Sie sind auf Seite 1von 14

Engineering Structures 99 (2015) 78–91

Contents lists available at ScienceDirect

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

Stress–strain analysis of dented rectangular plates subjected to uni-axial


compressive loading
S. Saad-Eldeen 1, Y. Garbatov, C. Guedes Soares ⇑
Centre for Marine Technology and Ocean Engineering (CENTEC), Instituto Superior Técnico, Universidade de Lisboa, Av. Rovisco Pais, 1049-001 Lisboa, Portugal

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

Article history: The aim of this work is to analyse the local and the global structural behaviour of rectangular steel plates
Received 25 June 2014 with a local dent. Nonlinear finite element analyses have been performed to explore the effect of different
Revised 13 April 2015 dent depths on the ultimate strength and the post-collapse behaviour. The post-collapse modes are
Accepted 22 April 2015
discussed and the change of the buckling mode for different plate thickness ranges is categorized. The
Available online 15 May 2015
behaviour in post-collapse regime is analysed using the defined stress–strain rate. Two relationships have
been developed to estimate the ultimate strength reduction as a function of the plate slenderness and
Keywords:
dent ratio. Based on the existing ultimate strength design formula and a new developed factor, a new for-
Stress–strain
Plates
mulation for ultimate strength of damaged plate has been introduced accounting for the effect of local
Damage dent. The new developed formulation is capable to account for an initial global imperfection, residual
Dent stresses, openings, corrosion deterioration and existence of dents.
Finite element analyses Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction influences the load carrying capacity of the structure and its ulti-
mate strength.
Plates, in either un-stiffened or stiffened are the most spread For plates, Dow and Smith [1] studied numerically the influence
structural components of thin-walled structures, such as ship of localized imperfections on the buckling and post-buckling
and offshore structures. Ship plates are generally subjected to sev- behaviour of long rectangular plates under uni-axial longitudinal
eral types of in-plane or lateral loads. The loads may be applied compression. It was concluded that the amplitude of the localised
separately or in a combination with each other. The in-plane loads imperfection is the governing factor in the collapse of plates.
may be tensile or compressive depending on the current loading Moreover, changing the position of the localized imperfection does
conditions. not significantly influence the strength of the plate.
In order to assess the ultimate hull girder strength, the capacity Paik et al. [2] and Paik [3] investigated the ultimate strength
of the ship’s hull girder against longitudinal bending moment has characteristics of dented steel plates under axial compressive and
to be defined. The longitudinal bending moment, due to hogging shear loads. The effects of shape, size (depth, diameter), and
and sagging loading conditions, causes compressive loads on the location of the dent on the ultimate strength behaviour of simply
main structural elements (plates) during the life cycle. Therefore, supported steel plates are studied. A closed-form formula for pre-
it is important to study the structural behaviour of plates under dicting the ultimate strength of dented steel plates is empirically
uni-axial compression. derived by a curve fitting based on the computed results.
Due to the operational conditions, the ship structural compo- Luís et al. [4] studied the effect of dimple imperfections caused
nents are subjected to different damage scenarios accounting for by local accidents, on the ability of a plate assembly to resist com-
corrosion degradation and fatigue cracking. As a result of dropping pressive loads. It was found that the effect of the dimple imperfec-
objects, collision and grounding, local dents may be formed. All of tion is higher when is positioned near the unloaded edge of the
these types of damages directly affect the structure, which in turn plate and depends upon its amplitude and on the slenderness of
the plate.
Raviprakash et al. [5] studied the influence of various dent
parameters (dent length, dent width, dent depth and angle of ori-
⇑ Corresponding author. Tel.: +351 218417957.
entation of the dent) on the static ultimate strength of thin square
E-mail address: c.guedes.soares@centec.tecnico.ulisboa.pt (C. Guedes Soares).
1
On leave from the Naval Architecture and Marine Engineering Department,
plates of different thicknesses under uni-axial compressive load-
Faculty of Engineering, Port Said University, 42526 Port Fouad, Egypt. ing. It was found that the longer dents with variation of size and

http://dx.doi.org/10.1016/j.engstruct.2015.04.041
0141-0296/Ó 2015 Elsevier Ltd. All rights reserved.
S. Saad-Eldeen et al. / Engineering Structures 99 (2015) 78–91 79

angle of orientation of dents drastically reduce the ultimate Table 1


strength. In general, the effect of variation of dent parameters on Plate element configurations.

the ultimate strength of the dented plate magnifies with an Item a b t GI LD Units
increase in the plate thickness. Group 1 600 400 10 1.7 0 mm
For stiffened panels, Witkowska and Guedes Soares [6] and Luís 1000 400 10 1.7 0 mm
et al. [7] investigated the behaviour and ultimate strength of stiff- Group 2 800 400 4 1.7 0 mm
ened panels with local imperfections. It was found that the dimple 800 400 6 1.7 0 mm
imperfection has a negligible effect on the rigidity of the model, 800 400 7 1.7 0 mm
but it induces a more violent collapse and the position of the dim- 800 400 8 1.7 0 mm
800 400 10 1.7 0 mm
ple imperfection is an important factor. Also, depending on geo- 800 400 20 1.7 0 mm
metrical characteristics, stiffener deformations may significantly
Group 3 800 400 4 8.50 [2,4,6,8] mm
reduce the ultimate strength. The size of the stiffeners proved to 800 400 6 5.63 [2,4,6,8] mm
be an influential factor to the collapse behaviour and ultimate 800 400 8 4.22 [2,4,6,8] mm
strength of dented panels, which agrees with the work done by 800 400 10 3.38 [2,4,6,8] mm
Amante et al. [8], where it was reported that the structural panels 800 400 12 2.82 [2,4,6,8] mm
800 400 14 2.41 [2,4,6,8] mm
should be designed with stiffeners presenting additional robust-
800 400 16 2.11 [2,4,6,8] mm
ness in the regions prone to the impact loads, in order to maintain 800 400 18 1.88 [2,4,6,8] mm
an acceptable level of integrity for a damaged panel. 800 400 20 1.70 [2,4,6,8] mm
Xu and Guedes Soares [9] studied the influence of a local dent
on the collapse behaviour of stiffened panels. The effect of residual
stresses caused by the local dent on the collapse behaviour of stiff-
ened panels was also investigated. It was concluded the residual where t is the plate thickness, ry is the Yield stress, E is the material
stresses caused by the indentation, reduce slightly the ultimate Young modulus. The plate slenderness of the analysed plates varies
strength of the dented stiffened panels. from 0.75 to 3.75. For ship plates, normally b varies from 1 to 5
The study presented here is a continuation of the research work [12,13]. The adequacy of b to represent the compressive strength
done by Saad-Eldeen et al. [10], in which another scenario for of rectangular plates has been demonstrated by various design
reduced the ultimate strength of plates was studied; the presence expressions and studies [14,15].
of openings. The effect of the opening ratio and orientation to the Numerical analyses of the ultimate strength of unstiffened rect-
ultimate strength of plates was investigated and an expression angular plates are performed with a general non-linear finite ele-
was developed to estimate the ultimate as a function of the plate ment commercial code – ANSYS. The FEA utilizes the full
slenderness and residual plate-breadth ratio. Newton–Raphson equilibrium iteration scheme, the large defor-
Most of the work done in the present literature review for mation option was activated to solve the geometric and material
dented plates is dealing with the ultimate strength point, without nonlinearities and pass through the extreme points. The automatic
analysing the behaviour of the plate before and after reaching it. It time stepping features are employed allowing ANSYS to determine
is important to analyse the post-collapse regime in order to under- appropriate load steps.
stand the complexity of the structural behaviour. Therefore, the A shell element was used to generate the entire FE model. The
present study investigates the local (ultimate) and global beha- shell element, SHELL 181 is defined by four-nodded element with
viour of rectangular steel plates accounting for the presence of a six degrees of freedom at each node: translations in the x, y and
local dent. Based on the performed analyses, several conclusions z directions, and rotations about the x, y and z axes. SHELL181 is
are derived and a new coefficient accounting for the effect of a local well-suited for linear, large rotation, and/or large strain nonlinear
dent on the ultimate strength of the rectangular plate element is applications and suitable for analysing thin to moderately-thick
developed. shell structures.
The material used in the FE model is of low carbon steel with
2. Finite element model the yield stress, ry of 290 [MPa], the Young modulus, E of 206
[GPa] and the Poisson’s ratio, t of 0.3. The stress–strain model is
The behaviour of rectangular plate elements with and without elastic-perfectly plastic, as defined by Kim et al. [11].
local dent, subjected to uni-axial compression will be analysed. At stresses below the yield stress, ry the material behaviour is
The analysed plates are categorised into three groups as a function linear with a tangent modulus of E ¼ ry =ey . At the level of the yield
of plate length, breadth, thickness, global imperfections and local stress, the material flows without strain hardening. When the
dent damage. The geometrical configurations of the analysed material is unloaded by reducing the stress below the yield stress,
plates are given in Table 1 and shown in Fig. 1, where a is the plate it behaves elastically in a manner unaffected by the plastic flow.
length, b is the plate width, t is the plate thickness, l is the dent The quadrilateral element size of 5 mm has been defined as a
length, s is the dent width, GI is the global initial imperfection good solution of the finite element model based on the compara-
and LD is the local dent depth. tive analysis with the experimental test and finite element results
The intact structural model configurations, boundary conditions presented in [11], for an intact plate with a thickness of 4 mm, as
and material stress–strain curve is based on a real experimental described briefly in [10] and shown in Fig. 2. The model calibration
test performed by Kim et al. [11]. The configurations of the intact leads to an element size-plate thickness ratio of 1.25, thus element
plate up to a 8 mm plate thickness are based on the experimental size of ES = 1.25 ⁄ t is used for analysing the rest of the plates with
test performed in [11]. Systematic variation of the plate thickness a thickness bigger than 4 mm.
up to 20 mm is also included in the present analysis, which may One shape of initial imperfection is considered in the analyses
cover the diversity of engineering applications. and modelled as one half downward wave, based on the Fourier
The principal parameter governing the buckling strength of series, as given by Eq. (2), and shown in Fig. 1.
intact plate is the plate slenderness, defined as:
rffiffiffiffiffiffi X
m
ipx X
n
jpy
b ry x ¼ x0 sin sin ð2Þ
b¼ ð1Þ i¼1
a j¼1 b
t E
80 S. Saad-Eldeen et al. / Engineering Structures 99 (2015) 78–91

Fig. 1. Initial imperfection shape and boundary conditions.

unloaded edges. For the second boundary condition, BC2, the


unloaded edges are simply supported, keeping them straight to
simulate the condition of no stiffeners along the unloaded edges.
In order to follow the experimental test conditions, as presented
in [11], where the axial force has been transmitted from the
hydraulic jack to the tested plate through a stiff solid beam, to
insure the uniform distribution of the applied loading, in the pre-
sent FE model, a uni-axial compressive loading, in the longitudinal
direction x, is generated by imposed axial nodal forces.

3. Finite element analysis

The finite element results presented here are in the same order
of the groups presented in Table 1. Group 1 analyses the effect of
different boundary conditions on the ultimate strength and the col-
lapse mode for different plate aspect ratios. Group 2 deals with the
effect of the varying the plate thickness with constant initial
imperfections on the ultimate strength as well as the final defor-
mation shape of rectangular plates. The last group, Group 3, which
covers the main objective of the presented work, investigates the
effect of the variation of the local dent depth as a function of the
plate thicknesses on the ultimate compressive strength, the buck-
Fig. 2. Finite element model calibration for t = 4 mm. ling mode and the collapse rate.

3.1. Group 1 – Effect of boundary conditions


where x is the longitudinal axis, y is the transversal axis, m is the
mode of the imperfections in (i.e. the number of the half-sine waves Two boundary conditions are considered: fully constrained
within the length of the plate, it is assumed here as m = 1), n is the (BC1) and free to rotate (BC2), as described before and presented
mode of the transverse direction of the plate (i.e. the number of in Fig. 1. The analysed rectangular plates are with a breadth (b),
half-sine waves within the width of one plate, n = 1) and x0 is the thickness (t), global imperfections (GI) of 400, 10 and 1.7 mm,
imperfection amplitude, x 2 [0, a] and y 2 [0, b]. respectively, and with a different plate length (a) of 600 mm and
The average initial imperfection amplitude for the analysed 1000 mm, as given in Table 1.
model, especially in group 3, see Table 1, is calculated by the fol-
lowing expression [14,16]: 3.1.1. Ultimate lateral deformation
2
x0 ¼ ab t ð3Þ The ultimate lateral deformation assumed here is the maximum
lateral deformation that occurs at the ultimate strength point. As
where a = 0.15 as given in [16], and b is the plate slenderness. may be seen from Fig. 3, a/t = 60, in case of constrained boundary
Two boundary conditions, BC1 and BC2, are considered in the conditions (BC1), the developed ultimate deformation is less than
analyses as shown in Fig. 1, in order to investigate the effect of dif- the initial global imperfection and this is a result of higher energy
ferent boundary conditions on the ultimate strength and the final absorbed to keep the edges flat due to the constrained rotation. On
deformation shape. For BC1, the unloaded edges are rotationally the contrary, for the free rotation (BC2), a more developed defor-
restrained, simulating the existence of two stiffeners along the mation was registered due to the free rotation near the edges.
S. Saad-Eldeen et al. / Engineering Structures 99 (2015) 78–91 81

Table 2
Ultimate strength and lateral deformation, Group 1.

Item Constrained Free rotation E⁄ = kE


L/t GI/t ru/ry Uz, m ru/ry Uz, m
60 0.17 1.09 0.0014 0.97 0.0036 0.86E
100 0.17 1.11 0.0008 1.03 0.0026 0.93E

Fig. 3. Ultimate deformation vs. different boundary, a/t = 60.

By increasing the plate length to 1000 mm, a/t = 100, keeping


the same initial global imperfection amplitude of 1.7 mm, the
developed deformation shape for the constrained boundary condi-
tions is different from a/t = 60 and local buckling starts to take
place as marked by circles in Fig. 4, leading to loss of the stiffness.
Therefore, the registered ultimate deformation for the ratio
a/t = 100 is less than the one of a/t = 60, as presented in Fig. 4 Fig. 5. Ultimate strength as a function of boundary conditions.
and given in Table 2.
In case of free rotation conditions, the developed ultimate
deformation shape is totally different from a/t = 60. The location (upper left corner). The coefficient k 2 [0, 1], which indicates how
of the maximum ultimate lateral deformation changed from the much the structural behaviour is close to the behaviour of the
middle to a location about 0.3a. This gives an indication to the fact material. As may be seen in Fig. 5 and Table 2, the apparent mod-
that the flexural stiffness of the plate reduced due to the high ulus of the constrained and free rotation boundary conditions in
length/breadth and length/thickness ratios. the case of a/t = 100 is bigger than the one of a/t = 60. This gives
indication that by increasing the a/t ratio, the effect of the free
rotation boundary conditions on reducing the post-buckling stiff-
3.1.2. Ultimate strength
ness decreases, as can be seen in Table 2.
The calculated normalized strength with respect to the normal-
From the ultimate strength point of view, as may be seen from
ized strain is plotted in Fig. 5 and shown in Table 2, for different
Fig. 5 and Table 2, by increasing the aspect ratio of length/ thick-
length/thickness ratios and different boundary conditions.
ness, the effect of boundary conditions on the ultimate strength
Elasto-plastic-post-buckling stiffness may be estimated in terms
decreases. Based on the previous observation regarding the dimin-
of the apparent structural modulus of elasticity E⁄ = kE (the ratio
ishing of different boundary condition effect, also, the constrained
of the average stress carried by the plate to average strain)
edge condition (BC1) is closer to the actual condition, where stiff-
extended from the proportional limit which is non-linear and
eners may exist along the unloaded edges as for the tested intact
elasto-plastic, up to the ultimate point which is plastic, see Fig. 5
plate structure done by Kim et al. [11], the constrained boundary
condition, BC1, will be used for the rest of the FE analyses in
Group 2 and 3.

3.2. Group 2 – Effect of plate thickness

The effect of different plate thickness (t) of 4, 6, 7, 8, 10 and


20 mm with a constant plate length (a), breadth (b) and global
imperfection (GI), as stated in Table 1, on the ultimate strength,
global behaviour and the failure mode, considering only the con-
strained boundary conditions (BC1), is investigated here.

3.2.1. Ultimate lateral deformation


The ultimate lateral deformation along the plate length is pre-
sented in Fig. 6. It is evident that the deformation shape registered
by 4 mm plate thickness is totally different from the initial global
imperfection, producing three half waves, 3rd order. The reason
for this buckling mode is the more flexibility, resulting from the
higher plate slenderness of 3.75, which consequently leads to
lower ultimate strength as given in Table 3. If a longitudinal strip
Fig. 4. Ultimate deformation vs. different boundary condition, a/t = 100. such as AB, as can be seen in Fig. 8, tends to form a single buckle;
82 S. Saad-Eldeen et al. / Engineering Structures 99 (2015) 78–91

Fig. 8. Buckling of plate under uni-axial compression.


Fig. 6. Ultimate deformation as a function of different plate thickness, longitudinal
direction.
directions. The ultimate lateral deformation occurs at 0.25a, with
a value of 0.0027 m as given in Table 3 and shown in Fig. 6.
It can be seen from Fig. 6 that the rest of the analysed thick-
Table 3
Ultimate strength and lateral deformation, Group 2. nesses shows a buckling mode from the 1st order, as the one of
the initial imperfection.
t, mm GI/t ru/ry Uz long, m Uz trans, m
For t 6 7 mm, the registered lateral deformation is bigger than
4 2.35 0.59 0.0027 0.0007 the initial imperfection regardless its location, which is on contrary
6 3.53 0.97 0.0026 0.0026
for t > 7 mm, as shown in Fig. 6 and given in Table 3. This can be
7 4.12 1.03 0.0021 0.0021
8 4.71 1.07 0.0014 0.0014
explained with increasing of the resistance of the transverse strips
10 5.88 1.11 0.0010 0.0010 to the longitudinal buckling as the plate thickness increases (less
20 11.76 1.15 0.0009 0.0009 slenderness), which may be also confirmed with the development
of lower order buckling mode, 1st, see Fig. 6. Also this can be as a
result of the small slope near the loaded edges as the plate thick-
ness increases, which in turn reduces the global deformation.

3.2.2. Ultimate strength


The registered normalized strength with respect to the normal-
ized strain is plotted in Fig. 9, for different plate thicknesses and
the estimated ultimate strength are given in Table 3. It is evident
that by increasing the plate thickness, the ultimate strength
increases with a sudden collapse due to the high rigidity, because
the plate becomes stocky resulting to a less plate slenderness, b.
This behaviour is more obvious for a 20 mm plate thickness where
b = 0.75 and the plate registered more strain at almost the same
load, indicating the more energy needed to develop the buckling
mode.

Fig. 7. Ultimate deformation as a function of different plate thickness, transverse


direction.

its curvature will be much less than the curvature of the transverse
strip CD, which tries to resist the buckling. This means that the
resistance is greater than the tendency to buckle and the strength
corresponding to this mode (m = 1) is very high. Therefore, the
plate tends to buckle in such way so that the curvatures of longitu-
dinal and transverse strips are as equal as possible. This leads to
multiple buckles in alternate directions as presented in Fig. 6, for
longitudinal deformation, and Fig. 7 for transverse deformation.
Therefore, the registered lateral deformation in the middle of the
plate is the about 0.0007 m, in longitudinal and transverse Fig. 9. Ultimate strength as a function of different plate thickness.
S. Saad-Eldeen et al. / Engineering Structures 99 (2015) 78–91 83

For a 4 mm plate thickness, the behaviour is opposite; the plate of the dent were selected from the series of the analysis done by
buckles smoothly with less energy, besides the changes occurred in the authors for a plate with large openings [10]. After that the same
the buckling mode, as discussed before, due to the higher slender- size has been used to compare the behaviour of the plate with a
ness of 3.75. dent or an opening as reported in [17].
The longitudinal stress distribution for a 4 and 20 mm plate
thickness respectively, is shown in Fig. 10. It is clear that the stress
3.3.1. Ultimate lateral deformation
distribution is different and the plate of a 20 mm thickness is
A series of ultimate lateral deformation along the longitudinal
highly stressed with a spread of plasticity. On the contrary, for a
and transverse directions for a dented plate with a thickness of 6,
plate of a 4 mm thickness, only specific locations are highly
10, 16 and 20 mm is presented in Fig. 12(a–d).
stressed, where bigger deformation occurs.
For a 6 mm plate thickness, two buckling modes are registered.
The first one for intact plate (0 mm LD), where the 3rd buckling
3.3. Group 3 – Effect of different local dent depth mode is developed due to the higher imperfection amplitude; Eq.
(3), besides the resistance of the transverse strips to the longitudi-
The analysis of this group is the main goal of the present work, nal buckling resulting in bulking, which establishes equal central
where the effect of the local dent on the ultimate strength, global displacement as shown in Fig. 12(a). The second one is for dented
behaviour and the failure modes is explored. Different local dent plate, where the 1st order buckling mode occurs, in which the local
depths (LD) of 2, 4, 6 and 8 mm in different plate thicknesses (t) dent forces the plate to deform in one direction. By increasing the
of 4, 6, 8, 10, 12, 14, 16, 18 and 20 mm, with the presence of an ini- depth of the local dent, from 4 to 8 mm, the registered central dis-
tial global imperfection for each plate thickness, as given in Table 1, placement is almost the same, as tabulated in Table 4. This indi-
are analysed, considering only the constrained boundary condi- cates that for a 6 mm plate thickness, the increase of the dent
tions. The local dent damage has been modelled using one half depth is not affecting the ultimate displacement.
downward wave in the same direction of the initial imperfection, For t = 10 mm (see Fig. 12b), only one buckling mode from the
based on the Fourier series, as given by Eq. (2) and shown in 1st order was registered, which is due to the imperfection and
Fig. 11 (right). The dent length l and breadth s are 250 and the lower plate slenderness of 1.5. By increasing the dent depth,
100 mm, respectively, as shown in Fig. 11 (left). The dimensions the same observation of matching the ultimate strength

1 1
NODAL SOLUTION NODAL SOLUTION
SEP 27 2013 SEP 27 2013
STEP=1 19:18:11 STEP=1 19:55:30
SUB =15 SUB =36
TIME=.421684 TIME=.835792
SX (AVG) SX (AVG)
RSYS=0 RSYS=0
DMX =.002774 DMX =.002996
SMN =-.313E+09 SMN =-.335E+09
SMX =.217E+08 SMX =-.307E+09

MX
MN
Y
Y
Z X
Z X

MX

MN

-.313E+09 -.239E+09 -.164E+09 -.898E+08 -.155E+08 -.335E+09 -.329E+09 -.322E+09 -.316E+09 -.310E+09
-.276E+09 -.201E+09 -.127E+09 -.527E+08 .217E+08 -.332E+09 -.326E+09 -.319E+09 -.313E+09 -.307E+09
Plate-4mm,1.7mmGI Plate-20mm,1.7mmGI

Fig. 10. Longitudinal stress distribution for 4 mm (left) and 20 mm plate thickness (right).

1 X
ELEMENTS
Z
Y NOV 1 2013
15:14:01

Plate with local dent

Fig. 11. Indentation of unstiffened plate, configurations (left), finite element model (right) .
84 S. Saad-Eldeen et al. / Engineering Structures 99 (2015) 78–91

(a) t = 6mm

(b) t = 10mm

(c) t = 16mm
Fig. 12. Ultimate deformation as a function of different local dent depth, longitudinal direction (left) and transverse direction (right), (d) t = 20 mm.
S. Saad-Eldeen et al. / Engineering Structures 99 (2015) 78–91 85

Fig. 12 (continued)

Table 4
displacement occurs as for t = 6 mm, which can be seen in
Ultimate strength and lateral deformation, Group 3. Fig. 12(b) and Table 4. For a 16 mm plate thickness, only one buck-
ling mode from the 1st order was observed as for t = 10 mm, see
t, mm GI/t LD, mm LD/t ru/ry Uz long, m Uz trans, m
Fig. 12(c). With a small dent depth of 2.11 mm, which equals to
4 2.11 0 0 0.80 0.0033 0.0015 the implemented global imperfection, the ultimate lateral defor-
2 0.5 0.79 0.0029 0.0019
4 1 0.78 0.0027 0.0023
mation is almost the same as for the intact plate, as can be seen
6 1.5 0.75 0.0024 0.0027 in Fig. 12(c). This indicates that for a thick plate of 16 mm, there
8.5 2.125 0.73 0.0027 0.0032 is no effect of the local dent damage with a depth less than or equal
6 0.94 0 0.00 0.93 0.0028 0.0022 to the global initial imperfection.
2 0.33 0.91 0.0030 0.0030 By increasing the local dent depth, the registered central lateral
4 0.67 0.87 0.0029 0.0029 deformation increases and the developed buckling mode becomes
5.63 0.94 0.85 0.0029 0.0029
stiff, showing small deformation at a distance of 0 to 0.275 m and
8 1.33 0.83 0.0029 0.0029
0.525 to 0.8 m along the plate length, see Fig. 12(c). This gives an
8 0.53 0 0.00 1.01 0.0021 0.0021
indication that the dent borders at this thickness acts as local
2 0.25 0.97 0.0025 0.0025
4.22 0.53 0.94 0.0027 0.0027 hinges.
6 0.75 0.92 0.0027 0.0027 For a thick plate, t = 20 mm, almost the same observations
8 1.00 0.90 0.0027 0.0027 (buckling mode, matching the ultimate displacement) occurs as
10 0.34 0 0.00 1.07 0.0016 0.0016 for t = 16 mm, see Fig. 12(d). By increasing the local dent depth,
2 0.20 1.03 0.0020 0.0020 LD = 8 mm, the central lateral deformation increases and the devel-
3.38 0.34 1.01 0.0020 0.0020 oped buckling mode becomes stiffer, with almost zero deformation
6 0.60 0.99 0.0024 0.0024
8 0.80 0.97 0.0024 0.0024
at a distance of 0–0.275 m and 0.525–0.8 m along the plate, as
shown in see Fig. 12(d). This gives an indication that the dent bor-
12 0.23 0 0.00 1.11 0.0011 0.0011
2.82 0.24 1.06 0.0015 0.0015
ders at this thickness acts as local hinges especially at a higher dent
4 0.33 1.05 0.0016 0.0016 depth, therefore, there is no flexibility to deform far away from the
6 0.50 1.02 0.0020 0.0020 dent location.
8 0.67 1.00 0.0021 0.0021 It can be concluded that based on the developed buckling
14 0.17 0 0.00 1.13 0.0011 0.0011 modes, for the analysed plate thicknesses, by increasing the plate
2.41 0.17 1.09 0.0013 0.0013 thickness, the effect of local dent depth on the ultimate central dis-
4 0.29 1.07 0.0015 0.0015
placement decreases.
6 0.43 1.05 0.0016 0.0016
8 0.57 1.03 0.0017 0.0017
16 0.13 0 0.00 1.15 0.0009 0.0009
3.3.2. Ultimate strength
2.11 0.13 1.11 0.0010 0.0010 The normalized strength vs. the normalized strain relationships
4 0.25 1.09 0.0012 0.0012 for the analysed plate thicknesses are presented in Fig. 13.
6 0.38 1.07 0.0015 0.0015 For a 6 mm plate thickness, see Fig. 13(a), by increasing the dent
8 0.50 1.07 0.0016 0.0016
depth, the ultimate strength decreases, as given in Table 4. From
18 0.10 0 0.00 1.15 0.0010 0.0010 the post-collapse point of view, it is visible that the plate with
2 0.11 1.13 0.0010 0.0010
and without dents collapsed in a soft way referring that the plate
4 0.22 1.11 0.0011 0.0011
6 0.33 1.09 0.0013 0.0013 has more flexibility due to the slenderness of 2.5. Also, the esti-
8 0.44 1.08 0.0014 0.0014 mated strain on the post-collapse regime of the plate with dent
20 0.08 0 0.00 1.15 0.0009 0.0009 is almost the same, which indicates that the global initial imperfec-
1.77 0.09 1.15 0.0009 0.0009 tion is dominating the post-collapse and the presence of the dent
4 0.20 1.13 0.0009 0.0009 only affects the ultimate strength, as given in Table 4. A clear dif-
6 0.30 1.11 0.0013 0.0013
ference in the collapse rate (defined as the slope of the stress–
8 0.40 1.09 0.0014 0.0014
strain curve after achieving the collapse point) between the plate
86 S. Saad-Eldeen et al. / Engineering Structures 99 (2015) 78–91

(a) t = 6mm (b) t = 10mm

(c) t = 16mm (d) t = 20mm


Fig. 13. Ultimate strength as a function of different local dent depth.

without dent (0 mm LD) and with 2 mm LD is noticeable, this is the dent depth increases. The formation of a hinge around the dent
because of the different buckling mode order as 3rd for 0 mm LD borders may be seen clearly from the consequences of the ultimate
and 1st 2 mm LD, as presented in Fig. 12(a). deformation in the longitudinal direction as plotted in Fig. 12(c,
For a 10 mm plate thickness, see Fig. 13(b), with increasing the left), at which, with increasing the dent depth, the deformed shape
dent depth, the ultimate strength decreases with respect to the is not smooth, showing a knuckle at the borders of the dent, which
intact one, as given in Table 4. From the post-collapse point of is not exist for thinner plates.
view, it is clear that the plate element with and without dent col- For a thick plate of 20 mm, the estimated ultimate strength of
lapses in a less soft way compared to the one of t = 6 mm and this is the plate without local dent is almost the same as with 1.77 mm
because of the smaller slenderness of 1.5. LD, which is equal to the global imperfections, as shown in
A coincide in the collapse rate between the plate element with Fig. 13(d) and Table 4. This is due to the smaller initial imperfec-
and without dent was registered after normalized strain of 1.6. The tions and dent depth. From the post-collapse point of view, the
reason for that is the same developed buckling mode of the 1st plate without dent shows high energy absorption, represented by
order, as may be seen from Fig. 12(b). higher strain at small load increment and this is as a result of the
By increasing the plate thickness from 10 to 16 mm, the col- presence of plasticity due to the lower slenderness of 0.75. With
lapse behaviour of a plate element with and without dent absorbs increasing the dent depth, the ultimate strength decreases with
more energy due to the small slenderness of 0.94 and the collapse the less collapse rate as may be seen in Fig. 13(d).
rate decreases with a smaller rate, which reflects the spread of
plasticity, as shown in Fig. 13(c). 3.3.3. Stress distribution
As may be seen from Fig. 13(a and b), the stress–strain curves in The von Mises stresses of the analysed plates with a local dent
the post-collapse regime coincide for the plate thicknesses of 6 mm depth of 8 mm LD at the ultimate load sub step of finite element
and 10 mm. However, this is not the same for a thickness of 16 mm analysis is presented in Fig. 14. As may be seen from Fig. 14(a),
(see Fig. 13c), in which the borders of the dent acts are a hinge as for a 6 mm plate thickness, a uniform distribution of high stresses
S. Saad-Eldeen et al. / Engineering Structures 99 (2015) 78–91 87

1 1
NODAL SOLUTION NODAL SOLUTION
STEP=1 DEC 10 2014 STEP=1 DEC 10 2014
SUB =19 16:31:39 SUB =24 16:28:27
TIME=.80203 TIME=.930029
SEQV (AVG) SEQV (AVG)
DMX =.002934 DMX =.002462
SMN =.653E+07 SMN =.766E+08
SMX =.290E+09 SMX =.290E+09

Y Y

Z X Z X
MN
MN

MX MX

.653E+07 .695E+08 .133E+09 .196E+09 .259E+09 .766E+08 .124E+09 .171E+09 .219E+09 .266E+09
.380E+08 .101E+09 .164E+09 .227E+09 .290E+09 .100E+09 .148E+09 .195E+09 .243E+09 .290E+09
Plate-6mm,5.6mmGI,8mmLD Plate-10mm,3.38mmGI,8mmLD

(a) t = 6mm (b) t = 10mm


1 1
NODAL SOLUTION NODAL SOLUTION
STEP=1 STEP=1 DEC 10 2014
DEC 10 2014
SUB =16 SUB =18 17:35:00
16:24:46
TIME=.613599 TIME=.631295
SEQV (AVG) SEQV (AVG)
DMX =.001755 DMX =.001604
SMN =.504E+08 SMN =.302E+08
SMX =.290E+09 SMX =.290E+09

MN
Y Y

Z X Z X
MN

MX MX

.504E+08 .104E+09 .157E+09 .210E+09 .263E+09 .302E+08 .879E+08 .146E+09 .203E+09 .261E+09
.770E+08 .130E+09 .184E+09 .237E+09 .290E+09 .590E+08 .117E+09 .175E+09 .232E+09 .290E+09
Plate-16mm,2.11mmGI,8mmLD Plate-20mm,1.7mmGI,8mmLD-Longitudinal-down dent

(c) t = 16mm (d) t = 20mm


Fig. 14. von Mises stresses distribution, 8 mm LD.

is located in the central region, extended in the transverse direc-


tion, where the location of the dent damage is less uniformly
stressed, especially in the middle part.
By increasing of the plate thickness from 6 to 10 mm, the stress
distribution is different; where the location of the dent is highly
stressed. The area around the dent, in the transverse direction, is
highly stressed. Away of the dent location, the plate is uniformly
stressed compared to the one of 6 mm thickness; see Fig. 14(b).
For a 16 mm plate thickness, the extension of the highly
stressed area increases and the plasticity spreads over all the plate,
more than that of 6 and 10 mm plate thickness. Minor locations
around the dent borders, in the longitudinal direction, show a dis-
turbance of the stresses with values less than the yield stress, as
may be seen from Fig. 14(c). As the plate becomes thicker,
t = 20 mm, the plate reaches the full plasticity, with stresses closer
to the yield stress of the material, except in some minor locations,
near the dent borders, with less stresses. The formation of hinges is
more visible at the borders of the dent, see Fig. 14(d), which con-
firms the post-buckling behaviour of the plate.
The normalized ultimate strain vs. the local dent depth is plot-
ted in Fig. 15. It can be seen that from a 4 mm to 10 mm plate Fig. 15. Ultimate normalized strain vs. local dent depth.
88 S. Saad-Eldeen et al. / Engineering Structures 99 (2015) 78–91

the ratio between the dent depth and the plate thickness,
 
dent ratio;dd ¼ LD
t
.
A relationship between the reduction of the ultimate strength
as a function of the plate slenderness and the dent ratio is pre-
sented in Fig. 17.
At a constant dent depth, by increasing the plate slenderness,
the reduction percentage of the ultimate strength increases up to
b = 2.15 (t = 7 mm), after that the behaviour changes, showing a
lower reduction. This observation coincides with the work done
by Amani et al. [18] for unstiffened intact mild carbon steel, stain-
less steel and aluminium plates under axial compression, where
the results shown that the transition from slender to stocky plates,
regardless of material type, occurs at the slenderness of 2.
Regarding the dent ratio (dent depth/plate thickness), the same
behaviour occurs as for the plate slenderness. By an increasing of
the dent ratio, the reduction of the ultimate strength increases
up to dd = 1.15, after that the behaviour changes, giving a lower
reduction at the higher dent ratios.
Fig. 16. Ultimate normalized strength vs. local dent depth. Using the regression analysis, based on the finite element
results, two relationships have been developed to estimate the
ultimate strength reduction percentage with respect to the intact
thickness, the ultimate strain at 0 mm LD is around 1. By increasing plate, as a function of the plate slenderness, b as well as the dent
the plate thickness, the registered ultimate strain is much bigger, ratio dd, as given in Eqs. (4) and (5), respectively and plotted in
which can be explained with the fact that the thicker plate needs Fig. 17, with coefficient of determination, R2 of 0.9986.
more energy to initiate buckling and the spread of plasticity is  
much higher than for the thinner plate. By increasing the dent ru0
¼ 0:56b3  5:81b2 þ 17:20b  4:24 ð4Þ
depth up to 4 mm LD, the ultimate strain decreases. After that, as ru8 b
the plate thickness increases with a dent depth, the effect of the
 
dent on reducing the ultimate strain decreases and this is due to ru0
the higher stiffness of the thick plate with respect to the dent ¼ 3:70d3d  20:44d2d þ 32:26dd  4:24 ð5Þ
ru8 dd
depth and the global initial imperfections.
Fig. 16 shows the normalized ultimate strength as a function of Several expressions have been proposed to estimate the ulti-
a local dent depth; it is noticeable that by increasing the local dent mate strength of intact plate under axial compression as the one
depth, the ultimate loading capacity that the plate can with stands proposed by Faulkner [14]. The equation accounts implicitly for
decreases, which in turn leads to lower ultimate strength. average levels of initial deflection and it can be complemented
with others that dealt explicitly with the effect of residual stresses.
Faulkner [14] developed an expression for ultimate strength,
fitting it with data on the ultimate plate strength leading to:
4. Ultimate strength assessment a1 a2
/p ¼  ; b  1:0 ð6Þ
b b2
Generally, the ultimate strength of intact steel plates may be
defined as a function of plate slenderness, b. For dented plates, where the constants a1 and a2 are given as a1 = 2.0 and a2 = 1.0 for
another important parameter that affects the ultimate strength is simple supports and a1 = 2.5 and a2 = 1.56 for clamped supports.

Fig. 17. Reduction of the ultimate normalized strength as a function of plate slenderness and dent ratio.
S. Saad-Eldeen et al. / Engineering Structures 99 (2015) 78–91 89

Guedes Soares [15] extended that formulation accounting for


the effect of residual stresses, initial distortion and boundary
conditions.
For a better approximation in simplified design equations some
authors (e.g. Guedes Soares [13]) decided to divide the ultimate
strength in various terms, accounting for initial geometrical imper-
fection and residual stresses as:
/u ¼ /p Rd Rr ð7Þ

where /p is the perfect ultimate strength, Rd is the reduction factor


due the initial geometry imperfection and Rr is the reduction factor
due to the residual stresses.
Dwight and Moxham [19] related the level of compressive
residual stress to the width of the tensile block gt:

rr 2gt
/r ¼ ¼ ð8Þ
ry b  2gt
where rr is the residual stress and ry is the yield stress of the mate- Fig. 18. Local dent correction factor as a function of dent ratio.
rial and the reduction factor due to residual stresses may be calcu-
lated as:
8 2
Rr ¼ ð1  /r Þ ð9Þ < 0:0018dd  0:0521dd þ 1:02 t  4 mm
>
Rd ðdd Þ ¼ 0:0514d2d  0:1611dd þ 1:01 4 mm < t  6 mm ð12Þ
This equation accounts implicitly for average levels of initial deflec- >
:
tion and it can be complemented with others that dealt explicitly 0:0723d2d  0:1851dd þ 1:00 t > 6 mm
with the effect of residual stresses. The analysed dent ratio dd varies from 0.085 to 2.125. It can be
Guedes Soares [20] has extended that formulation by deriving a seen from Fig. 18 that for all defined ranges of the plate thickness,
strength assessment expression for the compressive strength of by increasing the dent ratio the corresponding local dent correc-
plate elements under uni-axial load, which deals explicitly with tion factor, Rd is decreasing, which in turn reduces the ultimate
initial defects as: strength.
/u ¼ ð/p Bp Þ ðRr Br Þ ðRd Bd Þ ð10Þ In a previous study, as reported by Saad-Eldeen et al. [10], the
effect of openings has been investigated and an expression was
where up is given by Eq. (6), Bp, Br and Bd are model uncertainty fac- developed to estimate the ultimate strength of plates subjected
tors and Rr and Rd are strength reduction factors which are due to to compressive loading as a function of the plate slenderness and
the presence of weld induced residual stresses and initial distor- residual plate-breadth ratio.
tions respectively. As may be seen from Fig. 19, a plot of systematic calculations of
That type of formulation has been extended to cover several ultimate compressive strength as a function of plate thickness and
damage scenarios that are affecting the ultimate strength of ships dent depth is presented. It can be seen that for a plate with a thick-
and offshore platforms, such as openings, corrosion and local dent ness less than 7 mm the behaviour is steeper and for a plate with a
as: thickness bigger than 7 mm the behaviour becomes less steep. This
indicates that the compressive strength for thinner plates up to
/ ¼ /0p R0d R0r Ro Rcorr Rd ð11Þ
7 mm is more sensitive to the dent depth rather than for the
where /0p is the perfect plate strength accounting for the model thicker plates. This observation coincides with the inflection point
uncertainty, R0d is a strength reduction factor due to initial distor- of the slenderness as may be shown in Fig. 17.
tions accounting for the model uncertainty, R0r is a strength reduc-
tion factor due to residual stresses, Ro is a strength reduction
factor due to openings, Rcorr is a strength reduction factor due to
corrosion and Rd is a strength reduction factor due to the presence
of a local dent. The correction factors have been proposed recently,
to model the effect due to corrosion degradation, Rcorr, by Silva et al.
[21,22], due to welding induced residual stresses, R0r , by [23,24] and
due to a large opening, Ro, by [10]. In the present analyses, only the
effect of local dent is studied and the other factors related to resid-
ual stresses, opening and corrosion degradation are neglected, i.e.
R0r Ro Rcorr ¼ 1.
Based on the analysis of the finite element results, a new addi-
tional coefficient, Rd has been identified to account for the presence
of local dent and its effect on the ultimate strength of the plate
under uni-axial compression as given by Eq. (12) and plotted in
Fig. 18. Three levels of thickness are defined accounting for the
observed behaviour with respect to the thicknesses and dent con-
figurations for constrained rectangular plates. The coefficient of
determinations R2, for the thickness ranges t  4 mm,
4 mm < t  6 mm and t > 6 mm are of 0.9873, 0.9998 and
0.9867 respectively. Fig. 19. Ultimate strength reduction as a function of plate thickness and dent depth.
90 S. Saad-Eldeen et al. / Engineering Structures 99 (2015) 78–91

The stress–strain rate, which is defined by Eq. (13), for each the plate thickness increases, more than 18 mm, leading to the
plate as a function of a local dent depth is plotted in Fig. 20. As fourth level, the stress–strain rate in the presence of dent is almost
may be seen, the behaviour of the plate can be categorized into constant, which reflects the higher stiffness of the plate element.
four thickness levels. The first one; up to 7 mm, the second one As discussed before, different behaviour of a plate with different
is from 8 mm to 12 mm, the third one is from 14 mm to 18 mm thicknesses coupled with a local dent is observed, which gives an
and the fourth one is of a thickness bigger than 18 mm. indication that there is a scaling effect, which leads to a different
effect to the plates of different thicknesses.
Dr ru  r2ey Using Eq. (6) given by Faulkner [14] and considering the coeffi-
r SS ¼ ¼ ð13Þ
De eu  2ey cients for the constrained BCs, the ultimate strength of the intact
plate, as a function of the plate slenderness is given in Fig. 21. As
For the first level, the response of the plate in the post-collapse
may be seen, the finite element model, based on an initial imper-
region is difficult to be predicted for the small dent depth of 2 mm
fection of 1 half wave demonstrates higher ultimate strength.
LD, and this is due to the flexibility of the plate as a result of the
The reason for that is that the finite element model is not consid-
higher slenderness and the stress–strain curve decreases with a
ering the residual stresses and assumes only one-half-wave as an
smaller rate, which confirms that the dent depth has little effect
initial imperfection.
on the post-collapse behaviour and the global initial imperfection
When the FE model is based on a 2 two-half-wave initial imper-
is dominating, as shown in Fig. 20. For the second level a normal
fection, the estimated ultimate strength is coinciding with Eq. (6),
behaviour is registered with one buckling mode with and without
which also may be seen in Fig. 21.
the presence of the dent and a bigger reduction of stress–strain
rate is observed due to the moderate plate slenderness.
For the third level, the plate slenderness is less than 1, therefore,
5. Conclusion
the behaviour of the plate without local dent is stiffer and more
energy is needed to initiate buckling. With the presence of local
A series of nonlinear finite element analyses for a rectangular
dent, buckling starts and the stress–strain rate becomes less. As
plate of different configurations, with and without local dent were
presented and classified into three groups.
Group 1 studied the effect of different boundary conditions on
the global behaviour of the rectangular plate and it was observed
that by increasing the aspect ratio and the length/thickness ratio,
the reduction of the ultimate strength with the free rotation
boundary condition decreases. Also, with an increasing of the
length/thickness ratio, the effect of the free rotation boundary con-
dition on reducing the post-buckling stiffness decreases.
Group 2 analyses the effect of different plate thicknesses on the
ultimate strength, global behaviour and failure modes. It was
noticed that for the higher plate slenderness, multiple buckling
orders are developed in alternate directions, resulting in lower
strength. By increasing the plate thickness, the ultimate strength
increases with a sudden collapse due to the high rigidity because
the plate becomes stocky.
Group 3 investigates the effect of different local dent depths on
different plate thicknesses combined with the initial global imper-
fection on the ultimate strength, global behaviour and collapse
modes. It was pointed out that by increasing the plate thickness,
Fig. 20. Stress–strain rate vs. local dent depth. the effect of the local dent depth on the ultimate displacement
decreases. For thin plates, the global initial imperfection is domi-
nating the post-collapse behaviour and the presence of the dent
only affects the ultimate strength. By increasing the plate thick-
ness, the effect of the local dent depth on the ultimate strength
decreases. An inflection point of the plate slenderness of a plate
with and without a dent is defined in which the behaviour of the
plate changed. Based on the achieved results, a relationship has
been developed to estimate the ultimate strength reduction as a
function of the plate and dent ratio. A new coefficient accounting
for the effect of a local dent on the ultimate strength of a rectangu-
lar plate element is developed as a function of the dent depth to
plate thickness ratio. For all the defined ranges of the plate thick-
ness, with increasing the dent ratio the corresponding local dent
correction factor is decreasing, which in turn reduces the ultimate
strength.
Based on the existing ultimate strength design formula and the
new developed factor, a new formulation for ultimate strength of
damaged plate has been introduced accounting for the effect of
local dent. The new developed formulation is capable to account
for an initial global imperfection, residual stresses, openings, corro-
sion deterioration and existence of dents. Four levels of plate thick-
Fig. 21. Ultimate strength as a function of plate slenderness, intact plate. nesses are categorized according to their behaviour with respect to
S. Saad-Eldeen et al. / Engineering Structures 99 (2015) 78–91 91

the local dent. It was identified that different behaviour of a plate [10] Saad-Eldeen S, Garbatov Y, Guedes Soares C. Ultimate strength assessment of
steel plates with a large opening. In: Guedes Soares C, Peña L, editors.
with different thicknesses coupled with a local dent exists, which
Developments in maritime transportation and exploitation of sea
gives an indication that there is a scaling effect, which leads to a resources. London, UK: Taylor & Francis Group; 2014. p. 373–80.
different effect to the plates of different thicknesses. [11] Kim UN, Choe IH, Paik JK. Buckling and ultimate strength of perforated plate
panels subject to axial compression: experimental and numerical
investigations with design formulations. Ships Offshore Struct 2009;4:337–61.
Acknowledgement [12] Smith CS. Compressive strength of welded ship grillages. Transact RINA
1975;117:325–59.
[13] Guedes Soares C. Design equation for ship plate elements under uniaxial
The first author has been funded by the Portuguese Foundation compression. J Constr Steel Res 1992;22:99–114.
for Science and Technology (Fundação para a Ciência e Tecnologia [14] Faulkner D. A review of effective plating for use in the analysis of stiffened
plating in bending and compression. J Ship Res 1975;19:1–17.
– FCT) under Contract SFRH/BPD/84823/2012.
[15] Guedes Soares C. Design equation for the compressive strength of unstiffened
plate elements with initial imperfections. J Constr Steel Res 1988;9:287–310.
[16] Smith CS, Davidson PC, Chapman JC, Dowling JP. Strength and stiffness of
References ships’ plating under in-plane compression and tension. Transact RINA
1988;130:277–96.
[1] Dow RS, Smith CS. Effects of localized imperfections on compressive strength [17] Saad-Eldeen S, Garbatov Y, Guedes Soares C. Compressive strength assessment
of long rectangular plates. J Constr Steel Res 1984;4:51–76. of rectangular steel plates accounting for the presence of a local dent or an
[2] Paik JK, Lee JM, Lee DH. Ultimate strength of dented steel plates under axial opening. In: Guedes Soares C, Santos TAR, editors. Maritime technology and
compressive loads. Int J Mech Sci 2003;45:433–48. engineering. London, UK: Taylor & Francis Group; 2015. p. 543–51.
[3] Paik JK. Ultimate strength of dented steel plates under edge shear loads. Thin- [18] Amani M, Alinia MM, Fadakar M. Imperfection sensitivity of slender/stocky
Walled Struct 2005;43:1475–92. metal plates. Thin-Walled Struct 2013;73:207–15.
[4] Luís RM, Guedes Soares C, Nikolov PI. Collapse strength of longitudinal plate [19] Dwight JB, Moxham KE. Welded steel plates in compression. Struct Eng
assemblies with dimple imperfections. Ships Offshore Struct 2008;3:359–70. 1969;47:49–66.
[5] Raviprakash AV, Prabu B, Alagumurthi N. Residual ultimate compressive [20] Guedes Soares C. Uncertainty modelling in plate buckling. Struct Saf
strength of dented square plates. Thin-Walled Struct 2012;58:32–9. 1988;5:17–34.
[6] Witkowska M, Guedes Soares C. Collapse strength of stiffened panels with local [21] Silva JE, Garbatov Y, Guedes Soares C. Ultimate strength assessment of
dent damage. In: Proceedings of 27th international conference on offshore rectangular steel plates subjected to a random non-uniform corrosion
mechanics and arctic engineering (OMAE 2008), Paper OMAE2008-57950: degradation. Eng Struct 2013;52:295–305.
ASME; 2008. [22] Silva JE, Garbatov Y, Guedes Soares C. Reliability assessment of a steel plate
[7] Luís RM, Witkowska M, Guedes Soares C. Collapse behaviour of damaged subjected to random and non-uniform corrosion wastage. Eng Struct
panels with a dimple imperfection. In: Proceedings of 26th international 2014;59:13–20.
conference on offshore mechanics and arctic engineering (OMAE 2007): ASME; [23] Tekgoz M, Garbatov Y, Guedes Soares C. Finite element modelling of the
2007. ultimate strength of stiffened plates with residual stresses. In: Guedes Soares
[8] Amante DO, Chujutali JA, Estefen SF. Residual strength of dented stiffened C, Romanoff J, editors. Marstruct Bk Series. London, UK: Taylor & Francis
panels. In: Proceedings of the ASME 2013 32nd international conference on Group; 2013. p. 309–17.
ocean, offshore and arctic engineering 0MAE2013. Nantes, France: ASME; [24] Tekgoz M, Garbatov Y, Guedes Soares C. Ultimate strength assessment of a
2013. p. OMAE2013-10343. stiffened plate accounting for welding sequences. In: Proceedings of the 11th
[9] Xu M, Guedes Soares C. Assessment of residual ultimate strength for wide international symposium on practical design of ships and other floating
dented stiffened panels subjected to compressive loads. Eng Struct structures (PRADS 2013). Changwon City, Korea: CECO; 2013. p. 1089–95.
2013;49:316–28.