Professional Documents
Culture Documents
ABSTRACT: This experimental study investigates the response of steel-concrete composite sandwich panel of various
configurations under quasi-static and dynamic loading. The differences in these configurations include the thickness of the steel
plates, the concrete properties of the sandwich core and the connectors to ensure the composite action of the panels. The latter
utilises a novel J-hook connector system [1] and the Eurocode [2] is used as the basis of design to predict the partial composite
capacity of the specimens. The quasi-static experiment series utilised a three-point laboratory load test and the dynamic load
experimental study involved the use of an actual explosive test setup. The differences in response of six configurations of
sandwich panels in the quasi-static and dynamic load series were investigated. Both experimental series showed the contributive
effect of increased steel plate thickness and the presence of the concrete core. The effect of the concrete core is observed to be
distinctively larger under blast loading. In addition, the use of ultra high strength concrete (fc ≈ 200MPa) for the core resulted in
contradictory contribution to the composite sandwich panels under blast loading regimes as compared to quasi-static loading.
Observations from the dynamic experimental series show hints that use of connectors may not benefit the blast design of steel-
concrete-steel sandwich panels.
Specimen
Top plate Steel I-beam to
Removable steel
clamp plates support
instrumentation
Ground zero
of blast load
(b)
Figure 2. Test setup which includes one specimen (for
Figure 1. Notation for (a) SCS sandwich composite panel and illustration) and design of RC supporting structure.
(b) Cellular steel panel (SP).
As shown in Figure 2, the removable steel clamp plates,
2.2 Material which were attached to the RC support structure by four bolts,
The LWAC and HSC were mixed in the laboratory while the enabled the tests will be repeated using the same support
NSC was procured from a commercial supplier. NSC had a structure and steel rods welded behind the plates created a
specific density of 2300kg/m3 and the cylinder strength on the simply support end support for the specimens. Under such
date of test was 35MPa. The LWAC had a specific density of boundary condition, the effective span of the specimen is 1m.
1250 kg/m3 and compressive cylinder strength of 20MPa. It
was mixed from expanded clay aggregates, sand, ordinary Figure 3 illustrates the instrumentation that was installed for
portland cement, silica fume and water. The Ducorit D4 each blast test, which included pressure sensors installed on
material was used in HSC and the compressive strength of the the front surface of the shear wall of the RC support structure,
material was found to be 190MPa. The steel plates were potentiometers mounted onto the steel I-beam behind the
acquired through a commercial supplier and the result from specimen (see Figure 2) to record the mid span deflection
coupon test based on the ASTM standard [8] showed that the histories, accelerometer welded at the midspan on the back
yield strength to be 325MPa and it attained an ultimate face of the specimens and strain gauges attached on the mid-
strength of 450MPa at 0.17 strain. span and quarter-spans of the specimens.
Proceedings of the 8th International Conference on Structural Dynamics, EURODYN 2011 3375
Simple
Support on
Roller
Strain gauge
Load Line Accelerometer
Accelerometer
Potentiometer
Strain gauge
Pressure gauge
(a) (b)
(b)
Figure 4. (a) Schematic and (b) laboratory setup for the quasi-
static three-point test.
Similar to the dynamic load test, the instrumentation was
essential to capture the data in this experiment. Figure 5
shows the instrumentation on each specimen which included
displacement and strain gauges at the quarter and mid spans
on the face plates of the specimen. In addition, force and
displacement readings were also acquired from the load cell of
the actuator and the supports respectively.
(a) (b)
Figure 6. (a) SP (left) and SCSN4 (right) after blast and (b)
elevation profile of deformed specimen SP.
Excessive local buckling was also observed on the top steel
face plate and side plates. Figure 7 shows that the top face
plate that buckled between the steel stiffeners which forced
Proceedings of the 8th International Conference on Structural Dynamics, EURODYN 2011 3376
the side plates near the midspan to bend towards the centre of 100
Actual Recording
the specimen. Filtered Recording
80
Displacement (mm)
60
40
20
-20
0 20 40 60 80 100 120
Time (msec)
Figure 7. Local buckling on the specimen SP. Figure 9. Potentiometer reading history of SCSN.
Rupture of steel was also observed at the side edge of the top Specimen SCSNE had less damage as compared to SCSN as
steel face place around the midspan. This could be attributed observed in Figure 6. The elevation profile view of the
to the result of excessive compressive response of the top deformed SCSNE specimen shows a predominantly flexural
plate due to the large deformation of the specimen. response (see Figure 10). The midspan permanent deflection
was 31 mm and this is similar to the data captured by the
Specimen SCSN4, which was positioned adjacent to SP potentiometer, which is plotted in Figure 10. However, the
sustained relatively less damage and the permanent deflection displacement time history by the potentiometer did not reflect
at the midspan of the specimen measured 27 mm. The the reading correctly as there was a sharp drop in
response was observed to be flexural. displacement at 153ms. A possible reason for this could be
traced to the functionality of the potentiometer.
Specimens SCSN and SCSNE is shown in Figure 8. The
60
specimen deformed in a flexural manner similar to that of Actual Recording
-20
0 20 40 60 80 100 120 140 160 170
Time (msec)
(a) (b)
which indicated the debonding between the J-hook connectors As mentioned in Section 2.3, the blast pressure was recorded
and the face plate. by two pressure sensors that were mounted onto the RC
support structure. The first blast test did not yield any readings
from the instrumentation due to the trigger failure.
14000
12000
10000
Overpressure (kPa)
8000
Blast 2
Blast 3
6000 CONWEP
4000
2000
-2000
0 0.4 0.8 1.2 1.6 2 2.4 2.8 3.2 3.6 4
Time (msec)
(a) (b)
Figure 11. (a) SCSL (left) and SCSH (right) after blast and (b) Figure 14. Pressure-time history comparison.
elevation profile of deformed specimen SCSL. 4000
125 3600
Actual Recording
3200
100 2800
Impulse (kPa-msec)
2400
Displacement (mm)
Blast 2
75 Blast 3
2000
CONWEP
1600
50
1200
800
25
400
0
0
0 0.6 1.2 1.8 2.4 3 3.6 4.2 4.8 5.4 6
0 20 40 60 80 100 120 Time (msec)
Time (msec)
Figure 15. Impulse-time history comparison.
Figure 12. Potentiometer reading history of SCSL.
From the data acquired from second and third blasts, it was
Specimen SCSH exhibited flexural response under the blast found that the pressure sensor on the right side of the RC
loading (see Figure 11). The midspan permanent deflection of support structure had malfunctioned. The blast pressure
the specimen was 79mm but the potentiometer for this recordings and subsequent calculated impulse based on the
specimen was found to be damaged after test. The top steel pressure data are plotted in Figures 14 and 15 respectively.
face plate buckled under the excessive compressive forces that Included in the figures are comparisons with CONWEP
resulted from the large displacement of the specimen (see predictions [9].
Figure 13).
It is observed that the recorded peak overpressure and the
decay of both blasts are almost identical and the same could
be said about impulses. In comparison with the blast pressure
curves predicted using CONWEP, the impulse variations from
the actual test seem to fit the CONWEP plots well,
particularly for Blast 2. This signifies the reliability of
CONWEP in predicting the impulse for close-in explosions
and that the CONWEP impulse curves may not be
conservative in this case. On the contrary, there seem to be a
considerable underestimation of the actual peak overpressure
by CONWEP. While the CONWEP plot peaked at 5600kPa,
Figure 13. Local buckling of the steel face plate of SCSH. the actual readings from the pressure sensors documented
Proceedings of the 8th International Conference on Structural Dynamics, EURODYN 2011 3378
12700kPa and 13200kPa peak overpressures for Blast 2 and tabulated in Table 2.The mode of failure of these three
Blast 3, respectively. It may be argued that these peaks may specimens are attributed to the buckling of the top steel face
be the result of noise but still a higher peak overpressure will plate and Figure 18 shows an example of this response.
be obtained after filtering the data. Furthermore, impulse is a
Table 2. Yield and ultimate points of specimens.
more important for the dynamic response as the panels
responded predominantly to the impulsive regime. The Yield Yield Ultimate Ultimate
response time (see Figure 9, 10 and 11) of the panels is more Specimen Load Deflection Load Deflection
than 10 times the load duration (see Figure 14) (kN) (mm) (kN) (mm)
3.2 Quasi-Static Load Test Result SCSN4 204 10.9 266 68.0
SCSN 142 9.45 194 38.9
Due to the manner in which the specimen panels were loaded, SCSL 147 10.21 163 25.2
all the specimens failed in a flexural manner. Figure 16 and 17 SP 202 6.95 - -
plots the full and partial load-deflection history respectively, SCSH 149 7.30 - -
in which the deflection refers to the midspan displacement and SCSNE 159 11.1 190 59.5
the partial load history is limited to the elastic material phase.
300
SCSN4
250
SCSH
200
Load (kN)
SCSNE
SCSN
150
SCSL
100
50
SP
0
0 50 100 150
Midspan Deflection (mm) Figure 18. Buckling of top face plate that initiated softening.
Figure 16. Plot of load against midspan deflection (full range). Specimen SCSH exhibited similar response as the three
300
specimens which are described in the preceding paragraph.
The specimen yielded at 149kN with a midspan deflection of
250 7.3mm. As the panel hardened, the top plate buckled, which is
SCSN4 not dissimilar to the response shown in Figure 18. However,
SP
200 instead of softening, the specimen continued to harden and
SCSH due to the time constraint of the experiment, the load was
Load (kN)
4 DISCUSSION SCSL was due to its lower shear capacity. Since such
shear response would result in the premature buckling of
4.1 Deductions from Dynamic Load Test
the steel plates which, in turn, would reduce the capacity
Table 3 compiles the permanent midspan deflections of the six of the wall panel to resist blast loading, the use of LWAC
specimens tested. By comparing the permanent deformations, is thus, not recommended to be use in such sandwich
extent of damage and the modes of failure of the specimens, structures against blast threats.
the following conclusions on the blast resistance of the • The use of HSC may have adverse effects on the SCS
specimens could be drawn. sandwich panels based on the comparison of specimens
Table 3. Comparison of permanent midspan deflections SCSN and SCSH, with the latter having a larger midspan
deflection. Hence, such material may not be suitable in
Specimen Permanent Deflection (mm) the design of blast resistant structures. This could be
SP 160 attributed to the constitutive behavior of the HSC which
SCSN4 27 exhibited a softening gradient which was considerably
SCSN 57 steeper than the other two concretes. Consequently, the
SCSNE 31 HSC may have been shattered as the shock wave
SCSL 86 propagates through the core. This process prior to the
SCSH 81 global bending response of the specimen might have
• By comparing SP and SCSN4, the concrete core provided rendered the concrete from developing its full ultimate
the SCS sandwich specimen with added mass and rigidity strength. Thus, further tests are needed to understand the
which help to reduce the effects of local buckling. dynamic response of the HSC.
Despite possessing bending moment capacity, SCSN4 4.2 Deductions from Quasi-static Load Test
exhibited much less damage as compared to SP. This
showed that the concrete core is more effective in Comparing the load variation and the failure modes of the
enhancing the residual capacity of such panels. specimens under quasi-static loading, the following
• The performance of J-hook connectors in maintaining the deductions were drawn:
integrity of SCS sandwich specimens, which responded • It was clear by comparing specimen SP and the rest of the
flexurally, was found to be satisfactory as little or no specimens that the concrete core contributed to the post
separation was observed between the steel face plates and yield hardening behavior of the SCS sandwich panel.
the core. SCSL underwent shear response and was an This translates to specimens with a higher energy
exception whereby several J-hook connectors debonded absorbent capacity as compared to panels which being
from the steel face plates at the areas of shear failure. reinforced by steel stiffener plates, even if they are stiffer,
However, it must be noted that the side plates of all the as observed in the current test series.
sandwich specimens appeared to play a significant role in • Another deduction is that the use of a thicker face plates
holding the two steel face plates together throughout the will produce an elastically stiffer sandwich panel that
blast response. This was evident from the deformed shape possesses higher yield and ultimate capacities.
of SCSNE which had no connectors and yet little or no • The use of various concrete cores did affect the quasi-
separation was observed between the steel face plates of static response of sandwich panels. Although SCSN and
this specimen. Since the specimens were scaled, the SCSL did not differ in terms of elastic modulus and yield
influence of the connectors may not be as significant as in capacity, the post yield responses were different with
the case of larger wall panels. SCSN possessing a higher hardening modulus and
• By comparing specimens SCSN4 and SCSN, it was found ultimate capacity. This is attributed to the higher strength
that the thickness of the steel face plates significantly of the concrete core. The same can be said by comparing
influenced the permanent deformation of the specimens. SCSN and SCSH as the latter displayed a higher ultimate
By increasing the thickness of the steel face plates by 1 capacity than the former. In addition to increasing the
mm, the permanent deflection of the SCS sandwich elastic modulus of the specimen panel, upon the buckling
specimen was reduced by almost half. of the top face plate, the HSC used allowed the tensile
• The damage incurred by specimen SCSN was more forces undertaken by the bottom steel plate to be balanced
severe than in SCSNE which has the exact specification with the high compressive forces undertaken by the high
sans J-hook connectors. More tests have to be carried out compressive strength of the concrete core. This is clearly
to investigate and confirm the reasons behind this as it is exhibited by the extensive plastic phase, as compared to
contradictory to pretest prediction and design philosophy. the rest of the specimens. These deductions show that,
One possible explanation is that there were more points under quasi-static loading, the influence of the concrete
of stress concentration within the concrete core with J- core is not significant in the elastic phase but, beyond
hook connectors which may increase crack formation. yield, the properties of the concrete play an important
• Specimen SCSL had larger deflection and deformed pre- role.
dominantly in shear. This could be attributed to two • Similar to the concrete core, the influence of steel J-hook
factors: the lower weight and shear strength of the connectors was only exemplified after yield. Comparing
concrete core. However, it was found that SP also the post-yield response of specimen SCSN and SCSNE,
underwent flexural response despite being lighter. Hence, the factor allowing SCSN to have a higher ultimate load
it may be deduced that the shear response of specimen but lower ultimate deflection can be attributed to the J-
Proceedings of the 8th International Conference on Structural Dynamics, EURODYN 2011 3380
hook connectors. By observing the buckling area, it can • The use of concrete core and an increment of the
be deduced that the connectors improves the composite thickness of the steel face plates improved the both static
action between the steel and concrete. This allowed and dynamic performance of the sandwich panel
specimen SCSN to have a higher ultimate load. The specimens.
reason behind the higher ultimate deflection of specimen • The use of LWC in SCS composite sandwich panel may
SCSNE could be due to the larger area of buckling. subject specimens to shear failure due to the low shear
Therefore, the specimen was able to deflect slightly more capacity of the core material.
before softening. • The use of J-hook connectors reduced the blast resistance
4.3 Deductions by Comparison Between Dynamic and of the SCS composite sandwich panel but this was not
Quasi-static Load Tests evident in the quasi-static test as panels with and without
these connectors exhibited similar response.
By comparing the deductions drawn from the preceding two • HSC may not be appropriate for use in blast design with
sections, the factors influencing the dynamic and static design the loading similar to the experiment as they did not
approaches are different. Certain parameters, which may perform as well as NSC. This was contrary to the
improve the performance under quasi-static loading, may not excellent capacity the material demonstrated in the quasi-
necessarily yield better blast resistance under dynamic loads. static test.
The following are the results by comparing the two
experimental series:
• The concrete core improved both static and dynamic ACKNOWLEDGMENTS
performance of the sandwich panel specimens. The quasi-
The research described herein was funded by the Defence
static test showed the lack of post yield response due to
Science and Technology Agency and National University of
extensive local buckling of the face plates and this was
Singapore. The authors sincerely express their gratitude for
also clearly demonstrated in the dynamic test as the
the supports.
specimen lacked any resistance beyond yield to counter
the blast load. REFERENCES
• The thickness of the steel face plates exhibited similar [1] J.Y. Richard Liew and K.M.A. Sohel. Lightweight steel-concrete-steel
influence in both static and dynamic responses as, with sandwich system with J-hook connectors. Engineering Structures,
increased thickness of these plates, the energy absorbent Volume 31, Issue 5, May 2009, Pages 1166-1178.
capacity of the panel is increased. [2] Eurocode 4. Design of composite steel and concrete structures—Part
1.1: General rules and rules for buildings. BS EN 1994-1-1:2004 .
• LWC has also shown similar deficiencies under both [3] T.S. Lok, J.R. Xiao, S. Lan and L Heng. Response of SFR concrete
loading regimes as compared to the other two concrete panels to air-blast loading. Proceedings of the Asia-Pacific Conference
materials. on Shock & Impact Loads on Structures, Singapore, 1996.
• The performance of the J-hook connectors did differ in [4] S. Lan and L. Heng. Composite structural components under blast
loading. Structures under Shock and Impact, 2002, Pages 13-21.
both test series. Under quasi-static loading, comparing [5] S. Lan, T.S. Lok and L. Heng. Composite structural panels subjected to
specimens with and without these connectors, the explosive loading. Construction and Building Materials, Volume 19,
differences were marginal. Prior to the test, it was thought Issue 5, 2005, Pages 387-395.
[6] H. Bowerman, N. Coyle and J.C. Chapman. An innovative
that the J-hook will enhance the blast performance of the
steel/concrete construction system. The Structural Engineer, Volume 80,
panels as they enhance the composite action between the Issue 20, 2002, Pages 33–38.
steel plates and concrete core. The dynamic test showed [7] http://www.corussecurity.com/en/design_and_innovation/blast_protecti
that this approach should be further studied as the on/
[8] ASTM standard. ASTM E8/E8M-08 Standard Test Methods for Tension
specimen without J-hook connectors had a lower
Testing of Metallic Materials.
deflection. [9] U.S. Department of the Army. Fundamentals of Protective Design for
• HSC performed exceptionally well under quasi-static Conventional Weapons. Technical Manual 5-855-1, 1986.
loading as the test seemed to exhibit a high energy
absorbent capacity by possessing an extensive plastic
hardening phase. This characteristic was not
demonstrated in explosive test as SCSH deformed the
most amongst the SCS sandwich panel specimens.
5 CONCLUSION
Two test series were conducted to characterize the static and
dynamic behavior of six different configurations of SCS
composite sandwich panels. The dynamic test subjected the
panels to blast loading at a scaled distance of 1.077m/kg1/3 and
the panels of the same configurations were loaded in a three-
point quasi-static load test till failure. The influence of the
steel face plates, J-hook connectors and concrete core
materials were investigated. The combined investigation of
the results of both series of experiments showed that: