Composites: Part B 68 (2015) 153–161

Contents lists available at ScienceDirect

Composites: Part B
journal homepage: www.elsevier.com/locate/compositesb

Composite and non-composite behaviors of foam-insulated concrete

sandwich panels
Junsuk Kang ⇑
Department of Civil Engineering and Construction Management, Georgia Southern University, Statesboro, GA 30460-8047, United States

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

Article history: The structural behaviors of foam-insulated concrete sandwich panels subjected to uniform pressure have
Received 11 September 2012 been evaluated. This study showed that the interface conditions such as composite and non-composite
Received in revised form 7 August 2014 had a significant effect on the response of foam-insulated concrete sandwich panels, indicating that
Accepted 20 August 2014
the simulated shear tie resistance should indeed be incorporated in numerical analyses. Finite element
Available online 27 August 2014
models were developed to simulate the detailed shear resistance of connectors and the nonlinear behav-
iors of concrete, foam and rebar components. The models were then validated using data from static tests
Keywords:
performed at the University of Missouri. The modeling approach used here was compatible with the
Sandwich panel
A. Foams
American Concrete Institute (ACI) Code and existing design practices. The results of this study will there-
B. Interface fore provide improved methodology for the analysis and design of foam-insulated sandwich panels under
B. Strength both static and blast loadings.
C. Finite element analysis (FEA) Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction between the interior and exterior concrete sections, referred to as

Generally, static tests are performed in advance in order to
define the resistance and failure mode of foam-insulated concrete
sandwich panels (FICSP) before the dynamic tests are conducted.
Several research programs have sought to evaluate the structural
performance of FICSP subjected to static loading [1–8]. It was
shown from several researchers [9,10] that FICSPs are also an
efficient way to mitigate the impact induced by blast or dynamic
loading. Analytical and experimental studies of glass fiber-rein-
forced polymer/steel concrete sandwich panels under out-of-
plane load and the developed analytical method predicted the
panel deflections observed in the experiments with reasonable
accuracy. However, even though the fundamental static test
results provide researchers with the data needed to evaluate
the static resistance functions for sandwich walls, most of the
results from static tests must be interpreted with caution due
to the limitations imposed by the force–displacement history of
the samples. It should be noted that FICSPs are constructed from
components made of various materials, including concrete, foam,
rebar, welded wire reinforcement (WWR) and shear ties, each of
which affects the failure mechanism of the structure. Especially
the shear ties have significant effects on ultimate flexural
strength of the sandwich panels and are used to provide integrity
⇑ Tel.: +1 912 478 7295; fax: +1 912 478 1853.
E-mail address: jkang@georgiasouthern.edu
withes [11]. The type and arrangement of the shear tie connec-
tors allowed the panels to act as partially to fully composite.
Therefore, the analytical studies to support static tests are a vital
part of efforts to efficiently evaluate the structural effects of all
components of the FICSP.
The primary objective of this paper was to evaluate the
structural behaviors of FICSP subjected to uniform pressure using
FE analyses. In this study, finite element (FE) models were devel-
oped and then validated by comparing the simulation results
with experimental data from static tests performed at the
University of Missouri [11]. The FE models in this study offer a
useful way of understanding the contribution of each component
to the failure mechanism of FICSPs subjected to uniform pressure
and should ultimately lead to better design documentation and
engineering level predictive tools for blast resistant sandwich
concrete walls.
We foresee that in the near future we will continue to design
and build civil (not military) defense shelters such as family shel-
ters in homes, common shelters in buildings or public shelters in
underground installations for the population in cases of emer-
gency for local envisages threats, levels of protection and rescue
procedures. In addition to the structural protection, these tilt-up
technologies are the most energy efficient wall systems. There-
fore, it is highly anticipated that these results make measurable
contributions to the increase of the application to the engineering
practice.

http://dx.doi.org/10.1016/j.compositesb.2014.08.034
1359-8368/Ó 2014 Elsevier Ltd. All rights reserved.
154 J. Kang / Composites: Part B 68 (2015) 153–161
2. Finite element modeling
2.1. Modeling methodology
Concrete is comprised of a wide range of materials, whose prop-
erties are quantitatively and qualitatively different. A numerical
strategy for solving any boundary value problem with location of
fracture, therefore, should consider complex constitutive model-
ing. Nonlinear concrete models were developed and validated
using testing data [11] from reinforced concrete beams and FICSPs.
Numerical investigations were executed by a general-purpose FE
analysis package, ABAQUS [12]. The concrete damaged plasticity
model in ABAQUS takes into consideration the degradation of the
elastic stiffness induced by plastic straining both in tension and
compression and provides a continuum, plasticity-based, damage
model for concrete. It assumes that the two main failure mecha-
nisms are tensile cracking and compressive crushing of the con-
crete material. The material parameters of the concrete damaged
plasticity model used in this study are presented in Table 1. The
foam (referred to as XPS hereafter) consisted of extruded polysty-
rene thermal board insulation. The concrete and foam portions
were modeled as solid elements. Rebar and WWR were modeled
using beam elements and the embedded element technique in
ABAQUS. The stress–strain relationships for rebar, WWR and XPS
are shown in Fig. 1.
Geometrically nonlinear static problems may involve buckling
or collapse behavior, where the load–displacement response shows
a negative stiffness and the structure must release strain energy to
remain in equilibrium. The Riks method uses the load magnitude
as an additional unknown and solves simultaneously for loads
and displacements. In ABAQUS [12], the ‘‘arch length,’’ l, along
the static equilibrium path in load–displacement space is used,
which offers the advantage of providing solutions regardless of
whether the response is stable or unstable.
2.2. Concrete code calibration and validation
The data from two test matrix, with two different sizes and/or
reinforcement conditions for the concrete beam (Table 2), were
employed to verify the concrete code and calibrate the parameters
of the nonlinear concrete model. The FE models for the concrete
code verification were simply supported and uniformly loaded
across a clear span shown in Fig. 2a. The concrete and reinforce-
ments (rebar and WWR) were modeled using solid elements
(C3D20; 20-node quadratic brick) and truss elements (T3D3; 3-
node quadratic truss), respectively. The interface properties
between concrete and reinforcements were assumed to be fully-
bonded.
Table 1
The material parameters used for the concrete damaged plasticity (CDP) model.
Concrete
E, modulus of elasticity GPa (psi) 24.8 (3.6E+6)
m, Poisson’s ratio 0.18
Density kg/m3 (pcf) 2403 (150)
Compressive strength MPa (psi) 28–34 (4000–5000)
Tensile strength MPa (psi) 2.8(400)
Concrete compression hardening
Yield stress, MPa (psi) Crushing strain
24 (3500) 0.0
28–34 (4000–5000) 0.002
17 (2500) 0.003
a
The ratio of initial equibiaxial compressive yield stress to initial uniaxial compressive yield stress.
b
The ratio of the second stress invariant on the tensile meridian, q (TM), to that on the compressive meridian, q (CM).
The load–displacement curves from the FE analyses were com-
pared with those of the test matrix. The vertical displacements at
the center of the bottom slab were monitored. As Fig. 3a and b
show, the results from the FE analyses were in good agreement
with those from the test matrix, although the initial stiffness from
the FE analyses was a little higher than that of the test matrix. This
is probably due to 1) the presence of a crack in the sample at mid-
span before testing and/or 2) insufficient information regarding the
tensile strength of the concrete.
2.3. Direct shear test
The ultimate flexural capacity of FICSP can highly be affected by
the stiffness and failure mode of its shear tie connectors [13]. Direct
shear tests were used to evaluate the shear resistance data (shear
force–displacement history) for composite and non-composite ties,
which was used as an input in the FE modeling in order to simulate
efficiently the shear resistance of the tie in the tilt-up sandwich
panel model. A multi-point constraints (MPC) approach was used
to model each shear tie. The test configuration consisted of three
concrete layers, two shear ties, and two layers of foam as shown
in Fig. 2b. The symmetrical test configuration was chosen to mini-
mize eccentricity. This study used the same material properties
for concrete and rebar as those of the concrete code verification.
The axial tensile forces were applied on the rebar in the central con-
crete panel and then the bottom portions of the rebar embedded in
the outside concrete layers were fixed, as shown in Fig. 2b. The
interface properties between concrete and XPS were assumed to
be frictionless. The resistance on the applied forces was, therefore,
provided only by spring elements. The nonlinear SPRING elements
in ABAQUS were used to model the actual shear resistances of ties.
SPRING1 was used to simulate the shear resistance of coupled nodes
between concrete and XPS. As shown in Fig. 2b, SPRING1 serves as
the intermediary between a node (coupled nodes) and ground.
SPRING2 was used to simulate the axial behavior of the ties,
although the axial forces were small relative to the shear forces.
The test data used as input data in SPRING1 and SPRING2 for the
shear and axial strength are shown in Fig. 3c and d. The shear
resistances from the tests were compared with those from the FE
models. Figs. 3c and d compares the tested shear resistances with
the shear resistance from the FE models and proved that the MPC
approach provides an efficient and accurate representation of the
shear resistance of various sandwich panel ties without having to
explicitly model intricate shear connector systems.
2.4. FE models of FICSP
FE models were developed for a FICSP system subjected to uni-
form pressure in this study. These sandwich panels can be either
Parameters of CDP model
w, dilation angle 30°
e, flow potential eccentricity 0.1
rb0/rc0a 1.16
Kc b 0.667
l, Viscosity parameter 0.0
Concrete tension stiffening
Remaining stress after cracking, MPa (psi) Cracking strain
2 (300) 0.0
0 0.002
– –
 J. Kang / Composites: Part B 68 (2015) 153–161 155

(a) (b)

Fig. 1. Stress–strain relationship: (a) rebar and WWR; and (b) XPS.

Table 2
Concrete code verification samples.

Name Depth, cm (in.) Width, cm (in.) Reinforcement

Concrete Beam 1 29 (11.5) 46(18) Welded-Wire W4  W4 @ 25.4 (10) # 8’s @ 24 (9.5)
Concrete Beam 2 15 (6) 46(18) Welded-Wire W4  W4 @ 8.3 (3.25) # 4’s @ 7.6 (3)

Fig. 2. FE models : (a) concrete beam for concrete code verification; (b) sandwich panel for shear test (unit: cm); and (c) deformed shape of sandwich panel for shear test
(P = tensile force applied in the rebar at the midpoint of the concrete panel; SPRING 1 = shear resistance; SPRING 2 = axial behavior).
156 J. Kang / Composites: Part B 68 (2015) 153–161

Fig. 3. Test results versus FE models: (a) concrete beam 1; (b) concrete beam 2; (c) sandwich panel with composite tie for shear test; and (d) sandwich panel with non-
composite tie for shear test.

Fig. 4. FICSP Specimen: (a) details [7]; and (b) FE model [11].
 J. Kang / Composites: Part B 68 (2015) 153–161 157

composite or non-composite, depending on the exact configura-

(a) (b)
tion. The MPC approach was utilized for the FE modeling in order
Concrete wythe σ σ
to adequately simulate the shear resistance of the ties and the
interface between the concrete and XPS, which was verified Rebar
through the simulation of the shear tie tests.
The details of the Tilt-up Concrete Association (TCA) Composite
3-2-3 Single Span Static Specimen used here are shown in Fig. 4a. Foam
The composite connector was used to provide the connection
between the concrete and XPS. The FE models of the FICSP are Composite tie Non-composite tie
shown in Fig. 4b. The same material properties were used for the
concrete, XPS and rebar as in the concrete code verification. Fig. 6. Stress distributions of composite and non-composite FICSP sections: (a)
The sandwich panel was simply supported and subjected to composite section and (b) non-composite section.
uniform pressure, as shown in Fig. 4b. The results from the static
tests of 3-2-3 FICSPs subjected to uniform pressure were compared
(a) εce εce
with those generated by the FE models. As shown in Fig. 5a, the (b)
ultimate strength of FICSP is highly affected by the interface prop-
erties between concrete and foam. It can be explained from the fact c
kd
that the stress distributions of composite sections are totally differ-
ent from those of non-composite sections (Fig. 6). The sandwich d − kd d−c
panel with composite ties has a rigid response under shear, which As f y As f y
experiences the stress distribution shown in Fig. 6a. The sandwich
panel with non-composite ties, however, can be considered as a ε s = εy ε s > εy
stacked plate where each plate is individually stressed (Fig. 6b)
Fig. 7. Strains and stress distributions in a reinforced concrete member: (a) at
[13]. Fig. 5a showed that the uniform pressure decreased abruptly yielding point and (b) at initial failure.
after it reached the structure’s ultimate strength. This was proba-
bly because several shear ties failed due to the slippage induced
by the shear forces, as shown in Fig. 5b, where the slippage
from the neutral axis to the tension face, and
between concrete and XPS in FICSP at the ultimate pressure is 0
f c ¼ the compressive strength of concreteðkPaÞ. The yielding
clearly visible. It is noticeable that the broken sections of FICSP
moment was calculated based on the elastic concrete stress
were not initiated at the middle but along a finite length adjacent
distributions shown in Fig. 7 as follows:
to the center of the span. These results confirm that the modeling
 
methodology applied in this study is indeed an efficient way to kd
model the shear resistance of the various type of ties used in M y ¼ As f y d  ð3Þ
3
sandwich panels.
where My = the yielding moment, As = the entire steel area, fy = the
yielding stress of the steel, d = the effective depth of the beam,
3. ACI Code and design practices
and kd = the distance from the compression face to the cracked elas-
tic neutral axis. The nominal moment capacity (Mn), based on Fig. 7,
The pressure–displacement and moment–curvature history
was calculated using the following expression:
from the ACI Code and standard design practices [14] were com-
 a
pared with those from the FE analyses. The initial cracking moment M n ¼ As f y d  ð4Þ
was evaluated from the following equations: 2
f y As
f r  Iut a¼ 0 ð5Þ
M cr ¼ ð1Þ 0:85f c b
yt
qffiffiffiffi c ¼ b1 a ð6Þ
0
f r ¼ 19:7 f c ð2Þ
where Mn = the nominal moment capacity, a = the depth of the
where Mcr = the cracking moment, fr = the modulus of rupture, Iut = equivalent constant stress, b = the width of the beam, and c = the
the moment of inertia of the uncracked section, yt = the distance distance to the neutral axis. The inelastic curvature was evaluated

Fig. 5. Test data versus FE results for FICSP subjected to uniform pressure: (a) pressure–displacement history and (b) deformed shapes of FE model.
158 J. Kang / Composites: Part B 68 (2015) 153–161

Table 3
Numerical results from the ACI Code and design practices.

Assumed fully composite Non-composite

qffiffiffiffiffi pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
fr (kPa) 0 fr = 3656
f r ¼ 19:7 f c ¼ 19:7 34; 450 ¼ 3656
h i h i
Iut (cm4) 1
Iut ¼ 2 12
3
ðbÞðhÞ þ ðb  hÞðys Þ2 1
Iut ¼ 2 12 ðbÞðhÞ
3
h i h i
1
¼ 2 12 ð40Þð7:6Þ3 þ ð40  7:6Þð6:35Þ2 ¼ 27; 900 1
¼ 2 12 ð40Þð7:6Þ3 ¼ 2997
Mcr (kN-cm) Mcr ¼ f ryIut ¼ ð0:3656Þð27;900Þ ¼ 1007 Mcr ¼ f ryIut ¼ ð0:3657Þð2997Þ ¼ 288
t 10:16 t 3:81
Dcr (cm) Dcr ¼ 5wL4
¼ 5ð0:087Þð305Þ4
¼ 0:14 5wL
Dcr ¼ 384EI
4 5ð0:025Þð305Þ
¼ 384ð2408Þð2997Þ
4
¼ 0:38
384EIut 384ð2408Þð27;900Þ ut
 
My (kN-cm) M y ¼ As f y ðd  kd
3Þ M y ¼ 2As f y d  kd 3
¼ ð1:42Þð48:03Þð16:51  2:72=3Þ ¼ 1065 ¼ 2ð1:42Þð48:03Þð3:81  1:14=3Þ ¼ 468
4 h i
Ict (cm ) Ict ¼ 13 ðbÞðcÞ3 þ nðAs Þðd  kdÞ
2
Ict ¼ 2 13 ðbÞðcÞ3 þ nðAs Þðd  kdÞ
2
h i
¼ 1=3ð40Þð2:72Þ3 þ 8ð1:42Þð16:5  2:72Þ2 ¼ 2428
¼ 2 1=3ð40Þð1:14Þ3 þ 8ð1:4Þð3:8  1:14Þ2 ¼ 202
Dy (cm) 5wL
Dy ¼ 384EI
45ð0:09Þð305Þ
¼ 384ð2408Þð2428Þ ¼ 1:7
4
5wL
Dy ¼ 384EI
4 4
5ð0:04Þð305Þ
¼ 384ð2408Þð202Þ ¼9
ct ct

a (cm) a¼
f y As
0 ¼ ð48:03Þð1:42Þ
¼ 0:57 a = 0.57
0:85f c b 0:85ð0:3445Þð40Þ
 
Mn (kN-cm) M n ¼ As f y ðd  2aÞ ¼ ð1:42Þð48:03Þð16:51  0:57
2 Þ M n ¼ 2As f y d  2a ¼ 2ð1:42Þð48:03Þð3:81  0:57
2 Þ
¼ 1107 ¼ 481
Dn (cm) 5wL
Dn ¼ 384EI
45ð0:095Þð305Þ
¼ 384ð2408Þð2428Þ ¼ 1:78
4
5wL
Dn ¼ 384EI
45ð0:041Þð305Þ
¼ 384ð2408Þð202Þ
4
¼ 9:3
ct ut
d L d
Du Du ¼ Dn þ Dp ¼ Dn þ  /n Þ  2 ð/u 2 Du ¼ Dn þ Dp ¼ Dn þ  /n Þ  2L
2 ð/u
¼ 1:78 þ 16:51 305
2 ð0:0037  0:000184Þ 2 ¼ 6:2 ¼ 9:3 þ 3:81 305
2 ð0:0037  0:001Þ 2 ¼ 10:1

*b = width of FICSP; h = depth of a concrete panel; n = ratio of modulus of elasticity of steel and concrete; ys = the distance from the neutral axis to the rebar.

by considering the plastic rotation, which is not confined to a single

cross section but is distributed over a finite length, referred to as the
plastic hinge length [15]. The plastic hinge length in this study was
assumed to be d/2. The deflections and curvatures were evaluated
from the following equations:

5wL4
Dcr;y;orn ¼ ð7Þ
384EI
8M
w¼ ð8Þ
L2
M
/cr;y;orn ¼ ð9Þ
EI
ec b1
/u ¼ ð10Þ
a
Du ¼ Dn þ Dp ð11Þ
d L
ð/  /n Þ  ð12Þ
2 u 2

where Dcr,y,n,oru = the deflection of the beam at the midpoint of the

span (at the stage of cracking, yielding, nominal moment or ulti-
mate), w = the applied uniform pressure, L = the span length of the
beam, E = the modulus of elasticity of the concrete, I = the moment
of the inertia of the beam section, M = the bending moment at the
middle span, /cr,y,n,oru = the curvature of the beam at the midpoint
of the span (at the stage of cracking, yielding, nominal moment or
ultimate), ec = the ultimate strain of the concrete, b1 = the fraction
of the distance to the neutral axis, Dp = the plastic deflection, and
hp = the inelastic rotation.
The panel shown in Fig. 4a was used in the numerical example.
The parameters and related numerical results from the ACI Code Fig. 8. Pressure–displacement and moment–curvature history: (a) pressure–
and design practices are listed in Table 3. Fig. 8 shows that the displacement history and (b) moment–curvature history (L = span length).
pressure–displacement and moment–curvature relationships from
the current ACI Code and design practices provide the upper and concrete walls and the deflection limits for FICSPs could thus be
lower boundaries for the FE analyses and tests. Evidently, the resis- increased considerably.
tance and sliding of the ties significantly affects the global behav-
iors of FICSPs, as shown in Fig. 8a and b. The ACI Code [14] limits 4. Structural performance of FICSP
the maximum deflection for slender concrete walls by L/180. As
shown in Fig. 8a, the pressures in the FE models and test at this The detailed structural behaviors of each component of a FICSP
deflection reach up to 50% of the ultimate pressure. No significant subjected to uniform pressure were evaluated based on the FE
cracks, however, were induced. This suggests that the current ACI models developed here. The results are presented and discussed
Code provides very conservative deflection limits for slender in this section.
 J. Kang / Composites: Part B 68 (2015) 153–161 159

(a) (b)

Stress at middle span (MPa)

10 4

Stress at bottom (MPa)

0

-10 Bottom wythe

Top wythe 2
-20

-30

-40 0
0 30 60 0.000 0.001 0.002 0.003
Vertical deflection (cm) Strain (cm/cm)

(c) 0
Stress at top (MPa)

-10

-20

-30

-40
-0.003 -0.002 -0.001 0.000
Strain (cm/cm)

Fig. 9. Longitudinal stresses and strains of concrete at the midpoint of the span: (a) stress–deflection history; (b) stress–strain curve at bottom wythe; and (c) stress–strain
curve at top wythe (tensile is positive).

4.1. Concrete panels 4.2. Rebar

The stress-vertical deflection and stress–strain curves of the top The structural behaviors of the rebar in the FICSPs were moni-
and bottom wythes at the midpoint of the span are plotted in Fig. 9. tored in detail, as shown in Fig. 10. Fig. 10a and b shows the stress
Interestingly, the cracking of the bottom wythe was induced before distributions of the top and bottom reinforcements along the span
the ultimate compressive stress was imposed. The stress–strain at the initial yielding and ultimate stages of the system. At the ini-
history after the ultimate compressive and tensile strength were tial yielding stage, the highest compressive stresses at the top
reached clearly shows the compression hardening and tension wythe occurred in the same location at which the FICSP experi-
stiffening effects. The close relationships between the pressure– enced the failure of sections, as shown in Fig. 5b. The tensile stres-
displacement curves of the experimental tests and the FE modeling ses at the bottom wythe from this point were increased
suggest that the simulation of the compression hardening and ten- substantially up to the midpoint of the span. The deformed shape
sion stiffening of the concrete is satisfactory for this application. of a FICSP subjected to uniform pressure, therefore, could be

(a) 9 (b) 600

Stress at initial yielding

Stress at ultimate stage

Bottom wythe Bottom wythe

6
Top wythe 300 Top wythe
(MPa)

(MPa)

3
0
0

-3 -300
Location Location
Middle of span Middle of span

(c) 600
Stress (MPa)

300

0
0.000 0.003 0.006
Strain

Fig. 10. Structural performance of rebar: (a) stress distribution at initial yield point; (b) stress distribution at ultimate stage; and (c) stress versus strain.
160 J. Kang / Composites: Part B 68 (2015) 153–161

(a) 12
200
Shear resistance (kN)

Stress at middle (kPa)

150

6
100

50
0
0 30 60
0
Vertical deflection (cm) 0.000 0.005 0.010 0.015
End of top wythe Middle of top wythe Strain (cm/cm)
End of bottom wythe Middle of bottom wythe
Fig. 12. Longitudinal stress–strain history induced on foam at the midpoint
(longitudinal direction is along the span).
(b) 15
Shear resistance (kN)

The data in Fig. 11, therefore, suggest that the enhancement of the
10 shear strengths at the ends will be a more efficient way to increase
the ultimate strength of FICSPs subjected to uniform pressure than
5 enhancing the shear strengths at the midpoint or at other loca-
tions. It is noticeable from Fig. 11a that the yield point of each
tie installed at one of the other locations coincided with the deflec-
0
0 1 2 tion point at the ultimate strength of the system, as shown in
Shear deformation (cm) Fig. 5a, which indicates that the shear resistance of its ties has a
significant effect on the ultimate strength of a FICSP.
Input from shear testing End of bottom wythe
End of top wythe Middle of bottom wythe
4.4. Foam
Middle of top wythe

As shown in Fig. 12, the foam insulation experienced no signif-

(c) 12
icant stress for any of the loading stages. The stress path of the
foam reversed after the ultimate strength of system was reached.
Shear resistance (kN)

It should be noted that the primary purposes of foam insulation

are not to reinforce the structure but to enhance energy efficiency
6 and mitigate impact damage.

5. Conclusions

0 This study investigated the structural behaviors of FICSPs sub-

0 180
Location of tie (cm)
Top wythe Bottom wythe
Fig. 11. Structural behavior of shear ties: (a) shear resistance of tie versus vertical
deflection of FICSP; (b) shear resistance of tie versus shear deformation; and (c)
distribution of shear resistances of tie along the span at ultimate stage of system.
360
jected to uniform pressure using both experiments and finite ele-
ment models. The modeling approach in this study was found to
be a good way to simulate the structural behaviors during all load-
ing stages (elastic, yielding, ultimate and post-failure). This study
also compared the new approach with the existing ACI Code and
current design practices. The following findings and conclusions
were drawn:

(1) Both composite and non-composite action had significant

expected to be associated closely with the stress distribution in the effects on the response of FICSPs, suggesting that the simu-
rebar along the span. The membrane action of the FICSP under the lated shear tie resistance should be incorporated in numerical
large deflection also affected these stress distributions shown in analyses of FICSPs. This study showed that the multi-point
Fig. 10a and b. The rebar did not reach to the ultimate strength constraint (MPC) approach is an efficient way to model the
of steel after it yielded, as shown in Fig. 10c, which explains that resistance of shear ties.
the concrete had already experienced severe tensile cracking and (2) The modeling approach developed in this study satisfactorily
compressive crushing failures. simulated the detailed structural behaviors of concrete,
rebar, foam and ties during all loading stages (elastic, yield-
4.3. Shear ties ing, ultimate and post-failure).
(3) The current ACI Code and design practices provided the
As Fig. 11 indicates, the shear resistances of the ties and the glo- upper and lower boundaries for the FE analyses and tests
bal behaviors of FICSPs are highly related. The shear resistances of and was found to apply very conservative deflection limits
the ties at the midpoint and end were evaluated along the vertical (L/180) for slender concrete walls. The deflection limits for
deflection. As shown in Fig. 11a, the resistances of the ties at the FICSP could, therefore, be significantly increased.
end and at the bottom were higher than those at the middle and (4) The pressure–displacement curves (Fig. 8) of the experimen-
the top, respectively. It was found from Fig. 11 that the largest tal test data and the FE model showed good agreement due
and smallest shear forces on the ties were induced at the end of to the satisfactory simulation of the compression hardening
the bottom wythe and at the middle of the top wythe, respectively. and tension stiffening of concrete.
 J. Kang / Composites: Part B 68 (2015) 153–161
(5) The deformed shape of the FICSP subjected to uniform pres-
sure conformed closely to the stress distribution in the rebar
along the span.
(6) The largest and smallest shear forces in the ties were
induced at the end of the bottom wythe and at the midpoint
of the top wythe, respectively, as shown in Fig. 11a, which
means that enhancing the shear strengths at the ends is a
more efficient way to increase the ultimate strength of FICSP
subjected to uniform pressure than enhancing the shear
strengths at the midpoint or other locations. It is interesting
to note that the yield point of each tie installed at another
location coincided with the deflection point in the ultimate
strength of the FICSP, from Figs. 11a and 5a, which suggests
that the shear resistance of the ties has a significant effect on
the ultimate strength of the FICSP.
Acknowledgments
The work reported herein was partially supported by Auburn
University and Georgia Southern University. The investigator is
grateful for their sponsorship. This study does not necessarily
reflect the opinions or conclusions of Auburn University or Georgia
Southern University.
References
[1] Caner FC, Bažant ZP. Size effect on strength of laminate-foam sandwich plates:
finite element analysis with interface fracture. Compos Part B: Eng
2009;40(5):337–48.
161
[2] Demelio G, Genovese K, Pappalettere C. An experimental investigation of static
and fatique behavior of sandwich composite panels joined by fasteners.
Compos Part B Eng 2001;32(4):299–308.
[3] Tarlochan F, Ramesh S, Harpreet S. Advanced composite sandwich structure
design for energy absorption applications: Blast protection and
crashworthiness. Compos Part B Eng 2012;43(5):2198–208.
[4] Bunyawanichakul P, Castanié B, Barrau JJ. Non-linear finite element analysis of
inserts in composite sandwich structures. Compos Part B Eng 2008;39(7–
8):1077–92.
[5] Einea A, Salmon DC, Tadros MK, Culp TD. A new structurally and thermally
efficient precast sandwich panel system. PCI J 1994;39(4):90–101.
[6] Precast/Prestressed Concrete Institute (PCI) Committee Report. State-of-the-
art of precast/prestressed sandwich wall panels. PCI J, vol. 56(2), 2011. p. 131–
76.
[7] Salmon DC, Einea A, Tadros MK, Culp TD. Full scale testing of precast concrete
sandwich panels. ACI Struct J 1997;94(3):354–62.
[8] Pantelides CP, Surapaneni R, Reaveley LD. Structural performance of hybrid
GFRP/steel concrete sandwich panels. ASCE 2008;12(5):570–6.
[9] Naito CJ, Dinan RJ, Fisher JW, Hoemann JM. Precast/prestressed concrete
experiments – series 1 (volume 1). Interim Report. AFRL-RX-TY-TR-2008-4616.
Air Force Research Laboratory, 2008.
[10] Naito CJ. Analytical assessment of the blast resistance of precast, prestressed
concrete components. Interim report. AFRL-ML-TY-TP-2007-4529. Air Force
Research Laboratory, 2007.
[11] Newberry CM, Davidson JS, Hoemann JM, Bewick BT. Simulation of Prestressed
Concrete Sandwich Panels Subjected to Blast Loads. Air Force Research
Laboratory, AFRL-RX-TY-TP-2010-0014, Preprint, Feb. 2010.
[12] SIMULIA. ABAQUS Analysis User’s Manual Version 6.7. 2007.
[13] Naito CJ, Hoemann JM, Bewick BT, Hammons MI. Evaluation of shear tie
connectors for use in insulated concrete sandwich panels. Interim report.
AFRL-Rx-TY-TR-2009-4600. Air Force Research Laboratory, 2009.
[14] American Concrete Institute (ACI). Building code requirements for structural
concrete and commentary. ACI 318, Farmington Hills, Mich, 2008.
[15] Nilson AH, Darwin D, Dolan CW. Design of concrete structures. 13th
ed. Boston: McGraw-Hill; 2004.