Curved and Layered Structures 2024; 11: 20240008

Review Article

Muhamad Imam Firdaus* and Ristiyanto Adiputra*

Deterioration and imperfection of the ship

structural components and its effects on the
structural integrity: A review
https://doi.org/10.1515/cls-2024-0008 structural strength, contributing to the sustainability and
received December 31, 2023; accepted April 11, 2024 effectiveness of vessel design in the global maritime
Abstract: The design of ship structural safety is crucial to industry.
ensure the ship’s survivability during the operation. Extensive Keywords: stiffened plate, stiffened panel, box girder, ulti-
research has been conducted on ship structural components, mate strength, structural imperfection
including box girders, stiffened panels, and plates, beyond the
ideal conditions by considering the implication of manufac-
turing processes, vessel usage, and aging in the form of defects
like cracks, corrosion, and imperfections, both locally and Nomenclature
globally. Previous research has also explored various meth-
odologies, conditions, and parameters to understand the 2c/b Ratio between crack length and section width
impact of damages and imperfections on ship structure A0i Amplitude of hungry horse imperfection mode
and strength. However, there is a significant need to bridge A1 , Aj Amplitude of ARE imperfection mode
the gap in prior research to advance technology and ship AB The total sectional area of the outer bottom
structural strength analysis. A comprehensive benchmark AB′ The total sectional area in the inner bottom
study specifically focused on improving ship structural AD The total sectional area of the deck
component needs, identifying differences and gaps among Aj + 1 High-order amplitude of ARE imperfection mode
existing studies as challenging. This article thoroughly As The total sectional area of the side shell
reviews ship structural components, such as box girders, a Span length
stiffened panels, and plates, while examining the effects of B0 Amplitude of torsional and column imperfec-
structural defects like corrosion, cracks, and imperfections tion mode
on ship structural integrity. It synthesizes the influence of b Plate width
various defect parameters, including crack length, angle, b1 The breadth of the longitudinal tensile residual
position, corrosion severity, pit corrosion, pit diameter, stress block
and pit models, using finite element modeling and experi- C0 Amplitude of local imperfection mode
mental investigations, particularly emphasizing ship struc- D The hull depth
tural components. The comparative analysis of methods and E Modulus young
parameters presented in this review will serve as a valuable g The neutral axis position above the baseline
reference for future investigations and studies related to H The depth of the hull section of the linear elastic
ship structural strength and design. The article’s contribu- state
tion is expected to enhance the understanding of ship hs The maximum height of the web
j Half-wave number of the buckling mode
Lw Weld leg length
 Mu Ultimate bending moment
* Corresponding author: Muhamad Imam Firdaus, Surabaya m Half-wave number of the buckling mode
Merchant Marine Polytechnic, Surabaya, 12760, Indonesia, Rf Ratio between cracked ultimate strength of box
e-mail: imam.fi[email protected]
girder and ultimate strength of box girder in
* Corresponding author: Ristiyanto Adiputra, Research Center
for Hydrodynamics Technology, National Research and Innovation
initial condition
Agency (BRIN), Surabaya, 60112, Indonesia, t Plate thickness
e-mail: [email protected] v Ratio Poisson

Open Access. © 2024 the author(s), published by De Gruyter. This work is licensed under the Creative Commons Attribution 4.0 International License.
2  Muhamad Imam Firdaus and Ristiyanto Adiputra

Wc Column imperfection mode transporting raw materials and passenger ships such as
Wl Local plate imperfection mode Ro–Ro ships are susceptible to structural failure due to
Wt Torsional imperfection mode grounding, collisions with other vessels, or fatigue resulting
wopl Plate distortion from continuous loading over a certain period [3–7]. Numerous
max
wopl Maximum magnitude of plate distortion accidents involving these vessels have resulted in environ-
x Coordinate mental damage, as highlighted by Brubaker [8], who estimated
xi Coordinate that between 600,000 and 1.8 million tons of oil spills routinely
y Coordinate pollute the seas each year. This unfortunate situation is una-
yi Coordinate voidable due to the recurring accidents that plague these ves-
z Elastic section modulus sels annually.
zi Coordinate Ship accidents resulting from structural failures were
β Plate slenderness quite common, with records showing significant incidents
ε Strain involving bulk carriers from the 1970s to the 1990s [9,10].
εy Yield strain The number of such accidents has continued to rise, as
εrcx Compressive residual strain indicated by a study conducted by Eliopoulou, which docu-
εY Compressive yield strain mented a total of 146 accidents caused by structural failure
η Constant involving various types of oil tankers, ranging from VLCCs
θ Angle to ultra-large crude carriers, from 1990 to 2020. It is worth
λ Column slenderness emphasizing that ship accidents caused by structural failure
λτ The ratio between the shear and longitudinal in- have historically contributed significantly to maritime acci-
plane stresses dents, even in the past two decades. As key players in the
σ Strength global economy, addressing this issue is essential, consid-
σ0 The magnitude of residual compression stress ering its far-reaching impacts, including loss of life, property
σcr Critical bending stress damage, and broader concerns such as marine and coastal
σE Bending stress environmental pollution [11].
σr The magnitude of residual tensile stress Through the common structural rules (CSR), safety and
σu Ultimate strength regulations regarding ship design are governed by the ulti-
σuD The ultimate buckling strength for deck mate limit state (ULS) parameter. Various approaches have
σuS The ultimate buckling strength for side been conducted to enhance the ship strength, including
σx Longitudinal stress experimental, analytical, and numerical studies. In 2020,
σxu Ultimate strength of corroded plates The International Association of Classification Societies
σxu0 Ultimate strength of initial plates (IACS) released IACS-CSR in Part III.1, which focuses on ulti-
σy Yield stress mate strength, compiles findings from researchers, and inves-
σy1 Transversal stress tigates recent developments in ship research. The widespread
σyB The yield strength outer bottom technological advancements have enabled us to analyze the
σyB
′ The yield strength Inner bottom ship strength not only through experimental methods but
σyS The yield strength side also through numerical methods using various commercial
ϕ Edge function software. The utilization of these methods has led to innova-
tions such as the development of formulas for calculating
strength, as demonstrated in the studies conducted by Kim
et al. [12] and Xu et al. [13], who used plate slenderness func-
1 Introduction tions to determine the ultimate strength of stiffened panels.
This extends beyond stiffened panels, as research on simpler
The safety and structural integrity of ships are paramount in structures like plates and more complex ones like box girders
ensuring secure transportation. Modes of maritime trans- has also been extensively conducted.
portation such as bulk carriers, oil tankers, and very large The development of ship safety and structural strength
crude carriers (VLCCs) require special attention due to their is achieved by incorporating the aspect of initial imperfec-
crucial role in distributing raw materials. Bulk carriers tion. In actual conditions, initial imperfections occur due to
transport commodities such as cement, grains, coal, and factors such as the influence of fabrication processes like
oil, while oil tankers play a vital role in the sea-based dis- welding, damage from usage, and age-related wear and
tribution of crude oil products [1,2]. Both bulk carrier vessels tear. Effects such as residual stress in the structure due
 Deterioration and imperfection of the ship structural components  3

to welding processes are addressed by reducing the yield elements typically form a larger amalgamation known as
stress value or decreasing the plate and stiffener thickness stiffened plates, any failure occurring within the stiffened
due to corrosion effects. These improvisations have been plate will only happen when all plates experience failure
integrated into ship strength analysis. Investigations and simultaneously [19]. Therefore, the upper limit of failure
studies conducted on cases involving the introduction of due to plates is the summation of the maximum load-bearing
initial imperfections have yielded accurate results that capacity of all plate elements while also taking into account
match real-world conditions. Therefore, this consideration the role of stiffeners.
is being evaluated for adoption as a standard in the IACS-
CSR with adjustment through a correction factor.
The advent of technological advancements, such as
numerical analysis, provides a more efficient solution for 2.1 Ultimate strength of stiffened plate
predicting ship strength. However, there is a need for vali-
dation between modeling parameters and experimental The design of ship structures can be considered as rectan-
analysis to ensure that the results align with real-world gular boxes enveloped by stiffened panels. A clearer depic-
conditions. The development of numerical analysis has tion of the ship’s structure can be observed in Figure 1.
given rise to various approaches to defining simulation para- Subsequent research in ship structure design emphasizes
meters, prompting [14] to compare different types of boundary buckling conditions and the ultimate strength of the ship’s
conditions, namely, symmetric and periodic boundary condi- structure. The ultimate strength of a structure or system is
tions. The findings revealed a significant deviation between the defined as a point where the incremental addition of
two under certain conditions and minor deviations under spe- applied loads cannot be sustained or supported by the
cific circumstances. Another parametric study was conducted structure so that when it surpasses the ultimate point,
by Hanif et al. [15], focusing on the configuration area and the the structure experiences failure and collapse. An illustra-
modeling of transverse stiffeners. Adiputra et al. [16] utilized tion of ultimate strength is presented in Figure 2.
numerical calculations to predict ship operational times under Buckling and the ultimate strength of plates them-
hogging and sagging conditions, considering various factors selves are fundamentally crucial for the strength of a
such as geometry and initial imperfections. More recently, in ship, both locally and globally. In this regard, the plate
larger ship structures, Denev [17] conducted research and slenderness ratio is a critical parameter. Plate slenderness
found that waste corrosion can affect the section modulus can be defined by the following equation:
and ultimate strength.
β = (b / t ) σ y / E , (1)
In this present article, a review of significant literature
related to the research advancements in ship structural
strength through various methods, including experiments,
analytics, and numerical analyses using various software,
was conducted. Various influences and factors, such as the
type, size, and distribution of initial imperfections on
plates, stiffened panels, and box girders are also discussed
to understand the postbuckling behavior of each variation.
In addition, differences in approaches to numerical studies
are explored to provide insights into future research trends
in this field.

2 Plate
The strength of a ship’s plate structure significantly influ-
ences the overall strength of the ship’s structure. A ship’s
structure can be regarded as a combination of reinforced
plates seamlessly joined with approximately equal spacing
between each other [18]. Plates themselves constitute the pri- Figure 1: Illustration of ship structure produced based on the informa-
mary structural components in ship structures. Since plate tion obtained from the study by Zhang [20].
4  Muhamad Imam Firdaus and Ristiyanto Adiputra

inner bottom and inner longitudinal bulkhead can experi-

ence pressure loads from cargo loads. This lateral pressure,
whether directly applied or through hull frame bending,
generates transverse loads within the plate plane.
Due to the crucial and intricate nature of ultimate strength
and buckling in plate structures, this field has been studied and
researched for over 100 years. Further research into the ana-
lysis of ultimate strength in the plate structures is primarily
conducted using axial compression loads, with various
enhancements and updates in terms of analysis methods
and geometric variations. In the axial loading analysis, the
Figure 2: Ultimate strength point illustration. Produced based on the
model considered is a simple, longitudinally supported plate
information obtained from the study by Zhang [20]. subjected to axial forces, as depicted in Figure 3.
In the initial loading stage, the plate experiences axial
compressive forces and behaves linearly. Subsequently, as
where σy represents the yield stress, E is the Young’s Modulus additional loads are applied, lateral deflection increases,
of the material employed, and t denotes the plate thickness. and simultaneously, minor yielding occurs on the long side
Further research on the structural design of modern of the plate. This results in the load-shortening curve
oil tankers and bulk carriers with lengths ranging from 150 becoming nonlinear. Structural failure occurs when the
to 400 m indicates that the ratio of plating, denoted as a/b, ultimate state is reached, followed by significant lateral
between stiffeners and transverse frame ranges from 3.5 to deflection and failure near the long side [20]. Examples
6.7 in both the deck and bottom sections of the mid-ship. of load-shortening curves for plates with thicknesses of
The ratio of width to thickness of plating, represented as 32, 18, and 12 mm can be seen in Figure 4.
b/t, falls within the range of 25–65. The plate thickness The classic equation regarding elastic buckling stress
ranges from 12 to 26 mm. The area of the plate itself for plates can be observed in Eq. (2):
accounts for approximately 60–80% of the cross-sectional
area of the stiffened panel [20]. σE π 2E ⎛ b ⎞2 3.62
= ≈ . (2)
In ship plate structures, the most common loading sce- σy 3(1 − v 2 ) ⎝ t ⎠ β2
nario typically involves a combination of longitudinal
However, using the aforementioned equation and the
loads, transverse loads, shear loads, and lateral loads
compression analysis method explained earlier, accurate
[18]. Several studies have conducted strength analyses of
results cannot be obtained for certain plate thicknesses.
ship plates under both combined and single loads, with the
For instance, when β = 1, the buckling stress will be 3.6
primary loading component on the deck structure and the
times the yield stress, resulting in inaccurate outcomes.
bottom structure being axial compressive loads. Therefore,
Therefore, an equation based on empirical corrections
in standard design analysis of the entire hull girder, the
for critical buckling at specific plate thicknesses is subse-
main loading component considered is the longitudinal
quently adopted, as depicted in Eq. (3):
load. However, the bottom plating and the lower portion
of the side shell can also be subjected to relatively high σcr 1 σy σy 1
= 1− , for > . (3)
additional loads from lateral pressure. Furthermore, the σy 4 σE σE 2

Figure 3: Example of loading set up on plate analysis [21].

 Deterioration and imperfection of the ship structural components  5

Figure 4: Examples of load-shortening curves at different plate thicknesses illustrated based on the information obtained from the study by
Zhang [20].

In the evolution of scientific knowledge in the field of σu 2 1

= − 2. (8)
plate strength analysis, the initial research on ultimate σy β β
strength and the formulation of plate strength equations
Widely adopted formulas for calculating the ultimate
were conducted by Box in 1883, followed by further studies
strength of plates under axial compression, such as the
carried out by von Karman in 1924. During the period from
Frankland and Faulkner equations, required significant
1883 to 1975, several formulas for analyzing plate ultimate
development time due to their inherently complex para-
strength were proposed, mainly through semi-analytical
meters. Both of these formulas, despite some differences,
and empirical approaches. Some of these can be seen in
have become standard equations utilized. Many researchers
Eqs. (4) and (5).
have subsequently proposed alternative equations for cal-
Box [22]:
culating the ultimate strength of plates, as suggested by Paik
σu 1 and Pedersen [27].
= 0.5 . (4)
σy β One of the studies on the analysis of plates under axial
Von karmam [23]: compression was conducted by Zhang and Khan [21]. The
research was performed using Faulkner’s equation and then
σu 1.9
= . (5) compared with nonlinear finite element analysis (FEA). In
σy β
the nonlinear FEA, a simple supported plate model with a
Schnadel [24]: length (a) of 4,300 mm and width (b) of 815 mm was used.
σu 1.81 This plate was referenced from a double-hull oil tanker with
= 0.5 + . (6) a length of 230 m. A parametric study was conducted with
σy β2
plate thickness ranging from 11.8 to 32 mm, equivalent to β
Frankland [25]: values ranging from 1 to 2.7. The plate was assumed to have
σu 2.25 1.25 five and a half waves in the longitudinal direction and one
= − . (7)
σy β β2 and a half waves in the transverse direction. The results of
the comparison between FEA and calculations using Faul-
Faulkner [26]:
kner’s equation showed that the FEA results aligned with
6  Muhamad Imam Firdaus and Ristiyanto Adiputra

Faulkner’s equation calculations. The difference between of lateral deformation, cracks also reduce the ultimate
the two was less than 0.5% for β values between 1.5 and strength of the panel as deformation outside the plane
2.5, demonstrating the reliability of the FEA. increases. The influence of crack damage on the ultimate
compressive strength of the plate is similar to its effect on
the ultimate tensile strength of the plate.
2.2 Analysis of plate with cracks Paik [29] then continued his research with experi-
mental and finite element methods (FEM) in two separate
Subsequently, research on plates was extended to include studies to determine the residual ultimate strength assess-
the introduction of cracks in plate structures. Paik et al. ment of cracked plates, focusing on longitudinal cracks.
[28] conducted research to determine the ultimate strength The geometry of the plates used is shown in Figure 6(a),
of plate structures with initial cracks subjected to axial loads, with the x-axis in the longitudinal direction and the y-axis
both compression and tension. Paik et al. [28] employed in the transverse direction. Four variations of longitudinal
experimental and numerical methods to determine the ulti- cracks are used, as shown in Figure 6(b). The experimental
mate strength of cracked steel plates. Numerical studies were results from the study by Paik [29] revealed that longitu-
conducted using nonlinear FEA. In his research, he varied the dinal center cracks, although relatively rare in real-world
size and location of crack damage experimentally and numeri- cases, exhibit a significant reduction, which is 7% of the
cally (Figure 5). ultimate compressive strength reduction when 2c/b = 0.5.
Paik et al. [28] showed that crack damage significantly In addition, it has been observed that transverse cracks in
reduces the ultimate strength of plates in experimental and steel plates have the potential to propagate when subjected
numerical analyses using ANSYS. Paik et al. [28] also stated to longitudinal axial tension, leading to a consequent
that the size of the crack is inversely proportional to the reduction in ultimate strength. The most common location
ultimate strength, meaning that larger cracks result in for crack occurrence is longitudinal edge cracks, particu-
lower ultimate strength. Furthermore, Paik et al. [28] noted larly at the weld intersection between the plate structure
that cracks in panels subjected to axial compression can and support elements.
close and then buckle. It is also possible that lateral deflec- In a subsequent study using FEA with the same con-
tion occurs due to initial deformation or additional loca- figuration as the experimental research, Paik [30] stated
lized loading outside the plane of the plates. After buckling, that longitudinal center cracks do not cause a significant
lateral deformation is almost certain to occur. In the case reduction in the ultimate strength of the plate, regardless

Figure 5: Experimental setup of Paik [28]. (a) A schematic view of the test set-up and (b) a photo of the test set-up.
 Deterioration and imperfection of the ship structural components  7

Figure 6: Configuration of (a) plate and (b) crack [29].

of the crack size when 2c/b < 0.5 under longitudinal com-
pression. A similar trend was observed for longitudinal
end cracks when 2c/b < 0.5. However, Paik [30] mentioned
that longitudinal end cracks result in a slightly more sig-
nificant decrease in the ultimate strength of the plate.
Further research on the analysis of ultimate strength
with imperfection cracks was conducted by Li et al. [31].
The analysis was performed using the FEM with ABAQUS
software, and the results were verified by comparing them
with the previous research conducted by Paik [30]. Li et al.
[31] concluded that larger crack sizes lead to a decrease in
the ultimate strength of the plate. Ultimate strength also
inversely correlates with the crack angle, whereas when
the crack approaches a transverse position, the ultimate
strength decreases, as shown in Figure 7. In addition, con-
sidering factors such as longitudinal and transverse cracks
and slenderness, it can be concluded that the reduction in
Figure 7: Ultimate strength graph on the effect of the crack angel [31]. ultimate strength due to transverse cracks is higher as the
8  Muhamad Imam Firdaus and Ristiyanto Adiputra

plate slenderness increases compared to longitudinal cracks and risk of corrosion increase. One type of corrosion is
with the same slenderness rate. pitting corrosion, which, by inducing stress concentration
Furthermore, Li et al. [32] continued their research by in a structure, can compromise a ship’s strength. Studies on
incorporating a new method for comparison. Li et al. [32] plates with pitting corrosion have been conducted, consid-
conducted analyses using the nonlinear FEM and artificial ering variations such as depth of penetration (DOP), pit
neural network (ANN). The research also considered the depth, pit shape, pit size, pit distribution, and location.
influence of crack orientation angles, lengths, longitudinal The corrosion patterns vary, including conical, hemisphe-
and transverse positions of multiple cracks, as well as the rical, and ellipsoidal shapes.
plate slenderness ratio on the plate strength characteris- Zhang et al. [33] conducted a study on the ultimate
tics. The geometry used by Li et al. [32] is shown in Figure 8, strength of hull structural plates, considering pitting cor-
with values of a = 2,400 mm and b = 800 mm. The elastic rosion. Zhang et al. [33] modeled and analyzed pitting
modulus and Poisson’s ratio were 205,800 MPa and 0.3, corroded plates using the FEM. In addition to pitting corro-
respectively. sion, Zhang et al. [33] considered residual stresses, taking
The research results indicated that the data obtained into account that ship hulls are welded structural plates. In
from the FEA method using ABAQUS were consistent modeling the distribution of corroded pits, Zhang et al. [33]
with and aligned with the results obtained from the ANN approximated uniformly distributed pits on the plate sur-
method. The formulas derived can also assess residual ulti- face and assumed that the plate only rusted on one side.
mate strength. From the data obtained, the ultimate strength The loading applied in the study was combined loading,
decreases significantly as the crack angle between the side with a schematic shown in Figure 9.
and longitudinal direction increases. The reduction in ulti- Zhang et al. [33] indicated that the total corroded
mate strength is most prominent when the crack is in the volume loss reduces ultimate strength. However, with the
transverse position. Li et al. [32] also stated that an increase same total corroded volume, pit depth, and surface area
in crack length reduces the plate’s strength. The increase in variations did not significantly affect ultimate strength.
plate slenderness also reduces the ultimate strength. The
position factor is also important, where if the longitudinal
crack is centered in the middle of the plate, the decrease in
ultimate strength will be drastic. In the case of a transverse
crack, the closer it is to the long side of the plate, the more
noticeable the reduction in ultimate strength.

2.3 Analysis of plate with corroded condition

In addition to cracks, corrosion is another imperfection

variable considered in ultimate strength analysis research.
Given that marine structures or ship structures are fre-
quently in direct contact with seawater, the likelihood Figure 9: Schematic combined longitudinal and transverse load [33].

Figure 8: Plat and crack configuration [32].

 Deterioration and imperfection of the ship structural components  9

The study also showed that the ratios between transverse random corrosion thickness distribution on ultimate strength.
and longitudinal stress and the ratios between shear and A total of 3,575 models of corroded plate surfaces were gen-
longitudinal stress influence the decrease in ultimate erated using the Monte Carlo method and subsequently ana-
strength alongside the corroded volume. As shown in lyzed using the FEM with dedicated finite element coding. An
Figure 10, it is evident that the shear-to-longitudinal stress equation was derived through regression analysis to predict
ratio parameter has a more significant impact on redu- the strength reduction in line with the analysis results, and it
cing ultimate strength. exhibited good accuracy.
One of the studies on the ultimate strength analysis of Experimental research was conducted by Zhang et al.
corroded plates was conducted by Silva et al. [34]. Silva [35] to compare the experimental ultimate strength results
et al. employed the FEM to investigate the influence of with those obtained from FEA. The experiments involved

Figure 10: Relationship graph of ultimate strength and (a) longitudinal and transversal stress ration; (b) shear stress and longitudinal stress ratio [33].
10  Muhamad Imam Firdaus and Ristiyanto Adiputra

Figure 11: Corroded plate model [35]. (a) Pits distribution and (b) finite element model.

applying axial compressive loads to several plate samples intensity, and general corrosion. A similar study was con-
using a specialized jig. The FEA models were assumed ducted by Nakai et al. [37]. The experiments showed that
based on corrosion parameter measurements and calcula- nominal tensile strength and elongation in corroded plates
tions of corroded volume loss. The finite element models decreased. It was also found that pitting corrosion cases
used are shown in Figure 11. would reduce tensile strength more significantly than speci-
Due to the parameter matching between experiments mens corroded overall. Nakai et al. [37] further stated that
and FEA, Zhang et al. [35] simplified the corroded pits as tensile strength is more vulnerable to pitting corrosion than
cylindrical pits with the same corroded volume loss, evenly compressive and bending strengths.
distributed on the plate model in the FEA. To simplify cal- In the case of general corrosion, comprehensive stu-
culations, the number of pits and their diameters were dies have also been carried out. General corrosion is a type
assumed to be the same. The results of this research showed of corrosion that frequently occurs in steel plate structures,
that the difference between the ultimate strength values of as shown in Figure 12. Methods for simulating general cor-
plates with corrosion using experiments and FEA was less rosion vary from uniformly reducing plate thickness to
than 10%. Therefore, Zhang et al. [35] concluded that the using complex surface roughness. Generally, it is widely
research results are reliable. The results of the comparison agreed that uniform corrosion patterns do not adequately
between experiments and FEA are presented in Table 1. reflect the local plastic hinge.
Research on the ultimate strength of corroded plates One study on general corrosion in plates was con-
has also been conducted under tensile loads. One such study ducted by Khedmati et al. [38] using the FEM with a
was performed by [36], using non-linear implicit three-dimen- random thickness surface model. General corrosion was
sional FEA with LS-DYNA coding. Ahmmad and Sumi [36] randomly distributed on one or both sides of the tested
focused on determining the strength and deformability of plate. The test results showed that variations such as slen-
steel plates with variations in pit sizes, degrees of pitting derness and aspect ratio affect the characteristics of

Table 1: Comparison between experimental and FEA result [35]

Sample No Experiment result FEA result Relative error

Corroded Ultimate Number Pit Pit Corroded Ultimate Corroded Ultimate

volume load (kN) of pits diameter depth volume load (kN) volume load (%)
loss (mm) (mm) (mm) loss (%)

1 155,152 220 144 20 3.4 153,812 228 −0.86 3.64

2 148,033 205 144 20 3.3 149,288 233 0.85 13.66
3 212,939 442 144 20 4.7 212,623 471 −0.15 6.56
 Deterioration and imperfection of the ship structural components  11

spacing between variation points was determined using the

Monte Carlo method.
The results obtained align with the findings in the
study by Khedmati et al. [38], where the reduction in ulti-
mate strength due to corrosion is quite sensitive to the
slenderness ratio and somewhat less sensitive to the aspect
ratio. Ahbar also stated that the strength values for random
corrosion are higher than or equal to those for uniform
corrosion at low corrosion levels.
While most ultimate strength analyses are conducted
using axial loading, such as compressive or tensile loads,
the ultimate strength of ships can also be determined using
combined loading, as demonstrated in a study by Cui et al.
[18]. Cui et al. [18] employed a simplified analytical method
proposed by Fujita et al. [41] by extending and modifying
the method to apply to plates subjected to combined
loading conditions. The equations used were then com-
pared with existing references. Cui et al. [18] concluded
that transverse compressive stress linearly reduces long-
itudinal ultimate strength. Cui et al. [18] stated that lateral
Figure 12: Corrosion on ship plate [39]. pressure has a minimal effect on ultimate strength, which
was compared with experimental results, showing good
agreement.
corroded plates. In this study, Khedmati et al. [38] stated In addition to the study by Cui et al. [18], Fujikubo et al.
that corrosion occurring on one side of the plate slightly [42] estimated the ultimate strength of bottom plating sub-
reduces ultimate compressive strength compared to corro- jected to transverse forces and lateral pressure. A deflec-
sion occurring on both sides of the plate. The corrosion on tion analysis was conducted by Fujikubo et al. [42] to assess
the analyzed plate in the study by Khedmati et al. [38] is the influence of lateral pressure on the plate. From the
similar to the corroded plate in the experiment by Zhang results obtained, Fujikubo et al. formulated an estimation
et al. [39], depicted in Figure 12. formula for ultimate strength. The analysis was performed
Another study by Rahbar-Ranji [40] focused on non- using a series of nonlinear FEM analyses. The analysis results
linear FEA to determine the ultimate strength of corroded showed a reduced normalized strength with increased trans-
plates with irregular surfaces. The results were then com- verse thrust and lateral pressure. Fujikubo et al. [42] com-
pared with the ultimate strength of corroded plates with pared between the FEM analysis and the equations derived in
uniform surfaces and thickness. Corroded plates with irre- the study demonstrated good agreement, especially for thin
gular surfaces were assumed to have undulated surfaces, as plate thicknesses.
shown in Figure 13. The plate thickness was varied, and the In its further development, the structure of ships with
better strength is continuously researched, and one of the
approaches involves the application of composite mate-
rials to achieve the desired properties and strength. One
prevalent application of composites is the utilization of
polymer composites for ships [43]. This practice has been
increasingly adopted since the 1980s, primarily for military
purposes aiming to reduce maintenance costs and attain
superior structural strength. Composite materials, such as
sandwich panel configurations, have been widely employed,
notably exemplified by the Visby Corvette ships. These ves-
sels incorporate carbon and glass fiber surfaces, resulting in
weight reduction of over 30%. One notable study on the use
of plate structures was conducted by Ma and Mahfuz [44].
Figure 13: Undulated model of plate surface [40]. They employed finite element simulation coupled with
12  Muhamad Imam Firdaus and Ristiyanto Adiputra

fluid–structure interaction on ship sandwich plate struc- advanced, especially with the development of numerical
tures. The core material of the sandwich structure utilized solutions facilitated by technological advancements.
in their research was polyvinyl chloride foam. In addition, Nevertheless, previous methods, particularly experimental
carbon/epoxy material layers were applied to withstand methods, continue to be developed to validate the results of
wave and blast loads. Stress analysis has shown that the other analysis methods. Experimental research on stiffened
core–skin interface near the girder is identified as the panels has been conducted by Xu and Soares [46] using wide
most susceptible area in sandwich plate structures. Interla- stiffened panels consisting of four stiffeners. The results
minar shear stress at this location exceeds the average core of these experimental tests were compared with numer-
shear stress by approximately 27%. ical method developments and methods specified in the
In general, considering the plate structure as a crucial CSR. Three different boundary conditions and three types
component in ships, a multitude of research studies have of measurable initial imperfections, namely local,
been conducted on plate structures encompassing various column, and torsional imperfections, were considered
parameters and research methodologies. Further research for all specimens in each testing method. In the study,
is imperative for analyzing and advancing plate structures the predicted ultimate strength values from CSR were
and their ultimate strength estimation in the future. All the the lowest, while the three results from numerical simu-
studies discussed earlier can be found in Table 2. lations or the FEM yielded higher values than the experi-
mental results due to the influence of welding residual
stress and fabrication.
Experimental research conducted by Xu and Soares
3 Stiffened panel [46] had previously been carried out by Xu and Soares
[47] using the stiffened panel geometry of an oil tanker
Stiffened panels have been widely employed in ship con- with a ½ + 1 + ½ bays model configuration. In Figure 14,
struction due to their ability to enhance structural strength 15 strain gauges were installed on the panel, with four of
and safety without significantly increasing weight [45]. The them positioned on the stiffener and the rest on the plate’s
addition of stiffened panels is crucial in reinforcing a ship’s surface.
structure against sagging and hogging effects. In such con- In the experimental method conducted by Xu and
ditions, the ship’s structure experiences compression, espe- Soares [47], a 300-ton hydraulic press was utilized to apply
cially in the longitudinal and transverse stiffener sections. uniaxial compression forces, as shown in Figure 15. A
Therefore, studies on stiffened panels can be categorized simply supported condition was applied to the edges of
within structural engineering. Stiffened panels come in the panel. With this setup, the maximum average stress
various geometric variations, primarily differentiated by during each loading cycle remained below the material’s
the type of stiffener used. Tee-bars, flat-bars, and L-bars yield stress.
are examples of geometric variations for stiffened panels. Hu et al. [48] conducted experiments using full-scale
Recently, extensive research has been conducted to eval- stiffened panels with three categories of variations: “as-
uate the structural integrity of stiffened panels. Several built,” “dented,” and “damaged.” Testing was performed
studies have aimed to assess the strength of stiffened by combining in-plane and lateral loads with symmetric
panels, with ultimate strength value being one of the key boundary conditions. The results differed among the three var-
parameters for evaluation. When calculating the strength iations applied to the stiffened panel, as shown in Table 3. The
or ultimate strength value of stiffened panels, several cri- study’s findings indicated that the “damaged” specimen dimen-
tical geometric parameters, such as plate and column slen- sions exhibited different and inconsistent behavior compared
derness, are important references that can significantly to the “undamaged” specimen.
influence the strength of stiffened panels. Over time, the development of experimental testing on
stiffened panels has evolved by considering various other
effects from actual conditions. Zhang et al. [35] studied the
influence of pitting corrosion on stiffened panels subjected
3.1 Various methods for analyzing stiffened to compression loads. Variations in pit location and dia-
panel meter affected strain distribution and led to changes in
buckling positions. Pit location principally serves as the
Various methods have been employed to calculate the ulti- point of initial deformation when a stiffened panel is sub-
mate strength value of stiffened panels. Currently, research jected to a load. The degree of pit corrosion (DOP) had the
on stiffened panels subjected to axial compression has most significant influence in reducing the ultimate strength
Table 2: Studies related to ultimate strength analysis of ship plate

Test type Material Analysis method Author Title Important remarks Year

Transversal and Steel Nonlinear FEM Fujikubo Estimation of ultimate strength of ship bottom plating under Fujikubo et al. conducted an estimation of finite 1999
lateral pressure et al. [42] combined transverse thrust and lateral pressure element strength and proposed an equation, then
compared it with the FEM. The results showed
alignment and acceptability
Combined loading Steel Analytical method Cui et al. [18] Strength of ship plates under combined loading The utilization of the analytical method proposed 2001
by Cui et al. is acceptable, considering the research
findings and comparisons
Transverse compressive stress linearly decreases
longitudinal ultimate strength
Lateral pressure has a minimal effect on ultimate
strength
Axial compression Steel Experiment Paik et al. [28] Ultimate strength of cracked plate elements under axial Crack damage, crack size, and position significantly 2005
and tension and FEM compression or tension reduce both tensile and compressive ultimate
strength
The utilization of experimental methods in
analyzing ultimate strength through the box
compression test
Axial compression Steel Experiment Paik et al. [29] Residual ultimate strength of steel plates with longitudinal Experiments using the same method as before but 2008
cracks under axial compression–experiments with an emphasis on the type and variations of
longitudinal crack locations
Axial compression Steel FEM Paik [30] Residual ultimate strength of steel plates with longitudinal Crack analysis using FEM with an emphasis on the 2009
cracks under axial compression – nonlinear FEM investigations types of cracks and variations in longitudinal crack
locations
Axial compression Steel Finite element Zhang and Buckling and ultimate capability of plates and stiffened panels FEA of ultimate strength in compressed plates. The 2009
method khan [21] in axial compression results are used for formula development and
compared with results obtained using Faulkner’s
equation for validation
Quasi-static Steel FEM Ahmmad and Strength and deformability of corroded steel plates under The analysis is conducted using non-linear implicit 2010
tensile load Summi [36] quasi-static tensile load three-dimensional FEA with LS-DYNA coding
The results of the experiments show that the
nominal tensile strength and elongation of
corroded plates decrease. It is also noted that
pitting corrosion cases result in a greater reduction
in tensile strength compared to specimens
corroded uniformly
Uniaxial compression Steel Finite element Khedmati A comparative computational investigation on the effects of The analysis is carried out using FEM with a random 2012
Deterioration and imperfection of the ship structural components

analysis et al. [38] randomly distributed general corrosion on the postbuckling thickness surface model, considering general
behavior of uniaxial loaded plates corrosion. This is done on both one and both sides


of the plate
13

(Continued)
Table 2: Continued
14 

Test type Material Analysis method Author Title Important remarks Year

The results indicate that corrosion occurring on one

side of the plate slightly reduces the ultimate
compressive strength compared to corrosion
occurring on both sides of the plate
In-plane Steel Non-linear FEA Rahbar- Ultimate strength of corroded steel plates with irregular Conducting ultimate strength analysis of plates 2012
compression Ranji [40] surfaces under in-plane compression with irregular surfaces using the Monte Carlo
method to generate surface patterns
Uniaxial compression Steel FEM Silva et al. [34] Ultimate strength assessment of rectangular steel plates Corroded plate modeling is conducted using Monte 2013
subjected to a random localized corrosion degradation Carlo simulations to obtain a randomized corroded
surface
Combined loading Steel Nonlinear finite Zhang et al. [33] Ultimate strength of hull structural plate with pitting corrosion The analysis is performed considering the total 2016
element analysis identification under combined loading corroded volume with pit surfaces simplified as
cylinders
Uniaxial compression Steel FEM and Zhang et al. [35] Ultimate strength experiment of hull structural plate with Analysis is conducted using the FEM and 2017
Muhamad Imam Firdaus and Ristiyanto Adiputra

experiment pitting corrosion damage under uniaxial compression experiments, with a comparison between
experiments and simulations for pitting corrosion
variables
Yan Experiments are conducted using a specially
designed jig
Pits distribution is done uniformly by equalizing
corroded volume loss
Axial compression Steel FEM Li et al. [31] A study on residual ultimate strength of steel unstiffened plate Ultimate strength analysis of ship plates using the 2020
with a crack FEM with variations in the angle and size of cracks
 Deterioration and imperfection of the ship structural components  15

Figure 14: Geometry and strain gauge configuration with transverse frames space (l) [47]; (a) l = 200, 300 mm, (b), l = 400, 450, 600 mm.

value, while the effect of corrosion volume loss was minimal as local and torsional imperfections, and residual stress, with
with the same DOP value. values of b/200, 0.0015, and 0.2σy , respectively.
Lin [49] formulated an equation that considered not only σU 1
plate slenderness but also column slenderness, as shown in = . (9)
σy 0.960 + 0.765λ2 + 0.176β 2 + 0.131λ2 β 2 + 1.064λ4
Eq. (9). Lin accounted for the influence of initial imperfections

Figure 15: Experimental set up; (a) Initial set-up, (b) Hydrolic press installation [47].
16  Muhamad Imam Firdaus and Ristiyanto Adiputra

Table 3: Summary of test result [48]

Specimen Test variable Test result

Dent/damage Plate edge restaurant Lateral load (kN) Peak load (kN) Buckling mode

SP 1.1 No Yes +10 1,572 Plate buckling followed by felxural buckling

SP 1.2 No Yes 0 1,736 Plate buckling followed by felxural buckling
SP 1.3 No Yes +25 1,453 Plate buckling followed by felxural buckling
SP 1.4 No Yes −25 1,275 Stiffener tripping
SP 1.5 No No −25 1,134 Stiffener tripping
SP 1.6 No Yes −10 1,673 Stiffener tripping
SP 1.7 No No +25 1,361 Plate buckling
SP 2.1 20 mm Yes 0 1,331 Plate buckling followed by felxural buckling
SP 2.2 35 mm Yes 0 1,116 Plate buckling followed by felxural buckling
SP 3.1 In web Yes 0 1,636 Local failure
SP 3.2 In flange Yes 0 1,773 Plate buckling followed by felxural buckling
SP 3.3 In flange Yes 0 1,683 Stiffener tripping

Paik and Thayamballi [50] provided a different for- stiffeners (flat, L, and T) were considered. The results of pre-
mula with limitations on its use, as shown in Eq. (10). dicting the ultimate strength for various types of stiffeners (flat,
Twelve longitudinal stiffened panels were analyzed with L, and Tee) based on numerical and analytical methods
different variations based on plate slenderness, accompa- through the formula were compared, showing that 83% were
nied by a single initial imperfection value of b/200. With the within a margin of error of ±5 and 95.5% were within a margin
same limitations and considering column slenderness, Zhang of error of ±10%, as shown in Figure 16.
and Khan [21] formulated a different formula to predict the σU 1
ultimate strength value. Using a 1 + ½ + 1 bay −1 + ½ + 1 span σy
=
0.995 + 0.936λ2 + 0.170β 2 + 0.188λ2 β 2 + 0.067λ4
model and providing local, torsional, and column imperfection (10)
1
modes with the same amplitude value, ultimate strength data ≤ 2,
λ
through numerical simulations were collected. A range of geo-
σU 1
metric parameters, including plate slenderness ranging from 1 = , for λ ≤ 2. (11)
σy β 0.28 1.0 + λ3.2
to 4.5 and column slenderness ranging from 0.25 to 0.9, were
considered. The results were then used for proposing the for- With the same goal of formulating a predictive for-
mula presented in Eq. (11). In addition, various types of mula for ultimate strength, Hanif et al. [51] introduced

Figure 16: Comparison of error bands between the proposed formula and all FE results [21].
 Deterioration and imperfection of the ship structural components  17

new variables. The research considered uncertainties in In numerical analysis with compression loads, the
initial imperfections, including local, torsional, and column FEM analysis is used to provide and modify various types
imperfection modes. The amplitude values for each imper- of imperfections that can reduce the ultimate strength of
fection mode varied, specifically 2.5, 25, 50, 75, and 100%, the structure. Shi et al. [53] examined the influence of pit-
resulting in a large dataset from cross-combinations of ting corrosion by varying pit location, pit diameter, and pit
amplitudes for each type of imperfection model. Through depth on the ultimate strength of stiffened panel. The pit
numerical analysis using Ansys APDL, the study also con- diameter ranged (dc) from 20 to 60 mm, while the depth (tc)
sidered aspect ratio and web slenderness values, resulting in ranged from 0 to 0.05t0 mm. The t0 means the original
a predicted ultimate strength formula with a standard error thickness of stiffened panel. Parameters and pit locations
of 2.08% compared to 750 data points analyzed using numer- are described in Table 4, and the pit shapes in the FEM
ical methods, as shown in Figure 17. Furthermore, a compar- illustration are shown in Figure 18.
ison was also made with models of different geometries based As shown in Figure 18, the pit has a cylindrical shape
on plate, web, and column slenderness values, resulting in Eq. and is located on only one side of the stiffened panel. The
(12) with a margin of error below 10%. introduction of pit corrosion generally reduces the ulti-
σU mate strength value by up to 20.7%. The largest reduction
= 0.682 − 0.322Wc − 0.154Wt − 1.580Wl − 0.025a / b occurs in the T1 graph, as shown in Figure 19, which dis-
σy
plays the load-end shortening curve of the stiffened panel
+ 0.370βw + 0.074λ3 − 0.01β 3βw3 (12) with and without the influence of pit corrosion.
+ 0.069βw3λ3 , for 1.9 ≤ ββw ≥ 3.3 and 0.37 Through numerical calculation, Rigo et al. [54] con-
≤ λβw ≥ 0.88. ducted a sensitivity analysis on the ultimate strength of
stiffened panels. In this study, L-type stiffened panels
As mentioned earlier, FEM is one of the methods used made of aluminum alloy were used under compressive
to analyze and study a topic by utilizing computational loading. The loading was applied by imposing displace-
capabilities in mechanics and mathematics. The use of ment in one direction while providing fixed constraints
numerical analysis can be performed without the need on the other side. The effect of reducing the yield stress
for complex manufacturing and preparation procedures, in the heat-affected zone (HAZ) by 10% resulted in a
as required in experimental methods. The accuracy of the decrease in ultimate strength ranging from 2 to 5%.
FEM method depends on various factors, and mesh den- Meanwhile, the effect of each millimeter of amplitude
sity is one example of a factor that can affect the accuracy applied caused a reduction in ultimate strength of
of the results obtained using the FEM method [52]. In their approximately 1.1%.
research, AbuBakar and Dow [52] used FEM to analyze
and compare penetration damage in various grounding
scenarios.
3.2 The influence of initial imperfection
mode on stiffened panel
1.2
FEM
Standard Error: 2.08% In both the studies by Shi et al. [53] and Rigo et al. [54], the
1.0
analysis of stiffened panels considered the influence of

0.8
σU/σy : FEM

Table 4: Parameter and location of the pit [53]

0.6

Model Pit Pit Number Pits

0.4
diameter thickness of pits distribution
dc (mm) tc (mm)
0.2
L1 60 2.5 12 Left
L2 60 2.5 12 Center
0.0
D1 20 2.5 36 All
0.0 0.2 0.4 0.6 0.8 1.0 1.2
D2 60 2.5 36 All
σU/σy : Derived Formula
T1 25 1 36 All
T2 25 3 36 All
Figure 17: Comparison of derived formula to FEM [51].
18  Muhamad Imam Firdaus and Ristiyanto Adiputra

Figure 18: FEA model of stiffened panel [53]. (a) FEA Model-L1, (b) FEA Model-L2, (c) FEA Model-D1, (d) FEA Model-D2, (e) FEA Model-T1, and (f) FEA
Model-T2.

initial imperfections. Using numerical methods, such ana- used a ½ + ½ bay + ½ + ½ span model, as it was capable
lyses are feasible because of the flexibility offered by the of modeling the specified buckling imperfections and applying
method in modifying geometry. Anyfantis [55] introduced simply supported conditions. Employing Ansys APDL, Any-
buckling imperfections into stiffened panels subjected to fantis [55] introduced critical plate imperfections, local plate
nonuniform thrust. Unlike Shi et al. [53], Anyfantis [55] imperfections, column imperfections, and torsional web

Figure 19: Stress–strain curves of the stiffened plates with pits [53]. (a) L1, (b) L2, (c) D1, (d) D2, (e) T1, and (f) T2.
 Deterioration and imperfection of the ship structural components  19

imperfections. Global buckling imperfections also included

tripping imperfection modes.
Similar to Anyfantis [55], Hanif et al. [51] conducted an
analysis of stiffened panels with the influence of initial
imperfections, including local, torsional, and column imper-
fections, as shown in Eqs. (13)–(15). The difference lies in the
magnitude of the amplitudes applied. The overall illustra-
tion of imperfection modes is shown in Figure 20.
In Hanif’s study [56], variations of amplitudes were
applied at 2.5, 25, 100, and 300%. The percentage levels of
initial imperfection severity refer to data provided by Fuji-
kubo et al. [57], as shown in Figure 21. Values of 2.5, 100,
and 300% indicate that the severity level falls within the
slight, average, and severe ranges. The amplitude of the
initial imperfection is determined by β 2t or the square of
plate slenderness multiplied by plate thickness. The numer-
ical results compared to analytical calculations outlined by Figure 21: Analysis of scatter data of initial geometric imperfections.
IACS-CSR [58] found that the IACS ultimate strength predic- Reproduced from the study by Fujikubo et al. [57].
tions fell within the ultimate strength range for severity
levels of 25–100%.
πyi zi πxi
mπxi Wt = B0 sin , (14)
Wl = C0sin sin , (13) hs a
a b

Figure 20: Illustration of the initial geometric imperfection modes applied to the stiffened panels: (a) column imperfection mode, (b) torsional
imperfection mode, (c) local imperfection mode, and (d) combination of the three imperfection modes [51].
20  Muhamad Imam Firdaus and Ristiyanto Adiputra

πxi results of the stochastic geometric imperfection model

Wc = B0sin . (15)
a demonstrate its ability to determine the imperfection model
Li et al. [59] introduced different types of imperfection based on the maximum distortion magnitude and combine
modes applied to stiffened panels, including hungry-horse the effects between them.
mode, Admiralty Research Establishment mode, and critical Gandhi et al. [62] studied the influence of initial deflec-
buckling mode. The study focused on assigning random tion on various types of stiffened panels based on their
amplitudes to each imperfection mode to understand the stiffener types (Tee and angle bar). In this study, only local
distribution of ultimate strength and buckling behavior and column imperfection modes were applied with a
exhibited by the stiffened panel. Equations for each imper- single-level severity amplitude of 100% referred to the first
fection mode are presented in Eqs. (16)–(18). buckling mode between two stiffeners. Simply supported
11 boundary conditions were applied to the model, where not
max
∑ A0i sin ⎛ iπx ⎞ ⎛ πy ⎞ (16) all edges were fixed, but some degree of freedom was
wopl = wopl sin ,
i=1
⎝ a ⎠ ⎝ b ⎠
coupled to apply continuous conditions similar to the
actual condition. The coupling provided freedom for the
max ⎧ πx ⎛ jπx ⎞
wopl = wopl ⎨A1 sin⎛⎝ ⎞⎠ + Aj sin edge to deform in specific rotational directions. Using a 1
⎩ a ⎝ a ⎠
(17) bay – 1 frame model, the study’s results were accurate,
⎡ ( j + 1)πx ⎤⎫ ⎛ πy ⎞ validated by comparing them with numerical results from
+ Aj + 1 sin ⎬sin ,
⎣ a ⎦⎭ ⎝ b ⎠
the ALPS - ULSAP and MSC MARC methods and also demon-
mπx ⎞ ⎛ πy ⎞ strated that variations in stiffener size had a significant
max
wopl = wopl sin⎛ sin , (18)
⎝ a ⎠ ⎝ b ⎠ impact on normalized strength values. The choice of stif-
a fener type, either Tee or angle Bar, did not show significant
j= + 1, (19) differences when observing the load-end shortening curve
b
analysis.
A1
= 4.0, (20) Rizzo and Caire [63] introduced variations in initial
Aj
imperfections in their stiffened panel analysis under com-
Aj + 1 = 0.01, (21) bined compression and shear by employing a 1 bay – 3
span model with 3D modeling of the transverse stiffener.
a
≤ m(m + 1) . (22) Before collecting data for analysis, a mesh convergence
b
study on the stiffened panel was conducted and subjected
The magnitude assigned in the study refers to research to pure compression to obtain accurate and efficient results,
conducted by Smith et al. [60], where the amplitude value as shown in Table 5. The combination of longitudinal com-
is a function of β 2t . Referring to the same graph provided pression and edge shear loading increased the ultimate
by Fujikubo et al. [57], the severity levels assigned are 25% strength value due to the influence of plastic hardening.
or slight condition, 100% or average condition, and 300% or The initial imperfections were introduced in the form of
severe condition. The stochastic geometric imperfections the first buckling mode on the damaged area, which was
applied are obtained from the study conducted by Geor- located in the middle of the plate. Further imperfections
giadis et al. [61] for local plating, as shown in Figure 22. The were introduced by applying a collision case caused by rigid

Figure 22: Schematics of stochastic geometric imperfection based on the information obtained from the study by Georgiadis et al. [61].
 Deterioration and imperfection of the ship structural components  21

Table 5: Mesh sensitivity study under pure compression – load pro- can occur due to corrosion. This is a potential issue consid-
portionality factor (LPF) [63] ering the long-term activity of ships, which can lead to cor-
rosion caused by air, seawater, and sea creatures. Feng et al.
Mesh Plate (mm) Stiffeners/frames (mm) LPF [64] analyzed the influence of pit degree (DOP), depth, loca-
1 100 × 100 100 × 100 4.573 tion, and distribution of pit corrosion on the ultimate
2 75 × 75 100 × 100 4.638 strength value when subjected to uniaxial compression.
3 50 × 50 100 × 100 4.625 Three types of stiffened panels adopted from the research
4 25 × 25 100 × 100 4.618
of Zhang and Khan [21] were used with a configuration of 11
5 50 × 50 50 × 50 4.621
6 50 × 50 75 × 75 4.623
bays and ½ + 1 + ½ span, as shown in Table 6. The transverse
stiffener modeling in Feng et al.’s [64] study was given only a
boundary condition in the form of fix constraints. In addi-
sphere indentation varied from 50, 100, 200, to 300 mm, as tion to corrosion effects, initial imperfections in the form of
illustrated in Figure 23. Subsequently, a combination of initial deflection in the form of local and global imperfection
loading was applied, with shear loading first, followed modes were also considered in the analysis. The pit dia-
by longitudinal compression loading. As expected, the meter was 60 mm when evenly distributed, but for ran-
magnitude of the initial imperfection decreased the ulti- domly distributed pits, the diameter ranged from 50 to
mate strength value, and in this case, the decrease in 70 mm. As for the pit depth, the variations ranged from
ultimate strength was more sensitive to compression 0.33t0, 0.67t0, to 1.0t0, where t0 represents the intact thick-
than to shear load. ness value, while DOP was varied in three magnitudes:
5.98, 11.96, and 23.72%. The addition of DOP was associated
with a significant reduction in ultimate strength, and this
effect increased as the pit depth increased, as shown in
3.3 The influence of corrosion on stiffened Figure 24 upper graph. In this figure, it can be seen that
panel with a low DOP of less than 7%, the effect of pit depth was
minimal. Similar to the DOP effect, increasing the pit depth
In addition to initial deflection and variations in buckling would reduce the ultimate strength value. However, based
imperfection modes, initial imperfections in stiffened panels on the trend shown, a larger DOP would not significantly

Figure 23: Stress distribution at maximum indentation – 200 mm displacement [63].

22  Muhamad Imam Firdaus and Ristiyanto Adiputra

Table 6: Geometric dimensions of the three stiffened panel extracted from [21]

ID Panel Stiffener web Stiffener flange Stiffener type

a (mm) b (mm) t (mm) h w (mm) tw (mm) h f (mm) tf (mm)

1 3,500 810 15 300 11 90 16 L

2 4,500 815 18 463 11 172 17 T
3 5,500 910 19 400 12 100 18 L

affect the reduction in ultimate strength in models with 3.4 The influence of residual stress on
different pit depths, as shown in Figure 24 lower graph. stiffened panel
The influence of random corrosion was analyzed on
stiffened panels by Feng et al. [65] using the nonlinear FEM Another type of initial imperfection that can affect the
(NLFEM). Stiffened panels on the deck portion were used in load-bearing capacity of stiffened panels is residual stress,
this analysis without modeling the transverse stiffener. The generally caused by heat imparted along the welded por-
randomness values of corrosion were obtained from 80 tion of stiffened panels. Li et al. [67] conducted research on
measurement data referring to 15 ships with different the influence of residual stress on the ultimate strength of
ages, with the statistical approach adopted from data stiffened panels. Figure 25 shows the idealized welding-
obtained in Wang et al.’s [66] research. The ultimate induced residual stress [67]. The distribution and magni-
strength values of randomly corroded panels followed a tude of residual stress have been previously formulated by
log-normal distribution. A Monte-Carlo approach was Smith and Anderson [68] through a simplified welding-
used to analyze the statistical characteristics of the ulti- induced residual stress distribution. With this configuration,
mate strength of stiffened panels affected by corrosion. residual stress was modeled as tension and compression
Corrosion data were obtained through the Monte-Carlo blocks, as shown in Figure 25(a). Using a similar model,
approach based on the log-normal model. the distribution of residual stress was formulated by Yao

Figure 24: Strength reduction caused by DOP and pit depth based on the information obtained from Feng et al. [64].
 Deterioration and imperfection of the ship structural components  23

Figure 25: Welding-induced residual stress in stiffened panels [67]; (a) Simplified model, (b) Model for fillet welding, and (c) Bi-axial distribution for
plating.

and Fujikubo [69] in Figure 25(b), but a uniform distribution the longitudinal tensile residual stress block is a function of
was assumed for compression stress upward along the stif- plate slenderness (β) and weld leg length (Lw). However, in
fener web. Paik and Thayamballi [70] formulated a different their study, the tensile residual stress was assumed to be
distribution of residual stress in Figure 25(c), with a bi-axial equal to the yield stress of the material. The distribution
pattern considering residual stress in the longitudinal and applied followed that formulated by Paik and Thayamballi
transverse directions. [70] but neglected the transverse direction. The results of
To account for the effects of residual stress, an edge the geometry illustrations affected by residual stress are
function was formulated by Gordo and Soares [71] to shown in Figure 26.
modify the material behavior of stiffened panels with resi- The analysis results show that a significant reduction
dual stress, resulting in the stiffness reduction by 1.0 – εrcx/ in ultimate strength occurs in almost all cases, and the
εYeq. The initial edge function was modified to simulate effect of residual stress can lead to a reduction of up to
material behavior as described sequentially in Eqs. (23a), 11.2% compared to stiffened panels without the influence of
(23b) and (24), which is also used in the CSR-H method. residual stress. Furthermore, the analysis results also indi-
cate that collapse in stiffened panels generally occurs at
ϕ = ε /εy , if ε /εy ≤ 1.0, (23a)
higher strains. This is due to the effect of residual stress
ϕ = 1 if , ε / εy > 1.0, (23b) and becomes more sensitive in stiffened panels with low
plate and column slenderness values. In addition, the
ϕ = ε /εy if , ε /εy ≤ 1.0 − εrcx /εy . (24)
application of residual stress does not alter the postcol-
In their study, stiffened panels were categorized into lapse behavior of stiffened panels. When applied only to
three types: stiffened panels without residual stress, stif- the local plate, residual stress reduces the average ultimate
fened panels with residual stress in the local plate, and strength by around 6%, but in combination with the plate
stiffened panels with residual stress in both the plate and and web, it ranges around 8.5%.
web. The initial residual stress in this case was controlled
b1 = c1 × L w + d1, (25)
using Eq. (21) developed by Yi et al. [72], where the width of
24  Muhamad Imam Firdaus and Ristiyanto Adiputra

Figure 26: Stiffened panel after relaxation of the initial residual stress [67].

where uniform bending and stiffened panel geometry were also

c1 = − 0.6212β 2 + 5.9145β − 3.3894, (26) introduced. Using ABAQUS, the flange and stem of the stif-
fener were divided into 96 elements, while the plate was
d1 = 1.9505β 2 − 16.234β + 18.711. (27)
divided into 384 elements to achieve the desired results.
In parametric studies conducted by Grondin et al. [73], Modeling of residual stress was realized in simulations by
beside the variation in residual stress in terms of its mag- imposing initial strains in the form of temperature distri-
nitude, initial deflection and the direction of applying bution and applying only longitudinal residual stress. The

Figure 27: Effect of residual stresses [73].

 Deterioration and imperfection of the ship structural components  25

imposition of initial strains in that direction was done by tends to have a gentler decline after experiencing collapse
setting the thermal expansion to zero in the orthotropic at high strains. With increasing magnitude of residual
temperature material property. The magnitude of residual stress, it can be observed in the graph that the “no load
stress at the junction between the plate and stiffener was shedding zone” also increases. This occurs in each varia-
given as the yield strength of the material in tension. To tion of plate thickness, whether it is 12.8, 17.8, or 21.3 mm.
understand the effect, variations in magnitude were per- 2η
σr
formed with values of 0, 15, and 30% of the yield strength. = .
σ0 ⎛ b ⎞ − 2η (28)
The effect of residual stress on the strength of the stiffened ⎝t ⎠
panel can be seen in Figure 27, which shows the normal-
ized strength against plate slenderness ratio. In stiffened
panel models with a plate slenderness greater than 1.7, a 3.5 The influence of crack on stiffened panel
significant decrease in strength due to the effect of residual
stress occurs. The influence of other types of initial imperfections in the
On the other hand, Khan and Zhang [74] introduced stiffened panel can also take the form of cracks. Typically,
the effect of residual stress on stiffened panels using the cracks are found along weld lines and intersections on
formula formulated by Faulkner [26], as shown in Eq. (28). stiffened panels. Analysis of stiffened panels with cracks
For practical design and simulation purposes, similar to has been conducted in both experimental and numerical
previous studies, the welding-induced stress distribution analyses. Bayatfar et al. [75] studied the influence on stif-
at the plate–stiffener junction was idealized by forming fened panels subjected to longitudinal compression. A one
tensile and compressive stress blocks. In Faulkner’s equa- bay–one span model was used in the analysis with a flat
tion [26], the magnitude of the residual stress depends on bar-type model of a stiffened panel. There were two varia-
the value of η, which in ship applications varies between 4, tions of cases given for the stiffened panel. In the first case,
5, and 6. In the shakedown value, the allowable value is 3 the crack was located in the middle of the plate transver-
and 4.5. sely and had a length of 10–50% of the plate width, while in
Unlike the research described earlier, in the study by the second case, there were two cracks on the stiffened
Khan and Zhang [74], residual stress values were set to be 0 panel. The first crack was in the middle of the plate with
and 0.25 of the yield strength. The value of η was set at 2.45, a length of 20% of the plate width, while the second crack
4.99, and 6.4. By introducing another variation in plate was vertically positioned in the middle of the longitudinal
thickness, as shown in Figure 28, the load-end shortening web-plate junction. The length of the second crack ranged
curve without residual stress exhibits a sharp decline after from 10 to 50% of the plate width. The graphs of the results
collapse, while the stiffened panel with residual stress from the first and second cases are shown in Figure 29. In

Figure 28: Load shortening curves of stiffened panel considering different values of compressive residual stress for 12.8, 17.8, and 21.3 mm plate
thickness based on the information obtained from the study by Khan and Zhang [74].
26  Muhamad Imam Firdaus and Ristiyanto Adiputra

Figure 29: Nondimensional average stress–average strain curves: (a) case 1 and (b) case 2 [75].

Figure 29(a), it can be seen that the effect of crack length In another study, Cui et al. [77] used cracks as the main
significantly reduces ultimate strength. In addition, local parameter alongside providing other parameters such as
plastic deformation is concentrated in the middle area of initial deflection. Using NLFEM, four types of crack models
the plate as the crack length increases. In the second case, were modeled, as shown in Figure 30. These models were
the increase in the crack length in the web area does not distinguished based on the location of the crack. Initial A,
significantly reduce the ultimate strength. B, P, and S have their meanings from the presented dia-
Using different loading test, Dexter and Pilarski [76] gram, where P means the crack is located only in the plate,
analyzed the effect of crack propagation on welded stif- while S means the crack location is in both the plate and
fened panels under four-point bending. Through experi- the stiffener. S refers to a symmetric crack location, while
mental testing, two 490-kN actuators under load control A refers to an asymmetric crack location relative to the
were used to provide a constant load range as the crack stiffener. The variations in the crack length ratio to plate
propagated. Ten different types of stiffened panels were width were 0.1–0.4, while the crack length ratio in the
tested in the study, with variations in stiffener type, stress stiffener to web height was 0.2.
range, notch length, and stiffened panel crack detail. The Figure 31(a) plots ultimate strength values, which show
results showed a reduction in the crack propagation rate a linear and significant decrease with the increasing crack
due to the effect of compressive residual stress between the length. It can be seen that adding a crack in the stiffener
stiffeners. significantly reduces the ultimate strength, while adding

Figure 30: Four types of cracked models [77].

 Deterioration and imperfection of the ship structural components  27

Figure 31: The ultimate compressive strength reduction characteristics of cracked stiffened plates as a function of (a) crack length and (b) longitudinal
location of crack [77].

symmetric cracks in the plate-only case (PA vs PS) has a boundary conditions concerning ultimate strength and
slight difference. In another case, variations in symmetric postbuckling behavior of stiffened panels. In this analysis,
and asymmetric cases with crack locations in both the a stiffened panel with a Tee-bar configuration was selected,
plate and stiffener result in a more significant gap as crack with material properties including a yield stress, Young’s
length increases. Meanwhile, in Figure 31(b), longitudinal modulus, and Poisson’s ratio of 313.6 MPa, 205.8 GPa, and
crack locations do not appear to have a considerable effect 0.3, respectively. Two different configuration models were
on ultimate strength. considered in the analysis: the ½ + 1 + ½ span and the ½ +
½ span. Variations were applied to the number of half-
wave critical modes in initial imperfection, as presented
in Table 7. The results obtained under periodic boundary
3.6 Various approach on applied modeling conditions showed significant differences compared to
parameter through numerical analysis symmetric boundary conditions when the buckling half-
wave number was even. Conversely, in cases where the
In the analysis of stiffened panels using the numerical number of half-waves was odd, the results between the two
method, the selection of model configuration and boundary boundary conditions were relatively similar due to the sym-
conditions is a crucial step that can significantly impact the metry exhibited in the initial imperfection modes. The results
analysis outcomes. Numerous studies have explored various from periodic boundary conditions did not overestimate
variations of model configuration. Some examples discussed the ultimate strength in various cases, whereas symmetric
include the model configuration of ½ + ½ bay – ½ + ½ span, boundary conditions could only be applied when the col-
which was utilized in studies by Anyfantis [55] and Hanif et al. lapse mode exhibited a symmetric shape. Therefore, the
[56], as well as the ½ + 1 + ½ bay – ½ + 1 + ½ span configuration
employed by Yu et al. [78]. In another parameter, the various
definitions and approaches of boundary condition are also Table 7: Number of half wave (m) in initial deflection in the longitudinal
direction [14]
shown in several cases. In 2023, Lutfi et al.’s study [79] demon-
strates the use of coupling conditions at the edges of the entire
Boundary Plate: a × Px/Py
stiffened panel to provide specific degrees of freedom for
condition b (mm)
deformation and rotation in accordance with the actual condi- 1:0 1:0.41 1:0.67 0:1
tions. In his analysis of OTEC seawater tank stiffened panels, BC 1 2,550 × 850 3 1 1 1
imposed displacements were also applied from two opposing 3,400 × 850 5 1 1 1
directions, while nodes at the center of the plate were sub- 2,890 × 850 3,4,5 — — —
jected to fixed constraint conditions. BC 2 2,550 × 850 3 1 1 1
3,400 × 850 5 1 1 1
In the different case, the research conducted by Xu
2,890 × 850 3,4,5 — — —
et al. [14] compared the use of symmetric and periodic
28  Muhamad Imam Firdaus and Ristiyanto Adiputra

selection of boundary conditions should be made carefully indicated by labels A1, A2, and A3 in Figure 33. In this figure,
to align with the structural characteristics under analysis. modeling the ½ + 1 + 1 ½ bay – ½ + 1 + 1 + ½ span config-
A similar study was also conducted by Hanif et al. [15] uration consumed 1,223 s in a single simulation, whereas the
to compare model configurations, boundary conditions, ½ + ½ bay – ½ + ½ span model only required 144 s, repre-
and methods for modeling transverse stiffeners when ana- senting more than eight times less time consumption per
lyzing the strength of stiffened panels. This study was a computation.
parametric investigation without cross-combining various As discussed earlier, stiffened panels play a crucial
variations. Apart from comparing the postbuckling beha- role in reinforcing ship structures during both hogging
vior for each variation, modifications were implemented to and sagging conditions. Analyses involving single loads,
assess the time efficiency during simulations using ANSYS such as compression, and combinations of loads have
APDL. The variations in model configurations included ½ + been extensively conducted with advanced improvisations,
½ bay – ½ + ½ span, ½ + 1 + ½ bay – ½ + 1 + ½ span, and ½ + including the introduction of initial imperfections like cor-
1 + 1 + ½ bay – ½ + 1 + 1 + ½ span. Regarding the modeling rosion, residual stress, and both global and local imperfec-
of transverse stiffeners, different approaches were employed, tion modes. However, continuous development is essential
such as modeling them in 3D or applying boundary condi- to enhance the safety of ship structures through various
tions as fixed constraints on the transverse stiffener. Geo- methods, particularly numerical techniques that show pro-
metric dimensions of transverse stiffeners were also adjusted mising potential in modeling different types of imperfec-
based on the research conducted by Platypodis and Anyfantis tions. All the studies mentioned earlier are summarized in
[80]. The results obtained by modifying the model configura- Table 8.
tions showed a trend of increasing ultimate strength values as
the scope of model configurations expanded. The largest dif-
ference in ultimate strength occurred in the smallest model
configuration, with variations reaching up to approximately 3.7 Stiffened curved plate structure
37 MPa. As for the modeling of transverse stiffeners, differ-
ences were observed between 3D modeling and not including After discussing various variations present in stiffened panels,
such modeling. However, these differences were relatively particularly flat stiffened panels, there exists another type of
small, not exceeding 10 MPa, as observed in Figure 32. In stiffened panel that constitutes the hull girder structure of a
addition, variations in geometric aspects of modeling trans- ship. Stiffened curved panels are reinforcement structures
verse stiffeners did not exhibit significant differences, with commonly found in the deck area with a camber, as well as
the largest variation being less than 0.5 MPa. When com- on the front and rear sides of the hull, as illustrated in
paring the time required for the analyses, significant differ- Figure 34. In comparison to flat stiffened panels, curved
ences were noted in the modeling of model configurations, as stiffened panels have received less attention in research.

Figure 32: Stress–relative strain curves of related transverse stiffener modeling parameters [15].
 Deterioration and imperfection of the ship structural components  29

Figure 33: Running time consumed on each model [15].

Typically, research efforts have been confined to curved between the increase in flank angle and web height with
plates. An example is the investigation conducted by Maeno respect to ultimate strength.
et al. [81], which focused on unstiffened curved plates to Similar research was also conducted by Seo et al. [86],
understand buckling and collapse behavior under axial varying the values of curvature, plate slenderness ratio, web
compression. Similar investigations were also carried out height, and three types of stiffeners. At a glance, the study
by Yumura et al. [82], varying the curvature of the plate. conducted by Seo et al. [86] appears simpler as it involves
Kwen et al. [83] undertook more diverse modifications invol- fewer geometric parameters and is subjected to only one
ving aspect ratio, slenderness ratio, curvature, and various type of loading, namely, axial compression. However, toward
loading conditions, including longitudinal compression, trans- the conclusion of their investigation, Seo et al. [86] were able to
verse compression, and shear loads. formulate an empirical calculation formula for determining
In the analysis of curved stiffened panels, different the ultimate strength of curved stiffened panels. The influence
approaches are employed, such as numerical analysis of initial imperfection in the form of plate and torsional imper-
[85] and empirical formulas [86]. Park et al. [87] conducted fection buckling modes was also considered, with a singular
an analysis on curved stiffened panels by modifying the value set at 25% or slight level.
type of stiffener, curvature, flank angle, and plate slender- Numerical results indicated that the choice of stiffener
ness ratio. Three different types of stiffeners, namely, flat, type did not affect the ultimate strength values but did
angle, and tee bars, were utilized in their study. The con- impact the collapse pattern. In the case of the flat bar, struc-
tainer ship type with a range of ½ + ½ bay – ½ + ½ span tural failure was attributed to the stiffener, while in the
range was subjected to simply supported conditions on all other two types of stiffeners, failure was caused by the plate
sides except the loading edge. The scope of the model range and local stiffener. As mentioned earlier, the empirical for-
area and the imposition of boundary conditions can be mula for ultimate strength was derived by considering the
observed in Figure 35. plate slenderness ratio parameter and limiting the accuracy
In addition to the aforementioned variations, geo- level of the results with the flank angle, as demonstrated in
metric modifications were made to web height, web and Eqs. (29) and (30).
flange thickness, and plate thickness, with respective var- σU 1
iations of 50–400, 12–15, and 12–26 mm, respectively. The = , (θ ≤ 5°) , (29)
σy 0.0683 + 0.994β′ + 0.0065β′ 2
plate width varied between 1,000 and 3,000 mm, and the
flank angle ranged from 50 to 450. Apart from geometric σU 1
= , (θ > 5°) . (30)
variations, Park et al. [87] also considered the influence of σy 1.1339 − 0.1291β′ − 0.0392β′ 2
initial deflection, such as plate and web imperfection buck-
ling modes. The research results on curvature and stiffener Further development of stiffened panel structure strength-
height variations indicated a proportional relationship ening endeavors has risen and demonstrated by Chen et al. [88].
Table 8: Milestone studies related to stiffened panel analysis

Test type Material Analysis method Author Title Important remarks Year
30 

Compression Steel Experiment, numerical, Lin [49] Ship longitudinal strength modeling Formulating a formula that takes into account not 1985
and analytical only plate slenderness but also column
slenderness, the influence of initial imperfections,
and residual stress
Lateral load ASTM A370 Experiment Hu et al. [48] Ultimate collapse tests of stiffened-plate ship The “damaged” specimen dimensions exhibit 1997
structural units different and inconsistent behavior compared to
the “undamaged” specimen
Combination of axial Steel Numeric Grondin Buckling of stiffened steel plates – a parametric The reduction in strength due to the effect of 1999
compression and et al. [73] study residual stress is significant because of the
bending potential combination of residual stress and
average buckling stress
Four point bending Steel Experiment Dexter and Crack propagation in welded stiffened panels The reduction in crack propagation rate is caused 2002
Pilarski [76] by the compressive residual stress effect between
stiffeners
Muhamad Imam Firdaus and Ristiyanto Adiputra

Compression Aluminum Numeric Rigo et al. [54] Sensitivity analysis on ultimate strength of The effect of reducing the yield stress in the HAZ 2003
aluminum stiffened panels by 10% results in a decrease in the ultimate
strength value ranging from 2% to 5%
Longitudinal stress Composite Composite column Chen et al. [88] Reliability analysis of a ship hull in composite The evaluation of ultimate longitudinal strength 2003
theory material involves the utilization of equations derived from
composite column theory. To assess structural
integrity, a hybrid approach combining the
response-surface method and the first-order
reliability method is utilized to ascertain safety
indices and the likelihood of failure
Axial compression HT32 steel Numerical and Zhang and Buckling and ultimate capability of plates and A semi-analytical formula for ultimate 2009
analytical Khan [21] stiffened panels in axial compression compressive strength assessments of stiffened
panels was proposed applied to the deck and
bottom structures for a range of various sizes oil
tankers and bulk carriers for flat, L, and Tee bar
type stiffened panel
Compression Steel Numeric Khan and Effects of welding-induced residual stress on The load-end shortening curve without residual 2011
Zhang [74] ultimate strength of plates and stiffened panels stress experiences a sharp drop after collapse,
while in the case of a stiffened panel with residual
stress, it tends to be smoother, and the “no-load
shedding zone” increases
Biaxial High strength Numeric Gandhi et al. [62] Ultimate strength analysis of stiffened plates Local and column imperfection modes with a 2012
steel with initial imperfections single amplitude of 100% on Tee and Angle Bar
type members provide a level of accuracy
consistent with the ALPS – ULSAP and MSC MARC
methods
(Continued)
Table 8: Continued

Test type Material Analysis method Author Title Important remarks Year

Axial Compression Steel Experiment, numerical, Xu and Comparisons of calculations with experiments The results of the comparison between the three 2013
and analytical Soares [46] on the ultimate strength of wide stiffened different numerical methods, experiments, and
through CSR panels CSR on wide stiffened panels with the influence of
initial imperfections showed that the CSR
predicted the lowest ultimate strength
Axial compression Steel Experiment Xu and Experimental study on the collapse strength of The test configuration with a ½ + 1 + ½ bays model 2013a
Soares [47] wide stiffened panels was able to avoid side bay collapse with simply
supported boundary conditions and verified the
use of FEM that aligns with it
Lateral displacement S235JR-EN10025 Experiment and AbuBakar and Simulation of ship grounding damage using Mesh density can affect the numerical result 2013
and S255JR- numeric Dow [52] the FEM analysis
EN10210
Compression No specified Numeric Xu et al. [14] Influence of boundary conditions on the Periodic boundary conditions do not overestimate 2013
collapse behavior of stiffened panels under the ultimate strength in various cases, while
combined loads symmetric boundary conditions can only be
applied when the collapse mode exhibits a
symmetric form
Compression Steel Numeric Bayatfar Residual ultimate strength of cracked steel The effect of crack length significantly reduces the 2014
et al. [75] unstiffened and stiffened plates under ultimate strength
longitudinal compression
Axial compression High tensile steel Numeric and empirical Seo et al. [86] Nonlinear structural behavior and design The numerical results indicate that the choice of 2016
formula formulae for calculating the ultimate strength stiffener type does not have an effect on the
of stiffened curved plates under axial ultimate strength values but does influence the
compression collapse pattern. In the case of the flat bar,
structural failure is attributed to the stiffener,
while in the other two types of stiffeners, failure is
caused by the plate and local stiffener
Compression Steel Q345 Experiment Zhang et al. [35] Experimental analysis of residual ultimate The variation in the location of pits and their 2017
strength of stiffened panels with pitting diameter has an effect on strain distribution and
corrosion under compression results in a change in the buckling position
Compression Steel Numeric Feng et al. [65] Reliability of the ultimate strength of ship The random corrosion wastage of the stiffened 2017
stiffened panel subjected to random corrosion panel on the bulk carrier mid-ship deck structures
degradation follows the log-normal distribution
Deterioration and imperfection of the ship structural components

Compression High tensile Numeric Cui et al. [77] Ultimate strength characteristics of cracked Variations in symmetric and asymmetric cases 2017
strength material stiffened plates subjected to uniaxial regarding the location of cracks in the plate and


compression stiffener result in a larger gap as the crack length

increases
31

(Continued)
Table 8: Continued
32 

Test type Material Analysis method Author Title Important remarks Year

Compression Steel Q345 Experiment and Shi et al. [53] Numerical assessment of experiments on the The influence of pit corrosion, in general, reduces 2018
numeric ultimate strength of stiffened panels with the ultimate strength value by up to 20.7%
pitting corrosion under compression
Combine loading (shear Steel Numeric Rizzo and Ultimate strength formulations for FPSO The magnitude of the initial imperfection reduces 2018
and compression) Caire [63] stiffened panels under combined compression the ultimate strength, and the reduction in
and shear with initial imperfections and ultimate strength is more sensitive to compression
damage than to shear load
Compression Steel AH36 Experiment and Yu et al. [78] Experimental and numerical investigation on Stiffened panels measured using 3D scanning 2019
numeric the ultimate strength of stiffened plates with technology can be applied in the ultimate strength
scanned initial geometrical imperfection assessment
Uniform and Steel Numeric and analytic Anyfantis [55] Ultimate strength of stiffened panels subjected Simplify formula by the influence of initial 2020
nonuniform to nonuniform thrust imperfection and non-uniform thrust
compression
Compression HT32 steel Numeric A parametric study on effects of pitting Increasing the rate of DOP is correlated with a 2020
Muhamad Imam Firdaus and Ristiyanto Adiputra

Feng et al. [64]

corrosion on stiffened panels’ ultimate strength significant decrease in ultimate strength, and this
effect intensifies as the pit depth increases
Axial compression High tensile steel Numeric Park et al. [87] Estimation of buckling and ultimate collapse The research findings on variations in curvature 2020
behavior of stiffened curved plates under and stiffener height indicate a proportional
compressive load relationship between the increase in flank angle
and web height with respect to ultimate strength
Compression No specified Numeric Georgiadis Influence of stochastic geometric imperfection The stochastic geometric imperfection model is 2021
et al. [61] on the ultimate strength of stiffened panel in capable of determining the imperfection model
compression based on the maximum distortion magnitude and
combining the effects of both
Compression Steel Numeric Li et al. [67] The influence of residual stress on the ultimate When applied only to the local plate, residual 2021
strength of longitudinally compressed stiffened stress reduces the ultimate strength by an
panels average of about 6%. However, when applied to a
combination of plate and web, it ranges
around 8.5%
Compression No specified Numeric Li et al. [59] A comparison of geometric imperfection Different geometric imperfection models lead to 2022
models for collapse analysis of ship-type an appreciable uncertainty in the ultimate
stiffened plated grillages compressive strength
Compression High strength Numeric Hanif et al. [15] Effect of design parameters on the ultimate The variation in the ultimate strength value 2023
steel strength and collapse behavior of stiffened resulted from differences in the range
panels configuration model, boundary conditions, and
transverse stiffener modeling was relatively
negligible, indicating the overall appropriateness
of the variations
(Continued)
 Deterioration and imperfection of the ship structural components  33

2023c
Their study primarily delves into the calculation methodology

Year

Proposes a simplified formula under variations of 2023

2023
for ultimate longitudinal strength and the reliability assessment
of stiffened panel structures. Chen et al.’s work utilizes composite

ultimate strength for a stiffened panel with a plate

of ultimate strength according to IACS falls within
reduces the ultimate strength and the prediction
imperfection modes, including column, torsional,
and local imperfection modes, with variations in

the ultimate strength range with severity levels

Effects of the geometrical imperfections on the The initial geometric imperfection significantly
materials with a Young’s modulus of 15 MPa in the panel sections.

aspect ratio of 6, which is approximately 7%

in the OTEC seawater tank design: Parametric 0.16 MPa results in the largest reduction in
Assessment of the stiffened panel performance The application of a transverse pressure of
initial geometric imperfections in three The assessment of ultimate longitudinal strength is conducted

the slenderness and span-to-bay ratio

through equations derived from composite column theory. To
evaluate structural integrity, a combination of the response-sur-
face method and first-order reliability method is employed to
determine safety indices and the probability of failure.

ranging from 25 to 100%

In a parallel investigation, Slater [89] directed his focus
Important remarks

toward the integration of composites into stiffened panels,

with particular consideration given to blast resistance.
Slater’s study utilizes GRP materials with the primary
objectives of mitigating fatigue cracking, reducing overall
weight, and enhancing fire resistance capabilities within
the stiffened panels. This dual research effort contributes
Assessment of the ultimate strength of stiffened

ultimate strength performances: A case study

geometrical aspects: Finite element approach

significantly to advancing the understanding and applica-

panels of ships considering uncertainties in

tion of composite materials in shipbuilding, facilitating

on the designed-steel stiffened panel

the development of lighter, stronger, and more resilient

marine structures.
study and sensitivity analysis
and simplified formula

4 Box girder
A box girder is a structural component of a ship shaped
Title

like a hollow rectangle or box, which serves to provide

strength to the ship’s structure, consisting of plates, stif-
feners, and transverse frames [90]. As a crucial structural
Hanif et al. [56]
Hanif et al. [51]

Lutfi et al. [79]

element in a ship, the strength of the box girder is a highly

critical subject in ship structure research. Therefore, inten-
Author

sive work has been conducted to analyze the behavior of

ship structures with box girder specimens subjected to
loads until reaching the ultimate strength limit. Box girders
of this type are also used in bridge structures, as reported
Analysis method

Numerical and

by Dowling [91], who conducted tests on box girders loaded

at their midsections. The test results showed that shear
analytical

Numeric

Numeric

loads and shear lag effects did not significantly impact

the ultimate strength reduction of the box girder.
Early research on box girders was conducted by Reckling
[92], who performed collapse tests on seven box girder models.
High strength

High strength

Ostapenko [93] conducted tests on three box girder models

Material

using bending, shear, and torque loads. Nishihara conducted

Steel
steel

steel

tests on eight different box girder models representing various

types of conventional ships such as tankers, bulk carriers, and
Lateral and transverse

container ships. Another study by Mansour and yang [94]

Table 8: Continued

involved experimental testing of two large-scale box girders

representing different types of ships. This research was
Compression

Compression

further extended with variations in loads, methods, analysis

Test type

pressure

approaches, and initial conditions of the box girder, including

factors like cracks and corrosion.
34  Muhamad Imam Firdaus and Ristiyanto Adiputra

Figure 34: Cylindrically curved plate panel of ship structures [84].

4.1 Analysis of box girder buckling matched the residual plastic strain distribution after
collapse and load removal.
Experimental testing was conducted by Gordo and Soares Experimental testing of box girders was conducted by
[95], using pure bending tests on a box girder to induce Gordo and Soares [97] with the aim of determining the
collapse with compressive loads on the upper flange and ultimate strength of box girders made from high-tensile
tensile loads on the bottom part. Mild steel material with a steel, specifically HTS 690 material with yield stress and
yield stress of 240 MPa and Young’s modulus of 210 GPa Young’s modulus values of 690 MPa and 200 GPa, respec-
was used in this study. To obtain realistic values, tension tively. In this research, Gordo and Soares [97] applied a
tests were initially conducted to determine the average pure bending moment load to the box girder, with tension
values. The experiments involved several loading cycles on the bottom part and compression on the upper part of
intended to release stress in the panel to very low levels. the box girder. The box was made of 4 mm thick plates, and
The resulting collapse is similar with the experiment con- the spacing between longitudinal stiffeners was 150 mm
ducted by Saad-Eldeen et al. [96] that can be observed in at the top, as illustrated in Figure 37. Using the same experi-
Figure 36. mental method as earlier, the experiments were conducted
Gordo and Soares [95] observed that the distribution of to apply multiple loading cycles aimed at obtaining stress
transverse strains at the center of the box girder indicated relief in the panels. From this study, Gordo and Soares [97]
a decrease in the effectiveness of the reinforced panels stated that the efficiency of HTS 690 material was quite
under tension conditions, especially in areas far from the impressive when used in box girders subjected to bending
main vertical plates (web) and when subjected to signifi- moments. Intense rotational and axial deformations occurred
cant moments. Strains detected in the middle of the panel in the middle frame, both on the upper and lower parts of the
were approximately half of what was observed at the box girder. The ultimate bending moment obtained was
corners. It is important to note that this strain pattern per- 1,526 kNm.
sisted even after the external load was removed. In addition, Gordo and Soares [98] experimental research on box
Gordo and Soares [95] stated that in panels subjected to girders continued, focusing more specifically on box gir-
compression, the strain distribution was less influenced by ders made of mild steel. Similar to the previous study, the
initial loading cycles, but significant plastic strains emerged box girders were tested using four loading cycles followed
during the collapse phase. The structural configuration after by unloading to identify and quantify the reduction in
 Deterioration and imperfection of the ship structural components  35

Figure 35: Double span/double bay model of stiffened curved plate [87].

Figure 36: Permanent deformation occurs when structure collapse [96].

36  Muhamad Imam Firdaus and Ristiyanto Adiputra

Figure 38: Graph of four loading cycles produced based on the infor-
mation obtained from the study by Gordo and Soares [98].

Figure 37: Geometry of Gordo and Soares’ experiment redrawn [98].

in vertical displacement. The energy dissipation in the early

residual stress resulting from welding. The results of the loading cycles of high-tensile steel box girders was consider-
four loading cycles are shown in Figure 38. ably lower compared to that of mild steel specimens. Gordo
Gordo and Soares [98] showed that the box girder per- attributed this difference to material properties related to
formed as expected, with a maximum applied load of 643.0 the ductile behavior of the two materials.
kN and a resulting displacement of 45.3 mm. Gordo and In a study conducted by Shi et al. [99], the focus was on
Soares [98] emphasized that residual stresses were an the dynamic response and ultimate strength of box girders
important variable in this experiment and could influence subjected to bending moments. Shi et al. [99] used box
the moment–curvature curves obtained. Furthermore, girders with the same scantling section and span length
Gordo and Soares [98] concluded that column slenderness as the experiments conducted by Paik [100] especially for
affected the type of structural failure, where a high MST-3 was selected as the object of analysis of that research.
column slenderness would lead to a sudden collapse, as The main reason for the selection is the used model is
seen in the experimental moment–curvature curves repre- convenient for showing the structural response differ-
sented by shedding patterns. ence between statics and dynamic bending moments.
In addition, Gordo and Soares [98] compared experi- The model has a 540 mm span length, with 3.05 mm
ments conducted on mild steel box girders and high tensile plating thickness, FB50 x 3.05 stiffener, and 180 mm stif-
steel box girders, noting differences in curvature, especially fener spacing. The boundary conditions of the model used

Figure 39: Boundary condition on Shi et al.’s research [99].

 Deterioration and imperfection of the ship structural components  37

are shown in Figure 39. Using FEMs in Abaqus, Shi et al. [99] subjected to torsional loads, and the analysis was per-
determined the dynamic ultimate bending moment and eva- formed using the FEM using the ABAQUS software. The
luation criteria. The study also considered the influence of various crack types considered in the study are illustrated
model length, plate thickness, and loading duration. in Figure 40. It was assumed that all cracks penetrate the
In this study, it was concluded that under higher applied thickness of the box girder, have no friction between their
dynamic moments, the vibration period of the box girder surfaces, and do not exhibit crack propagation.
increases due to local buckling and plastic deformation. In Shi and Wang [101] conclude that when the ratio of crack
dynamic loading scenarios, several sections exhibit signifi- length to cross-sectional width (2c/b) is less than 0.1, cracks will
cant deformation simultaneously, unlike in static loading have a small effect on the ultimate strength of the box girder.
cases where usually only one section fails. In addition, Shi However, a significant reduction in the ultimate strength of the
et al. [99] concluded that the length of the model has a sig- box girder will occur when the ratio of crack length to cross-
nificant influence on the dynamic ultimate moment when sectional width of the box girder is greater than 0.1. Shi and
used for shorter durations. Longer models require a higher Wang [101] also stated that when the crack size is less than 2c/b
dynamic moment to reach a dynamic state. = 0.3, edge cracks will cause the greatest reduction in torsional
ultimate strength. On the other hand, double-edge cracks will
result in the highest reduction in torsional ultimate strength
4.2 Analysis of box girder with crack when 2c/b is greater than 0.3. Shi and Wang [101] formulated
condition equations to predict the ultimate strength of cracked box gir-
ders, which have been verified with FEA.
Research on box girders with initial cracks was conducted In another study on box girders with cracks, the focus
by Shi and Wang [101]. In this study, box girders were was on the variable of crack inclination in the box girder,

Figure 40: Crack type configuration; (a) Center crack, (b) Section A-A, (c) Edge crack, and (d) Double edge crack [101].
38  Muhamad Imam Firdaus and Ristiyanto Adiputra

as conducted by Ao and Wang [102]. This research employed

the FEM by ABAQUS with variations in crack inclination
located in the middle of the box girder plate and subjected
to torsional loads at both ends of the box girder. In addition,
Ao and Wang [102] also considered crack size as a variable.
The results of Ao and Wang [102] research are shown
in Figure 41. It can be seen that when the ratio of crack
length to box girder section width (l/b) is less than 0.2,
changes in the crack angle do not significantly impact the
reduction in the ultimate strength of the box girder. How-
ever, a change in angle has a significant effect on ultimate
strength when the ratio of crack length to box girder width
exceeds 0.5. Ao and Wang [102] also stated that there is a
linear relationship between ultimate strength and crack
size, regardless of the angle variation. Meanwhile, there
is a trigonometric relationship between ultimate strength
and angle variation, especially when the ratio of crack
length to box girder width is greater than 0.2.
Figure 42: Experimental setup on Saad-Eldeen et al.’s research [96].

4.3 Analysis of box girder with corrosion the results of this research are referred to as the “initial cor-
condition roded box girder.”
Saad-Eldeen et al. [104] conducted experimental testing
Saad-Eldeen et al. [96,103] conducted a comprehensive study on box girders with a focus on moderate corrosion levels.
on corroded box girders. The initial research involved Furthermore, this research aimed to study the compressive
experiments on box girders subjected to four-point bend- failure strength of corroded box girders when subjected to
ing with a zero to initial corrosion condition, as depicted in four-point loading, resulting in a constant vertical bending
Figure 42. In this study, it was found that the maximum ver- moment. To induce corrosion, the box girder specimens
tical bending moment recorded on the test specimen was were exposed to Baltic seawater without any corrosion pro-
580.55 kN, with a displacement of 28.57 mm. Subsequently, tection. The water temperature and oxygen depolarization

Figure 41: Relationship of ultimate strength and (a) crack length and (b) crack angle. Produced based on the information obtained from the study by
Ao and Wang [102].
 Deterioration and imperfection of the ship structural components  39

process were increased to accelerate the corrosion rate. This compared with existing equations, and Saad-Eldeen et al.
resulted in corrosion that reduced the weight of the box [104] suggested that there should be consideration of the
girder specimens by approximately 42 kg or 15% of their direct effect of corrosion on the mechanical properties of
original weight. It was determined that the corrosion on the material.
the specimens was equivalent to 17.9 years of natural In a subsequent study, Saad-Eldeen et al. [105] con-
corrosion. ducted another assessment of the strength of severely
The testing was conducted in two loading cycles, where corroded box girders. They used the FEM with nonlinear
the first load was used to reduce residual stress, and the analysis and applied uniform vertical bending moment
second load was the maximum load until collapse occurred loads. Two corrosion degradation models were employed:
in the box girder. Saad-Eldeen et al.’s [104] research findings one based on average general corrosion with thickness
indicated that the ultimate strength value of moderately reduction and the other using thickness measurements at
corroded box girders decreased more significantly, by approxi- node locations in the finite element model. Initial imper-
mately 27.79%, compared to box girders that only experi- fections in the finite element model were based on the
enced initial corrosion. The experimental results were also initial imperfections by altering the vertical node positions

Figure 43: Existing stress–strain model [105]. (a) Elastic-plastic (EP), (b) elasto-plastic modified (EPM), (c) elasto-plastic modified 1 (EPM1), and
(d) elasto-plastic modified 2 (EPM2).
40  Muhamad Imam Firdaus and Ristiyanto Adiputra

in the finite element without affecting or adding stress to with existing equations used in methods for calculating ulti-
the specimen. The stress–strain relationship using existing mate bending moments, as shown in Eqs. (31)–(35).
equations and the experimental results from corroded box Faulkner and Sadden [107]:
girders were also developed as shown in Figure 43. 2
⎡ ⎛ σu ⎞ ⎤
The research results indicate a notable difference in Mu = 1.15zσy ⎢− 0.1 + 1.4465 − 0.3465 (31)
⎜ ⎟
⎥,
ultimate bending behavior, ranging from 10 to 14.4%, ⎣ ⎝ σy ⎠ ⎦
between the outcomes obtained through FEA using the
average corrosion thickness model and the real corrosion Viiner [108]:
thickness model. This discrepancy underscores variations Mu = ∝Z σu, (32)
in the finite element model for deteriorated structures. In
Frieze and Lin [109]:
addition, among the available stress–strain models, Model
2
EPM2 was proved to be a suitable stress–strain model for Mu σu ⎛ σu ⎞
= 0.172 + 1.548 − 0.368⎜ ⎟ . (33)
predicting curvature behavior and ultimate bending moment Mp σy ⎝ σy ⎠
based on the available data from both the moment–curvature
relationships in average corrosion and real corrosion Caldwell modified by Paik and Mansour [110]:
scenarios. ′
Mu = − AD (D − g )σuD − AB gσyB − AB′ (g − D)σyB
Furthermore, in a subsequent experimental study,
As (34)
Saad-Eldeen et al. [90] compared the degrees of severity − [(D − g )2σuS + g 2σyS,
in box girders with initial corrosion, moderate corrosion, D
and severe corrosion. By cross-referencing with the corro- Paik and Mansour [110]:
sion data used in Saad-Eldeen et al.’s [96] research, it was AS
determined that the box girder specimens used in this Mu = − AD (D − g )σuD − (D − H )(D + H − 2g1)σuS
D
research were equivalent to 0.2 years of corrosion for A′ B
initial corroded specimens, 17.9 years for moderately cor- − AB gσyB + (g − D B)[D BσuS − (H − D B)σyS] (35)
H 1
roded specimens, and 23.3 years for severely corroded box As H
girder specimens. − x[(2H − 3g1)σuS − (H − 3g1)σuS].
3D
In the subsequent research conducted by Saad-Eldeen
where AD, AB, AB′ , and AS are the total sectional area of the
et al. [106], FEA was performed, and the results were com-
deck, outer bottom, inner bottom, and side shell, respec-
pared with the experimental findings from their previous
tively; D is the hull depth; g is the neutral axis position
research. The same finite element model used in the pre-
above the baseline in the sagging condition; H is the depth
vious study was employed. FEA was carried out, along with
calculations using existing approaches and equations. The of the hull section of the linear elastic state; σuD and σuS are
stress–strain relationship was developed by taking into the ultimate buckling strength for deck and side, respec-
account residual stresses and the effects of corrosion. tively; σyB, σyB
′ , and σyS are the yield strength of outer
Initial imperfections were added to the model, with respec- bottom, inner bottom, and side, respectively.
tive depths of imperfection for the initially corroded, mod- By using Eqs. 29–(33), the values of the ultimate
erately corroded, and severely corroded cases being 1.72, bending moment for box girders under initially, moder-
3.94, and 3.92 mm, respectively. Mechanical properties under- ately, and severely corroded conditions can be calculated
went degradation due to corrosion deterioration, particu- and compared, as shown in Table 9. The highest error is
larly in the reduced value of the Young’s modulus, which evaluated at 33.4 and 33%, which are the comparisons
decreased by approximately 54.3%. between calculations using the Faulkner and Sadden
The results obtained using various stress–strain rela- [107] equation and experimental results for severely cor-
tionships, such as elastic–plastic and elastoplastic, consid- roded box girders and experimental results with calcu-
ering hardening and weld toe effects, were compared with lations using the Frieze and Lin equation [109] for moder-
the ultimate bending moment results from experiments. It ately corroded box girders. The highest agreement is seen in
can be concluded that the results obtained from nonlinear the calculation using the Caldwell equation for initially cor-
finite element analysis with the utilized model are reason- roded box girders, with an error of only 0.9%. It is also
ably consistent with the experimental results, with dif- evident that the results of the analysis for the three types
ferences or errors ranging from approximately 1–10% of box girders show an average change of 13.07%, with a
compared to the experimental results at each corrosion 5.7% change in comparison between initially corroded, mod-
level. In addition, the comparison study was conducted erately corroded, and severely corroded box girders.
 Deterioration and imperfection of the ship structural components  41

Tabel 9: Calculation result using existing formula and based on corrosion severity [106]

Ultimate bending moment Initially corroded Moderately corroded Severely corroded

Mu Mu (%) Mu Mu (%) Mu Mu(%)

Experimental result 568.9 — 382 — 201.1 —

Faulkner and sadden 687.5 20.8 382.6 0.16 268.1 33.4
Viner 558.2 1.9 305.8 20 213.7 6.3
Frieze and Lin 486.4 14.5 256.1 33 186.1 7.4
Caldwell modified, Paik and Mansour 563.7 0.9 268.9 29.6 199.1 0.9
Paik and Mansour 559.2 1.7 295.8 22.6 236.6 17.7

4.4 Analysis on box girder under combined by coupled damage. However, the difference is too small so
crack and corrosion condition it could be neglected.
Further research has been conducted to ascertain the
Following previous separate studies on the strength ana- strength of other ship structures, such as box girders,
lysis of corroded box girders with pit corrosion and cracks, through the application of composite materials, as evi-
Li et al. [111] conducted research by finite element analysis denced by the study conducted by Soares et al. [112]. Soares
using the Abaqus software to analyze box girders subjected et al. [112] conducted experimental and numerical investi-
to vertical bending moments. Li et al. [111] also considered gations into the behavior of box girder structures using glass
variables such as the bending mode, crack length, crack fiber-reinforced plastic (GFRP) materials. Specifically, the
angle, and crack positions on both the plate and the posi- material employed was WR/polyester GRP. In the experi-
tion indicated on the cross-section of the specimen, and the mental phase, the study utilized four-point bending tests
diameter of pitting corrosion damage. to assess the material strength of the structure. Meanwhile,
From this research, Li et al. [111] concluded that the numerical methods employed nonlinear finite element
variation in pitting corrosion diameter significantly affects codes to predict the responses obtained from the box girder.
ultimate strength. Similarly, as in the study of pitting cor- The research revealed that the comparison between numer-
rosion, the total volume loss due to corrosion plays a sig- ical results and experimental data demonstrated consistent
nificant role in the reduction of ultimate strength. It is also findings, particularly regarding buckling moment, buckling
observed that the presence of bottom or side cracks has an mode, and the prebuckling and postbuckling behavior of the
impact on the reduction. For bottom cracks, the ultimate box girder structure.
bending moment of the box girder in the hogging condition As one of the structural components of a ship, the box
is much smaller than in the sagging condition. This is due girder also plays a significant role in the overall strength of
to buckling that occurs at the bottom, enhanced by the the vessel. However, the diversity of research on box gir-
presence of cracks. The influence of other crack para- ders is relatively limited, indicating the need for further
meters is then considered. It is evident that the projected studies to advance ship structural development. A sum-
length of transverse and inclined cracks and the actual mary of the discussed research can be found in Table 10.
length of longitudinal cracks located below the stiffener
have a significant impact on the reduction in ultimate
strength. However, longitudinal cracks have a smaller 5 Conclusion
effect and almost no impact on the strength of the box
girder. This article provides a comprehensive summary of research
Regarding cracks on the sides of the box girder, the on ship structures, including plates, stiffened panels, and box
reduction in ultimate strength in the sagging condition is girders. It encompasses various research methods, such as
much greater than in the hogging condition. When cracks experiments, FEMs, and analytical calculation methods. The
are in vertical or inclined positions, the projected length of article also discusses the different loading conditions applied
the cracks has a significant impact on the strength reduc- to the tested specimens. In addition, it considers the condition
tion of the box girder. When cracks shift toward the deck, of ship structures, including initial states, corrosion, and the
the reduction becomes more significant. In general, it was presence of cracks. Overall, the article demonstrates how
shown that the reduction caused by either cracking or pit- these various testing conditions can significantly reduce the
ting corrosion is slightly larger than the reduction caused ultimate strength of ship structures.
 42

Table 10: Milestone studies related to box girder analysis



Test type Material Analysis method Author Title Important remarks Year

Pure bending Steel box Experimental Gordo and Experimental evaluation of the ultimate bending The initial experiments on the box girder, which used 1996
moment Soares [95] moment of a box girder bending moments, focused on residual stress and
ultimate bending moment
Four point bending GFRP Experiment and FEM Soares An experimental and numerical study on GFRP box Analysis conduct on GFRP material made box girder 2007
et al. [112] girder under pure bending
Pure bending Mild steel box Experimental Gordo and Experimental evaluation of the behavior of a mild Experiments on mild steel box girder and its comparison 2008
moment Soares [97] steel box girder under bending moment with the ultimate strength of high tensile steel box girder
Pure bending High tensile Experimental Gordo and Tests on ultimate strength of hull box girders made of Experimental research on box girder with HTS 690 2009
moment steel Soares [97] high tensile steel J.M. material, involving three different span variations
Vertical bending Steel Experimental Saad-Eldeen Experimental assessment of the ultimate strength of a Vertical bending moment loading on a box girder 2010
moment load et al. [103] box girder subjected to four-point bending moment
Vertical bending Steel Experimental Saad-Eldeen Compressive strength assessment of a moderately Loading under moderate corrosion conditions on a box 2011
Muhamad Imam Firdaus and Ristiyanto Adiputra

moment load et al. [96] corroded box girder girder, equivalent to a ship that has been in service for
17.9 years
Torsional loading Steel FEM Shi and Residual ultimate strength of cracked box girders The finite element analysis used the software Abaqus, 2012
Wang [101] under torsional loading with the parameters of crack size and crack location. The
ultimate strength reduction formula is also expressed as
a function of these parameters
Uniform bending Steel Experimental Saad-Eldeen Experimental assessment of corroded steel box- To compare the experimental results of three different 2013
moment et al. [90] girders subjected to uniform bending severity levels of corrosion on box girders: initial,
moderate, and severe
Vertical Bending Steel Experimental, FEM, Saad-Eldeen Effect of corrosion severity on the ultimate strength of To compare the results from numerical, experimental, 2013
Moment numerical et al. [106] a steel box girder and FEMs for three box girders with initial, moderate, and
severe conditions
Uniform vertical Steel FEM Saad-Eldeen Strength assessment of a severely corroded box Two types of corrosion reduction were used: average 2014
bending moment et al. [105] girder subjected to bending moment general corrosion thickness reduction and the real
thickness of the corroded plates
Torsional load Steel FEM Ao and Ultimate strength of box girders with incline cracks Finite element analysis of the box girder with crack 2015
Wang [102] conditions, followed by crack parameters
Vertical Steel FEM Li et al. [111] Residual ultimate strength of stiffened box girder with In this study, an analysis is conducted on box girders 2021a
coupled damage of pitting corrosion and a crack under conditions of cracks, corrosion, and their
under vertical bending moment combinations, along with various parameters
Bending moment Steel Experiment and FEM Shi et al. [99] Analysis of dynamic response and ultimate strength The analysis was conducted by also observing the 2023
for box girder under bending moment dynamic response of the box girder using six models with
variations in plate thickness and length
 Deterioration and imperfection of the ship structural components  43

Numerous investigations into ship plates and their load- collaboration with the Research Center for Hydrodynamics
testing capabilities have highlighted the considerable impact Technology, National Research and Innovation Agency.
of several types of initial imperfections. These initial deflec-
tion and crack have been altered to include variations in Funding information: Authors state no funding involved.
angles, dimensions, and potential locations on ship plates. In
addition, researchers have taken into account other factors, Author contributions: All authors have accepted responsibility
like pit corrosion, affecting one or both sides of the plate, for the entire content of this manuscript and consented to its
aligning with the outcomes of corrosion-induced degrada- submission to the journal, reviewed all the results and
tion. To gauge the unpredictability of these initial imperfec- approved the final version of the manuscript. MIF: formal
tions, Monte Carlo simulations were employed, enabling the analysis, writing – original draft, writing – review and editing,
assessment of the distribution of reductions in ultimate visualization, project management. RA: conceptualization,
strength. methodology, formal analysis,writing – original draft, writing
In the case of stiffened panels, research has been more – review & editing, visualization.
diverse and comprehensive, encompassing experimental,
analytical, and numerical studies. Variations in numerical Conflict of interest: Authors state no conflict of interest.
methods have led to different approaches in model config-
urations, boundary conditions, and options for modeling Data availability statement: The authors declare that the
transverse stiffeners in numerical analyses. Similar to ship data supporting the findings of this study are available
plates, the introduction of initial imperfections influences within the article.
the reduction in ultimate strength values. Moreover, dif-
ferent types of initial imperfections also affect the postbuck-
ling behavior observed in stress–strain graphs, including
the position of the ultimate strength and strain distribution References
differences. The development of numerical methods has
provided variations in modeling, types, and magnitudes of [1] Zheng K. Analysis of the freight of handysize ships on the global
bulk shipping market. The Maritime Commons: Digital Repository
buckling modes, cracks, and corrosion. In the realm of stif-
of the World Maritime University; 2018.
fened panel studies, numerical methods have been devel- [2] Okubo M, Kuwahara T. Prospects for marine diesel engine emis-
oped to extract data from random variations, allowing for sioncontrol. New Technologies for Emission Control in Marine
the creation of analytical formulas that enhance the effi- Diesel Engines. 1st edn. New York: Springer; 2020. p. 211–66.
ciency of predicting ultimate strength values. [3] Prabowo AR, Sohn JM, Putranto T. Crashworthiness performance
of stiffened bottom tank structure subjected to impact loading
Various studies have been conducted on box girders,
conditions: Ship-rock interaction. Curved Layer Struct.
with the majority focusing on modifications in geometry 2019;6(1):245–58.
and materials. However, the influence of initial imperfec- [4] Prabowo AR, Sohn JM, Bae DM, Zakki AF, Harsritanto BI.
tions in box girders, specifically cracks and corrosion, Investigating crashworthy single and double skin structures
remains relatively underexplored despite their prevalence against accidental ship-to-ship interaction. Curved Layer Struct.
in research on box girders under load. To date, the effects 2018;5(1):180–9.
[5] Fajri A, Prabowo AR, Muhayat N. Assessment of ship structure
of initial deflection, as seen in ship plates and stiffened
under fatigue loading: FE benchmarking and extended perfor-
panels, have not been identified in the context of box gir- mance analysis. Curved Layer Struct. 2022;9(1):163–86.
ders. While there has been extensive research on plates [6] Mursid O, Tuswan T, Samuel S, Trimulyono A, Yudo H, Huda N,
and stiffened panels with numerous studies and data col- et al. Effect of pitting corrosion position to the strength of ship
lection efforts as mentioned earlier, there is a notable gap in bottom plate in grounding incident. Curved Layer Struct.
2023;10(1):20220199.
the number of studies and variations in parameters when it
[7] Prabowo AR, Sohn JM, Bae DM, Cho JH. Performance assessment
comes to ship box girder structures. To advance and enhance on a variety of double side structure during collision interaction
our knowledge of ship structures comprehensively, future with other ship. Curved Layer Struct. 2017;4(1):255–71.
research efforts should focus on conducting more compre- [8] Brubaker D. Marine pollution and international law: principles
hensive studies on ship box girder structures. This will help and practice. London: Belhaven Press; 1993.
[9] Wei C, Wang G, Cridland M, Olson DL, Liu S. Corrosion protection
bridge the existing gap and contribute to a deeper under-
of ships. Handbook of Environmental Degradation of Materials.
standing of ship structural integrity as a whole.
Norwich, New York: William Andrew Publishing; 2018. p. 533–57.
[10] Nwigwe T, Kiyokazu M. Statistical analysis of bulk carrier accident
Acknowledgements: This work is part of the research activity from 2011 to 2020. TransNav: Int J Mar Navig Saf Sea Transp.
conducted by Surabaya Merchant Marine Polytechnic in 2022;16:1.
44  Muhamad Imam Firdaus and Ristiyanto Adiputra

[11] Du K, Monios J, Wang Y. Green port strategies in China. Green [31] Li D, Feng L, Xiao W. A study on residual ultimate strength of steel
PortsInland and Seaside Sustainable Transportation Strategies. unstiffened plate with a crack. Appl Ocean Res.
Cambridge, MA: Elsevier; 2019. p. 211–29. 2020;103(July):102336. doi: 10.1016/j.apor.2020.102336.
[12] Kim DK, Lim HL, Kim MS, Hwang OJ, Park KS. An empirical for- [32] Li D, Chen Z, Li J, Yi J. Ultimate strength assessment of ship hull
mulation for predicting the ultimate strength of stiffened panels plate with multiple cracks under axial compression using artificial
subjected to longitudinal compression. Ocean Eng. neural networks. Ocean Eng. 2022;263(Sep):112438. doi: 10.1016/j.
2017;140:270–80. oceaneng.2022.112438.
[13] Xu MC, Song ZJ, Zhang BW, Pan J. Empirical formula for predicting [33] Zhang Y, Huang Y, Zhang Q, Liu G. Ultimate strength of hull
ultimate strength of stiffened panel of ship structure under structural plate with pitting corrosion damnification under com-
combined longitudinal compression and lateral loads. Ocean Eng. bined loading. Ocean Eng. 2016;116:273–85. doi: 10.1016/j.
2018;162:161–75. oceaneng.2016.02.039.
[14] Xu MC, Yanagihara D, Fujikubo M, Soares CG. Influence of [34] Silva JE, Garbatov Y, Guedes Soares C. Ultimate strength assess-
boundary conditions on the collapse behavior of stiffened panels ment of rectangular steel plates subjected to a random localised
under combined loads. Mar Struct. 2013;34:205–25. corrosion degradation. Eng Struct. 2013;52:295–305. doi: 10.1016/
[15] Hanif MI, Adiputra R, Prabowo AR, Muhayat N, Marta AS, Huda N, j.engstruct.2013.02.013.
et al. Effect of design parameters on the ultimate strength and [35] Zhang Y, Huang Y, Wei Y. Ultimate strength experiment of hull
collapse behaviour of stiffened panels. J Appl Eng Sci. structural plate with pitting corrosion damage under uniaxial
2023;21(3):940–56. compression. Ocean Eng. 2017;130(Dec 2015):103–14. doi: 10.1016/
[16] Adiputra R, Yoshikawa T, Erwandi E. Reliability-based assessment j.oceaneng.2016.11.065.
of ship hull girder ultimate strength. Curved Layer Struct. [36] Ahmmad MM, Sumi Y. Strength and deformability of corroded
2023;10(1):20220189. steel plates under quasi-static tensile load. J Mar Sci Technol.
[17] Denev Y. Corrosion waste impact on ship girder strength through 2010;15(1):1–15. doi: 10.1007/s00773-009-0066-1.
life cycle. TransNav, Int J Mar Navig Saf od Sea Transp. [37] Nakai T, Matsushita H, Yamamoto N. Effect of pitting corrosion on
2024;18(2):415–8. the ultimate strength of steel plates subjected to in-plane com-
[18] Cui W, Wang Y, Pedersen PT. Strength of ship plates under pression and bending. J Mar Sci Technol. 2006;11(1):52–64. doi: 10.
combined loading. Mar Struct. 2002;15(1):75–97. doi: 10.1016/ 1007/s00773-005-0203-4.
S0951-8339(01)00009-0. [38] Khedmati, MR, Nouri, ZHME, Roshanali MM. A comparative
[19] Guedes Soares C. Design equation for ship plate elements under computational investigation on the effects of randomly distrib-
uniaxial compression. J Constr Steel Res. 1992;22(2):99–114. doi: uted general corrosion on the post-buckling behaviour of uni-
10.1016/0143-974X(92)90010-C. axially loaded plates. J Mech Sci Technol. 2012;26(3):767–83. doi:
[20] Zhang S. A review and study on ultimate strength of steel plates 10.1007/s12206-011-1222-1.
and stiffened panels in axial compression. Ships Offshore Struct. [39] Zhang J, Shi XH, Soares CG. Experimental analysis of residual
2016;11(1):81–91. doi: 10.1080/17445302.2014.992610. ultimate strength of stiffened panels with pitting corrosion under
[21] Zhang S, Khan I. Buckling and ultimate capability of plates and compression. Eng Struct. 2017;152:70–86.
stiffened panels in axial compression. Mar Struct. [40] Rahbar-Ranji A. Ultimate strength of corroded steel plates with
2009;22(4):791–808. doi: 10.1016/j.marstruc.2009.09.001. irregular surfaces under in-plane compression. Ocean Eng.
[22] Box TA. Practical Treatise on the Strength of Materials. London: 2012;54:261–9. doi: 10.1016/j.oceaneng.2012.07.030.
Spon; 1883. [41] Fujita Y, Nomoto T, Niho O. Ultimate strength of stiffened plates
[23] Karman Tv. Die mittragende Breite (The effective width). Berlin: subjected to compression. J Soc Nav Archit Jpn. 1977;141:190–7.
Springer; 1924. [42] Fujikubo M, Yao T, Khedmati MR. Estimation of Ultimate Strength
[24] Schnadel G. Proceedings 3rd International Congress on Applied of Ship Bottom Plating under Combined Transverse Thrust and
Mechanics. 1930, Stockholm: 1930. Lateral Pressure. J Soc Nav Archit Jpn. 1999;186:621–30. doi: 10.
[25] Frankland JM. The strength of ship plating under edge com- 2534/jjasnaoe1968.1999.186_621.
pression. US EMM Report. Vol. 469; 1940. [43] Mouritz AP, Gellert E, Burchill P, Challis K. Review of advanced
[26] Faulkner D. A review of effective plating for the analysis of stif- composite structures for naval ships and submarines. Compos
fened plating in bending and compression. J Ship Res. Struct. 2001;53(1):21–42.
1975;19(1):1–17. [44] Ma S, Mahfuz H. Finite element simulation of composite ship
[27] Paik JK, Pedersen PT. A simplified method for predicting ultimate structures with fluid structure interaction. Ocean Eng.
compressive strength of ship panels. Int Shipbuild Prog. 2012;52:52–9.
1996;43(434):139–57. [45] Mouhat O, Khamlichi A, Limam A. Reliability assessment of
[28] Paik JK, Kumar YVS, Lee JM. Ultimate strength of cracked plate buckling for stiffened panels by considering localized geometric
elements under axial compression or tension. Thin-Walled Struct. imperfections. Aust J Basic Appl Sci. 2013;7(8):616–24.
2005;43(2):237–72. doi: 10.1016/j.tws.2004.07.010] [46] Xu MC, Soares CG. Comparisons of calculations with experiments
[29] Paik JK. Residual ultimate strength of steel plates with longitu- on the ultimate strength of wide stiffened panels. Mar Struct.
dinal cracks under axial compression-experiments. Ocean Eng. 2013;31:82–101.
2008;35(17–18):1775–83. doi: 10.1016/j.oceaneng.2008.08.012. [47] Xu MC, Soares CG. Experimental study on the collapse strength of
[30] Paik JK. Residual ultimate strength of steel plates with longitu- wide stiffened panels. Mar Struct. 2013;30:33–62.
dinal cracks under axial compression-Nonlinear finite element [48] Hu SZ, Chen Q, Pegg N, Zimmerman TJ. Ultimate collapse tests of
method investigations. Ocean Eng. 2009;36(3–4):266–76. doi: 10. stiffened-plate ship structural units. Mar Struct.
1016/j.oceaneng.2008.12.001. 1997;10(8–10):587–610.
 Deterioration and imperfection of the ship structural components  45

[49] Lin YT. Ship longitudinal strength modelling. Doctoral disserta- [67] Li S, Kim DK, Benson S. The influence of residual stress on the
tion. Ann Arbor, Michigan: ProQuest; 1985. ultimate strength of longitudinally compressed stiffened panels.
[50] Paik, JK, Thayamballi, AK. An empirical formulation for predicting Ocean Eng. 2021;231:108839.
the ultimate compressive strength of stiffened panels. ISOPE [68] Smith CS, Anderson N. Strength of stiffened plating under com-
International Ocean and Polar Engineering Conference; 1997 bined compression and lateral pressure. Trans RINA.
May. ISOPE; 1997. p. ISOPE-I. 1992;13(1):131–48.
[51] Hanif MI, Adiputra R, Prabowo AR, Yamada Y, Firdaus N. [69] Yao T, Fujikubo M. Buckling and ultimate strength of ship and
Assessment of the ultimate strength of stiffened panels of ships ship-like floating structures. Oxford: Butterworth-
considering uncertainties in geometrical aspects: Finite element Heinemann; 2016.
approach and simplified formula. Ocean Eng. 2023;286:115522. [70] Paik JK, Thayamballi AK. Ultimate limit state design of steel-plated
[52] AbuBakar A, Dow RS. Simulation of ship grounding damage using structures. Hoboken, New Jersey: John Wiley & Sons; 2003.
the finite element method. Int J Solids Struct. 2013;50(5):623–36. [71] Gordo JM, Soares G. Approximate load shortening curves for
[53] Shi XH, Zhang J, Soares CG. Numerical assessment of experiments stiffened plates under uniaxial compression. Integr Offshore
on the ultimate strength of stiffened panels with pitting corrosion Struct. 1993;5:189–211.
under compression. Thin-Walled Struct. 2018;133:52–70. [72] Yi MS, Hyun CM, Paik JK. An empirical formulation for predicting
[54] Rigo P, Sarghiuta R, Estefen S, Lehmann E, Otelea SC, welding-induced biaxial compressive residual stresses on steel
Pasqualino I, et al. Sensitivity analysis on ultimate strength of stiffened plate structures and its application to thermal plate
aluminum stiffened panels. Mar Struct. 2003;16(6):437–68. buckling prevention. Ships Offshore Struct. 2019;14(Suppl
[55] Anyfantis KN. Ultimate strength of stiffened panels subjected to 1):18–33.
non-uniform thrust. Int J Nav Archit Ocean Eng. 2020;12:325–42. [73] Grondin GY, Elwi AE, Cheng JJ. Buckling of stiffened steel
[56] Hanif MI, Adiputra R, Prabowo AR, Muhayat N, Erwandi E, Huda N. plates – a parametric study. J Constr Steel Res. 1999;50(2):
Effects of the geometrical imperfections on the ultimate strength 151–75.
performances: A case study on the designed-steel stiffened panel. [74] Khan I, Zhang S. Effects of welding-induced residual stress on
Procedia Struct Integr. 2023;47:125–32. ultimate strength of plates and stiffened panels. Ships Offshore
[57] Fujikubo M, Yao T, Khedmati MR, Harada M, Yanagihara D. Struct. 2011;6(4):297–309.
Estimation of ultimate strength of continuous stiffened panel [75] Bayatfar A, Khedmati MR, Rigo P. Residual ultimate strength of
under combined transverse thrust and lateral pressure Part 1: cracked steel unstiffened and stiffened plates under longitudinal
Continuous plate. Mar Struct. 2005;18(5–6):383–410. compression. Thin-Walled Struct. 2014;84:378–92.
[58] International Association of Classification Societies [76] Dexter RJ, Pilarski PJ. Crack propagation in welded stiffened
(IACS). Common Structural Rules for Bulk Carriers and Oil panels. J Constr Steel Res. 2002;58(5–8):1081–1102.
Tankers; 2022. [77] Cui C, Yang P, Li C, Xia T. Ultimate strength characteristics of
[59] Li S, Georgiadis DG, Samuelides MS. A comparison of geometric cracked stiffened plates subjected to uniaxial compression. Thin-
imperfection models for collapse analysis of ship-type stiffened Walled Struct. 2017;113:27–38.
plated grillages. Eng Struct. 2022;250:113480. [78] Yu YZ, Feng GQ, Li CF, Ren HL. Experimental and numerical
[60] Smith CS, Davidson PC, Chapman JC. Strength and stiffness of investigation on the ultimate strength of stiffened plates with
Ships’plating under in-plane compression and tension. R Inst Nav scanned initial geometrical imperfection. China Ocean Eng.
Archit Trans. 1988;130:1–20. 2019;33:446–58.
[61] Georgiadis DG, Samuelides MS, Li S, Kim DK, Benson S. Influence [79] Lutfi YM, Adiputra R, Prabowo AR, Utsunomiya T, Erwandi E,
of stochastic geometric imperfection on the ultimate strength of Muhayat N. Assessment of the stiffened panel performance in the
stiffened panel in compression. Developments in the Analysis and OTEC seawater tank design: Parametric study and sensitivity
Design of Marine Structures: Proceedings of the 8th International analysis. Theor Appl Mech Lett. 2023;13(4):100452.
Conference on Marine Structures (MARSTRUCT 2021, 7-9 June [80] Platypodis EL, Anyfantis KN. On the modeling of ship stiffened
2021, Trondheim, Norway). Vol. 95; CRC Press; 2021. panels subjected to uniform pressure loads. Appl Mech.
[62] Gandhi PR, Doshi K, Vhanmane S. Ultimate strength analysis of 2022;3(1):125–43.
stiffened plates with initial imperfections. Int J Innov Res Dev. [81] Maeno Y, Yamaguchi H, Fujii Y, Yao T. Buckling/plastic collapse
2012;1(10):190–206. behaviour and strength of bilge circle and its contribution to ultimate
[63] Rizzo NA, Caire M. Ultimate strength formulations for FPSO stif- longitudinal strength of ship’s hull girder. In ISOPE International
fened panels under combined compression and shear with initial Ocean and Polar Engineering Conference. ISOPE; 2004.
imperfections and damage. Lat Am J Solids Struct. 2018;15:e73. [82] Yumura K, Katsura S, Ijima K, Yao T. Simulation of buckling col-
[64] Feng L, Hu L, Chen X, Shi H. A parametric study on effects of lapse behaviour of cylindrically curved plates under axial com-
pitting corrosion on stiffened panels’ ultimate strength. Int J Nav pression. In Proceedings of TEAM Conference; 2005.
Archit Ocean Eng. 2020;12:699–710. [83] Kwen YW, Park YI, Paik JK, Lee JM. Buckling and ultimate strength
[65] Feng GQ, Hu BN, Ren HL. Reliability of the ultimate strength of characteristics for ship curved plate structures. In Proceedings of
ship stiffened panel subjected to random corrosion degradation. the Annual Autumn Meeting. Sancheong Korea: SNAK; 2004.
China Ocean Eng. 2017;31:11–8. p. 351–6.
[66] Wang G, Lee AK, Ivanov L, Lynch TJ, Serratella C, Basu R. A sta- [84] Park JS, Paik JK, Seo JK. Numerical investigation and development
tistical investigation of time-variant hull girder strength of aging of design formula for cylindrically curved plates on ships and
ships and coating life. Mar Struct. 2008;21(2–3):240–56. offshore structures. Thin-Walled Struct. 2018;132:93–110.
46  Muhamad Imam Firdaus and Ristiyanto Adiputra

[85] Oh YC, Kim KT, Ko JY. Investigation for Collapse Mode of Stiffened [100] Paik JK. Ultimate limit state analysis and design of plated struc-
Curved Plate with Tee Shaped Stiffeners. J Korean Soc Mar tures. Chichester, UK: John Wiley & Sons; 2018.
Environ & Saf. 2011;17(3):295–300. [101] Shi GJ, Wang DY. Residual ultimate strength of cracked box gir-
[86] Seo JK, Song CH, Park JS, Paik JK. Nonlinear structural behaviour ders under torsional loading. Ocean Eng. 2012;43:102–12.
and design formulae for calculating the ultimate strength of [102] Ao L, Wang D. Ultimate strength of box girders with incline
stiffened curved plates under axial compression. Thin-Walled cracks; 2015. p. 1–6.
Struct. 2016;107:1–17. [103] Saad-Eldeen S, Garbatov Y. Experimental assessment of the ulti-
[87] Park JS, Ha YC, Seo JK. Estimation of buckling and ultimate col- mate strength of a box girder subjected to four-point bending
lapse behaviour of stiffened curved plates under compressive moment. 11th International Symposium on Practical Design of
load. J Ocean Eng Technol. 2020;34(1):37–45. Ships and Other Floating Structures, PRADS 2010. Vol. 2; 2010. p.
[88] Chen NZ, Sun HH, Soares CG. Reliability analysis of a ship hull in 1134–43.
composite material. Composite Struct. 2003;62(1):59–66. [104] Saad-Eldeen S, Garbatov Y, Guedes Soares C. Compressive
[89] Slater JE. Selection of a blast-resistant GRP composite strength assessment of a moderately corroded box girder. Mar
panel design for naval ship structures. Mar Struct. Syst & Ocean Technol. 2011;6(1):27–37.
1994;7(2–5):417–40. [105] Saad-Eldeen S, Garbatov Y, Guedes Soares C. Strength assess-
[90] Saad-Eldeen S, Garbatov Y, Guedes Soares C. Experimental ment of a severely corroded box girder subjected to bending
assessment of corroded steel box-girders subjected to uniform moment. J Constr Steel Res. 2014;92:90–102.
bending. Ships Offshore Struct. 2013;8(6):653–62. [106] Saad-Eldeen S, Garbatov Y, Guedes Soares C. Effect of corrosion
[91] Dowling PJ. Strength of steel box-girder bridges. J Struct Div. severity on the ultimate strength of a steel box girder. Eng Struct.
1975;101(9):1929–46. 2013;49:560–71.
[92] Reckling KA. Behaviour of box girders under bending and shear. [107] Faulkner D, Sadden JA. Toward a unified approach to ship struc-
Proceeding of the ISSC. Paris, France; 1979. p. II–246. tural safety. Trans RINA. 1979;121:1–28.
[93] Ostapenko A. Strength of ship hull girders under moment, shear [108] Viner AC. Development of ship strength formulation. Proc. Int.
and torque. Proc. SSC-SNAME Symposium on Extreme Loads Conf. on Advances in Marine Structures; 1986. p. 152–73.
Response; 1981. p. 149–66. [109] Frieze PA. Ship longitudinal strength modelling for reliability
[94] Mansour A, Yang JM, Thayamballi A. An experimental investiga- analysis. Proceedings of the Marine Structural Inspection,
tion of ship hull ultimate strength. SNAME. 1990;1(13):13–38. Maintenance, and Monitoring Symposium, SSC/SNAME.
[95] Gordo JM, Soares CG. Experimental evaluation of the ultimate Arlington; Mar 1991. p. 18–9.
bending moment. 1:1–14. [110] Paik JK, Mansour AE. A simple formulation for predicting
[96] Saad-Eldeen S, Garbatov Y, Soares CG. Experimental assessment the ultimate strength of ships. J Mar Sci Technol. 1995;
of the ultimate strength of a box girder subjected to severe 1:52–62.
corrosion. Mar Struct. 2011;24(4):338–57. [111] Li D, Feng L, Huang D, Shi H, Wang S. Residual ultimate strength
[97] Gordo JM, Soares CG. Experimental evaluation of the behaviour of of stiffened box girder with coupled damage of pitting corrosion
a mild steel box girder under bending moment. Ships Offshore and a crack under vertical bending moment. Ocean Eng.
Struct. 2008;3(4):347–58. 2021;235(July):109341.
[98] Gordo JM, Soares CG. Tests on ultimate strength of hull box girders [112] Soares CG, Chen NZ, Santos FM, Santos C. An experimental and
made of high tensile steel. Mar Struct. 2009;22(4):770–90. numerical study on GFRP box girder under pure bending.
[99] Shi GJ, Wang DY, Wang FH, Cai SJ. Analysis of dynamic response Proceedings of MARSTRUCT’07, the 1st International Conference
and ultimate strength for box girder under bending moment. on Marine Structures. Glasgow, United KingdomAdvancements in
J Mar Sci Eng. 2023;11(2):373. Marine Structures. Paper: P2007-3 Proceedings; 2007.