Engineering Structures 132 (2017) 249–260

Contents lists available at ScienceDirect

Engineering Structures
journal homepage: www.elsevier.com/locate/engstruct

Static and dynamic analyses of insulated concrete sandwich panels using

a unified non-linear finite element model
Paul M. Hopkins a, An Chen b,⇑, Mostafa Yossef b
a
TD&H Engineering, Media, PA 19063, USA
b
Department of Civil, Construction and Environmental Engineering, Iowa State University, Ames, IA 50010, USA

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

Article history: Insulated concrete sandwich panels are comprised of two outer concrete wythes and an inner layer of
Received 3 November 2015 foam insulation. They have been increasingly used because of their advantages of light weight and energy
Revised 4 November 2016 efficiency. Various shear connectors can be used to connect the two outer concrete wythes. More
Accepted 5 November 2016
recently, Fiber-Reinforced Polymer (FRP) shear connectors have been used, which can eliminate thermal
bridging and improve the thermal performance. Typical approaches to Finite Element (FE) analysis treat
static and dynamic analyses separately. However, due to the flexibility of the FRP shear connectors and
Keywords:
the cracking of the concrete in insulated concrete sandwich panels, a nonlinear static analysis model
Finite element analysis
Insulated concrete sandwich panel
would often diverge early based on a preliminary FE study conducted by the authors. To address this
Fiber-Reinforced Polymer shear connector issue, a nonlinear explicit dynamic FE model using ABAQUSÓ was developed, which can study both
Bending behavior the panels’ static behavior under typical flexural loading and dynamic behavior under blast loading.
Blast loading Nonlinear material properties were incorporated and damaged plasticity model was used to model con-
crete in both compression and tension. In order to simulate the static behavior, the time loading control
was applied to the FE model to slow down the rate of loading to smoothly capture the response. For
dynamic analysis under blast loading, a verification study was conducted first using the developed FE
model, where good correlations can be obtained between the FE and test results on a panel tested in a
previous study. The FE model was further used to study the dynamic behavior of two panels under blast
loading: one is a solid concrete panel and the other is an insulated concrete sandwich panel. It can be con-
cluded that, although the insulated concrete sandwich panel is lighter, it still performs relatively well
compared to the solid panel under blast loading. Therefore, it is promising to use insulated concrete sand-
wich panels for both conventional and blast-resistant structures.
Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction More recently, Fiber-Reinforced Polymer (FRP) shear connectors

have been used, which can eliminate thermal bridging and
As shown in Fig. 1, insulated concrete sandwich panels are com- improve the thermal performance. Various types of FRP connectors
prised of two layers of concrete, known as wythes, separated by a have been studied, including C-grid by Frankl et al. [1], G-grid by
layer of rigid foam plastic insulation. The two wythes are con- Soriano and Rizkalla [2], and Glass FRP bar connector by Woltman
nected by some form of shear transferring mechanism, generally et al. [3]. More recently, the authors have developed an innovative
using solid concrete web zones, metal connectors, plastic connec- FRP plate shear connector based on comprehensive bending tests
tors, or a combination of these elements. The panels can provide and linear finite element analysis (FEA) [4,5]. Four types of panels
dual function of transferring load and insulating the structure were built and tested which included sandwich panels with dis-
among other desirable characteristics of normal concrete panels, crete FRP shear connectors, segmental FRP shear connectors
such as durability, low-cost, fire resistance, etc. [19–25] Therefore, (Fig. 1), and continuous FRP shear connectors, and solid reinforced
they have been increasingly used in building industry. panels (Fig. 2). The dimensions of the segmental FRP shear connec-
tors are shown in Fig. 3 and a photo of the assembly of the insula-
tion with the segmental FRP connector is shown in Fig. 4. All the
above studies showed that the sandwich panels performed well
⇑ Corresponding author. compared to solid reinforced concrete panels and can be used as
E-mail addresses: paul.hopkins@tdhengineering.com (P.M. Hopkins), achen@ structural members.
iastate.edu (A. Chen).

http://dx.doi.org/10.1016/j.engstruct.2016.11.017
0141-0296/Ó 2016 Elsevier Ltd. All rights reserved.
250 P.M. Hopkins et al. / Engineering Structures 132 (2017) 249–260

Fig. 1. Insulated concrete sandwich panel.

Fig. 2. Solid reinforced concrete panel.

Fig. 3. Segmental FRP shear connector.

Constructing and testing panels, although a necessity to prove

and verify strength and durability, can be expensive. Therefore, a
reliable FEA model is needed so that parametric studies and
advanced analyses can be performed, such as the ones initiated
in the research [18], which is the objective of this study. ABAQUSÓ
[6] is used for this purpose. However, there are many types of sol-
Fig. 4. Assembly with segmental FRP connector.
vers and elements to choose from. This paper provides a summary
on the nonlinear analysis approaches which can be taken to obtain
a solution to the insulated concrete sandwich panel modeling, of insulated concrete sandwich panels under blast loading, which
based on which an explicit dynamic model is proposed. will also be studied in this paper.
The United States Army Corps of Engineers have years of pub-
lished research on the topic of blast loading and structural harden-
ing [14,15]. Many typical structural building materials, such as 2. Finite element analysis
masonry, reinforced concrete and steel were labelled as preferred
and tested materials to be used in buildings designed for blast The goal of this paper is to develop a reliable FEA solver that can
resistance. Precast concrete materials, specifically sandwich pan- predict the strength and deflection of the insulated concrete sand-
els, were mentioned in these publications. But they were mainly wich panels. Simplifying the model into shell elements or utilizing
used as envelope or cladding type materials, which are secondary other simplistic forms was eliminated so that we can capture
structures. Limited studies are available on the dynamic behavior cracking patterns and full failure mechanisms. ABAQUS offers a
 P.M. Hopkins et al. / Engineering Structures 132 (2017) 249–260 251

variety of solvers that would fit the problem presented, which is a wich panel has two-FRP segmental shear connectors which con-
simply supported beam with a static three-point load at the centre nect the top and bottom concrete wythes. The solid panel is
span. However, nonlinear static analysis models may diverge early 254 mm (10 in.) thick, 610 mm (24 in.) wide and 2743 mm (9 feet)
and cannot provide a complete solution to ultimate failure, as will long, with (2) 13 mm (#4) bars at top and (2) 16 mm (#5) bars at
be shown next. bottom, as shown in Fig. 2. The tension reinforcement ratio for the
solid panel is 0.0029 and the compression reinforcement ratio is
2.1. Panel construction 0.0018. Since the sandwich panel has the same width, depth, and
reinforcement layout as the solid panel, the reinforcement ratios
The sandwich panel (Fig. 1) is 254 mm (10 in.) thick, 610 mm for the two types of the panels are the same.
(24 in.) wide and 2743 mm (9 feet) long. The distance between
the supports are 2438 mm (8 feet). The panel has two 76 mm
(3 in.) thick concrete wythes and a 102 mm (4 in.) thick insulation 2.2. Materials
layer with extruded polystyrene (XPS). The bottom concrete wythe
has (2) 16 mm (#5) longitudinal bars and (5) 13 mm (#4) trans- There are four materials for the sandwich panel: concrete, steel
verse bars. The top concrete wythe has (2) 13 mm (#4) longitudi- reinforcement, insulation and FRP shear connectors. The material
nal bars and 13 mm (#4) transverse bars, which are situated at properties are shown in Tables 1 and 2. The compressive strength
the holes in the FRP shear connectors as shown in Fig. 1. The sand- of the concrete was obtained through testing [5]; and the theoret-
ical values were used for the complete non-linear stress-strain
curve [7]. The properties of steel, rigid expanded polystyrene insu-
Table 1
Material properties. lation, and FRP were from [12], [11], and [13], respectively.
Linear elastic properties were used for the insulation and FRP
Material Mass density Young’s Poisson’s
shear connectors; and concrete and steel used non-linear material
modulus ratio
values incorporating their respective stress-strain curves. The rein-
Concrete 2400 kg/m3 25.2 GPa 0.15
forcing steel is ASTM A615 Grade 60 and a typical stress-strain
(0.000225 lb s2/in.4) (3659 ksi)
Steel ASTM A615 7800 kg/m3 200 GPa 0.3
curve with the required logarithmic strain and plastic stress was
(0.000783 lb s2/in.4) (29,000 ksi) used, as shown in Fig. 5. For concrete, a different approach was
Insulation expanded 22 kg/m3 2.34 MPa 0.35 taken as the concrete material is a brittle, quasi-static material
polystyrene [11] (22.06E
6 lb s2/in.4) (340 psi) and the Damaged Plasticity model in ABAQUS was best suited for
this representation.

Table 2
Fiber-Reinforced Polymer (FRP) material properties.

Typical values
Property EATR 0.08500 | 2.2 mm Test method
Flexural strength 33  103 psi 228 MPa ASTM – D790
Flexural modulus 1.0  106 psi 6895 MPa ASTM – D790
Tensile strength 45  103 psi 310 MPa ASTM – D638
Tensile modulus 2.0  106 psi 13,790 MPa ASTM – D638
Barcol hardness 45 45 ASTM – D2583
Coefficient of linear thermal expansion 0.80  10
5 in./in./°F 14 lm/m/°C ASTM – D696
Thermal conductivity 0.4 Btu in./h ft2 °F 5.0 cal cm/h m2 °C ASTM – C177
Water absorption 0.2%/24 h@77 °F 0.2%/24 h@25 °C ASTM – D570
Specific gravity 1.75 1.75 ASTM – D792

ASTM A615 Gr. 60 Steel Reinforcement

900

800

700

600
Stress, MPa

500

400

300
ASTM A615 Gr. 60 Rebar
200
ASTM A615 Gr. 60 True Stress/Log Strain
100

0
0 0.05 0.1 0.15 0.2 0.25
Strain, mm/mm

Fig. 5. Stress/strain data for various steel [12].

252 P.M. Hopkins et al. / Engineering Structures 132 (2017) 249–260

Fig. 6. Compressive stress-strain curves for concrete.

Compressive cylinder tests were performed on concrete sam- good correlations can be observed between the theoretical model
ples for the panels per ASTM C39 and the results for the stress and FE results in Fig. 6. Therefore, the theoretical model was used
and strain are plotted in Fig. 6. The compressive strength of the in this study.
concrete was 28.4 MPa (4120 psi) [5]. In order to capture full The Damaged Plasticity Model in ABAQUS also has tension
degradation and stress strain curve for the concrete sample, the degradation component along with the compression damage. The
theoretical model based on the 28 day compressive strength by modified tension stiffening curve as shown in Fig. 7 was first devel-
Hsu & Hsu [7,26] is also plotted in Fig. 6, with damage parameters oped by Gilbert and Warner [8], then modified by Nayal and
shown in Table 3. An FE analysis based on the damage parameters Rasheed [9] and finalized by Wahalathantri et al. [10]. The theoret-
shown in Table 3 was conducted on a cylinder with a diameter of ical curve for tension is based on an assumed tensile strength of
150 mm and height of 300 mm under compressive load, where 10% of compressive strength. The cracking strain (ecr t ), damage

Table 3
Concrete damaged plasticity values (compression) (unit for the stress: MPa).

Concrete damaged plasticity

Compression damage Compression behavior Check Check
Damage Inelastic Yield Inelastic Plastic
Parameter Strain Stress Strain Strain
d ein
c sc ein
c epl
c (1-dc)E0
0.000 0 0 0 0 26,167
0.000 0 5.68148 0 0 26,167
0.000
3.28881E
05 11.46675442 0
3.28881E
05 26,167
0.000 0.000217112 16.44031471 0.000217112 0.000217112 26,167
0.000 0.000467112 20.39739695 0.000467112 0.000467112 26,167
0.000 0.000717112 23.37914656 0.000717112 0.000717112 26,167
0.000 0.000967112 25.50835223 0.000967112 0.000967112 26,167
0.000 0.001217112 26.93672502 0.001217112 0.001217112 26,167
0.000 0.001467112 27.81295463 0.001467112 0.001467112 26,167
0.000 0.001967112 28.40706742 0.001967112 0.001967112 26,167
0.012 0.002467112 28.06124848 0.002467112 0.002453896 25848.45095
0.041 0.002967112 27.24001741 0.002967112 0.002922512 25091.97887
0.078 0.003467112 26.20506414 0.003467112 0.00338296 24138.63786
0.116 0.003967112 25.09858818 0.003967112 0.003840675 23119.41416
0.155 0.004467112 23.99555384 0.004467112 0.004298521 22103.36069
0.228 0.005467112 21.93069587 0.005467112 0.00521961 20201.32913
0.292 0.006467112 20.12038973 0.006467112 0.006150428 18533.77648
0.347 0.007467112 18.56107159 0.007467112 0.007090837 17097.41991
0.394 0.008467112 17.22025666 0.008467112 0.008039596 15862.33627
0.435 0.009467112 16.06230802 0.009467112 0.008995344 14795.69882
0.501 0.011467112 14.17420573 0.011467112 0.010923188 13056.48471
0.553 0.013467112 12.70641043 0.013467112 0.012867094 11704.43385
0.594 0.015467112 11.534752 0.015467112 0.014822318 10625.16772
0.628 0.017467112 10.57800719 0.017467112 0.016785755 9743.867964
0.656 0.019467112 9.781662588 0.019467112 0.018755322 9010.319901
0.685 0.021967112 8.955494629 0.021967112 0.021223749 8249.300235
 P.M. Hopkins et al. / Engineering Structures 132 (2017) 249–260 253

Table 5
FEA element types.

Material Element Comment

component type
Concrete C3D8R Linear hexahedral element with enhanced
stiffness hourglass control and reduced
integration
Insulation C3D8R Linear hexahedral element with enhanced
stiffness hourglass control and reduced
integration
Rebar T3D2 Linear truss bar element
FRP Plate S4R Linear shell element
FRP Shear S4R Linear shell element
Connector

before obtaining the full solution, as will be shown in Section 2.5.

Fig. 7. Tensile stiffening model for concrete. This is probably due to the flexibility of the FRP shear connectors
and concrete cracking for insulated concrete sandwich panels.
Therefore, the work around was to use a dynamic explicit solver
parameter (d) and the plastic strain (epl
t ) are then calculated and in ABAQUS by realizing that the concrete failure is a quasi-static
incorporated in the ABAQUS Damaged Plasticity model, which mechanism and the Dynamic Explicit solver in ABAQUS is excellent
are shown in Table 4 [18]. for modeling brittle materials and obtaining reliable solutions and
results. Since the actual experimental test is not a dynamic event,
2.3. Finite element model the load in the FEA model required an amplitude or rate of loading
that slowed it down enough to create a more static response, as
The FEA model is comprised of solid, bar, and shell elements to shown in Fig. 10.
represent the concrete and insulation, the steel reinforcement and
the FRP shear connectors, respectively. The constituent materials
and element types used in the FE model are listed in Table 5. The 2.5. FEA results
size of the elements used were around 13–50 mm in length. Where
there was contact, such as insulation to concrete, contact was The dynamic explicit solver in ABAQUS using the Damaged Plas-
incorporated, and the remaining materials were tied such as the ticity model allowed the user to capture the full failure of the pan-
rebar to the concrete and the FRP to the concrete. The part els. It was used to study the sandwich and solid panels as shown in
instances for the sandwich panel with segmental shear connector Fig. 1 and Fig. 2, respectively, which was tested by the authors
are shown in Fig. 8. The reinforcing steel bars were assigned as [4,5].
truss element stringers embedded in the concrete part, as shown During the analysis, it was observed that the load tends to
in Fig. 9. bounce or rebound after it is applied during the early elastic phase,
Further explanations of the FE models regarding the modeling as shown in Fig. 11. Once the panel starts to deform and the con-
technique, materials used, sequence of construction, boundary crete begins to crack, the inelastic response absorbs this applied
conditions, along with materials properties and loads can be found loading and the quasi-static mechanism is captured. In Fig. 11,
from [18]. the zero to 1.0 step time is the applied self-weight of the concrete
panel to represent the panel placed on the testing blocks, which is
2.4. Solver techniques a static load. Then starting at time 1.0, the dynamic model com-
mences and the immediate rebounding of the load is seen as it
ABAQUSÓ V6.11 [1] was used for the pre- and post-processing begins to be applied and ramped up on the panel. Other FEA results
and the solver for the analysis. Four main techniques were investi- are shown in Figs. 12–15. Tensile damage has occurred as the top
gated: Static-General, Static-RIKS, Implicit Quasi-Static, and because of the localized effect of the load applied at the mid-span.
Dynamic-Explicit. The Static-General and the Static-RIKS methods Shear lock effect appears when there are not enough elements
provided good stable results for the panels during the loading of through the thickness. When shear lock happens, deformation is
the mostly linear to early nonlinear response stages. The solver under-predicted and shear stress is over-predicted. In this study,
then had difficulty progressing and the solution took a long com- fine mesh with quadrilateral element with hourglass mode were
putational time with small increments. Ultimately it stopped used to avoid shear locking. As shown in Fig. 12, the panel failed

Table 4
Concrete damaged plasticity (tension stiffening) values. (Unit for the stress: MPa).

Concrete damaged plasticity (tension stiffening)

Tensile behavior Tension Damage Check
Yield Eng. Cracking Damage Cracking Plastic
Stress Strain Strain Parameter Strain Strain
st et eck
t d eck
t epl
t

0 0 0.00000 0
2.84074 0.000108556 0.00000 0.000 0.00000 0
2.18737 0.000135695 0.00005 0.230 0.00005 2.71E
05
1.278333 0.000434224 0.00039 0.550 0.00039 0.000326
0.284074 0.000944437 0.00093 0.900 0.00093 0.000836
254 P.M. Hopkins et al. / Engineering Structures 132 (2017) 249–260

(a) Shear FRP Grid (Segmental-Up) (b) EPS Insulation

(c) Hollow Concrete Slab (d) FRP Top Plate

Fig. 8. FEA part instances in ABAQUS.

in excessive deflection, which indicates that shear lock did not as it cracks and the load is transferred to the reinforcing steel. As
occur. the panel becomes severely cracked and the FRP shear connectors
Throughout the testing and analysis of these panels, displace- and reinforcing steel carry most of the load, the curve becomes
ment has been the control. The testing showed that the panels more smooth and continuous. The advantage of using such a FEA
can displace up to 45 mm while the FEA models can have success- solver is that it allows the model to capture the excessive deflec-
ful solutions beyond 50 mm. The results from the dynamic explicit tions and the cracking of the concrete and ultimately the flexural
FEA model correlated well with the static experimental results failure.
from [5], as shown in Figs. 14 and 15 in terms of load-deflection A parametric study on the effect of viscosity parameter on the
curves. Both of these panels failed in flexure during testing with behavior of the sandwich panels was conducted. The parameters
excessive deflection and cracking of the concrete, where the rebar were varied from 0 to 0.01, with results shown in Fig. 16, which
yielded before the concrete crushed. The failure points or ultimate indicates that the viscosity parameter has limited effect. Therefore,
displacement in the curves shown in Fig. 15 are located at different a viscosity parameter of zero was used in the FE model.
values, because of the bouncing and rebounding of the load due to The contour plot of shear stress of the FRP connectors is shown
dynamic nature of the FEA. in Fig. 17. It is reported that the tensile strength of the FRP shear
Using the Static-General or Static-RIKS model, the FEA solution connector is 310 MPa, as shown in Table 2. The shear strength of
reached a fraction of completion before stopping, as circled in the FRP plate is about 55% of the tensile strength [27]. Therefore,
Fig. 15. The FEA curve using the explicit analysis is not smooth the shear strength of the FRP connectors is about 171 MPa. The
due to the quasi-static response of the concrete sandwich panel maximum shear stress in the FRP shear connectors is 134 MPa

Fig. 9. Reinforcing steel sections assigned as stringers.

 P.M. Hopkins et al. / Engineering Structures 132 (2017) 249–260 255

Fig. 13. Tension damage for sandwich panel.

3. FEA blast model verification study

As shown above, an FEA model was developed to study the

behavior of insulated concrete sandwich panels. It was shown that
this model can be used to predict the behavior of sandwich panels
under static load. In order to obtain confidence in using the FEA
model for dynamic study, the test from another published report
[16,17] was used for verification purpose. This particular panel
was a solid reinforced concrete panel with Normal Strength Con-
crete – Normal Strength Rebar (NSC-NR), which was reinforced
with 10 mm (#3) rebars, as shown in Fig. 18. The panel was tested
at the Army Research Labs in Virginia, USA in 2012. A mechanical
device at the Army Research Lab created a load similar to that of
13.608 kg (30,000 lb s) of TNT equivalent at 68.58 m (225 feet)
Fig. 10. Loading rate for static analysis. standoff distance. The strength of the steel and concrete and other
details can be found from [16,17].
The reference panel tested at the Army Research Lab [16,17]
(19.4 ksi), which is lower than the strength of the FRP. Therefore,
was recreated in ABAQUS by the authors. The blast pressure and
the FRP shear connectors did not fail during the test, which is con-
impulse diagram for the NSC-NR model was traced by the authors
sistent with the post-test inspection, where no damage of the FRP
in excel and then incorporated into the ABAQUS dynamic explicit
connectors was observed.
model, as shown in Fig. 19. The graph shown in Fig. 19a is the orig-

Force smoothes out as concrete

begins to crack and fail

Fig. 11. FEA measured support reaction versus step time for sandwich panel.

Fig. 12. Tension damage for solid panel.

256 P.M. Hopkins et al. / Engineering Structures 132 (2017) 249–260

Test results

FEA results

Fig. 14. Nonlinear FEA vs. experimental results for solid panel.

Explicit Dynamic
FEA results
Test results

Static Riks
FEA results

Fig. 15. FEA vs. experimental results for sandwich panel.

inal data pressure time plot [16,17] and the graph shown in from Thiagarajan et al. [16]. Deflection shape and crack patterns
Fig. 19b is the traced data plot by the authors in order to replicate from the authors’ ABAQUS model are shown in Fig. 21(b) and (c),
the data. The loading was applied and the same solution and ele- respectively, where good correlations with Fig. 21(a) can be
ment types as described in Section 2 were used, so that verification observed. Therefore, the ABAQUS model can be used with confi-
can be made. The results from the numerical analysis using ABA- dence in the following study.
QUS and the experimental results are shown in Fig. 20. Good cor-
relation can be seen in the deflection versus time plot, even 4. FEA blast modeling for insulated sandwich panel
though there is some difference for the initial slope of the curves,
probably due to approximation of boundary conditions and loading Due to the favorable comparisons obtained between the FEA
pressures. The experimental panel damage is shown in Fig. 21(a) and test results described above, the FEA model was further used

Viscosity parameter=0.01

Viscosity parameter=1E-5

Viscosity parameter=0

Fig. 16. Effect of the viscosity parameter.

 P.M. Hopkins et al. / Engineering Structures 132 (2017) 249–260 257

Fig. 17. Shear stress of FRP connectors (unit: psi; 1 psi = 6.8948 kPa).

Fig. 18. Reinforcing plan and section of Thiagarajan panel [16,17].

Peak Pressure Peak Impulse Peak Pressure Peak Impulse

Impulse (kPa-msec)
Pressure (kPa)

Pressure (kPa)

(a) Pressure and Impulse (Courtesy: US (b) Traced pressure vs. time plot for
Army ERDC) [16, 17] recreation of blast model in this study
Fig. 19. Recreation of blast pressure vs. time plot.
258 P.M. Hopkins et al. / Engineering Structures 132 (2017) 249–260

4.1. FEA model

ABAQUS was used for the numerical analysis and the dynamic
explicit solver was implemented. The same FEA properties that
were used in the static test panel as explained in Section 2 are
adopted. The differences between the static FEA model in Section 2
and the dynamic FEA model herein are that the static model uti-
lizes a ramped load approach to slow the applied load down to
an approximate static load.

4.2. Numerical analysis results

The blast load (dynamic) analysis was faster compared to the

Fig. 20. FEA comparisons with experimental test. explicit (static) analysis. The dynamic analysis took about 10 min
to finish, while the static analysis took about 7 h to finish, where
both analyses were run by the same computer.
Fig. 22 shows the damage to the solid and sandwich panel with
2.3 kg (5 lb m) TNT. The deflection versus time curve of the sand-
to study two panels: sandwich and solid panels as shown in Fig. 1
wich panel is compared to that of the solid reinforced concrete
and Fig. 2, respectively, which was studied in Section 2 under static
panel in Fig. 23, where the sandwich panel has more permanent
loading. The only difference is that a distance of 2743 mm was
deflection. Improvement can be made to the sandwich panel to
used between the two supports for the sandwich panels, instead
decrease this deflection, such as increasing the number of FRP
of 2438 mm in Section 2, so that the results can be better compared
shear connectors, providing external FRP plates to provide confin-
with the solid panel. Structures that are to resist blast loading need
ing effect, etc. Overall, the results are promising for the sandwich
to have energy absorption characteristics, maintain structural
panel, especially considering their 40% weight reduction and added
integrity, not collapse, and limit flying debris.
insulation properties.

(a) Experimental results (b) FE results-deflected shape (c) FE results-crack initiation

Fig. 21. FEA comparisons with experimental test.

(a) Solid Panel (b) Sandwich Panel

Fig. 22. Damage results to FRP sandwich panel, 2.3 kg (5 lb m) TNT.
 P.M. Hopkins et al. / Engineering Structures 132 (2017) 249–260 259

SOLID Panel

Sandwich Panel

Fig. 23. Deflection vs. time for sandwich and solid panel under 2.3 kg (5 lb m) TNT.

Fig. 24. Deflection versus time for 1000 solid concrete panel under blast loading.

deflected the most had the 2.3 kg (5 lb m) TNT blast load applied
to it and likewise the panel that deflected the least had the
0.45 kg (1 lb m) TNT blast load. The maximum mid-span deflection
of the 2.3 kg (5 lb m) TNT blast load was approximately 43 mm and
124 mm for the solid and sandwich panels, respectively.

6. Conclusions

The failure of a reinforced concrete panel, and especially an

insulated concrete sandwich panel with FRP shear connectors, is
a complicated and quasi-static mechanism with various failure
stages. The authors have performed FEA studies to show that the
full strength and deformation for a sandwich panel can be captured
with a dynamic explicit analysis solver using the Damaged Plastic-
ity material model in ABAQUS. Good correlation between the
Fig. 25. Deflection vs. time for the sandwich panel.
experimental results and the dynamic model can be achieved.
In addition to static loading, the dynamic explicit model can
also be used for blast loading. The panels showed good results
for blast attenuation and damage control when compared to con-
5. Parametric study ventional reinforced concrete panels, such as those recommended
in the US Army Corps of Engineers guidelines for blast resistant
A parametric study was conducted using the FE model devel- structures [14,15]. Based on the findings from this study, these
oped above by varying the amount of TNT. Figs. 24 and 25 show panels are promising for blast-resistant structure with respect to
four blast models and mid-span deflection responses for the solid energy absorption, limited deflection, and continuity of structural
and sandwich concrete panels, respectively. The panel that integrity and limitation of flying debris. The panels can be further
260 P.M. Hopkins et al. / Engineering Structures 132 (2017) 249–260

developed and tested as their economical design and energy effi- [8] Gilbert RI, Warner RF. Tension stiffening in reinforced concrete slabs. J Struct
Div Am Soc Civil Eng 1978;104:1885–900.
ciency characteristics allow for more advantageous construction
[9] Nayal R, Rasheed HA. Tension stiffening model for concrete beams reinforced
and should ultimately be included in design standards for future with steel and FRP bars. J Mater Civ Eng 2006;18(6):831–41.
blast resistant buildings as primary structural members. [10] Wahalathantri BL, Thambiratnam DP, Chan THT, Fawzia S. A material model for
The material properties for the concrete and the constituent flexural crack simulations in reinforced concrete elements using abaqus. In:
Proceedings of the first international conference on engineering, designing and
materials used in this study are the same for both the static and developing the built environment for sustainable wellbeing; 2011. p. 260–4.
dynamic analyses. High strain rate properties of the materials were [11] FMI-EPS, LLC, Post Falls, ID 83854.
not considered. While it does not affect the static analysis too [12] Lowes L. Finite element modeling of reinforced concrete beam-column bridge
connections. A dissertation submitted to the office of graduate studies of
much, high strain rate is more important for dynamic analysis. Fur- University of California, Berkeley, CA; 1995.
ther study is recommended on this topic, where dynamic (blast) [13] CRANE Composites Inc, 23525 W. Eames, Channahon, IL 60410.
tests can be conducted. FE models using materials with and with- [14] U.S. Army Corps of Engineers (1990) TM 5-1300, ‘‘Structures to Resist the
Effects of Accidental Explosions”, U.S. Army Corps of Engineers, Washington, D.
out high strain rate can be evaluated. Detailed parametric analysis, C. (also Navy NAVFAC P-397 or Air Force AFR 88-22).
including the amount and locations of TNT, different FRP shear [15] U.S. Army Corps of Engineers. Design of buildings to resist progressive
connectors, etc., will be conducted in a following study, which will collapse. U.S. Army Corps of Engineers, Washington, D.D. UFC 4-023-03; 2013.
[16] Thiagarajan G, Kadambi AV, Robert S, Johnson CF. Experimental and finite
be useful for designing blast resistant panels. element anlaysis of doubly reinforced concrete slabs subjected to blast loads.
Int J Impact Eng 2015;75:162–73.
[17] Vasudevan AK. Finite element analysis and experimental comparison of
Acknowledgements doubly reinforced concrete slabs subjected to blast loads. A thesis submitted to
the Faculty of The University of Missouri-Kansas City in partial fulfillment of
We gratefully acknowledge the Higher Education Research the requirements for the degree of Master of Science. University of Missouri-
Kansas City; 2013.
Council (HERC), Idaho State Board of Education for financial sup- [18] Hopkins PM. Non-linear finite element analysis of FRP-precast concrete
port; Missouri Structural Composites, LLC for technical support; sandwich panels (Doctoral dissertation.) University of Idaho; 2015.
and Crane Composites, Inc. for material donations. [19] PCI Committee on Pre-cast Sandwich Wall Panels. State-of-the-art of precast/
pre-stressed sandwich wall panels. PCI J 1997;42(2):92–134.
[20] Jun Lee B, Pessiki S. Experimental evaluation of precast, prestressed concrete,
References three-wythe sandwich wall panels. PCI J 2008;53(2):1–21.
[21] Pavese A, Bournas DA. Experimental assessment of the seismic performance of
a prefabricated concrete structural wall system. Eng Struct 2011;33
[1] Frankl Bernard A, Lucier Gregory W, Hassan Tarek K, Rizkalla Sami H. Behavior (6):2049–62.
of precast, prestressed concrete sandwich wall panels reinforced with CFRP
[22] Bai F, Davidson James S. Analysis of partially composite foam insulated
shear grid. PCI J 2011:42–54. concrete sandwich structures. Eng Struct 2015;91:127–209.
[2] Soriano J, Rizkalla S. Use of FRP grid for the composite action of concrete
[23] Benayoune A, Abdul Samad AA, et al. Flexural behaviour of pre-cast concrete
sandwich panels. In: Proceedings of the 11th international symposium on fiber sandwich composite panel – experimental and theoretical investigations.
reinforced polymer for reinforced concrete structures (FRPRCS11); 2013. Constr Build Mater 2008:580–92.
[3] Woltman GD, Tomlinson DG, Fam A. A comparative study of various FRP shear
[24] PCI Committee on Precast Sandwich Wall Panels. State of the art of precast/
connectors for sandwich concrete walls. Adv FRP Compos Civil Eng prestressed concrete sandwich wall panels. PCI Committee Rep
2011:237–40.
2011;2011:131–42.
[4] Norris TG. Bending behavior of insulated FRP-confined concrete sandwich [25] Einea Amin, Salmon David C, Fogarasi Gyula J, Culp Todd D, Tardos Maher K.
panels with FRP plate shear connectors. MS thesis, University of Idaho, State-of-the-art of precast concrete sandwich panels. PCI J 1991;1991:78–98
Moscow, ID, USA; 2014. [8].
[5] Chen A, Norris T, Hopkins P, Yossef M. Experimental investigation and finite [26] Xu Bin, Zou Degao, Kong Xianjing, Hu Zhiqiang, Zhou Yang. Dynamic damage
element analysis of flexural behavior of insulated concrete sandwich panels
evaluation on the slabs of the concrete faced rockfill dam with the plastic-
with FRP plate shear connectors. Eng Struct 2015;98:95–108. damage model. Comput Geotech 2015;65:258–65.
[6] Abaqus Inc, ABAQUS user manual, V6.11-1; 2011. [27] Chen A, Davalos JF. Strength evaluations of sinusoidal core for FRP sandwich
[7] Hsu LS, Hsu CT. Complete stress-strain behavior of high-strength concrete bridge deck panels. Compos Struct 2010;92:1561–73.
under compression. Mag Concr Res 1994;46(169):301–12.