the characteristic of
the “catenary action”. A clear demonstration of the internal force
mechanism is shown in Fig. 18. The time at which the frame beam
began to be the catenary action and the duration of each phase can
be easily distinguished on the basis of this criterion.
The curves of the rotational angles and axial displacement at the
connection with temperature for the three joints after impact
loading are shown in Fig. 19, where the rotational angle j was
calculated as in Eq. (8):

    
p ðDLD  DLC Þ ðDLA  DLB Þ
j¼ arctan  arctan (8)
180 LCD LAB

where DLA, DLB, DLC and DLD are the axial displacements of points A
to D in Fig. 2, respectively; and LCD and LAB are the distances be-
tween the two points.
Fig. 19 shows a positive correlation between the rotational angle
and the temperature for the three joints when the temperature was
within a certain range. However, the significant and varying de-
grees of reduction in the rotational angles of all joints suddenly
occurred at a certain temperature. Then, the rotational angle
slightly increased with the further increase in temperature. The
three processes of rotational angle were divided into compressive-
bending stage (CBS), coupling compressive-bending and tension-
bending stage (CCB-TBS) and tension-bending stage (TBS) based
on the characteristic of the forces at the ends of the frame beams in
each phase. Tcr1 and Tcr2 were defined as critical temperatures for
the aforementioned three stages.
The rotational angle of the joints was related to the internal
force mechanism variation of the frame beam. During the phase of
the compression arch action, a greater rotational angle was
observed due to the larger bending moment at the connection. At
the phase of the catenary action, a smaller rotational angle could be
found because of the smaller bending moments at the connection.
Nevertheless, a greater deformation occurred due to the increase in
the lateral displacement at the mid-span of the beam, which did
not mean that the rotational angle could be increased at the phase
of the catenary action. Based on the above analysis, a more
convincing explanation can be provided for the phenomenon of
rotational angle changes. The frame beams were mainly in the in-
ternal force mechanism of the compression arch action during the
CBS. At this moment, a decrease in stiffness and strength of the steel
due to the effect of temperature resulted in a certain reduction in
the bending moment at the beam end. However, the rotational
Fig. 13. Comparison of the deformations from the experiment and FE models: (a) EEC- angle tended to increase because the effect of increasing deflection
16-400; (b) WC-16-800; (c) WC-16-400; (d) EEC-16-400; (e) WBC-16-400. was stronger than the decrease in bending moment. In the CCB-TBS,

183
Y.-X. Guo, F. Xi, Y.-H. Tan et al. Defence Technology 25 (2023) 174e191

Fig. 14. Contours of damage for three joint types.

Fig. 15. In plane displacement field diagrams: (a) WC and WBC; (b) EEC.

a significant degradation of the compression arch action was TBS compared with the ambient temperature. At this stage of the
observed in the early stages of the impact, followed by a progres- catenary action, the tensile force at the beam end has caused an
sion towards the catenary action, which made a sudden reduction axial displacement of the column, allowing the occurrence of
in the rotational angle. The bending moment reduced to 25% at the increased rotational angle. Fig. 19 shows that the horizontal

184
Y.-X. Guo, F. Xi, Y.-H. Tan et al. Defence Technology 25 (2023) 174e191

4.3.2. Energy distribution

The characteristic of energy transformation can be employed to
further analyse the state of energy absorption in beams and col-
umns during the impact and the transformation of the mechanism
in the steel frames. In this section, an energy analysis was carried
out for the main energy absorbing components (beams and other
components) during the impact, with the WC as an example. A
fundamental physical law, the Law of Conservation of Energy, ap-
plies to a confined system and is also valid for FE calculations,
satisfying Eq. (9) and Eq. (10).

EI þ EFD þ EK ¼ ET (9)

EI ¼ ES þ EP (10)

where EFD, EK and ET are total energy dissipated through frictional

effects, total kinetic energy and total system energy, respectively;
and EI, ES and EP are the total strain energy, the total elastic strain
energy and the total plastic strain energy, respectively. In this study,
PEb and PEc were defined as the total plastic strain energy of the
beams and columns, respectively.
Fig. 16. Schematic of the TRC. Fig. 22 shows the normalized histogram about the ratio of the
value of PEb and PEc as a percentage of ET, in which the red and blue
colors represent beams and other components, respectively. The
displacement at the column end was essentially the same at tem-
figure demonstrates that PEb tended slowly decrease at tempera-
peratures below Tcr1 and rapidly increased when the temperature
tures below Tcr1 (basically in the range of 0.7e0.8). However, a
was above Tcr1, which could also provide an explanation for the
sudden rising trend of PEb was observed at temperatures above Tcr1,
above-mentioned phenomenon. In addition, a significant tilting of
which ranged from 0.8 to 0.9. The conclusions obtained in the
the frame columns was observed, which was the hallmark char-
previous section regarding the transformation of the mechanisms
acteristic of the catenary action.
can be adopted to explain the above-mentioned phenomenon. In
WC was used as an example to analyse the relationship between
the CCB-TBS, less axial forces and bending moments are exerted on
the bending moment of the column sections at the connection
the column end because of the initial development of ‘catenary
during the impact process and the bending moment degradation
action’ when the temperature varies from Tcr1 to Tcr2, allowing the
coefficient b with temperature, which is shown in Fig. 19(d). The
column to be less involved in energy absorption. The above analysis
maximum and average bending moments decreased with the in-
can be taken as an explanation for the increase of PEb. When the
crease in temperature, which also supported the above conclusion
temperature was above Tcr2, the one step further development of
that the degradation of compression arch action has occurred. The
the catenary action made it evident that axial tension forces
main feature of the ‘compression arch action’ was the development
increased. However, the bending moment continued to decrease,
of bending moments and axial force at the beam end to work in
resulting in an increase in PEc since the tipping of columns.
coordination with the columns. With regard to the bending
An energyetime analysis of the frame columns was carried out
moment degradation coefficient, the figure shows that the value of
to further investigate the energy distribution of the steel frame
the WC was 0.7 at ambient temperature, which was 28% higher
during impact. The time curves of PEc/ET are shown in Fig. 23. Due to
than that at 600
C. Accordingly, a negative correlation exists be-
the transformation of response mechanism, the energy dissipation
tween temperature and b. The effect of temperature, so to speak,
under the influence of temperature is not monotonic. The figure
plays an important role in the performance degradation of the steel
demonstrates that a significant variation in the energy absorption
frame.
process of the column occurred, especially when the temperature
The history curves of the impact force and the curves of the peak
was higher than Tcr1. A hysteresis phenomenon about the devel-
and duration of the impact force with temperature are shown in
opment of plastic was also observed in the column. The higher the
Fig. 20 and Fig. 21, respectively. The comparison in Fig. 21 indicated
temperature was, the longer the duration of hysteresis, and the
that the peak and duration of the impact force were negatively and
hysteresis phenomenon disappeared at the temperature below Tcr1.
positively correlated with temperature, respectively. In the three
The reasons for the above phenomenon can be explained as follows.
phases of the impact (impact, stabilisation and rebound), the in-
First, less axial forces and bending moments were exerted on the
crease in temperature had essentially no effect on the duration of
columns at the beginning of the impact in the small-deformation
the impact phase. However, the duration of the rebound phase had
stage, leaving the frame columns uninvolved in the energy ab-
been increased due to the effect of temperature. In this section, the
sorption. When entering the large-deformation phase, a gradually
relationship between impact force and duration with temperature
increasing axial tension was applied to the column end, causing the
has been described. Furthermore, the lower efficiency in absorbing
frame columns to begin to engage in the overall energy absorption
transient impact energy was exposed at elevated temperatures. The
at this time. The explanations also resulted in a time delay for the
FE results, associated with impact forces, were expected to be in
frame column to be involved in energy absorption. The hysteresis
good agreement with the previous studies [35], validating the ac-
phenomenon has been allowed to be exposed in combination with
curacy of the numerical simulation.
the above-mentioned explanations. Moreover, the hysteresis

185
Y.-X. Guo, F. Xi, Y.-H. Tan et al. Defence Technology 25 (2023) 174e191

Fig. 17. History curves of the axial force and bending moment.

phenomenon was a characteristic of the development for catenary mid-span and bending moment degradation coefficient b for
action. different connection types with diverse temperatures and SDR. In
terms of the size of the rotational angle, the order from largest to
4.3.3. Effects of temperature and SDR smallest was WBC, WC and EEC. A trend of varying degrees of in-
In this section, the impact response of the steel frames under the crease for TCR1 and TCR2 was observed with the increase in SDR,
influence of temperature and SDR would be investigated for three where the TCR1 of WC was highest, and the EEC was lowest.
different connection forms, namely, WC, WBC and EEC. Fig. 24(c) demonstrates that the vertical displacement at the mid-
Fig. 24 shows the rotational angle, vertical displacement at the span of the EEC, WBC and WB decreased in turn. At ambient

186
Y.-X. Guo, F. Xi, Y.-H. Tan et al. Defence Technology 25 (2023) 174e191

Fig. 18. Internal force variation mechanisms of frame beam: (a) Compressive arch action and (b) catenary action.

Fig. 19. Curves of the rotational angle and lateral displacement: (a) WC; (b) WBC and (c) EEC and (d) Curves of bending moment and b with temperature.

temperature, the difference of deflection in the mid-span beam smallest, as shown in Fig. 24(d). The effect of SDR had a weak in-
between the three connections was narrow when the SDR was fluence on b. However, a significant decreasing trend in b occurred
small. A widening difference emerged with the increase in tem- with the increase d decline by 29%, 32% and 21% at 600
C
perature and SDR. This phenomenon indicated that the SDR compared with the ambient temperature. At ambient temperature,
exhibited a stronger influence on deflection than temperature. the b values of the WC, WBC and EEC were approximately 0.7, 0.65
A parametric analysis based on bending moment degradation and 0.5, respectively, showing that even the performance of WC fell
coefficient b was carried out. The results showed a significant far short of the TRC. The following comprehensive conclusions
disparity in the performance of the steel frames with different about the performance of the different connection types could be
connection types d the b of WC was the largest, and EEC was the derived combined with the above-mentioned parameter analyses.

187
Y.-X. Guo, F. Xi, Y.-H. Tan et al. Defence Technology 25 (2023) 174e191

Fig. 20. History curves of the impact force. Fig. 23. History curves of PEc/ET.

For the WC steel frame, the impact resistance and synergistic

operation of the beams and columns were strongest, and the EEC
steel frame was the weakest. In addition, a significant difference in
performance to fully rigid connection was also observed, even
compared to the WC steel frame.

4.3.4. Effects of impact velocity and mass

The effects of impact velocity and mass on the mid-span
displacement of the frame beams with different end constraints
in high temperature environment are investigated via FE with the
assumption that the impact momentum remains unchanged (set
up as 500 kg m/s).
Fig. 25 shows the results of the maximum vertical displacement
at the mid-span with the various impact velocity. It can be seen that
higher impact velocity compared to larger impact mass make the
steel beam produce greater midspan displacement. The maximum
vertical displacement of large-mass impact at ambient temperature
Fig. 21. Curves of the peak and duration of impact force. is approximately 1/7 of that of high-velocity impact. Meanwhile,
the maximum vertical displacement of large-mass impact at
elevated temperature is approximately 1/8 of that of high-velocity
impact. Fig. 26 shows the ratio of the maximum vertical displace-
ment at the mid-span under high temperature (400
C and 600
C)
and normal temperature (20
C), which is recorded as d400 20
V =dV and
d600 20
V =dV (i.e., the dynamic displacement amplification factor in high
temperature environment). CB and PB in the figures are the confi-
dence band and prediction band of the dynamic displacement
amplification factor, respectively. Fig. 26 shows that the connection
types have a great influence on the dynamic displacement ampli-
fication factor. At 400
C and 600
C, the dynamic displacement
amplification factor for rigid connections (WC and WBC) is appro-
dimately approximately 1.25e1.5 and 1.5e2.0 respectively. The
dynamic displacement amplification factor for semi-rigid connec-
tion (EEC) is approximately 2.25e2.75 and 2.75e3.25 respectively.
The dynamic displacement amplification factor of the RC exposes a
higher sensitivity to SDR than that of SRC. The dynamic displace-
ment amplification factor under different impact velocities is more
stable. The SRC is relatively stable within the recommended value
range of the SDR structural design (8e20) [36], approximately 2.5
Fig. 22. Normalized histogram of (PEc/ET)/(PEb/ET). and 3.0e3.25 respectively. When the SDR is greater than 20, the
dynamic displacement amplification factor slightly decreases (2.25
and 2.75e3.0).
188
Y.-X. Guo, F. Xi, Y.-H. Tan et al. Defence Technology 25 (2023) 174e191

Fig. 24. Parametric analyses based on the connection type, temperature and SDR: (a) and (b) Influence on the rotational angle; (c) Influence on the vertical displacement; (d)
Influence on b.

Fig. 25. Maximum vertical displacement at the mid-span: (a) 400
C; (b) 600
C.

189
Y.-X. Guo, F. Xi, Y.-H. Tan et al. Defence Technology 25 (2023) 174e191

Fig. 26. Dynamic displacement amplification factor in fire: (a) 400
C; (b) 600
C.

5. Conclusions The temperatures effect had driven the development of the

internal force mechanism from the compression arch action
In this study, the dynamic behaviours and internal force to the catenary action.
mechanisms of restrained beams subjected to localised fire and (5) The effect of temperature on the dynamic displacement
sequent impact loading were systematically investigated. The amplification factor is not negligible compared to the pa-
following conclusions are drawn based on the experimental ob- rameters of the impulsive. The values of the dynamic
servations and parametric analysis by numerical simulations. displacement amplification factor in high-temperature
environment have been proposed for the different joints,
(1) The temperature effect has an important influence on the temperatures and SDRs.
deformation of the restrained beams. Only a certain degree of
local buckling occurred at the mid-span section for the three The adopted structure may not conform to the reasonable scale
connection types of restrained beams at 20
C. The endplates proportion due to the limitation of experimental conditions. The
of extended endplate connection were sheared off at 400
C, large deformation and failure of the structure may not be observed
which could not been not observed at 20
C. Meanwhile, due to the small impact mass and kinetic energy. All these de-
different degrees of torsionaleflexural buckling were ficiencies need to be improved by further experimental research in
observed at the mid-span section at 400
C. When the tem- the future.
perature was up to 800
C, distortionary buckling was
evident at the mid-span, and the load-carrying capacity was Declaration of competing interest
almost completely lost.
(2) At elevated temperature, two different failure modes were The authors declare that they have no known competing
observed in the joints, namely, the connection failure and the financial interests or personal relationships that could have
beam end failure. The failure modes of the welded connec- appeared to influence the work reported in this paper.
tion and welded-bolted connection were first damaged at
the junction of the web and the outer flange of the structural Acknowledgements
column, which belonged to “the connection failure”. The
extended endplate connection was first damaged at the This work was supported by the National natural Science
connection between the endplate and the upper flange of the Foundation of China [grant numbers 12172198, 11272189and
beam, which belonged to “the beam end failure”. 52078283]; Youth Innovation Technology Project of Higher School
(3) At elevated temperature, the welded connection and in Shandong Province [grant number 2019KJG015].
welded-bolted connection frame columns would suffer a
greater degree of damage and have lower residual load ca- References
pacity due to the different failure modes. The extended
endplate connection steel frames retained the structural [1] Federal Emergency Management Agency (FEMA. World Trade Center building
performance study. USA: FEMA; 2002.
stiffness of each floor in multi-storey frames because part of
[2] Wang YC. Steel and composite structures: behaviour and design for fire safety.
the load-bearing capacity was retained, which allowing the Spon press; 2002.
overall structure to be damaged but not collapsed and [3] Ramesh S, Choe L, Zhang C. Experimental investigation of structural steel
avoiding the progressive collapse. beams subjected to localized fire. Eng Struct 2020;218:110844. https://
doi.org/10.1016/j.engstruct.2020.110844.
(4) Different phases of force interaction were observed in the [4] Wu YW, Fan SG, He BB, et al. Research on the fire resistance design of high-
restrained beams with increasing temperature. In addition, strength steel hollow columns under axial compression. Eng Struct
critical temperatures (TCR1 and TCR2) of the three stages 2021;234:111943. https://doi.org/10.1016/j.engstruct.2021.111943.
[5] Jones Norman. Structural impact. Cambridge University Press; 1989.
existed. During the impact, the internal force mechanism [6] Nawar MT, Arafa IT, Elhosseiny OM. Numerical damage evaluation of perfo-
included “compression arch action” and “catenary action”. rated steel columns subjected to blast loading. Defence Technology 2021.

190
Y.-X. Guo, F. Xi, Y.-H. Tan et al. Defence Technology 25 (2023) 174e191

https://doi.org/10.1016/j.dt.2021.03.019. In press, https://www.sciencedirect. [21] Yu HX, Richard Liew JY. Steel framed structures subjected to the combined
com/science/article/pii/S2214914721000453. effects of blast and fire - Part 2: case study. International journal of Advanced
[7] Babaei H, Darvizeh A. Analytical study of plastic deformation of clamped Steel Construction 2005;1:85e104.
circular plates subjected to impulsive loading. J Mech Mater Struct 2012;7(4): [22] Richard Liew JY, Chen H. Explosion and fire analysis of steel frames using fiber
309e22. https://doi.org/10.2140/jomms.2012.7.309. element approach. J Struct Eng 2004;130:991e1000. https://doi.org/10.1061/
[8] Guo Z, Huang SS. Behaviour of restrained steel beam with reduced beam (ASCE)0733-9445(2004)130:7(991).
section exposed to fire. J Constr Steel Res 2016;122:434e44. https://doi.org/ [23] Chen H, Richard Liew JY. Explosion and fire analysis of steel frames using
10.1016/j.jcsr.2016.04.013. mixed element approach. J Eng Mech 2015;130:606e16. https://doi.org/
[9] Allam A, Nassif A, Nadjai A. Behaviour of restrained steel beam at elevated 10.1061/(ASCE)0733-9399(2005)131:6(606).
temperature e parametric studies. Journal of Structural Fire Engineering [24] Sun YX, Wang X, Ji C, et al. Damage effect of steel circular tube subjected to
2019;10(3):324e39. https://doi.org/10.1108/JSFE-11-2018-0036. fire and blast. J Constr Steel Res 2021;176:106389. https://doi.org/10.1016/
[10] Babaei H, Mostofi TM, Alitavoli M. Study on the response of circular thin plate j.jcsr.2020.106389.
under low velocity impact. Geomechanics & Engineering 2015;9(2):207e18. [25] Forni D, Chiaia B, Cadoni E. Strain rate behaviour in tension of S355 steel: base
https://doi.org/10.12989/gae.2015.9.2.207. for progressive collapse analysis. Eng Struct 2016;119:164e73. https://
[11] Babaei H, Mostofi TM. New dimensionless numbers for deformation of circular doi.org/10.1016/j.engstruct.2016.04.013.
mild steel plates with large strains as a result of localised and uniform [26] Forni D, Chiaia B, Cadoni E. Blast effects on steel columns under fire condi-
impulsive loading. Proc IME J Mater Des Appl 2020;234(2):231e45. https:// tions. J Constr Steel Res 2017;136:1e10. https://doi.org/10.1016/
doi.org/10.1177/1464420716654195. j.jcsr.2017.04.012.
[12] Zhao B, Kruppa J. Structural behaviour of an open car park under real fire [27] Xi F, Li QM, Tan YH. Dynamic response and critical temperature of a steel
scenarios. Fire Mater 2004;28:269e80. https://doi.org/10.1002/FAM.867. beam subjected to fire and subsequent impulsive loading. Comput Struct
[13] Chen L, Wang YC. Efficient modelling of large deflection behaviour of 2014;135:100e8. https://doi.org/10.1016/j.compstruc.2014.01.014.
restrained steel structures with realistic endplate beam/column connections [28] Eurocode 1. Actions on structures d Part 1-2:General actions d actions on
in fire. Eng Struct 2012;43:194e209. https://doi.org/10.1016/ structures exposed to fire. UK British Standards Institution; 2002.
J.ENGSTRUCT.2012.05.030. [29] Eurocode 3. Design of steel structures d Part 3-2:Mechanical properties of
[14] Chen L, Wang YC. Methods of improving survivability of steel beam/column carbon steel. UK British Standards Institution; 2006.
connections in fire. J Constr Steel Res 2012;79:127e39. https://doi.org/ [30] ABAQUS Inc. ABAQUS analysis User's manual v.2019. 2019.
10.1016/j.jcsr.2012.07.025. [31] Technical specification for high strength bolt connections of steel structures
[15] Lindberg B, Pedersen JB. Plastic deformation of impact loaded frames. Int J d Part 3.2.5: material and design index. China Construction Industry Press;
Impact Eng 1987;6:101e8. https://doi.org/10.1016/0734-743X(87)90013-3. 2011.
[16] Wang H, Yang B, Zhou XH, et al. Numerical analyses on steel beams with fin- [32] Mostofi TM, Babaei H, Alitavoli M. Theoretical analysis on the effect of uniform
plate connections subjected to impact loads. J Constr Steel Res 2016;124: and localized impulsive loading on the dynamic plastic behaviour of fully
101e12. https://doi.org/10.1016/j.jcsr.2016.05.016. clamped thin quadrangular plates. Thin-Walled Struct 2016;109:367e76.
[17] Yang B, Wang H, Yang Y, et al. Numerical study of rigid steel beam-column https://doi.org/10.1016/j.tws.2016.10.009.
joints under impact loading. J Constr Steel Res 2018;147:62e73. https:// [33] Symonds P, Jones N. Impulsive loading of fully clamped beams with finite
doi.org/10.1016/j.jcsr.2018.04.004. plastic deflections and strain-rate sensitivity. Int J Mech Sci 1972;14:49e69.
[18] Wang H, Tan KH, Yang B. Experimental tests of steel frames with different https://doi.org/10.1016/0020-7403 (72)90006-9.
beamecolumn connections under falling debris impact. J Struct Eng [34] Guo ZT, Shu KO, Gao B, et al. J-C model based failure criterion and verification
2020;146:04019183. https://doi.org/10.1061/(ASCE)ST.1943-541X.0002469. of Q235 steel. Explos Shock Waves 2018;38:1325e32 [in Chinese)].
[19] Richard Liew JY. Survivability of steel frame structures subject to blast and [35] Fujikake K, Li B, Soeun S. Impact response of reinforced concrete beam and its
fire. J Constr Steel Res 2008;64:854e66. https://doi.org/10.1016/ analytical evaluation. J Struct Eng 2009;135:938e50. https://doi.org/10.1061/
j.jcsr.2007.12.013. (ASCE)ST.1943-541X.0000039.
[20] Yu HX, Richard Liew JY. Steel framed structures subjected to the combined [36] Wang H, Yang B, Chen K, et al. Parametric analysis and simplified approach for
effects of blast and fire - Part 1: state-of-the-art review. International journal steel-framed subassemblies with reverse channel connection under falling-
of Advanced Steel Construction 2005;1:67e84. https://doi.org/10.18057/ debris impact. Eng Struct 2020;225:111263. https://doi.org/10.1016/
ijasc.2005.1.1.4. j.engstruct.2020.111263.

191