International Journal of Impact Engineering 145 (2020) 103688

Contents lists available at ScienceDirect

International Journal of Impact Engineering

journal homepage: www.elsevier.com/locate/ijimpeng

Factors influencing impact force profile and measurement accuracy in drop T

weight impact tests
⁎ ⁎
Huawei Li, Wensu Chen , Hong Hao
Center for Infrastructural Monitoring and Protection, School of Civil and Mechanical Engineering, Curtin University, Perth, Australia

A R T I C LE I N FO A B S T R A C T

Keywords: Drop weight tests on RC beams have been intensively reported in literature. Load cells are commonly used to
Drop weight measure the impact force acting on the beam. Different researchers adopted different configurations, e.g., lo-
Load cell cation of load cells in the test, which could affect the impact load measurement. Although various impact force
Mass ratio profiles have been reported by different researchers, there is no systematic study regarding the influences of the
Impact force profile
test setup on the measured impact forces. Therefore, this study numerically investigates the influences of test
RC beam
setups on impact force measurement accuracy and impact force profile of RC beam under drop weight impact. It
is found that when the load cell is embedded into drop weight, the mass distribution of drop weight causes the
measured impact force to deviate from the actual contact force acting on the beam. To obtain the true impact
force recorded by the load cell mounted at the rear of drop weight head, the drop weight mass ratio αd (i.e., mass
of the drop-weight to mass of the head) should be lower than 20. Placing load cell between drop weight and
beam changes the local contact stiffness of impact zone and thus leads to different impact force profiles. In
addition, the drop weight to beam mass ratio within the range of 0.5 and 4.0 affects the relative velocity between
drop weight and beam after the first impulse and hence results in different impact force profiles.

1. Introduction impact interlayer, other factors may also affect the test results. In the
drop weight impact test results reported in literature, the methods of
Drop weight impact test is a widely used approach to study impact measuring the impact force in different tests are not necessarily the
behavior of RC beams [1–5]. In these tests, RC beams are impacted by a same, which could affected the measurement accuracy. These mea-
drop weight falling from a certain height according to the desired im- surement setups can be generally classified into two types, i.e., indirect
pact velocity. The impact force, reaction force and midspan displace- method and direct method as shown in Fig. 1. One of the indirect
ment are commonly recorded to analyse the dynamic responses of methods is to calculate the impact force via multiplying the accelera-
specimens. The drop weight test setups used in the previous studies of tion by the mass of drop weight based on Newton's Second Law. The
RC beams by different researchers have different configurations, which acceleration of drop weight is obtained by attaching accelerometers to
affect the test observations. A previous study investigated the influences drop weight [7–9] or by differentiating the velocity of drop weight
of inclination of drop weight, geometry of drop weight head, and im- measured by laser Doppler velocimeter (LDV) system [10,11]. Another
pact interlayer on the test data [6]. It was found that these factors af- indirect method is to sum the reaction force at the supports and the
fected the peak impact force, impact duration, reaction force, and beam integration of the acceleration and mass of specimen along its length
failure modes because drop weight inclination angle and head geometry [8,12]. The measurement of impact force by the indirect methods de-
affect the contact of the drop weight and RC beams. Similarly, placing a pends on the accuracy of acceleration and velocity measurement. In
different impact interlayer such as steel plate and rubber pad between addition, most impact tests adopt the direct measurement method
drop weight and RC beam affects the contact stiffness hence also leads [1,3–5,13,14], in which the impact force is recorded directly by a dy-
to different peak impact force profiles onto RC beams. Therefore, it was namic load cell (strain gauge type [15–17] or piezoelectric type
concluded that careful analyses are needed when designing the drop [18,19]) mounted at the rear of drop weight head [3,14,20,21] or
weight test and analysing the test data to achieve the desired scenario placed between the drop weight and tested specimen [5,22,23] as
and obtain reliable test data [6]. shown in Fig 1. Obviously, the impact force acting on the tested spe-
In addition to the influences of drop weight head geometry and cimens is different from the impact force recorded by the load cell

⁎
Corresponding authors.
E-mail addresses: wensu.chen@curtin.edu.au (W. Chen), hong.hao@curtin.edu.au (H. Hao).

https://doi.org/10.1016/j.ijimpeng.2020.103688
Received 5 March 2020; Received in revised form 1 August 2020; Accepted 7 August 2020
Available online 09 August 2020
0734-743X/ © 2020 Elsevier Ltd. All rights reserved.
H. Li, et al. International Journal of Impact Engineering 145 (2020) 103688

Fig. 1. Measurement methods of impact force in previous impact tests [1,3–5,7–14,20–23].

installed at the rear of the drop weight head. That is, the impact force instead of the entire impact force profile. There has been no systematic
measured by the load cell would be lower than the actual contact force study to correlate the impacting conditions with the respective impact
on the tested specimen due to the inertia force of the drop weight head. force profiles that are important to quantify the impulse onto the spe-
On the other hand, inserting a load cell between the drop weight and cimen and to determine the dynamic response.
tested specimen changes the contact stiffness of impact zone, which, This study numerically investigates the influences of different im-
like placing an interlayer between the drop weight and beam, would pact test setups and the effect of mass ratio of drop weight to beam on
result in different impact force profiles as reported in Ref. [6]. There- the impact force profile and its measurement accuracy in drop weight
fore, different setups of load cell measurement affect the value of im- impact tests. The accuracy of impact force measurement is quantified
pact force, and those reported in literature were not necessarily the analytically by comparing the contact force between drop weight and
actual impact force acting on the tested RC beams. Since the impact beam and the impact force measured by the load cell obtained in nu-
force acting on RC beam determines the dynamic response of the tested merical simulations. The actual impact force acting on structures can be
specimen, an accurate measurement of impact force is essential to re- determined based on the analytical derivations, which leads to a more
flect the capacity of the RC beam. In addition, the accurate impact force accurate assessment of structural performance in drop weight impact
is also important for numerical model calibration. The difference be- tests. In addition, simulations are also carried out to examine the mass
tween impact force recorded by load cell and the actual contact force ratio of drop weight to the beam on the impact force profiles. With the
acting on specimens would mislead the development of reliable nu- results presented in this study, the impact force profiles can be ex-
merical models. To date, although the load cell has been widely used in plained and controlled.
drop weight impact test, the investigations about the effects of the mass
distribution of drop weight and the load cell location on impact force 2. Numerical model calibration
are very limited. More studies about the influences of impact test
configurations on accuracy of impact force measurements are deemed 2.1. Drop weight impact test
necessary.
Various profiles of impact force time history of RC beams under Without loss of generality, the drop weight impact test of RC beams
drop weight impact have been observed in the previous studies. Based conducted by Fujikake et al. [3] is employed to calibrate the numerical
on intensive literature review, the impact force profiles can be generally model in this study. This experimental study reports detailed testing
categorized into three types (Type I to III) based on their characteristics data and has been used for the calibration of numerical models
as shown in Fig. 2. Type I is the impact force profile with only one [32,34–37]. Fig. 3 shows the drop weight impact test setup. The RC
primary force peak [24]. Type II has a primary impact force peak fol- beam is simply supported at both ends over a clear span of 1.4 m. The
lowed by multiple secondary peaks [2,7,14,25,26]. The secondary force RC beam is impacted by a drop hammer with hemispherical head
peaks gradually decrease with time. Type III has a primary impact force falling from various heights, i.e., 0.15 m, 0.3 m, 0.6 m, and 1.2 m. The
peak followed by a plateau [3,4]. Type III impact force profile was impact force is recorded by the load cell installed at the rear of the drop
observed on the RC beams [3,4], RC columns [27] and concrete-filled weight head. The total mass of drop hammer is 400 kg. A laser dis-
steel tube columns [28,29] in drop weight impact tests. By comparing placement sensor is located below the RC beam to measure the midspan
and analysing the existing drop weight testing data of RC beam displacement. The dimension and rebar configuration of RC beam are
[2–4,14,24,30,31] as summarized in Table 1, it is found that the impact illustrated in Fig. 4. The total length of the beam is 1.7 m. The width
force profile of type I and type II were observed when the impact mass and depth of the beam section are 150 mm and 250 mm, respectively. A
was lower than the mass of impacted specimen, while the impact force total of four longitudinal rebars with a diameter of 16 mm are placed
profile type III occurred when the tested specimen was impacted by a symmetrically at the compressive and tensile sides. The 10 mm-dia-
drop weight heavier than the specimen. Therefore, the mass ratio be- meter stirrups are arranged along the beam length at a space of 75 mm.
tween the drop weight and the impacted specimen affects the impact The yield strength of longitudinal rebar and stirrups is 426 MPa and
force profile. However, most of the previous studies on RC beams focus 295 MPa, respectively. The compressive strength of concrete is
on the effect of drop weight mass on the peak impact force [32,33] 42.0 MPa.

2
H. Li, et al. International Journal of Impact Engineering 145 (2020) 103688

Fig. 2. Different impact force profiles of RC beam under drop weight impact.

2.2. Numerical model 0.2 in this study.

The keyword *MAT_PIECEWISE_LINEAR_PLASTICITY (MAT_24) is
Numerical models of RC beam under drop weight impact are de- employed for the steel longitudinal rebars and stirrups. The DIF for steel
veloped and calibrated in LS-DYNA based on the testing data [3] as rebars proposed by Malvar [44] is defined by the keyword *DEFINE_-
shown in Fig. 5. The constant stress solid element with a single in- CURVE and combined with MAT_24 model. The failure strain of steel is
tegration point is used for the concrete and drop weight. Longitudinal determined as 0.12. In addition, the elastic material model *MAT_E-
rebars and stirrups are simulated by Hughes-Liu beam element with LASTIC (MAT_1) is used for drop weight.
2 × 2 Gauss quadrature integration. The radius of hemispherical head
of drop weight is set as 90 mm according to the test setup. Different
2.2.2. Contact and boundary constraint
impact velocities are assigned to the drop weight by using the keyword
The surface to surface contact is defined between drop weight and
*INITIAL_VELOCITY_GENERATION according to the corresponding
concrete beam and the standard penalty formulation is employed
falling height. The acceleration of gravitation is set as 9.8 m/s2 by using
(SOFT = 0) [45]. The contact stiffness scale factors (SFS/SFM) for
the keyword *LOAD_BODY. The mesh size of 10 mm for the numerical
master and slave elements are determined as 0.2. Moreover, the beam is
model is adopted after conducting a mesh convergence study to obtain
constrained at both ends to achieve the simply supported boundary
reliable results with reasonable computational efficiency.
conditions by using the keyword * BOUNDARY_SPC_SET. The degrees
of freedom of nodes in the boundary sets are defined to prevent the
vertical movement of beam but allow free rotation of beam ends. In
2.2.1. Material model
addition, the rebars are embedded into concrete by using the keyword
The concrete material is simulated by
*CONSTRAINED_BEAM_IN_SOLID (CBIS).
*MAT_CONCRETE_DAMAGE_REL3 (MAT_72R3) in which material
parameters can be generated automatically by determining the un-
confined compressive strength of concrete and unit conversion factors. 2.2.3. Comparisons between numerical and test results
The concrete model has been proven able to accurately predict the The failure modes and dynamic responses of RC beams are com-
dynamic behavior of RC structures under extreme dynamic loads pared between numerical and test results as shown in Figs. 6 and 7,
[36,38,39]. Moreover, the dynamic increase factors (DIFs) for concrete respectively. Accurately predicting the concrete cracks is still a chal-
compressive and tensile strength [40] are employed. The keyword lenge for researchers [46,47]. In this numerical, the concrete cracks are
*MAT_ADD_EROSION is used along with the concrete model MAT_72R3 presented by the numerical damage contours [32,35]. It can be ob-
to delete over distorted concrete elements. The maximum principal served from Fig. 6 that the numerical concrete damage contours agree
strain criterion has been employed in RC structures under extreme loads well with the concrete cracks in the test results. Vertical cracks at
in the previous studies [41–43] and it is also adopted and determined as midspan extended from the bottom surface of beams to the impact zone

3
H. Li, et al. International Journal of Impact Engineering 145 (2020) 103688

Table 1
Summary of impact force profile type in previous drop weight impact tests of
RC beam.
Reference Specimen Mass ratio Impact force profile type

Chen and May, 2009 [2] A1 0.726 Type II

A2 0.726 Type III
A3 1.307 Type III
B1 0.726 Type II
B2 1.307 Type III
B3 0.726 Type II
B4 1.307 Type III
Fujikake et al., 2009 [3] S1616 3.05 Type III
S2222 3.05 Type III
S1322 3.05 Type III
Fig. 4. Dimension and rebar configuration of RC beam (unit: mm) [3].
Xu and Zeng, 2014 [30] BD1-1 1.17 Type III
BD2-1 1.17 Type III
BD3 1.17 Type III numerical model is consistent with that in the test result. The impact
BD4 2.67 Type III
force plateau can be also seen in the numerical impact force profile.
BD5-1 2.67 Type III
Yilmaz et al., 2014 [24] S1 0.16 Type I Moreover, the maximum displacement in the numerical models is
S2 0.16 Type I comparable to the experimental maximum displacement response.
S3 0.16 Type I These comparisons demonstrate the reliability of the numerical model,
S4 0.16 Type I
which can be used in the subsequent numerical studies about the RC
S5 0.16 Type I
Zhao et al. 2017 [4] B-1700-4.6 1.42 Type III
beams under drop weight impact.
B-1052-6.4 0.875 Type III
B-868-7.14 0.723 Type II 3. Impact force measurement
C-1700-4.6 2.36 Type III
C-1300-5.56 1.81 Type III
C-868-7.14 1.21 Type III Load cells have been widely used in the drop weight impact test to
D-1700-4.6 2.36 Type III measure the impact force. As reviewed above, the load cell can be
D1300-5.56 1.81 Type III mounted at the rear of drop weight head or placed between drop weight
D-868-7.14 1.21 Type III and beam. In this section, the influences of drop weight mass dis-
Yan et al., 2018 [14] B1a 0.45 Type II
B1b 0.45 Type II
tribution and load cell location on the impact force measurement are
Guo et al. 2019 [31] S5 2.67 Type III investigated. The drop weight configuration, geometry and reinforce-
S6 2.67 Type III ment layout of RC beam are kept the same as those presented in Figs. 3
S7 2.67 Type III and 4.
S8 2.67 Type III

3.1. Effect of drop weight mass distribution

When the load cell is installed at the rear of the drop weight head as
illustrated in Fig. 8, the mass distribution of drop weight may affect the
recorded impact force. In order to investigate the effect of drop weight
mass distribution on the measured impact force, the drop weight mass
ratio αd is defined as the ratio of the mass of weight above the load cell
(mw) to the mass of head (mh), as expressed in Eq. (1).
mw
αd =
mh (1)

Different drop weight mass ratios of 0.33, 0.5, 1.0, 2.0, 3.0, 5.0,
10.0, 20.0, 30.0, and 50.0 are employed in the study. In the simulation,
the mass ratio of drop weight is adjusted by modifying the density of
the weight above the load cell and the head below the load cell. It is
worth mentioning that the total mass of the drop weight is kept as
400 kg and the impact velocity is 4.85 m/s. The measured dynamic
force in the load cell is obtained from the axial force at the middle
height of load cell in the numerical model and is defined as the mea-
sured impact force as illustrated in Fig. 8. The actual force acting on the
beam from drop weight is identified as contact force and is compared
with the impact force measured by the load cell. Fig. 9 compares the
Fig. 3. Drop weight test setup (unit: mm) [3]. time histories of contact force and measured impact force. The overall
profiles of measured impact force and contact force are similar, that is,
the peak impact force is followed by a force plateau. However, it can be
are well predicted in the numerical models. With the increase of drop
seen that there are some discrepancies between the contact force and
height, some inclined cracks appear in both the test and numerical
the measured impact force of the specimens impacted by the drop
results. Moreover, the beam impacted by the drop weight falling from a
weight when the mass ratios are low. The impact force measured by
height of 1.2 m suffers a severe local concrete spalling at the side of
load cell is lower than the contact force directly acting on the beam.
impact zone, which is well captured in the numerical model as pre-
With the increase of drop weight mass ratio, the discrepancy becomes
sented in Fig. 6. In addition, the time histories of impact force and
smaller and the measured impact force closes to its corresponding
midspan displacement are illustrated in Fig. 7. The impact force in the
contact force. This can be explained by the force equilibrium according

4
H. Li, et al. International Journal of Impact Engineering 145 (2020) 103688

Fig. 5. Numerical model of RC beam.

to the D'Alembert's principle as shown in Fig. 8. During the course of enough with an accumulated error less than 10% as compared to the
impact, the impact force measured by load cell (Flc) is equal to the contact force applied onto the specimen. An equation showing the re-
inertia force of weight (Fiw) which depends on the mass of weight above lationship between the drop weight mass ratio (αd) and Df is proposed
the load cell. Flc is also equal to the subtraction of inertia force of head via the regression analysis in Eq. (3) (R2 = 0.9849).
(Fih) from contact force (Fc). Therefore, the contact force (Fc) is higher
Df = 0.4836α d−0.575 (3)
than the measured impact force (Flc) from load cell. As each specimen is
subjected to the identical impact mass and velocity of drop weight, each Therefore, in the drop weight impact test, the discrepancy between
specimen has very similar contact force as shown in Fig. 9. That is to the measured impact force by load cell and the actual contact force
say, the mass distribution does not affect the contact force acting on the acting on the specimen can be assessed by using Eq. (3) based on the
specimens but the measured impact force is affected by the mass dis- drop weight mass ratio αd.
tribution. In addition, the impact force measured by load cell could be cor-
The discrepancy (Df) between contact force and meansured impact rected to get better measurement of the actual impact force acting on
force in each group as presented in Fig. 9 is assessed by the root mean the beam according to the drop weight mass ratio as shown in Figs. 9
square error (RMSE) as illustrated in Eq. (2). and 11. The commonly used load cell in drop weight impact test is
2
composed of a steel cylinder attached with several strain gauges and the
n
1 Fc, i − Flc, i ⎞ steel cylinder works in an elastic state. Therefore, the drop weight with
Df =
n
∑⎛ ⎜
Fc, i
⎟

i=1 ⎝ ⎠ (2) the embedded load cell can be simplified as a mass-spring system in
Fig. 12 based on the force equilibrium as illustrated in Fig. 8. The
where Fc,i is the actual contact force acting on the specimen and Flc,i is weight and head of drop weight are illustrated by two mass blocks. The
the impact force measured by load cell at the corresponding ith instant load cell is presented by the spring with elastic stiffness k as shown in
as shown in Fig. 8. The Df of each group with different drop weight Fig. 12 and thus the spring force is the measured impact force in load
mass ratios is presented in Fig. 10. When the drop weight mass ratio is cell. According to the D'Alembert's principle, the dynamic equilibrium
not higher than 1.0, i.e., the mass of weight is not larger than that of equations are expressed as follows,
head, Df is 0.76, 0.67, and 0.54 for the drop weight mass ratio of 0.33,
0.5, and 1.0, respectively. The values of Df are higher than 0.5, in- m w x¨ w − k (x w − xh ) = 0 (4)
dicating a larger discrepancy between the contact force and meansured
mh x¨h + k (x w − xh ) = Fc (t ) (5)
impact force. With the increase of the drop weight mass ratio, Df de-
creases significantly. When the drop weight mass ratio is higher than where xw and xh are the vertical displacement of weight and head,
20.0, Df gradually reduces to be lower than 0.1, which indicates the while ẍ w and ẍh are the acceleration of weight and head. Fc(t) is the
meansured impact force curve agrees well with the contact force curve time history of contact force between drop weight and RC beam. As
as shown in Fig. 9. Therefore, when the drop weight mass ratio is higher repoted in [48], the hard impact is defined as the striking object is more
than 20.0, the impact force measured by load cell is deemed accurate rigid than the impacted object and the impact energy is dissipated by

Fig. 6. Comparison of failure modes between numerical and test results [3].

5
H. Li, et al. International Journal of Impact Engineering 145 (2020) 103688

Fig. 7. Comparisons of time histories of impact force and displacement between the present numerical results and the test results in Ref. [3].

Fig. 8. Illustration of drop weight mass ratio and force equilibrium analysis.

6
H. Li, et al. International Journal of Impact Engineering 145 (2020) 103688

Fig. 9. Comparison between contact force and meansured impact force.

the deformation and damage of impacted object. The drop weight head When the mass ratio of drop weight in Eq. (1) is introduced into
can be considered as rigid compared to the RC beam in this study, and Eq. (7), the relationship between the actual contact force and the
since the head, load cell and weight move together during impact, the measured impact force is found to be
acceleration of drop weight head ( ẍh ) is approximately the same as that
1 ⎞
of weight above the load cell ( ẍ w ), which can be expressed as follows, Fc (t ) = ⎛1 +
⎜ ⎟Flc (t )
⎝ α d⎠ (8)
Fc (t )
x¨ w = x¨h = Thus, the impact force correction factor βc can be expressed as
m w + mh (6)
follows,
The measured impact force Flc(t) in load cell illustrated by the
1
spring force is given by βc = 1 +
αd (9)
mw
Flc (t ) = k (x w − xh ) = m w x¨ w = Fc (t ) The measured impact force in load cell can be corrected by multi-
m w + mh (7)
plying the correction factor βc to obtain the actual contact force. The

7
H. Li, et al. International Journal of Impact Engineering 145 (2020) 103688

Fig. 13. Relationship between the mass of drop weight head and the period of
Fig. 10. The relationship between drop weight mass ratio and discrepancy of
drop weight system.
force Df.

For the simplified mass-spring system of drop weight as shown in

Fig. 12, varying the mass ratio of drop weight changes the mass dis-
tribution of the mass-spring system and thus causes the variation of the
natural frequency of the system. The natural frequency of the system is
expressed as,

k (m w + mh )
ω=
m w ·mh (10)

In this study, the mass of drop weight (m w + mh ) is kept as a con-

stant of 400 kg and the spring stiffness for load cell is also constant. By
normalizing the spring stiffness k and the mass of whole drop weight as
1, the relationship of ω and mh can be derived as,
Fig. 11. Df of measured impact force, corrected impact force and smoothed 1
ω=
impact force. mh ·(1 − mh ) (11)

and the period of the mass-spring system is

corrected impact force is also presented in Fig. 9. It can be seen that the
corrected impact force is close to the actual contact force by amplifying T = 2π mh ·(1 − mh ) (12)
the measured impact force with the correction factor βc. Fig. 11 com-
pares Df of the meansured impact force and the corrected impact force. The relationship between the period of the system and mh is illu-
Df of the corrected impact force is calculated according to the RMSE of strated in Fig. 13. It is found that the period of system is the largest
the actual contact force and the corrected impact force. As shown in when mh is 0.5, i.e. drop weight mass ratio αd = 1.0. The period of
Fig. 11, Df of the corrected impact force is lower than that of original system becomes shorter for other mass ratios. Since the contact force
measured impact force for all specimens, indicating a smaller dis- acting on the mass-spring system is identical in all the specimens with
crepancy between the corrected impact force and the actual contact different drop weight mass ratios, the system with the longest period
force. However, the Df of the corrected impact force in most cases is still (mass ratio αd = 1.0) would lead to the largest spring response, which
higher than the expected threshold of 0.1 and there are still some dis- causes the largest fluctuation between the measured and corrected
crepancies at the peak impact force, owing to the significant fluctua- impact forces. The difference between the corrected peak impact force
tions in the measured impact force in load cell. and the actual peak contact force is shown in Fig. 9.
In order to obtain more accurate actual contact force, the Savitzky-
Golay smoothing method is employed to reduce the fluctuation of im-
pact force. It can be seen from Fig. 9 that the smoothed impact force in
general agrees well with the actual contact force. The more accurate
peak force also can be well obtained. The Df of smoothed impact force is
lower than the expected threshold of 0.1 as presented in Fig. 11.
However, it is found that the smoothed peak force is lower than that of
actual contact force when the mass ratio of drop weight is higher than
20.0. Since Df of these specimens is lower than 0.1, the corrected impact
force is accurate enough to represent the actual contact force. However,
if the smoothing method is used for these specimens with the mass ratio
higher than 20.0, the smoothed impact force would lead to a higher Df
than that of the measured and corrected impact force as shown in
Fig. 11. A flowchart to obtain the actual contact force acting on beam in
the drop weight impact test is presented in Fig. 14.
To sum up, the mass distribution of drop weight has a significant
effect on the accuracy of measured impact force. When the mass of drop
weight head is larger than that of weight above load cell, the measured
impact force deviates significantly from the actual contact force acting
on the specimen. Thus, the drop weight mass ratio should be carefully
designed to obtain more precise impact force in drop weight impact
Fig. 12. Simplified mass-spring system for drop weight. tests, especially in the case of drop weight with relatively small mass.

8
H. Li, et al. International Journal of Impact Engineering 145 (2020) 103688

beam (setup II) are shown in Fig. 15. The impact mass and impact
velocity of drop weight are 400 kg and 4.85 m/s, respectively. Fig. 16
illustrates the shape and dimension of load cell used in the previous
drop weight impact tests [5]. The mass of load cell is about 18.1 kg and
the density of load cell is adjusted to match the mass of load cell.
Moreover, the contact between drop weight and load cell and the
contact between load cell and RC beam are defined as the surface to
surface contact.
In the numerical model as shown in Fig. 15(a), a drop weight with
load cell mounted at the rear of drop weight head impacts the RC beam
directly. As presented above, when the mass ratio is large, the mass
distribution of drop weight has insignificant influence on the mea-
surement of impact force, the drop weight mass ratio (αd) of 50.0 is
therefore adopted in this section, and the impact force measured by the
embedded load cell is compared with that measured in setup II as
shown in Fig. 15(b). It is worth mentioning that the measured impact
force in the load cells is obtained from the section axial force at the
middle height of load cells as illustrated in Fig. 15. Fig. 17(a) shows the
time histories of the measured impact forces in the numerical models.
During the first impact pulse, the peak impact force measured in the
embedded load cell (setup I) is lower than that in the load cell placed
between drop weight and beam (setup II). The impact duration of the
first impulse obtained by the setup I is 3.5 ms while that of dynamic
forces in the setup II is 2.4 ms. This is because the steel load cell in-
creases the contact stiffness and leads to a higher force peak and a
shorter duration. Moreover, it is worth noting that the measured impact
force reaches its peak at 1.3 ms and 0.9 ms for the setup I and setup II,
respectively, indicating that using steel load cell between drop weight
and beam leads to a higher loading rate. In addition, a force plateau
following the first impact pulse is measured by the load cell mounted at
the rear of drop weight head, while multiple secondary peaks are ob-
served by using the placed load cell. That is to say, given the same input
impact energy, the force measured by the load cell in the setup I shows
the impact force profile of type III, which is different from that of type II
in the setup II. Therefore, placing load cell between drop weight and
beam changes the local contact stiffness and thus affects the impact
force profile including the primary impact pulse and the subsequent
secondary force peaks or plateau.
For the case with load cell being placed between drop weight and
Fig. 14. Flowchart to obtain actual contact force.
beam (setup II), three dynamic forces, i.e., the contact force between
the drop weight head and load cell (F1), the measured impact force in
The measured impact force in the load cell mounted at the rear of drop load cell (F2), and the contact force between load cell and RC beam (F3),
weight head is suggested to be corrected if the drop weight mass ratio is are compared as shown in Fig. 17(b). It is found that F1 has the highest
lower than 20.0. The actual force acting on beam can be obtained ac- first peak of 522.45 kN, followed by F2 of 476.94 kN and F3 of
cording to the steps in flowchart as shown in Fig. 14. Moreover, in the 443.33 kN. During the course of impact, the drop weight first contacts
numerical simulation of drop weight impact, the numerical impact the load cell, which mobilizes an upwards inertia force of load cell.
force is usually determined by the contact force between the drop According to force equilibrium as shown in Fig. 18, the contact force
weight and specimen [49–51] and is compared with the measured ex- between drop weight and load cell (F1) is balanced by the upward in-
perimental impact force in load cell. If the experimental impact force is ertia force and the contact force between load cell and beam (F3). The
measured by load cell mounted at the rear of drop weight head and the inertia force of load cell results in a 15.14% difference between F1 and
drop weight mass ratio is lower than 20.0, the effect of drop weight F3. Moreover, the impact force (F2) measured by the load cell experi-
mass distribution should be also considered in the numerical model to encing an upward inertia force is usually higher than the actual contact
obtain the correct impact force and to calibrate the numerical model force (F3) applied onto the beam. For example, the difference in peak
accurately. impact force between the measured F2 and F3 acting on the beam is
7.05%. Since the load cell placed on the beam suffers an upward inertia
3.2. Effect of load cell location force which is mainly governed by its weight, it is suggested to employ
the load cell as light as possible in the setup II.
In addition to load cell embedded into the drop weight, load cell
may be placed on the beam directly, that is, located between drop 4. Effect of mass ratio
weight and impacted specimen [5,22,23]. The previous study reported
that placing an interlayer between drop weight and beam resulted in Based on the previous testing data [2–4,7,14,24–26], the impact
different impact force profiles [6]. Therefore, load cell placed between force profiles can be categorized into three types as illustrated in Fig. 2.
drop weight and RC beam can affect the impact force profile. In this The impact force depends on the interaction of drop weight with the
section, the effect of load cell location on the impact force acting on the beam. Upon impact, drop weight with a lighter mass would bounce off
RC beam is investigated. The numerical models of locating load cell at quickly from a heavier beam while a heavier drop weight would move
the rear of drop weight head (setup I) or between drop weight and together with a lighter beam. The masses of drop weight and beam

9
H. Li, et al. International Journal of Impact Engineering 145 (2020) 103688

Fig. 15. Different load cell locations.

affect the motion state of beam and drop weight and further influence 2.4 m (higher impact velocity). In addition, it is worth noting that the
the impact force profile. The mass ratio (α) is defined as, impact force drops to a small amplitude after the first peak as high-
md lighted in Fig. 19(c) and (d). The small minimum force amplitude
α= changes from zero to a certain value with the increase of mass ratio. The
mb (13)
higher mass ratio yields a larger minimum force amplitude. For ex-
where md is the mass of drop weight and mb is the mass of beam be- ample, when the mass ratio is less than 2.0, the force drops to zero as
tween two supports. In this study, the dimension and rebar layout are shown in Fig. 19(c). However, the minimum force is 12. 44 kN,
identical to those as shown in Fig. 4 and the mass of beam is kept the 39.70 kN, and 42.03 kN for the mass ratio of 2.0, 3.0, and 4.0, re-
same as 131.25 kg. The mass of drop weight varies from 32.81 kg to spectively. It can be concluded that the mass ratio has a significant
525.0 kg to have the corresponding mass ratio (α) in the range of 0.25 effect on the impact force profile. The impact force profile with a force
to 4.0. The drop weight falls from a height of 0.3 m, 0.6 m, 1.2 m or plateau occurs when the mass ratio is larger than 2.0. Under the same
2.4 m to reach an impact velocity of 2.42 m/s, 3.43 m/s, 4.85m/s, or mass ratio, the lower impact velocity results in the fluctuations of force
6.86 m/s, respectively. plateau while the higher impact velocity leads to an almost constant
impact force plateau.
4.1. Numerical results with different mass ratios
4.2. Discussion and analysis of different impact force profiles
Fig. 19 shows the impact force acting on beams impacted by drop
weight with different mass ratios and falling from different heights. In According to the results shown in Fig. 19, the impact force profile
general, increasing the mass ratio increases the first peak of impact can be summarized as three types, i.e., only one primary force peak
force, impact duration of first impulse and total impact duration. For (Type I as shown in Fig. 2(a)), primary force peak with one or multiple
the drop weight impact with the minimum mass ratio of 0.25 at dif- secondary peaks (Type II as shown in Fig. 2(b)), and primary force peak
ferent velocities, only one primary pulse is observed in the impact force followed by a force plateau (Type III as shown in Fig. 2(c)). In order to
profile (Type I), indicating that the drop weight and beam separate after better demonstrate the effect of mass ratio on the motion state of drop
the impact. The drop weight rebounds upwards and the beam moves weight and beam and on the impact force profile, the time histories of
downwards. In terms of the mass ratios of 0.5, 0.75, and 1.0, the impact impact force, velocities of drop weight and beam, and displacements of
force profile presents a primary pulse followed by one or more sec- drop weight and beam at the impact zone are presented in Fig. 20. It is
ondary pulses (Type II) with the increase of mass ratio. The higher mass worth mentioning that the velocity and displacement of beam are ex-
ratio yields more secondary pulses. This can be explained that the tracted from the bottom of beam at midspan. The velocity and dis-
heavier drop weight has larger inertia and thus is more difficult to placement of drop weight are extracted from rigid body velocity and
bounce off the beam, leading to an increased number of contact and displacement of drop weight because the drop weight is considered as
impact duration. Moreover, regarding the mass ratio of 2.0, 3.0, and rigid as compared to the RC beam, and the velocity and displacement of
4.0, a force plateau appears (Type III) after the primary impact. How- the entire drop weight is deemed as the same although some minor
ever, the impact force plateau presents fluctuations under the drop deformation occurs at the tip of drop weight head during the impact.
heights of 0.3 m and 0.6 m (relatively low impact velocity) or remains The numerical results with the mass ratios of 0.25, 0.5 and 3.0 at the
almost constant at about 100.0 kN under the drop weights of 1.2 m and drop height of 1.2 m representing the three impact force profiles are

Fig. 16. Dimension of load cell.

10
H. Li, et al. International Journal of Impact Engineering 145 (2020) 103688

Fig. 17. Effect of load cell location on the impact force.

employed for the comparison.

In terms of the mass ratio of 0.25, the impact force profile has only
one pulse (type I) as presented in Fig. 20(a). During the course of im-
pact, the displacement of drop weight is slightly larger than that of
beam at the impact zone, indicating the drop weight and beam are in
contact. The interaction between drop weight and beam decelerates the
drop weight but accelerates the beam downwards, resulting in a de-
crease of drop weight velocity but an increase of beam velocity. The
velocity of drop weight decreases and turns into positive (i.e., the drop
weight moves upwards) as shown in Fig. 20(a) because of the upward
reaction force from the beam applying on the drop weight. This leads to
the separation between drop weight and beam and consequently the
decrease of impact force to zero. At the end of the primary impulse, the
velocity of drop weight is upward and the drop weight is separated from
the beam. Therefore, it can be seen that the drop weight rebounds off
the beam after the first impact, leading to only one primary force peak
Fig. 18. Force equilibrium of load cell. in the impact force profile.
For the case with mass ratio of 0.5, the impact force, velocity, and

Fig. 19. Time history of impact force for different mass ratios.

11
H. Li, et al. International Journal of Impact Engineering 145 (2020) 103688

Fig. 20. Time histories of impact force, velocity, and displacement.

displacement are illustrated in Fig. 20(b). During the first impact pulse, accelerates to its maximum velocity of 4.69 m/s and then decelerates to
the changes of beam velocity and drop weight velocity as well as the the velocity of 2.99 m/s, which is close to the velocity of drop weight as
impact force are similar to those for the mass ratio of 0.25. However, shown in Fig. 20(c). Then the impact force increases to a plateau with
the velocity of drop weight at the end of primary impulse drops to about an almost constant value. The velocity of drop weight is close to that of
0.22 m/s which is much lower than its initial velocity of 4.85 m/s. After beam during the force plateau, that is to say, the velocity difference
the first impact pulse, the displacement of beam becomes larger than between drop weight and beam is very small. Therefore, the beam and
that of drop weight, which implies that the beam and the drop weight drop weight keep in contact and move together with a close and gra-
are separated. The drop weight moves downwards with the velocity of dually declining velocity. When the velocity of beam and drop weight
0.22 m/s until the second impact occurs. The beam decelerates due to becomes zero, the initial kinetic energy of drop weight is dissipated by
its flexural stiffness and reaches the maximum displacement. Then the the elastic and plastic deformation of beam. At this instant, the dis-
beam moves upwards to recover its elastic deformation. The opposite placement of beam reaches its maximum and the impact force plateau
movement direction of beam and drop weight results in the second begins to decrease. Then the drop weight and beam move upwards
impact as shown in Fig. 20. This impact process may continue several together until the elastic deformation of beam is recovered. The velo-
times until the drop weight rebounds off the beam and the multiple city of beam and drop weight reaches the upward maximum. Due to the
secondary force peaks appear in the impact force profile type II as flexural stiffness of beam, the velocity of beam decreases gradually and
shown in Fig. 2. is lower than that of drop weight. The velocity difference between beam
For the case with mass ratio of 3.0, the impact force profile presents and drop weight results in the final separation of beam and drop weight
a force plateau after the primary impact as shown in Fig. 20(c). After and thereby impact force drops to zero at the end of impact.
the first force peak, the impact force decreases due to the reduction in Primary impact pulse appears in the impact force profile regardless
the relative velocity of drop weight and beam. The impact force drops of its type. After the primary impact pulse, the drop weight and beam
to a certain value instead of zero, implying that the drop weight and with various mass ratios exhibit different velocities, which cause dif-
beam are still in contact and moving in the same direction. The beam ferent interactions between drop weight and beam and thus generate

12
H. Li, et al. International Journal of Impact Engineering 145 (2020) 103688

various impact force profiles. In addition, the beam at midspan moves Acknowledgements
downwards and experiences a certain vertical displacement during
impact. The vertical displacement mobilizes the flexural stiffness of The authors acknowledge the financial support from the Australian
beam which provides a certain resistance and reduces the beam velo- Research Council (ARC) via Australian Laureate Fellowship
city. The beam moves downwards to its maximum displacement. After (FL180100196). The first author also gratefully acknowledges the fi-
that, the beam returns and moves upwards to recover its elastic de- nancial support from Curtin International Postgraduate Research
formation. Therefore, the flexural stiffness of beam affects the beam Scholarship (CIPRS) and Curtin Strategic International Research
velocity after the first impact. In addition, the contact stiffness at the Scholarship (CSIRS).
impact zone has effect on the whole impact force profile. It can be
concluded that the primary impact peak is governed by the impact References
energy and local contact stiffness while the remaining part of impact
force profile is related to other factors such as mass ratio, contact [1] Kishi N, Mikami H, Matsuoka KG, Ando T. Impact behavior of shear-failure-type RC
stiffness and flexural stiffness of beam. beams without shear rebar. Int J Impact Eng 2002;27(9):955–68.
[2] Chen Y, May IM. Reinforced concrete members under drop-weight impacts. Proc
Inst Civ Eng-Struct Build 2009;162(1):45–56.
5. Conclusions [3] Fujikake K, Li B, Soeun S. Impact response of reinforced concrete beam and its
analytical evaluation. J Struct Eng 2009;135(8):938–50.
[4] Zhao D, Yi W, Kunnath SK. Shear mechanisms in reinforced concrete beams under
This study investigates various impact force measurement methods impact loading. J Struct Eng 2017;143(9):04017089.
and the effect of mass ratio on the impact force profile of RC beam [5] Pham TM, Hao Y, Hao H. Sensitivity of impact behaviour of RC beams to contact
stiffness. Int J Impact Eng 2018;112:155–64.
under drop weight impact. Different mass distributions of drop weight [6] Li H, Chen W, Hao H. Influence of drop weight geometry and interlayer on impact
on the measured impact force are studied to reveal their influence on behavior of RC beams. Int J Impact Eng 2019;131:222–37.
the accuracy of measuring impact force. The impact forces measured by [7] Kishi N, Bhatti AQ. An equivalent fracture energy concept for nonlinear dynamic
response analysis of prototype RC girders subjected to falling-weight impact
load cell mounted at the rear of drop weight head and directly placed
loading. Int J Impact Eng 2010;37(1):103–13.
on the beam are compared. In addition, the effect of mass ratio (α) of [8] Huynh L, Foster S, Valipour H, Randall R. High strength and reactive powder
drop weight mass to beam mass on the impact force profile is discussed. concrete columns subjected to impact: Experimental investigation. Constr Build
The major conclusions drawn in this study are summarized as follows. Mater 2015;78:153–71.
[9] Yoo D-Y, Banthia N, Kim S-W, Yoon Y-S. Response of ultra-high-performance fiber-
reinforced concrete beams with continuous steel reinforcement subjected to low-
(1) The drop weight mass distribution has a significant effect on the velocity impact loading. Compos Struct 2015;126:233–45.
measured impact force by the load cell mounted at the rear of the [10] Al-Rifaie A, Guan ZW, Jones SW, Wang Q. Lateral impact response of end-plate
beam-column connections. Eng Struct 2017;151:221–34.
drop weight head. The measured impact force by load cell agrees [11] Al-Rifaie A, Jones SW, Wang QY, Guan ZW. Experimental and numerical study on
well with the contact force when the mass ratio of weight to drop lateral impact response of concrete filled steel tube columns with end plate con-
weight head (αd) is higher than 20.0. In contrast, if the mass ratio of nections. Int J Impact Eng 2018;121:20–34.
[12] Saatci S, Vecchio FJ. Effects of shear mechanisms on impact behavior of reinforced
weight to drop weight head (αd) is less than 20.0, the measured concrete beams. ACI Struct J 2009;106(1):78–86.
impact force deviates from the actual contact force acting on the [13] Tachibana S, Masuya H, Nakamura S. Performance based design of reinforced
beam by more than 10%. The measured impact force by load cell concrete beams under impact. Nat Hazards Earth Syst Sci 2010;10(6):1069–78.
[14] Yan Q, Sun B, Liu X, Wu J. The effect of assembling location on the performance of
can be corrected by using the proposed correction method to derive precast concrete beam under impact load. Adv Struct Eng 2018;21(8):1211–22.
more accurate contact force acting on the beam. [15] Isaac P, Darby A, Ibell T, Evernden M. Experimental investigation into the force
(2) Load cell mounting at the rear of the drop weight head or placing propagation velocity due to hard impacts on reinforced concrete members. Int J
Impact Eng 2017;100:131–8.
between the drop weight and beam results in different impact force
[16] Yoo D-Y, Banthia N. Size-dependent impact resistance of ultra-high-performance
profiles. Placing load cell between the drop weight and beam in- fiber-reinforced concrete beams. Constr Build Mater 2017;142:363–75.
creases the local stiffness and leads to higher impact force peak, [17] Dey V, Bonakdar A, Mobasher B. Low-velocity flexural impact response of fiber-
shorter duration, and multiple secondary force peaks as compared reinforced aerated concrete. Cem Concr Compos 2014;49:100–10.
[18] Wang W, Wu C, Li J, Liu Z, Lv Y. Behavior of ultra-high performance fiber-re-
with the impact forces from load cell mounted at the rear of drop inforced concrete (UHPFRC) filled steel tubular members under lateral impact
weight. loading. Int J Impact Eng 2019;132:103314.
(3) The mass ratio of drop weight mass to beam mass (α) affects the [19] Ulzurrun GSD, Zanuy C. Enhancement of impact performance of reinforced concrete
beams without stirrups by adding steel fibers. Constr Build Mater 2017;145:166–82.
relative velocity between drop weight and beam after the first im- [20] Zhou X, Zhang R, Xiong R, Zhang G, Wang X. An experimental study of the impact
pact pulse and thus influences the impact force profiles. In this mechanical properties of RC beams following replacements of stainless steel re-
study, with the mass ratio of 0.25, only one primary impact force inforcements of equal strength. Adv Mater Sci Eng 2019:2019.
[21] Lee J-Y, Shin H-O, Yoo D-Y, Yoon Y-S. Structural response of steel-fiber-reinforced
(type I) can be observed in the impact force profile. For the cases concrete beams under various loading rates. Eng Struct 2018;156:271–83.
with mass ratios of 0.5, 0.75, and 1.0, the impact force profile ex- [22] Pham TM, Hao H. Impact behavior of FRP-strengthened RC beams without stirrups.
hibits a primary impact pulse followed by one or multiple sec- J Compos Constr 2016;20(4):04016011.
[23] Wu M, Zhang C, Chen Z. Drop-weight tests of concrete beams prestressed with
ondary pulses (type II). With the mass ratio of 2.0, 3.0, and 4.0, the unbonded tendons and meso-scale simulation. Int J Impact Eng 2016;93:166–83.
impact force profile consists of a primary pulse followed by a force [24] Yilmaz M, Anil Ö, Alyavuz B, Kantar E. Load displacement behavior of concrete
plateau (type III). beam under monotonic static and low velocity impact load. Int J Civ Eng
2014;12(4):488–503.
[25] Jin L, Zhang R, Dou G, Xu J, Du X. Experimental and numerical study of reinforced
CRediT authorship contribution statement concrete beams with steel fibers subjected to impact loading. Int J Damage Mech
2018;27(7):1058–83.
[26] Bhatti AQ, Kishi N, Konno H, Mikami H. Elasto-plastic dynamic response analysis of
Huawei Li: Conceptualization, Methodology, Software, Formal prototype RC girder under falling-weight impact loading considering mesh size
analysis, Investigation, Data curation, Writing - original draft. Wensu effect. Struct Infrastruct Eng 2012;8(9):817–27.
Chen: Conceptualization, Validation, Formal analysis, Data curation, [27] Liu B, Fan W, Guo W, Chen B, Liu R. Experimental investigation and improved FE
modeling of axially-loaded circular RC columns under lateral impact loading. Eng
Writing - review & editing. Hong Hao: Validation, Data curation, Struct 2017;152:619–42.
Writing - review & editing, Supervision, Funding acquisition. [28] Wang R, Han L-H, Hou C-C. Behavior of concrete filled steel tubular (CFST)
members under lateral impact: Experiment and FEA model. J Constr Steel Res
2013;80:188–201.
Declaration of Competing Interest [29] Deng Y, Tuan CY, Xiao Y. Flexural behavior of concrete-filled circular steel tubes
under high-strain rate impact loading. J Struct Eng 2012;138(3):449–56.
[30] Xu B, Zeng X. Experimental study on the behaviors of reinforced concrete beams
There is no conflict of interest in this study.

13
H. Li, et al. International Journal of Impact Engineering 145 (2020) 103688

under impact loadings. China Civ Eng J 2014;47(2):41–51. subjected to blast loading. Mater Des 2015;65:662–74.
[31] Guo J, et al. Dynamic behaviour and energy dissipation of reinforced recycled ag- [42] Li J, Hao H, Wu C. Numerical study of precast segmental column under blast loads.
gregate concrete beams under impact. Constr Build Mater 2019;214:143–57. Eng Struct 2017;134:125–37.
[32] Guo J, Cai J, Chen W. Inertial effect on rc beam subjected to impact loads. Int J [43] Do TV, Pham TM, Hao H. Dynamic responses and failure modes of bridge columns
Struct Stab Dyn 2017;17(04):1750053. under vehicle collision. Eng Struct 2018;156:243–59.
[33] Li H, Chen W, Hao H. Dynamic response of precast concrete beam with wet con- [44] Malvar LJ. Review of static and dynamic properties of steel reinforcing bars. Mater
nection subjected to impact loads. Eng Struct 2019;191:247–63. J 1998;95(5):609–16.
[34] Jin L, Xu J, Zhang R, Du X. Numerical study on the impact performances of re- [45] Hallquist JO. LS-DYNA theory manual. Livermore software Technology corporation;
inforced concrete beams: A mesoscopic simulation method. Eng Fail Anal 2006. p. 531.
2017;80:141–63. [46] Rabczuk T, Zi G, Bordas S, Nguyen-Xuan H. A simple and robust three-dimensional
[35] Jiang H, Wang X, He S. Numerical simulation of impact tests on reinforced concrete cracking-particle method without enrichment. Comput Meth Appl Mech Eng
beams. Mater Des 2012;39:111–20. 2010;199(37-40):2437–55.
[36] Wongmatar P, Hansapinyo C, Vimonsatit V, Chen W. Recommendations for de- [47] Rabczuk T, Bordas S, Zi G. On three-dimensional modelling of crack growth using
signing reinforced concrete beams against low velocity impact loads. Int J Struct partition of unity methods. Comput Struct 2010;88(23–24):1391–411.
Stab Dyn 2018:1850104. [48] Bischoff PH, Perry SH, Eibl J. Contact force calculations with a simple spring-mass
[37] Zhang C, Gholipour G, Mousavi AA. Nonlinear dynamic behavior of simply-sup- model for hard impact: a case study using polystyrene aggregate concrete. Int J
ported RC beams subjected to combined impact-blast loading. Eng Struct Impact Eng 1990;9(3):317–25.
2019;181:124–42. [49] Zhao D, Yi W, Kunnath SK. Numerical simulation and shear resistance of reinforced
[38] Li J, Hao H. Numerical study of concrete spall damage to blast loads. Int J Impact concrete beams under impact. Eng Struct 2018;166:387–401.
Eng 2014;68:41–55. [50] Adhikary SD, Li B, Fujikake K. Low Velocity Impact Response of Reinforced
[39] Pham TM, Hao H. Influence of global stiffness and equivalent model on prediction Concrete Beams: Experimental and Numerical Investigation. Int J Protect Struct
of impact response of RC beams. Int J Impact Eng 2018;113:88–97. 2015;6(1):81–111.
[40] Hao Y, Hao H. Influence of the concrete DIF model on the numerical predictions of [51] Pham TM, Hao H. Effect of the plastic hinge and boundary conditions on the impact
RC wall responses to blast loadings. Eng Struct 2014;73:24–38. behavior of reinforced concrete beams. Int J Impact Eng 2017;102:74–85.
[41] Chen W, Hao H, Chen S. Numerical analysis of prestressed reinforced concrete beam

14