Influence of application of sandwich panel on static and

dynamic behavior of ferry ro-ro ramp door

Tuswan Tuswan1, Achmad Zubaydi1*, Bambang Piscesa2, Abdi Ismail1, Rizky
Chandra Ariesta1
1 Department of Naval Architecture, Institut Teknologi Sepuluh Nopember
Jl. Teknik Kimia, Keputih, Sukolilo, Surabaya, Jawa Timur, Indonesia
Email: zubaydi@na.its.ac.id

2 Department of Civil Engineering, Institut Teknologi Sepuluh Nopember

Jl. Teknik Kimia, Keputih, Sukolilo, Surabaya, Jawa Timur, Indonesia

Abstract. All articles

Keywords: guidance, author, ICOMERA 2020 (Font Calibri, size 9, flush justify, line spacing: single,
before 0, after 30 pt, indentation: left 2.75 cm, right 0.2 cm)

1. Introduction
Sandwich panel has been extensively developed in engineering application fields such as
aerospace and ship structure (Ardhyananta et al., 2019).

Sandwich structures can be defined as a subset of multilayered composite structures, A typical

sandwich structure consists of the outer facings and the core embedded between them. The
facing-core interface is often the most vulnerable part of the sandwich structure. This interface is
often bonded. The choice of sandwich materials depends on the function of the structure, lifetime
loading,
availability and cost. The facings are built of stiff and strong materials and they are much thinner
than the light and relatively compliant core Victor Birman, George A. Kardomateas.

Low density and the capability to keep damage localized are two important characteristics of
marine composite materials. A typical sandwich structure is similar to an I-beam where the flanges
transport the bending and in-plane loads, while the web sustains transverse shear, redistributes
concentrated normal to the surface forces and defends the structure integrity. V. Birman, G. A.
Kardomateas.

Specifically, marine sandwich panels are prone to debonding because of several reasons such as
different adhesion properties of the interfaces layers, difficulties in controlling the proper bonding
during the manufacturing process (Chen et al., 2017; Fatt and Sirivolu, 2017), significant
distinctions between thickness and elastic moduli of the constitutive layers (Gaiotti and Rizzo,
2013; Tuswan et al., 2020). The presence of debonding can severely degrade the load-carrying
capacity (Bragagnolo et al., 2020) and the vibration properties of the sandwich (Savic-Barcan et al.,
2018).

This underlying principle of this method is to evaluate the modal parameters of the damaged
structures such as the natural frequencies (Yang et al., 2016; Zhao et al., 2016), mode shapes
(Kaveh and Zolghadr, 2015),

Reduction in the structural weight of ships will increase their cargo-carrying efficiency. This
increase in efficiency is obtained by either carrying more cargo with the same displacement or by
increasing the speed of the ship. As a result of this, the focus in the marine industry has been
shifted toward the structural designs and optimization of panels, either by means of modifying the
dimensions or utilizing alternative configurations for the panel structures. The number of panel
stiffeners still to be installed by the yard is reduced when applying sandwich panels because of the
high sectional modulus compared to a single plate panel. Production costs at a shipyard can be
reduced by applying sandwich panels instead of conventional stiffened panels due to the reduction
of panel stiffeners and therefore the reduction of brackets, cut-outs, collar plates and buckling
stiffeners. The number of panel stiffeners still to be installed by the yard is reduced when applying
sandwich panels because of the high sectional modulus compared to a single plate panel.
Production costs at a shipyard can be reduced by applying sandwich panels instead of
conventional stiffened panels due to the reduction of panel stiffeners and therefore the reduction
of brackets, cut-outs, collar plates and buckling stiffeners.

2. Modeling Description of application of sandwich panel

2.1 Application of sandwich panel on ramp door

The ramp door is operated to facilitate an efficient cargo flow and quick turnaround in the port. It
is usually assembled of a watertight axial ramp door with entering and bridging flaps. The
application of sandwich panel is focused on designing the safe, efficient, and economic structure,
while comprehensively reaching the design requirements. The objective is to optimize cargo
carrying capacity and reducing structural weight.
The reference model used is a ramp door of Ferry Ro-Ro 300 GT. The existing model is entirely
constructed by two stiffened steel plates connected by stiffeners. The modification is arranged by
replacing the top and bottom stiffened plates with the sandwich panel and eliminating the use of
stiffeners, as illustrated in figure 1. The modification is calculated based on DNV-GL rules of steel
sandwich panel construction to calculate the configuration of sandwich thickness and stiffeners
(DNV-GL, 2016). Strength index criteria are used to calculate the configuration of sandwich
thickness to meet the equivalent strength between the existing and modified models (DNV-GL,
2016).
The dimension of top and bottom stiffened plates of the existing ramp door is 8 m in length, 8.6
m in width. The thickness of top and bottom stiffened plates are 10 mm and 12 mm, respectively.
The longitudinal and transverse stiffeners are constructed by T profile with T 300x14 + 100x12 mm
in dimension. The modification is arranged by applying a sandwich panel in the top and bottom
sandwich, eliminating the use of longitudinal and transverse stiffeners, and the change of T profile
to I profile in the stiffeners. The sandwich panels use steel material as the faceplate and the
combination of unsaturated polyester resin (UPR) and clamshell powder as core material. Based
on DNV-GL strength index calculation, the proposed model variation and complete configuration
of the ramp door model are fully presented in table 1.
 Figure 1. Illustration of the application of sandwich panel on ramp door

Table 1. Model configuration of the ramp door

Top Bottom
Total of longitudinal
Models thickness thickness Total of transverse stiffener
stiffener
(mm) (mm)

Existing 10 12 11 (T 300x14 + 100x12 mm) 16 (T 300x14 + 100x12 mm)

Model A 5-15-5 5-15-5 11 (I 300x14 mm) 16 (I 300x14 mm)
Model B 5-15-3 5-15-3 11 (I 300x14 mm) 16 (I 300x14 mm)
Model C 5-15-3 5-15-3 12 (I 300x14 mm) 12 (I 300x14 mm)
Model D 5-15-3 5-15-3 9 (I 300x14 mm) 9 (I 300x14 mm)
Model E 5-15-3 5-15-3 6 (I 300x14 mm) 6 (I 300x14 mm)

25
24
23
23.06
22.99

22
21
W (tons)

20
20.63

19.76

19
18
18.11

17
16.47

16
15
Existing Model A Model B Model C Model D Model E
Model Variations
Figure 2. Weight estimation of different configuration models

In order to conduct preliminary analysis regarding the application of sandwich panel, rough
weight estimation is conducted to ensure that the specific sandwich application will lead to
substantial benefits for decision making to the shipyard as well as the shipowner. Weight
reduction should be significant to consider the unforeseen technical, practical, and financial
problems during the engineering and manufacturing of the sandwich panel. Weight comparison
between existing and modification model is illustrated in figure 2. Structure weight ( m) is the
relationship between the density (ρ) and how much space it takes up (V). From the diagram, it can
be reviewed that the application of the sandwich panel considerably decreased the weight.
However, the proposed model A results marginally increases weight by about 0.3%. Model B,
model C, and model D produce the weight-saving about 10.3%, 14.1%, and 21.2%, serially.
Moreover, proposed model E achieves the highest weight saving about 28.3%.
Compared with similar research regarding the application of sandwich panel with the same
material properties in different ship structure is stated. Sujiatanti et al. stated that the application
of the sandwich panel in Ro-Ro’s car deck comprehensively reduce the structural deck about 12%.
Another research (Tuswan, car deck senta) shows that the weight savings can be obtained in the
range of 8.87% - 11.6% in the car deck model. The application of sandwich panel in the side hull
structure is also stated in [Tuswan, material today), The weight saving can be reached up to 17%.
Compared to similar research applied in other ship types, Momčilovic and Motok [21] studied the
application of sandwich panel in general cargo barge offers a weight saving from 5 to 15%. In the
container barge is even less, about 4% - 13%. However, even though this application proposes
valuable benefits, the rough weight reduction, even greater than 50%, can be occasionally
discovered. In contrast, weight-saving for a composite sandwich can reach up to 70% that is higher
than steel sandwich application (Kortenoeven et al.) Further investigation should be performed to
obtain preliminary strength assessment based on loading conditions to analyze the comparison of
strength to weight ratio of whole configuration models.

2.2 Finite element development for structural strength analysis

The structural strength analyses are a substantially important aspect to investigate the
implementation of the sandwich panel. The finite element simulation is implemented to analyze
structural strength by comparing the von Mises (σv), normal stress (σ11), shear stress (τ), and
deformation (U) between existing and modification models. The objective of the application of the
sandwich panel on the ramp door model is to achieve the optimum configuration model, which
has the highest strength to weight ratio. In the finite element model discretization of the models,
the sandwich panel can be modeled using the layer-wise solid/shell element (Krueger and O’Brien,
2001). Based on Lloyd’s Register, 2020), top and bottom plates of the steel sandwich panel are to
be represented by shell elements, and the core material by solid elements with aspect ratio does
not exceed 1:4. In this analysis, the faceplate is modeled using the eight-node quadrilateral shell
element with reduced integration (SC8R). The core material is modeled using the eight-node
hexahedral element (3D8I). The longitudinal and transverse stiffeners are modeled using the eight-
node quadrilateral shell element with reduced integration (SC8R). Contact modeling between
parts of the structure is modeled by tie constraints based on surface-surface contact. Meanwhile,
the boundary condition should be organized in such a way that it could be the same as the real
condition. In this analysis, CFCF (clamped in after and front ramp door, and free in the side ramp
door) is applied.
Design load scenarios for strength calculation is calculated. The total static load is the sum of
local load represented vehicle load and uniformly distributed load from the weight of the ramp
door itself. The model is loaded with a uniformly distributed self-weight based on figure 2. The
vehicle load is the sum of chassis weight and the cargo. In this analysis, the vehicle has a single
wheel configuration in the front side and double-double axis in the rear side with the same load
print. Both front and rear load print areas are 1 x 0.3 m based on DNV-GL, 2015 calculation. Both
front and rear wheel load is 60 kN and 90 kN, respectively. The material used in modeling ramp
door is illustrated in table 2.

Table 2. Mechanical constant of the ramp door

Parameter Faceplate Core material
 Shear modulus (Pa) 8.077 × 1010 1.692 × 109
Young’s Modulus (Pa) 2.1 x 1011 4.4 x 109
Poisson’s ratio 0.30 0.30
Density (kg m-3) 7850 1465
Yield strength (MPa) 275 24.8
Tensile strength (MPa) 292 24.8

2.3 Finite element development for modal dynamic analysis

The numerical investigation of free vibration analysis is conducted to provide an understanding of
the free oscillation behavior of ramp door partially debonded at the skin-core adhesive layer. The
free vibration analyses are performed using a linear perturbation load step where Lanczos
iteration method for eigenvalues extraction is applied in the first ten-mode. The influence of
debonding presence, debonding geometry, debonding ratio, debonding location, debonding
thickness, and the number of debonding on the modal characteristics are presented in detail by
comparing the parameters between an initial and damaged ramp door model. Configuration
model C is used in this analysis. The finite element discretization and material properties are
explained in section 2.2. The study can be divided into three steps: (1) meshing of the model and
assembling stiffness and mass matrices of elements to model the equation of motion, (2)
extracting the eigenvalues and eigenvectors, and (3) performing post-processing to obtain the
required dynamic behavior of the model.
Several theories of debonding modeling are used. Debonding is located on the interface layer
between the faceplate and core material in the bottom sandwich panel. The debonding is modeled
as an artificial imperfection at the interface between two layers. The planar size of prescribed
debonding is defined by a damage parameter (D%) denoting the ratio of the debonded area (Ad) to
the entire interface layer area of the sandwich (Atotal).
Further, debonding is treated as an artificial imperfection at the interface layer and activated
before the oscillation. During the pre-processor, debonding is modeled by creating a small gap
between the faceplate and core material. In the present investigation, to model the debonding
problem, two different simplified modeling strategies will be investigated. In order to prevent the
probability of the two surfaces overlapping each other, a spring contact element modeling is
applied. In (ABAQUS, 2014), the spring element (SPRING2) is implemented to connect the node
between core and faceplate. The detailed spring element modeling and its constitutive modeling is
illustrated in figure 3. To model spring element behavior, as shown in Fig. 3b, the spring stiffness is
set to zero (k=0 N m-1) when in tension and is set to a high value (k= 210 x 109 N m-1) when in
compression and the displacement value between two nodes in the spring element is zero
(Burlayenko and Sadowski, 2009). Another modeling technique is assumed by removing the spring
contact element in the debonded region and assuming as avoid.

Figure 3. (a) Debonding modeling technique with spring element (b) the law of spring element
contact
3. Result and discussion

3.1 Mesh convergence study

First, mesh convergence study is an important issue that needs to be conducted in most of the
static analysis. The objective is to obtain a mesh that thoroughly balances accuracy and computing
time. The mesh convergence in the existing ramp door model with CFCF boundary condition is
analyzed. Material properties used in this analysis are stated in table 2. The mesh convergence
study is carried out by investigating nonlinear static analysis to obtain the relation between stress,
deflection, and mesh size. The step is by creating and analyzing a mesh using the fewest and
reasonable number of elements. Then, recreate the mesh with a denser element distribution, re-
analyze it. Keep increasing the mesh density and re-analyzing the model until the results converge
satisfactorily. Therefore, several mesh sizes between 0.1 m and 0.03 m are used to get the
optimum mesh sizes. The mesh element size 0.05 m with the number of elements and nodes
210,861 and 279,157, respectively, is chosen as a reliable result. Good agreement indicating
successful modeling of eigen vibration problem can be recognized. Therefore, it validates the
correctness of the model discretization used in this analysis. The method of establishing mesh
convergence required a curve of a critical parameter of von Mises stress and displacement to be
plotted against the number of mesh, as can be seen from figure 4.
50 0.00130
48
46 σ (MPa)
0.00125
44 U (m)
Mesh size: 0.05 m
Convergence Line

42
σv (MPa)

U (m)
40 0.00120
38
36
0.00115
34
32
30 0.00110
0 50000 100000 150000 200000 250000
Number of Element
Figure 4. Convergency analysis in the existing model

3.2 Numerical result of structural strength analysis

One of the problems with the application of the sandwich panel is its sensitivity to concentrated
load, which may cause damage. Therefore, the influence of the use of sandwich panel on
structural strength analyses is substantially vital as a thorough advance check to study the
implementation of the sandwich panel. In this section, the finite element results are presented to
thoroughly observe the comparison of structural strengths between existing ramp door and
modified models with different sandwich thickness and number of stiffener configurations. The
model configurations used in this analysis are stated in table 1. All the input of geometric
boundary conditions and material properties is similar to the models used in the mesh
convergence.
The result from nonlinear static analysis is shown in figure 5. Figure 5a presents in sequence the
comparison between von Mises stress and displacement of entire models. From the result, it
shows that the application of sandwich panel with reduction of stiffener mostly reduces the von
Mises stress and displacement, except in model E. Comparing between existing model and model
A which have similar weight, can reduce both von Mises stress up to 32% and displacement up to
35.6%. It can be seen from figure 5a that the reduction of stiffener for increasing weight-saving
causes the increase of von Mises and displacement of the models. So, only model E, which has the
 highest weight saving, experiences the higher maximum stress and displacement than the existing
model. Another result depicted in figure 5b. The application of sandwich material can reduce both
the normal stress and shear stress, as described in model A, model B, and model C. In contrast,
model D and model E experience higher normal and shear stress compared to the existing model.
A sandwich panel is similar to an I-beam where the basic idea is that the surfaces plates carry
loads in bending action while the core transfers the load in shear action. The reduction of stress by
application of sandwich panel is affected by the difference in thickness configuration between
stiffened steel plate and sandwich panel, and separation of the faceplate by a core material causes
to the significant increase in the sectional modulus and sectional area which can improve bending
stiffness. Its application will also remove the sources of stress concentrations, so that will decrease
the stress occurring in the model. The stiffness of the structure is dependent on the cross-section,
and the lower the stiffness is, the higher will the displacement occurring in the model.
Optimization of model configuration leads to the searched result, which best fits the
optimization target under applied constraints. Figure 6 presents the total reduction percentage
accumulated from the reduction average of total maximum stress and weight for different
modification models. The highest values among these are determined to be the optimal model
configuration. It can be decided that the optimal proposed model configuration for the application
of the sandwich panel in the ramp door is model C with a total reduction percentage of 19.24%, as
depicted in figure 6.
a 70 0.0030 b 70 25
σ U σ11 τ
60 0.0025 60

50 20
50
0.0020
σ11 (MPa)
σv (MPa)

40 40
U (m)

0.0015 15

τ (MPa)
30 30
0.0010
20 20
10
10 0.0005 10

0 0.0000 0 5
Existing Model A Model B Model C Model D Model E Existing Model A Model B Model C Model D Model E
Model Variations Model Variations

Figure 5. a) von Mises stress and displacement (U) between existing and modified model b)
Normal stress and shear stress between existing and varied models
35%
18.94%
19.24%
25% 12.42%
15.87%
Reduction percentage

15% 2.73%
32.04%

27.61%
10.28%

24.42%
14.06%

21.20%

28.35%
3.63%

5%
-0.29%

-22.89%

-5%

-15% Weight Reduction

Weight Reduction
Total reduction
-25%
Model A Model B Model C Model D Model E
Model variations
Figure 6. Reduction percentage between existing and modified models

3.3 Numerical result of modal dynamic analysis

The numerical results of the finite element analyses are reviewed to provide insight and
understanding of the free oscillation behavior of the ramp door model partially debonded at the
faceplate/core interface. The influence over a wide range of debonding ratio, geometries,
numbers, thicknesses, and boundary conditions have been presented to study the effects of the
parameters on the modal behavior of the model. Previously, the initial study is firstly reported to
give information on natural frequency differences of different configuration models.

3.3.1 Comparison of the dynamic behavior of different configuration models

This section discusses the dynamic response between the conventionally stiffened and four
modified models assembled sandwich panel. All the geometry, material properties, and boundary
condition are similar to the model used in mesh convergence analysis. Figure 7 shows the
comparison of eigenfrequency in the first ten-mode for different configuration models. From the
obtained results, all configuration models installed sandwich panel (model A – model E) have
lower natural frequency compared to the existing model in almost mode number. The natural
frequency differences between the model in the lower mode number are insignificant, but it can
be significant in the higher mode number. In summary, it can be concluded that all configuration
models installed sandwich panels have low stiffness than the existing model. It also can be seen
that the model which has the lowest stiffness is model E. The more the stiffener reduced, the more
the stiffness decreased.
a 85
Existing b Existing
75 Model A 150 Model A
Natural frequency (Hz)

Model B Model B
Natural frequency (Hz)

65 Model C 130 Model C

Model D Model D
55 Model E Model E
110
45 75
73 90
35 71
69 70
25 67
3.7 4 4.3
15 50
1 2 3 4 5 6 7 8 9 10
Mode number Mode number

Figure 7. Comparison of natural frequency between existing and modified model a) mode 1-5 b)
mode 6-10

3.3.2 Effect of debonding ratio

This section firstly analyzes the effect of the debonding ratio of model C in the first ten-mode. All
the material properties and the applied boundary condition are similar to the above convergence
study. The ramp door containing a square debonded region at the center with a damage ratio of
2.5%, 5%, 7.5%, and 10% are analyzed. The damage ratio is computed by taking the ratio of the
debonding area to the total surface area of the interface layer. In this case, debonding is assumed
to be located on the interface between the top faceplate and the upper part of the core material
of the bottom sandwich panel. The free vibration analysis using the Lanczos method is used. Both
of debonding modeling techniques: with spring element and without spring element, are
evaluated. Comparing natural frequencies between the existing and model with pre-existing
debonding, shifts in natural frequencies are usually discovered. Figure 8a presents the comparison
of natural frequency as a function of debonding ratio modeling with the spring element. It can be
seen that the natural frequencies predominantly decrease with the increase of the debonded
ratio, while the first three-mode is practically insensitive to the debonding presence. It is
noticeable that the shift on the higher modes is greater than the lower ones, as clearly shown in
figure 9a. Besides, the frequencies shift is more rapidly as mode number increases. Although the
trend of frequency change is violated due to local thickening phenomenon caused by debonding
Paolozzi, that for some lower mode numbers have the frequencies of the debonded model even
higher than the initial model, as can clearly be seen in the damage ratio of 2.5 % in figure 9a.
The frequency shifts of the model containing debonding increases due to a loss in stiffness caused
by initial debonding, and the mode shapes contain local deformation in the discontinuity area
(Tuswan, ICIMECE). Besides, from figure 9a, it obviously can also be seen that the presence of a
small debonding ratio (D < 5%) does not almost change the lower natural frequencies, only
decreases the higher natural frequencies. Another research also states this phenomenon in
[Burlayenko and Sadowski, 2010, the influence of skin/core debonding]. One of the interesting
findings in figure 9a is that the natural frequency of the model with 2.5% debonding ratio is slightly
higher compared to the intact model. The reduction of natural frequency is caused by the
reduction of the stiffness of the model. It can also be seen that the natural frequency shift, as
observed in figure 9, does not exhibit monotonous trends as the mode number increases. Hence, it
can be summarized that the debonding can influence natural frequencies and is mode dependent.
The assessment to determine the debonding presence in the model should be taken in high
modes.
To obtain more insight regarding debonding modeling technique, the model without spring
contact is also reviewed. Debonding causes a high natural frequency shift even in the lower mode.
Compared with debonding modeling with spring elements, the frequency shift without spring
contact modeling is more significant in all evaluated modes. Removing the contact may lead to
incorrect results, which significantly overestimate the result. Thereby, using the spring contact
model, the calculated responses of the model with debonding are dampened compared because
contact elements are used to prevent the detached models from overlapping each other.
Therefore, the model of the spring element is necessary to accurately represent the dynamic
response of the models (Burlayenko and Sadowski, 2011b).

a 140
Mode 1 b 140
Mode 1

Mode 2 120 Mode 2

120
Natural frequency (Hz)

Mode 3
Natural frequency (Hz)

Mode 3 100
100 Mode 4 Mode 4
80 Mode 5
Mode 5
80 Mode 6
Mode 6 60
Mode 7
60 Mode 7
40
Mode 8
Mode 8
20 Mode 9
40 Mode 9
Mode 10
Mode 10 0
20 0% 2.5% 5% 7.5% 10%
0% 2.5% 5% 7.5% 10% Damage parameter
Damage parameter

Figure 8. Comparison of natural frequencies as a function of damage parameter between a) with

spring element contact b) without spring element contact
a 58
b
98
88
48 2.5% 5% 7.5% 10% 2.5% 5% 7.5% 10%
Natural Frequency shift (Hz)
Natural Frequency shift (Hz)

78
68
38
58
28 48
38
18
28
18
8
8
-2 -2
1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10
Mode number Mode number
 Figure 9. Comparison of natural frequencies shift between a) with spring element modeling b)
without spring element modeling

3.3.3 Effect of debonding geometry

Comparative assessment of the modal characteristics of the ramp door model to the debonding
geometry is further examined. Four types of debonding geometry corresponding to the case of
interfacial damage such as circular, square, through-the-length, and through-the-width, are
previously investigated in (Tuswan ICIMECE). Debonding geometry is wholly illustrated in (Tuswan
ICIMECE). In a previous analysis (Tuswan ICIMECE), the effect of debonding geometry is analyzed
in 5% debonding ratio. Hence, further study needs to be conducted to compare the influence of
debonding geometry on the natural frequency using different damage ratio. In this section, the
effect of the equally sized debonding geometry is evaluated by both 5% and 10% debonding ratio.
All the model geometries, material properties, damage location, and the applied boundary
conditions are similar to the model in section 3.3.2. the natural frequency in the first ten-mode of
both with and without spring contact conditions, as previously stated, are compared.
The comparison of natural frequencies as a function of debonding geometry with different
debonding ratios modeled by spring element modeling and without spring element modeling is
depicted in figure 10 and figure 11. Figure 10 presents the effect of the debonding ratio, which has
the same ratio of damage (5% debonding ratio). As observed in figure 10, interfacial debonding
induces a decrease in natural frequency, especially in the square and circular debonding
geometries. In contrast, there is no significant natural frequency decrease in the through-the-
length, and through-the-width debonding geometry. In the lower mode, the natural frequency is
altered by small damage but not as significant as in higher modes. Although, it can be seen from
figure 10 that the effect of debonding geometry generally has no regular trend on the natural
frequency shift. Several researchers also mention that debonding detection in the sandwich
structure in small damage is more sensitive in higher modes (Burlayenko, 2011, Ismail et al., 2020,
Tuswan OPEN ENGINEERING, TUSWAN ICIMECE). Debonding decreases the shear stress transfer
area between the faceplate and core, which can reduce the stiffness and strength of the model.
Moreover, figure 11 interprets the effect of debonding geometry modeled by same 10%
debonding ratio on free vibration response. The influence of debonding geometry with a higher
debonding ratio on frequency shift is more significant, as compared between figure 10a and figure
11a. As observed in figure 11a, interfacial debonding reduces the natural frequency even in the
lower mode, especially in the square and circular debonding geometries. In contrast, there is no
significant frequency shift in the through-the-width debonding geometry. It also can be reviewed
that there is a frequency shift in high mode in through-the-width geometry. This shift does not
occur in the variation of debonding geometry with damage ratio 5%.
Compared with debonding modeling with spring element contact, the frequency shift without
spring contact modeling is more significant. Figure 10b and 11b show the comparison of natural
frequency without spring contact modeling. Debonding causes a decrease in natural frequency
even in the lower mode, especially in the square and circular debonding shape. However, it seems
there is an insignificant natural frequency decrease in the through-the-length, and through-the-
width debonding. This phenomenon also occurs in modeling with a spring element. It proves that
the localized debonding shape is more sensitive to the natural frequency changes.
Removing the contact may lead to incorrect results, which significantly overestimate the value of
natural frequency. Thereby, using the spring contact model, the calculated responses of the model
with debonding are dampened compared to the model without contact. As summarized, the
resulting study of the effect of debonding ratio and debonding geometry between two different
modeling strategies results in the same phenomena. Debonding modeling without spring element
produces higher natural frequency than modeling with spring element contact. Therefore, in the
 next section, analyzing the debonding behavior is only modeled by inserting the spring element in
the debonding area.
As can be summarized, the frequency shifts for all analyzed geometry of damage increase
between two debonding modeling strategies with the increase of the damage ratio. In the other
results given by (Burlayenko, Numerical modeling EURODYN 2011) (Burlayenko, Influence of
skin/core debonding on free vibration) the circular debonding geometry is more sensitive than
other damage geometry to the free oscillation behavior.
Figure 10. Comparison of natural frequencies as a function of debonding
a 140 b 140
Mode 1 Mode 1

120 Mode 2 120 Mode 2

Natural frequency (Hz)

Natural frequency (Hz)

Mode 3 Mode 3
100
100 Mode 4 Mode 4

Mode 5 80 Mode 5
80
Mode 6 60 Mode 6

60 Mode 7 Mode 7
40
Mode 8 Mode 8
40 Mode 9 20 Mode 9

Mode 10 Mode 10
20 0
initial Through-the- Through-the- Square Circular initial Through-the- Through-the- Square Circular
length width length width
Debonding geometry Debonding geometry

geometry with 5% debonding ratio between a) with spring element modeling b) without spring
element modeling
a 140 140
Mode 1 b Mode 1

120 Mode 2 120 Mode 2

Mode 3 Mode 3
Natural frequency (Hz)

Natural frequency (Hz)

100
100 Mode 4 Mode 4
Mode 5 80 Mode 5
80
Mode 6 Mode 6
60
60 Mode 7 Mode 7
40
Mode 8 Mode 8
40 Mode 9 20 Mode 9
Mode 10 Mode 10
20 0
initial Through-the- Through-the- Square Circular initial Through-the- Through-the- Square Circular
length width length width
Debonding geometry Debonding geometry

Figure 11. Comparison of natural frequencies as a function of debonding geometry with 10%
debonding ratio between a) with spring element modeling b) without spring element modeling

3.3.4 Effect of number of debonding

In this section, the effect of the number of debonding is further examined by using the FE model of
the ramp door with spring element contact. The effect of the number of debonding on the modal
parameters in the first ten-mode is discussed. All the model geometries, material properties,
damage location, and the applied boundary conditions are similar to the model used in section
3.3.2. The comparison of natural frequency between single debonding and multiple debonding
with same damage ratio is compared. Firstly, circle debonding is modeled as a single debonding
with 10% damage ratio. Secondly, debonding is modeled by two and three equally sized circular
debonding. The damage inflicted by these two circular debonding zones is of 10% of the total area.
The two and three debonding zones are symmetrically located with respect to the longitudinal
midlines of the sandwich panel. As seen from figure 12, the frequencies are more affected by the
single debonding than the equally sized two and three debonding zones with same debonding
ratio. Model containing single debonding results lower natural frequency than model containing
 multiple debonding. It can be analyzed that the higher the mode number increased, the more the
frequency shift increased. However, there is no significant frequency shift in lower mode, as
shown in mode number 1 and mode number 2, as depicted in figure 12. The natural frequencies
are the most affected by debonding zones located along the longitudinal midline of the sandwich
plate, as also stated in (Burlayenko 2011b). Further research is needed to be addressed with
various debonding ratio and location. As mentioned in (Burlayenko, 2011b), using only the natural
frequency as a damage parameter is impossible to foresee the location of multiple debonding. To
overcome this issue, the associated mode shapes can provide more reliable information. By
contour changes in the mode shapes, which are different from those for the model using single
debonding, multiple debonding can be detected. Therefore, further analyses need to be addressed
with various damage location to analyze the location of multiple debonding.
a 85
b 140
Initial Initial
1 circle 130
75 1 circle
2 circle 120 2 circle
Natural frequency (Hz)

Natural frequency (Hz)

3 circle 3 circle
65 110
100
55 90
80
45
70

35 60
50
25 40
1 2 3 4 5 6 7 8 9 10
Mode number Mode number

Figure 12. Natural frequencies of intact and debonding models with different number of
debonding (a) mode number 1- (b) mode number 6-10

3.3.5 Effect of debonding depth

To study the effect of debonding depth on the free vibration response, three circulars debonding
containing different debonding depth will be analyzed. Damage is modeled by circular debonding
located in the bottom sandwich panel with 10% damage ratio. Spring element contact is applied to
model debonding. All the model geometries, material properties, damage location, and the
applied boundary conditions are similar to the model used in section 3.3.2. Debonding depth is
modeled by creating a small gap in the interface layer. The small gap is modeled by creating 10%
of core thickness, 20% of core thickness, and 30% of core thickness.
The finite element predictions for the first ten-modes associated with the corresponding
debonding depth are compared in figure 13. It can be recognized from figure 13 that increasing
debonding depth will decrease the natural frequency. Compared with the initial model, the effect
of debonding becomes more visible with the increase in the debonding depth. One can see that
the model with debonding depth 30% core thickness has the lowest natural frequency compared
to other models. Moreover, the frequencies change more rapidly as mode number increases while
the first bending mode practically has no significant effect on the debonding presence. The
frequency change of the debonded model increases due to a loss in stiffness caused by initial
debonding, and the mode shapes contain local deformation in the debonded area.
a 85
Initial b 170
Initial
10% Core Thickness 10% Core Thickness
75 150
20% Core Thickness 20% Core Thickness
Natural frequency (Hz)

Natural frequency (Hz)

30% Core Thickness 30% Core Thickness

65 130

110
55
90
45
70
35
50

25 30
1 2 3 4 5 6 7 8 9 10
Mode number Mode number
Figure 13. Natural frequencies of intact and debonding models with different debonding thickness
(a) mode number 1- (b) mode number 6-10

3.3.6 Effect of boundary condition

Next, It is crucial to examine the influence of applied boundary conditions on the modal
parameters due to the existence of localized damage. The influence of boundary conditions on the
free vibration of ramp door model damaged by debonding is studied by comparing the natural
frequency shift for four different types of the boundary condition which represent a loading that
may occur on the ramp door, such as CCCC (clamped in all side), CFFF (clamped in the rear side,
free in other sides), CFCF (clamped in front and rear side and free in other sides), and SFSF
(supported in front and rear side and free in other sides). `F' is denoted the free edge, `S' and `C'
represent the simply supported and clamped edges, serially. Damage is modeled by circular
debonding located in the bottom sandwich panel with 10% damage ratio. Spring element contact
is applied to model debonding. All the model geometries, material properties, damage location,
and the applied boundary conditions are similar to the model used in section 3.3.2.
Figure 14 displays the comparison of frequency shift between initial and debonded models with
different boundary conditions. The highest natural frequency shift is achieved by fully clamped in
all sides (CCCC). This trend remains the same for all ten eigenfrequencies. Other result
investigations are also mentioned by (Lou J & Lou L.). It is well-known that a fully clamped
boundary condition in all sides produces higher stiffness compared to others with the same
geometry parameters, as also mentioned by Arunkumar MP. Moreover, the CFCF and SFSF
boundary conditions have a similar frequency shift. In contrast, CFFF boundary condition has the
lowest natural frequency shift among other boundary conditions.
Further study needs to be handled to obtain the influence of different boundary conditions
analyzed on various debonding ratios. Thereby, the existence of debonding within the sandwich
panel can be detected by imposing of the corresponding boundary conditions more efficiently,
when the damage ratio is not very small (D% > 5%), as recommended by Burlayenko Influence of
skin/core debonding

140
CFFF CFCF SFSF CCCC
120
Natural Frequency shift (Hz)

100

80

60

40

20

0
1 2 3 4 5 6 7 8 9 10
Mode number
Figure 14. Natural frequencies of intact and debonding shift with different boundary conditions

3.3.7 Effect of debonding location

The final part, The investigated sandwich plates with different damage locations can cause changes
in natural frequency, even for the same damage ratios. It can be concluded that the farther the
location of the damage from the clamped edge, the lower the natural frequency deviation. The
results are confirmed by other studies [24], which states that the farther away from the clamped
edge, the effect of damage on the natural frequency becomes smaller. Thus, the natural frequency
deviation parameter can be used as a reference to detect the location of damage in the ship sandwich
plate.

a 40
Initial b
70
Initial
35 0.25 L 65 0.25 L
0.5 L 0.5 L
Natural frequency (Hz)

Natural frequency (Hz)

30 0.75 L 60 0.75 L
25 55
20 50

15 45

10 40

5 35

0 30
1 2 3 4 5 6 7 8 9 10
Mode number Mode number

Figure 15. Natural frequencies of intact and debonding models with different debonding location
a) mode number 1-5 b) mode number 6-10

4. Conclusions

5. Acknowledgement
This research is fully funded by Directorate of Research and Community Services, Ministry of
Research, Technology and Higher Education, The Republic of Indonesia under The Master’s Degree
Program Leading to Doctoral Degree for Excellent Bachelor Graduates (PMDSU) research scheme.

6. Reference

[1] Victor Birman, George A. Kardomateas, Review of current trends in research and applications
of sandwich structures, Composites Part B: Engineering, Volume 142, 2018, Pages 221-240,
[2] S. Sujiatanti, A. Zubaydi and A. Budipriyanto, "Finite Element Analysis of Ship Deck Sandwich
Panel," Applied Mechanics and Materials, vol. 874, pp. 134-139, 2018
[3] Tuswan, K Abdullah, A Zubaydi, A Budipriyanto 2019 Finite-element Analysis for Structural
Strength Assessment of Marine Sandwich Material on Ship Side-shell Structure. Materials
Today: Proceedings 13 (2019) 109–114
[4] Tuswan, Achmad Zubaydi, Agung Budipriyanto, Septia Hardy Sujiatanti. Comparative study
on ferry ro-ro’s car deck structural strength by means of application of sandwich materials.
Proceedings of The 3rd International Conference on Marine Technology (SENTA) 2018
Research Center Building 11F, Institut Teknologi Sepuluh Nopember (ITS) Surabaya,
Indonesia, December 5 – 6, 2018.
[5] N. Momčilović and M. Motok, “Estimation of ship lightweight reduction by means of
application of sandwich plate system,” FME Transactions, vol. 37, pp. 123-128, 2009.
[6] J. Kortenoeven, B. Boon and A. De Bruijn, “Application of Sandwich Panels in Design and
Building of Dredging Ships,” Journal of Ship Production, pp. 125-134, 2008.
[7] V. Birman, G. A. Kardomateas, Review of current trends in research and applications of
sandwich structures. Composites Part B: Engineering. 142 (2018) 221-240.
[8] Krueger, R., O’Brien, T.K., 2001. A shell/3D modeling technique for the analysis of
delaminated composite laminates. Composites Part A: Applied Science and Manufacturing
32(1), 25-44.
[9] DNV-GL: Steel sandwich panel construction, DNV-GL, Norway, 2016
[10] DNV-GL, Rules For Classification High speed and light craft Part 5 Ship types Chapter 2 Car
ferry,
[11] Norway: DNVGL, 2015
[12] Tuswan, Zubaydi, A., Piscesa, B., Ismail, A., 2020. Dynamic characteristic of partially
debonded sandwich of ferry ro-ro’s car deck: a numerical modeling. Open Engineering 10(1),
424-433.
[13] Burlayenko, V.N., Sadowski, T., 2010. Influence of skin/core debonding on free vibration
behavior of foam and honeycomb cored sandwich plates. International Journal Non-Linear
Mechanics 45(10), 959-968.
[14] Burlayenko, V.N., Sadowski, T., 2011b. Numerical Modeling of Sandwich Plates with
Partially Dedonded Skin-to-Core Interface for Damage Detection, The 8th International
Conference on Structural Dynamics. Leuven, Belgium.
[15] Paolozzi A, Peroni I. Detection of debonding damage in a composite plate through natural
frequency variations. J. Reinf. Plast. Compos. 1990;9(4):369-88. doi:10.1177/073168
[16] 449000900405
[17] Waldron K, Ghoshal A, Schulz MJ, Sundaresan MJ, Ferguson F, Pai PF, et al. Damage
detection using finite element and laser operational deflection shapes. Finite Elem Anal Des.
2002;38:193–226.
[18] Tuswan Tuswan, Achmad Zubaydi, , Bambang Piscesa, Abdi Ismail, Muhammad Fathi Ilham,
2020, Free vibration analysis of interfacial debonded sandwich of ferry ro-ro’s stern ramp
door, 6th International Conference on Industrial, Mechanical, Electrical and Chemical
Engineering (ICIMECE 2020) Surakarta
[19] Lou J, Wu L, Ma L, Xiong J, Wang B. Effects of local damage on vibration characteristics of
composite pyramidal truss core sandwich structure. Compos Part B-Eng. 2014;62:73–87
[20] Lu L, Le J, Song H, Wang Y, Huang C. Damage detection of sandwich panels with truss core
based on time domain dynamic responses. Compos Struct. 2019;211:443-454
[21] Arunkumar MP, Jagadeesh M, Pitchaimani J, Gangadharan KV, Babu MCL. Sound radiation
and transmission loss characteristics of a honeycomb sandwich panel with composite facings:
effect of inherent material damping. J Sound Vib. 2016;383:221–232
[22] Bragagnolo, G., Crocombe, A.D., Ogin, S.L., Mohagheghian, I., Sordon, A., Meeks, G.,
Santoni, C., 2020. Investigation of skin-core debonding in sandwich structures with foam
cores. Materials & Design 186, 2020, 1-10.
[23] Chen, Y., Hou, S., Fu, K., Han, X., Ye, L., 2017. Low-velocity impact response of composite
sandwich structures: modelling and experiment. Composite Structures 168, 322–334.
[24] Fatt, M.S.H., Sirivolu, D., 2017. Marine composite sandwich plates under air and water
blasts. Marine Structures 56, 163–185.
[25] Gaiotti, M., Rizzo, C.M., 2013. Finite element modeling strategies for sandwich composite
laminates under compressive loading. Ocean Engineering 63, 44-51
[26] Savin-Barcan, M., Beznea, E.F., Chirica, I., 2018. Influence of fabrication imperfections on
dynamic response of a sandwich composite panel of a ship deck structure. IOP Conference
Series: Materials Science and Engineering 400, 1-6.
[27] Kaveh, A., Zolghadr, A., 2015. An improved CSS for damage detection of truss structures
using changes in natural frequencies and mode shapes. Advances in Engineering Software 80,
93–100.
[28] Yang, C., Hou, X.B., Wang, L., Zhang, X.H., 2016. Applications of different criteria in
structural damage identification based on natural frequency and static displacement. Science
China Technological Sciences 59, 1746–1758.
[29] Zhao, B., Xu, Z., Kan, X., Zhong, J., Guo, T., 2016. Structural damage detection by using
single natural frequency and the corresponding mode shape. Shock and Vibration 2016, 1-8.
[30] Ardhyananta, H., Sari, E.N., Wicaksono, S.T., Ismail, H., Tuswan, Ismail, A., 2019.
Characterization of Vinyl Ester Bio-resin for Core Material Sandwich Panel Construction of
Ship Structure Application: Effect of Palm Oil and Sesame Oil. AIP Conference Proceedings
2202, 1-5.