Ocean Engineering 101 (2015) 182–200

Contents lists available at ScienceDirect

Ocean Engineering
journal homepage: [Link]/locate/oceaneng

Ultimate strength of steel brackets in ship structures

Sang Eui Lee a, Anil Kumar Thayamballi a, Jeom Kee Paik a,b,n
a
The Korea Ship and Offshore Research Institute (The Lloyd's Register Foundation Research Centre of Excellence),
Pusan National University, Busan 609-735, Republic of Korea
b
Department of Mechanical Engineering, University College London, Torrington Place, London WC1E 7JE, UK

art ic l e i nf o a b s t r a c t

Article history: Steel brackets are customarily used to prevent sideways deformation or lateral-torsional buckling in the
Received 3 March 2014 supporting components of structures such as ships and offshore platforms. The aims of this study are to
Accepted 4 April 2015 undertake nonlinear finite element analysis to examine the ultimate-strength characteristics of steel
brackets, and to develop a simple design formula to predict the ultimate strength of a steel bracket. The
Keywords: structural features of steel brackets in real ship structures are investigated. Finite element modelling
Ship structures techniques are developed to compute the ultimate-strength behaviour of steel brackets with different
Sideways deformation design variables, such as material type and breadth to height ratio. The findings of the research and the
Ultimate strength above-mentioned design formula have the potential to enhance the structural design and safety
Steel bracket
assessment of steel brackets in ship structures.
Nonlinear finite element analysis
& 2015 Elsevier Ltd. All rights reserved.
Ultimate strength design formula

1. Introduction et al., 1983; Krishnakumar, 1988; Wang and Liew, 1994; Vaaraniemi
et al., 2003; Safar and Machaly, 2005; Aung, 2006) under compres-
Steel-plated structures are widely used in structural systems sive, shear force or combination of them.
such as ships and offshore platforms. They are composed of plate In the early days, the buckling analysis of triangular plates
panels supported by beam members such as longitudinal girders, analytically involved. In particular, Salmon (1962) studied the
transverse frames and stiffeners. As these supporting members elastic stability characteristics of the connection utilising the
tend to deform sideways, brackets are attached to prevent lateral- Rayleigh–Ritz method under the assumption that the load is
torsional buckling or ‘tripping’ (Paik and Thayamballi, 2003; linearly applied on the loaded edge of the bracket with no
Hughes and Paik, 2013). horizontal displacement. Further, Salmon et al. (1964) conducted
The regulations established by various classification societies a series of laboratory tests to investigate the behaviour of eighteen
(DNVGL, 2012; IACS, 2006a, 2006b; LR, 2012) can be used to connections with aspect ratio ranging from 0.75 to 2.0 for small
determine the scantling requirements for steel brackets at the and large plate dimensions to include plates where buckling and
structural-design stage. However, no detailed guidelines for predict- yielding are anticipated. It was confirmed that the maximum
ing the strength performance of steel brackets are available. compressive stress takes place at the free edge which is found
It is noted that research efforts to investigate the strength of a on his previous analytical work (Salmon, 1962). Wang and Liew
steel bracket which has a triangular shape are relatively far less than (1994) utilised the pb-2 Rayleigh–Ritz method to investigate the
rectangular plates such as plates and stiffened panels (Paik and triangular plates under isotropic in-plane compressive load.
Thayamballi, 2003; Vhanmane and Bhattacharya, 2008; Zhang et al., Further the study was extended to buckling of triangular thick
2008; Paik and Seo, 2009a, 2009b; Shi and Wang, 2012; Paik et al., plates based on the Mindlin plate theory (Xiang et al., 1994; Wang
2012). For over the last century, there were a number of researches et al., 1994). Jaunky et al. (1995) studied the buckling of general
related to buckling analysis of a equilateral triangular plate triangular anisotropic plates subjected to combined in-plane loads
(Woinowsky-Krieger, 1933; Conway and Leissa, 1960; Wakasugi, utilising the Rayleigh–Ritz method. Xiang (2002) further investi-
1960b, 1961) and a isosceles triangular plate (Burchard, 1937; gated the buckling behaviour of triangular plates with both
Wittrick, 1954; Li, 1959; Cox and Klein, 1955; Han, 1960; Wakasugi, translational and rotational elastic edge constraints using the p-
1960a; Salmon, 1962; Salmon et al., 1964; Ueda et al., 1976, 1977; Tan Ritz and presented extensive buckling factors for several selected
isosceles and right-angled triangular plates at various edge
n
boundary conditions under isotropic in-plane compressive load.
Corresponding author at: The Korea Ship and Offshore Research Institute (The
Lloyd's Register Foundation Research Centre of Excellence), Pusan National University,
Aung (2006) also used Mindlin plate theory to investigate the
Busan 609-735, Republic of Korea. Tel.: þ 82 51 510 2429; fax: þ 82 51 518 7687. elastic–plastic buckling of various isosceles and right-angled
E-mail address: jeompaik@[Link] (J.K. Paik). triangular plates under combined compression and shear force.

[Link]
0029-8018/& 2015 Elsevier Ltd. All rights reserved.
 S.E. Lee et al. / Ocean Engineering 101 (2015) 182–200 183

Nomenclature uy Displacement on the y-axis

uz Displacement on the z-axis
b1 Breadth of the steel bracket w Reduced area ratio
b2 Breadth of steel-bracket toe wo Initial imperfection of the steel bracket on the x-axis
E Elastic modulus of material y0 y-axis along with the diagonal line c-c
g Increment factor of ultimate bending moment zi Vertical distance from the origin to the i-th node
f yi Force component on the y-axis at the i-th node ZP Plastic section modulus of the steel bracket
h1 Height of the steel bracket α Aspect ratio of steel bracket
h2 Height of the steel bracket toe β1 Slenderness ratio (height to thickness) of the steel
MP Plastic bending moment of the steel bracket bracket
Mu Ultimate bending moment of the steel bracket β2 Slenderness ratio (breadth to thickness) of the steel
ðM u =M P ÞF Nondimensionalised ultimate bending moment with bracket
fixed boundary condition γ i , ηi , κ i Coefficients of design formula for a simply supported
ðM u =M P ÞS Nondimensionalised ultimate bending moment with boundary condition
simply supported boundary condition θx Rotational restraints on the x-axis
ðM u ÞTriangular Ultimate bending moment of the triangular steel θy Rotational restraints on the y-axis
bracket θz Rotational restraints on the z-axis
ðM u ÞRadius Ultimate bending moment of the radiused steel λ1 Radius ratio (radius to height) of the steel bracket
bracket λ2 Radius ratio (radius to breadth) of the steel bracket
R Radius of the steel bracket μi Coefficient of design formula for a fixed boundary
Rx Rotational loading on the vertical edge of the steel condition
bracket ν Poisson's ratio
Ru Reduction factor of ultimate bending moment σx Axial compressive load on the x-axis
s Reduced area of the steel bracket with the radius at σY Yield stress of the material
the free edge σ z0 Stress component on the z-axis perpendicular to the
S Area of the triangular steel bracket diagonal line c-c
tb Thickness of the steel bracket ςi , ξi , ψ i , ζ i Coefficients of design formula for a fixed boundary
ux Displacement on the x-axis condition

As computing speeds and capabilities of numerical tools advance, thickness of a corner bracket in relation to buckling and/or plastic
it is expected that numerical simulations will play an important role strength (Ueda and Yao, 1987). The fundamental idea of the proposed
and contribute to accelerate the level of researches higher than method was that the collapse of a frame and a bracket takes place at
before. Particularly, the numerical simulations to examine the buck- the same time, since it is of no use for a bracket to carry more loads
ling analysis of triangular plates have been employed by lots of after the frame has collapse. However, in case of brackets attached to
researchers (Ueda et al., 1976, 1977; Tan et al., 1983; Vaaraniemi et al., prevent lateral-torsional buckling or ‘tripping’ of stiffeners, it is not
2003; Safar and Machaly, 2005). always true. If a bracket carries more loads, a stiffened-plate panel
The most distinguished numerical and experimental works are would stand more.
Ueda's series of studies (Ueda et al., 1976, 1977; Ueda and Yao, 1987) Furthermore, Safar and Machaly (2005) conducted experimental
for the triangular corner brackets using finite element method (FEM). and analytical research work on triangular bracket plates considering
A series of buckling analysis, elastic–plastic large deflection analysis both material and geometric nonlinearities. It was experimentally
and elastic–plastic thermal stress analysis were conducted for the confirmed that yielding along the free edge usually takes place prior
triangular corner brackets subjected to compression to clarify the to buckling and the distribution of contact stresses between the
effects of initial imperfection due to welding (Ueda et al., 1977). It beam and the bracket was triangular in shape with the peak stresses
was observed that initial deflection decreases the rigidity and at free edge of bracket at buckling. It was concluded that the
ultimate strength of a triangular corner bracket and this tendency connection possesses a significant amount of post-buckling strength
is more remarkable when the side length to thickness ratio such that the limit load is almost twice the critical load. The results of
decreases. It was found that the welding residual stresses in the all studies offer useful insights into the design of steel brackets. To
triangular corner bracket are usually in tension, and these stresses the best of the authors' knowledge, however, a limited number of
increase the buckling strength and the ultimate strength of the researches have been conducted on either the load-carrying capacity
bracket. At last, a method was proposed to determine the optimum of brackets or their ultimate strength.

Fig. 1. A schematic representation of typical steel bracket attached to a strong support member.
184 S.E. Lee et al. / Ocean Engineering 101 (2015) 182–200

The aim of this study is to use nonlinear finite element analysis in Fig. 1. The bracket is assumed to be attached to one side of the
to examine the ultimate-strength characteristics of steel brackets. support member, i.e., on the positive side of the y-axis.
The structural features of steel brackets in real ships are investi- The following five parameters are considered: aspect ratio
gated using data collected on six commercial ships currently in ðα ¼ b1 =h1 Þ; height to thickness ratio ðβ1 ¼ h1 =t b Þ; breadth to thick-
service: three tankers, two bulk carriers and one liquefied natural ness ratio ðβ2 ¼ b1 =t b Þ; radius to height ratio ðλ1 ¼ R=h1 Þ; and radius
gas carrier (LNGC). Finite element modelling techniques are to breadth ratio ðλ2 ¼ R=b1 Þ.
developed to calculate the ultimate-strength behaviour of steel
brackets with a range of design variables, such as material type
2.2. Geometrical features of steel brackets
and breadth to height ratio. The validation of the developed
modelling is conducted. Numerical computation is used to derive
Data on 52 steel brackets were collected from six merchant
a plausible design formula that predicts the ultimate strength of a
ships, namely three oil tankers, two bulkers and one LNGC. The
steel bracket.
principal features of the six ships are displayed in Table 1. Brackets
with h1 values greater than 1000 mm require additional strength-
ening members in the form of flat or angled bars, and were thus
2. Structural features of steel brackets in ship structures excluded from the analysis. Only the structural characteristics of
unstiffened brackets were considered.
2.1. Definition of geometrical parameters The geometrical parameters defined in Section 2.1 were then
calculated for the 52 steel brackets collected. Figs. 2 and 3 show
The geometrical attributes of a typical steel bracket attached to the statistical distribution of the parameters, and Table 2 sum-
a strong support member composed of plate panels are defined as marises the results of the analysis. These findings were used to
identify the geometrical parameters of a standard steel bracket, as
Table 1 follows: h1 ¼ b1 ¼ 600 mm, tb ¼12 mm and R¼800 mm. For sim-
Principal dimensions of ships.
plicity, it was assumed that h2 ¼ b2 ¼ 0.
Oil tankers Bulk carriers LNGC

Panamax Suezmax VLCC 82 K 151 K 213 K 2.3. Full plastic bending capacity of steel brackets

LBP (m) 219.0 261.0 315.0 222.0 265.0 302.4 In the steel bracket with uniform material properties, the first yield
Breadth (m) 32.2 48.0 58.0 32.26 46.0 50.0 occurs at the outer fibre of the cross-section where the highest
Depth (m) 20.6 23.2 29.0 20.2 22.7 27.0
Draft (m) 12.5 17.0 28.0 12.2 15.2 12.0
compression develops in the span. With further loading, the cross-
section will become entirely plastic. The entire bending capacity of the

0.5 0.5

0.4 0.4
Relative frequency

Relative frequency

0.3 0.3

0.2 0.2

0.1 0.1

0.0 0.0
200 400 600 800 1000 1200 200 400 600 800 1000 1200
h1 (mm) b1 (mm)

0.5 0.5

0.4 0.4
Relative frequency

Relative frequency

0.3 0.3

0.2 0.2

0.1 0.1

0.0 0.0
200 400 600 800 1000 1200 1400 1600 8 10 12 14 16 18 20 22 24
R (mm) tb (mm)
Fig. 2. Geometrical characteristics of steel brackets in real ships: (a) height, (b) breadth, (c) radius, and (d) thickness.
 S.E. Lee et al. / Ocean Engineering 101 (2015) 182–200 185

1.0

Relative frequency
0.8

0.6

0.4

0.2

0.0
0.6 0.8 1.0 1.2 1.4 1.6
α= h1/b1

0.5 0.5

0.4 0.4
Relative frequency

Relative frequency
0.3 0.3

0.2 0.2

0.1 0.1

0.0 0.0
10.0 20.0 30.0 40.0 50.0 60.0 10.0 20.0 30.0 40.0 50.0 60.0
β1= h1/tb β2= b1/tb

0.5 0.5

0.4 0.4
Relative frequency

Relative frequency

0.3 0.3

0.2 0.2

0.1 0.1

0.0 0.0
0.4 0.8 1.2 1.6 2.0 0.4 0.8 1.2 1.6 2.0 2.4
λ1= R/h1 λ2= R/b1
Fig. 3. Geometrical characteristics of steel brackets in real ships: (a) aspect ratio of height to breadth; (b) slenderness ratio of height to thickness; (c) slenderness ratio of
breadth to thickness; (d) radius ratio of radius to height; and (e) radius ratio of radius to breadth.

steel bracket typically depends on the cross-sectional geometry as well Table 2

Actual range and the most probable dimensions of steel brackets.
as material properties.
As an example, the schematic definition of stress distribution in the Variable Range Most probable
steel bracket along the diagonal line, c-c, under rotational compression
is illustrated in Fig. 4 (Ueda et al., 1977). The full plastic bending h1 (mm) 300–1000 600
b1 (mm) 300–1000 600
capacity, M P , is calculated by the first moment of axial stresses with
R (mm) 300–1500 600, 700
regard to the O-O axis when the cross-section entirely yields, namely: t b (mm) 11–22 12
α 0.75–1.5 1.0
Z l β1 20–55 30
MP ¼ σ z0 t b y0 dy0 ¼ Z P σ Y ð1Þ β2 20–55 30
0
λ1 0.7–1.7 1.2
2
λ2 0.7–2.0 1.2
where Z P ¼ l t b =8 is the plastic section modulus.
186 S.E. Lee et al. / Ocean Engineering 101 (2015) 182–200

z
A A A 1st buckling mode
z y
l
M c
h1

O B y O B O B
c
b1

c
σY σY

tb l

O O O O O O
c
Fig. 4. Stress distribution of a rectangular cross-section at first and full yielding.

z
uy z z
b2
A R A A′ A A′
Rx Rx
h1
h2
y y y
O B O B O B
b1
Fig. 5. Schematic representations of typical and applied geometries for steel brackets: (a) nomenclature; (b) typical geometry; and (c) applied geometry.

Fig. 6. Examples of the applied lowest eigen buckling mode for h1 ¼ b1 ¼600 mm, t b ¼ 12 mm, σ Y ¼ 235 MPa, w0 ¼ 0:3t b : (a) simply supported and (b) fixed.

3. Nonlinear finite element modelling response of the bracket is then as shown in Fig. 5(b or c). For
simplicity, only the latter case is considered here. It is assumed that
3.1. Finite-element model the vertical and horizontal edges of the steel bracket remain straight,
and that the in-plane degrees of freedom at the horizontal edge are
A bracket with the standard parameters detailed in Section 2.2 is restrained.
regarded as subject to in-plane bending moments arising from Nonlinear finite element analysis is performed using ANSYS (2012),
tripping or sideways deformation, as shown in Fig. 5(a). The structural to accommodate both geometrical and material nonlinearities.
 S.E. Lee et al. / Ocean Engineering 101 (2015) 182–200 187

The SHELL181 element, which has four nodes with six degrees of As a result, the whole stiffened panel may collapse with little increase
freedom at each node, is used to model the steel bracket. The bracket in the load, as shown in Fig. 8. Therefore, a bracket affixed to the
is modelled in elastic perfectly plastic, without considering strain point where the web is most vulnerable to tripping plays a significant
hardening. The material under study is mild steel with an elastic role in preventing the overall collapse of the panel.
modulus E of 205.8 GPa; a Poisson's ratio v of 0.3; and a yield strength In a continuous steel-plated structure subject to the sideways
σY of 235 MPa. deformation or tripping of strong support members, the general
The maximum magnitude of the bracket's fabrication-related deformation of a plate-beam combination can be described as
initial deflection w0 is assumed to be 30% of the thickness of the shown in Fig. 9(a). The tripping of the web may involve both
bracket; that is, w0 ¼ 0:3t b . The eigenvalue buckling mode is used sideways deformation and vertical deflection; and the rotation of
to determine the shape of the initial deflection. Fig. 6 provides an the stiffener web may be coupled with local buckling in the
example of the smallest buckling mode obtained from the eigen- attached plating. Nonlinear finite element analysis provides an
value buckling analysis with simply supported or fixed boundary accurate representation of tripping behaviour in specific cases.
conditions. For simplicity, the residual stress caused by welding is However, it is difficult to use this method to investigate the
not considered in this study.
The ultimate strength of the bracket is then taken as the maximum
load-carrying capacity under in-plane bending moments arising from
tripping. When the bracket is schematically defined, as in Fig. 7, its z
ultimate strength can be computed as follows:
X
n uy
Mu ¼ f yi zi ð2Þ A A′
i¼1

where f yi is a force component parallel to the y-axis at the i-th node of

the finite element; zi is the vertical distance from the origin, O; and n B
is the number of nodes along the vertical edge of the bracket. O B′ uz
y
3.2. Loading conditions
z
Assemblies of plating and stiffeners in ships and other offshore
structures may buckle if they are subjected to compressive loads uy
A A′
beyond a critical value. The buckling pattern of a stiffened panel
can be categorised into two major types, namely, overall buckling
and local buckling (Paik and Thayamballi, 2003). Forms of local
buckling, such as the sideways deformation or tripping of a web, B
occur relatively suddenly, causing the stiffened panel to unload. O uz
According to Paik and Thayamballi (2003), webs that are too high y
or too thin may experience local buckling, much like plate elements.
If support members of a web have insufficient torsional rigidity, the
z
web may twist sideways. This process, known as lateral-torsional
buckling, leaves the buckled or collapsed plating with no stiffeners.
uy
A A′
z
A
O B uz
fyi y
ri Fig. 9. General and idealised tripping deformations for a plate-stiffener combina-
tion: (a) general deformations of flexible stiffener web with plate rotational
O B y restraints; (b) simplified deformations of flexible web without plate rotational
restraints; and (c) simplified deformations of rigid stiffener web without plate
Fig. 7. Nomenclature for defining ultimate strength of the steel bracket. rotational restraints.

Fig. 8. Schematic of stiffened panel collapse modes associated with failure of support members (Paik and Thayamballi, 2003): (a) collapse by local buckling of stiffener web
and (b) collapse by tripping of stiffener.
188 S.E. Lee et al. / Ocean Engineering 101 (2015) 182–200

z z z
θz

free free
ux= uy= uz= 0 ux= uy= uz= θz= 0

θy
θx
y y y
x x ux= uy= uz= 0 x ux= uy= uz= θy= 0

Fig. 10. Coordinate system and applied boundary conditions of the steel bracket: (a) coordinate system; (b) simply supported; and (c) fixed.

250 250
Bending moment (kNm)

Bending moment (kNm)

200 200

150 150

100 100

50 50

0 0
0 0.1 0.2 0.3 0.4 0.5 0 0.1 0.2 0.3 0.4 0.5
Angle (deg) Angle (deg)
Fig. 11. Results of mesh-convergence study for h1 ¼ b1 ¼600 mm, t b ¼ 12 mm, σ Y ¼ 235 MPa, w0 ¼ 0:3t b : (a) simply supported and (b) fixed.

250 20

200 16
Force (kN)
Mu (kNm)

150 12

100 8

50
4

0
0
0 400 800 1200 1600 2000
0 10 20 30 40
Fig. 12. Results of mesh-convergence study for h1 ¼b1 ¼ 600 mm, t b ¼12 mm,
Deflection (mm)
σ Y ¼ 235 MPa, and w0 ¼ 0:3t b .
Fig. 13. Validation study on developed FE modelling technique (Ueda et al., 1977).

ultimate strength of a bracket supporting a plate–stiffener combi-

nation in general before tripping.
To solve this problem, tripping deformation can be idealised by Table 3
approximating to the most general case. The following three Variables for the series analysis associated with the effect of radius ratio.
idealisations are available (Paik and Thayamballi, 2003):
h1 b1 R (mm) t b (mm) σ Y Boundary
 A flexible web without plate rotational restraints (see Fig. 9(b)); (mm) (mm) (MPa) condition

 a rigid web with plate rotational restraints; or 300 300 300, 400, 500, 600, 900 12, 16, 235.0 S.S., fixed
 a rigid web without plate rotational restraints (see Fig. 9(c)). 20
600 600 600, 800, 1000, 1200, 12, 16, 235.0 S.S., fixed
1800 20
1000 1000 1000, 1200, 1400, 2000, 12, 16, 235.0 S.S., fixed
To ensure the simplicity and stability of the computational
3000 20
solutions, the distribution of the normal displacement along the
vertical edge of the steel bracket is assumed to be linear, allowing Note: S.S. represents the simply supported boundary condition.
 S.E. Lee et al. / Ocean Engineering 101 (2015) 182–200 189

the load-carrying capacity of the bracket alone to be assessed as

z z shown in Fig. 9(c).
Triangular
A A λ=1.5
s
λ=1.2 3.3. Boundary conditions
λ=1.0
Various types of support member are positioned along the edges of
the plate components of steel-plated structures. Each support member
y y has a finite degree of torsional rigidity, which determines the extent to
O B O B which it restrains the rotation of the plate edges. Zero rotational
restraint corresponds to a simply supported edge condition; the edge
Fig. 14. Schematic definition of steel brackets for varying radius ratios: (a) area
reduction and (b) radius ratio variation. condition is fixed when rotational restraint is infinite (Paik and

Simply supported Fixed

120 120

90 90
Mu (kNm)

Mu (kNm)
60 60

30 30

0 0
0.4 0.6 0.8 1.0 0.4 0.6 0.8 1.0
w w
Simply supported Fixed
400 400

300 300
Mu (kNm)

Mu (kNm)

200 200

100 100

0 0
0.4 0.6 0.8 1.0 0.4 0.6 0.8 1.0
w w
Simply supported Fixed
1200 1200

900 900
Mu (kNm)

Mu (kNm)

600 600

300 300

0 0
0.4 0.6 0.8 1.0 0.4 0.6 0.8 1.0
w w
Fig. 15. Effects of the radius ratio on the ultimate strength of steel brackets for α ¼ 1.0, t b ¼ 12 mm and σ Y ¼ 235 MPa: (a) simply supported and (b) fixed.
190 S.E. Lee et al. / Ocean Engineering 101 (2015) 182–200

Simply supported Fixed

1.0 1.0

0.8 0.8

0.6 0.6
Ru

Ru
0.4 0.4

0.2 0.2

0.0 0.0
0.4 0.6 0.8 1.0 0.4 0.6 0.8 1.0
w w
Simply supported Fixed
1.0 1.0

0.8 0.8

0.6 Ru 0.6
Ru

0.4 0.4

0.2 0.2

0.0 0.0
0.4 0.6 0.8 1.0 0.4 0.6 0.8 1.0
w w
Simply supported Fixed
1.0 1.0

0.8 0.8

0.6 0.6
Ru

Ru

0.4 0.4

0.2 0.2

0.0 0.0
0.4 0.6 0.8 1.0 0.4 0.6 0.8 1.0
w w
Fig. 16. Effects of the radius ratio on the ultimate strength reduction rate of steel brackets for α¼ 1.0, t b ¼12 mm and σ Y ¼ 235 MPa: (a) simply supported and (b) fixed.

Thayamballi, 2003). Therefore, the boundary condition of the plating  Simply supported boundary condition; see Fig. 10(b).
in a continuous stiffened plate structure such as a ship is neither Vertical edge: translational restraints in the x-, y- and
simply supported nor fixed, because the torsional rigidity of the z-directions, ux ¼ uy ¼ uz ¼ 0.
support members at the plate edges is neither zero nor infinite. Horizontal edge: translational restraints in the x-, y- and
Accordingly, simply supported and fixed boundary conditions are used z-directions, ux ¼ uy ¼ uz ¼ 0.
in this study to enable consideration of extreme cases (zero and  Fixed boundary condition, as shown in Fig. 10(c).
infinite rotational restraint), given that the boundary condition must Vertical edge: translational restraints in the x-, y- and
fall between the two extremes. z-directions, ux ¼ uy ¼ uz ¼ 0; rotational restraint in the
The boundary conditions investigated in this study are z-direction, θz ¼ 0.
described in Fig. 10. The coordinate system used for their mea- Horizontal edge: translational restraints in the x-, y- and
surement is shown in Fig. 10(a). The restraints are described in z-directions, ux ¼ uy ¼ uz ¼ 0; rotational restraint in the
detail below. z-direction, θy ¼ 0.
 S.E. Lee et al. / Ocean Engineering 101 (2015) 182–200 191

Simply Supported Fixed

1.0 1.0

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0.0 0.0
h1 R t h1 R t
(mm) (mm) (mm) (mm) (mm) (mm)
Fig. 17. Sensitivity of height, radius and thickness of steel brackets on the ultimate strength.

3.4. Mesh-convergence analysis Table 4

Variables for the series analysis associated with the effect of aspect ratio.
This section presents the results of mesh-convergence analysis of
α h1 :b1 h1 :b1 (mm) Boundary condition
the targeted bracket when σ Y ¼235 MPa and w0 ¼ 0:3t b , as shown
in Fig. 11. The ultimate bending moments for different boundary 0.50 2.0:1.0 600:300, 800:400, 1000:500 S.S., fixed
conditions are summarised in Fig. 12. Approximately 230 elements are 0.67 1.5:1.0 450:300, 600:400, 750:500, 900:600 S.S., fixed
sufficient to estimate the ultimate bending moment of a steel bracket. 0.75 4.0:3.0 400:300, 600:800 S.S., fixed
1.33 3.0:4.0 300:400, 600:800 S.S., fixed
1.50 1.0:1.5 300:450, 400:600, 500:750, 600:900 S.S., fixed
3.5. Validation with experiment 2.00 1.0:2.0 300:600, 400:800, 500:1000 S.S., fixed

The FE modelling technique developed in the present study

was validated with the experimental result (Ueda et al., 1977).
Fig. 13 shows the result of validation study for h1 ¼ b1 ¼ 320 mm, variables used in this section are summarised in Table 3. Ninety
tb ¼3.2 mm, h1 ¼ b1 ¼ R ¼ 0, wo ¼ 0:03t b and wo ¼ 1:11t b at fixed series are analysed.
boundary condition. It is confirmed that the developed FE model- A schematic definition of steel brackets is described in Fig. 14.
ling technique is effective for simulating ultimate strength of the The area ratio, w, defined as follows, is used to investigate the
steel bracket. effects of reduced area on the ultimate strength.
 s
w ¼ 1 ð3Þ
4. Effects of design variables S
In the above, s is the reduced area of the steel bracket with
In this section, three sets of parameters are analysed and the differences in radius and S is the total area of the triangular steel
results of the analysis are presented. First, to investigate the effects bracket, where S ¼ ðh1 b1 Þ=2.
of the slenderness ratio on the ultimate strength, five radius ratios The reduction factor Ru , defined as follows, is used to calculate
are considered. Second, the aspect ratio is taken as the design the ultimate-strength reduction rate.
variable; its effect on ultimate strength is evaluated with reference
ðM u ÞTriangular  ðM u ÞRadius
to six aspect ratios. Finally, to consider possible dimensional Ru ¼ ð4Þ
ðM u ÞTriangular
variations in the full range of bracket characteristics, four aspect
ratios, seven slenderness ratios and three types of material are In the above, ðM u ÞTriangular is the ultimate bending moment of a
considered. In addition, the sensitivity of each variable is analysed. triangular steel bracket and ðM u ÞRadius is the ultimate bending
The initial deflection of the bracket is assumed to be 30% of the moment of a radiused steel bracket.
bracket's thickness. The boundary conditions are either simply The ultimate strength of a steel bracket with a varying radius
supported or fixed. The analysis covers 834 cases. ratio is illustrated in Fig. 15. As the radius increases, the ultimate
bending moment of the radiused bracket approaches that of the
4.1. Effects of the radius ratio triangular steel bracket, regardless of the thickness and boundary
conditions. The thicker the bracket, the more rapidly the ultimate
In this section, nonlinear finite element analysis is conducted with bending moment increases.
an aspect ratio of 1.0 to investigate the effects of the slenderness ratio Fig. 16 illustrates different rates of reduction in the ultimate
on the ultimate strength of a steel bracket. The radius ratio is strength of a steel bracket. At heights of both 300 mm and
considered to range from 1.0 to 3.0, which contains the range 600 mm, the ultimate bending moment is significantly (up to
identified in Fig. 3(d and e). Bracket height is set at 300, 600 and 50%) smaller than the ultimate strength of the triangular bracket
1000 mm, with corresponding values for the radius. To further the when λ1 ¼ 1:3, which is very close to the most probable value
test on the effects of the slenderness ratio, three degrees of thickness ðλ1 ¼ λ2 ¼ 1:2Þ. Regardless of the bracket's height, the maximum
(12, 16 and 20 mm) of mild steel (σ Y ¼235 MPa) are considered. The reduction rate occurs between 0.5 and 0.7 when λ1 ¼ 1:0.
192 S.E. Lee et al. / Ocean Engineering 101 (2015) 182–200

Simply supported Fixed

200 400

Bending moment (kNm)

Bending moment (kNm)

160
300

120
200
80

100
40

0 0
0 0.1 0.2 0.3 0.4 0.5 0 0.1 0.2 0.3 0.4 0.5
Angle (deg) Angle (deg)

150 400
Bending moment (kNm)

120 Bending moment (kNm)

300

90
200
60

100
30

0 0
0 0.1 0.2 0.3 0.4 0.5 0 0.1 0.2 0.3 0.4 0.5
Angle (deg) Angle (deg)

Simply supported Fixed

150 400
Bending moment (kNm)

Bending moment (kNm)

120
300

90
200
60

100
30

0 0
0 0.1 0.2 0.3 0.4 0.5 0 0.1 0.2 0.3 0.4 0.5
Angle (deg) Angle (deg)
Fig. 18. Effects of the aspect ratio on the ultimate strength of steel brackets for t b ¼12 mm and σ Y ¼235 MPa: (a) simply supported and (b) fixed.

Therefore, radiused steel brackets are not recommended for 4.2. Effects of aspect ratio
general use, because their ultimate bending moment is signifi-
cantly smaller than that of triangular brackets. In this section, nonlinear finite element analysis is performed
Fig. 17 reveals the sensitivity of the bracket's ultimate strength with the most probable thickness of mild steel, tb ¼ 12 mm
to height, radius and thickness. Height (¼ breadth; aspect (σ Y ¼ 235 MPa), to investigate the effect of aspect ratio on the
ratio ¼1.0) is the main factor enhancing the ultimate strength of ultimate strength of a steel bracket. The aspect ratio is assumed to
the steel bracket; the effect of radius is weaker than that of the vary from 0.5 (h1:b1 ¼2.0:1.0) to 2.0 (h1:b1 ¼ 1.0:2.0). The max-
other variables. imum height and breadth are set within the identified range
 S.E. Lee et al. / Ocean Engineering 101 (2015) 182–200 193

(greater than or equal to 300 mm and less than or equal to larger than 1.0 (signifying greater breadth), the differences tend to
1000 mm). be larger as the height increases. Table 5 summarises the effects of
In light of the findings outlined in Section 4.1, the radius of the aspect ratio on the ultimate bending moment. The maxi-
steel bracket is not considered in this analysis. The variables used, mum increment is 18.8% greater when α ¼2.0 (h1 ¼1000 mm,
and their ranges, are summarised in Table 4. Thirty series are b1 ¼500 mm) than when α ¼0.5 (h1 ¼500 mm, b1 ¼1000 mm).
analysed. Fig. 19 describes the sensitivity of a bracket's ultimate strength
The factor g, defined as follows, is used to calculate the increment to height and breadth. The effect of breadth is shown to be more
of ultimate strength. significant than that of height. Therefore, to ensure the reliable
structural behaviour and integrity of ship structures, an aspect
ðM u Þα1  ðM u Þα0 ratio greater than or equal to 1.0 is recommended. This finding
g¼  100 ð5Þ
ðM u Þα0 suggests that increasing the breadth of a steel bracket is an
efficient means of increasing its ultimate strength.
In the above, ðM u Þα1 is the ultimate bending moment at aspect
ratio α1 , and ðM u Þα0 is the ultimate bending moment at aspect
ratio α0 . ðα0 ; α1 Þ is considered in the following combinations: 4.3. Combined effects of both radius ratio and aspect ratio
(0.75, 1.33), (0.67, 1.50) and (0.50, 2.00).
Fig. 18 illustrates the influence of aspect ratio on the ultimate In this section, the dimensional characteristics of the steel
bending moment. Under the simply supported boundary condi- brackets investigated in real commercial ships are used as design
tion, but not the fixed boundary condition, the ultimate bending variables. To consider possible dimensional variation in the whole
moment varies considerably with aspect ratio. For aspect ratios range of bracket characteristics, four aspect ratios, seven

Table 5 Table 6
Effects of the aspect ratio on the ultimate strength. Variables for the series analysis.

α h1 (mm) b1 (mm) M u (kNm) g (%) α t b (mm) σ Y (MPa) Boundary

h1 b1
(mm) (mm) condition
S.S. Fixed S.S. Fixed
1.00 300 300 10, 12, 14, 16, 18, 20, 235.0, 313.6, S.S, fixed
0.50 600 300 51.631 94.095 – –
400 400 22 352.8
2.00 300 600 53.483 95.129 3.5 1.1
500 500
0.50 800 400 74.094 159.230 – –
600 600
2.00 400 800 80.587 161.260 8.8 1.3
700 700
0.50 1000 500 96.617 234.640 – –
800 800
2.00 500 1000 114.780 237.290 18.8 1.1
900 900
0.67 450 300 45.419 82.080 – –
1000 1000
1.50 300 450 46.023 82.534 1.3 0.6
0.67 600 400 66.303 140.090 – – 1.33 300 400 10, 12, 14, 16, 18, 20, 235.0, 313.6, S.S, fixed
1.50 400 600 68.498 140.640 3.3 0.4 600 800 22 352.8
0.67 750 500 56.503 207.310 – –
1.50 300 450 10, 12, 14, 16, 18, 20, 235.0, 313.6, S.S, fixed
1.50 500 750 92.584 208.410 7.0 0.5
400 600 22 352.8
0.67 600 900 124.980 282.610 – –
500 750
1.50 900 600 109.980 280.500 13.6 0.8
600 900
0.75 400 300 42.578 76.183 – –
1.33 300 400 42.901 76.432 0.8 0.3 2.00 300 600 10, 12, 14, 16, 18, 20, 235.0, 313.6, S.S, fixed
0.75 800 600 103.310 263.190 – – 400 800 22 352.8
1.33 600 800 112.270 265.210 8.7 0.8 500 1000

Simply Supported Fixed

1.0 1.0

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0.0 0.0
h1 b1 h1 b1
(mm) (mm) (mm) (mm)
Fig. 19. Sensitivity of height and breadth of steel brackets on the ultimate strength.
194 S.E. Lee et al. / Ocean Engineering 101 (2015) 182–200

Simply supported Fixed

1000 2000

800 1600
Mu (kNm)

Mu (kNm)
600 1200

400 800

200 400

0 0
0 20 40 60 80 100 0 20 40 60 80 100
β1 β1
Simply supported Fixed
1000 2000

800 1600
Mu (kNm)

Mu (kNm)
600 1200

400 800

200 400

0 0
0 20 40 60 80 100 0 20 40 60 80 100
β1 β1

Simply supported Fixed

1000 2000

800 1600
Mu (kNm)

Mu (kNm)

600 1200

400 800

200 400

0 0
0 20 40 60 80 100 0 20 40 60 80 100
β1 β1
Fig. 20. Ultimate strength of steel brackets for α ¼1.0 (h1 : b1 ¼1.0:1.0): (a) simply supported and (b) fixed.

slenderness ratios and three types of material are analysed. In light respectively. Accordingly, materials with equivalent degrees of
of the results outlined in Sections 4.1 and 4.2, the effect of four yield stress are analysed. Nonlinear finite element analysis is
aspect ratios greater than or equal to 1.0 on the ultimate strength performed with the variables listed in Table 6, yielding 714 cases
of a triangular bracket is considered. for analysis.
The applied thickness ranges from 10 mm to 22 mm for every Figs. 20, 21, 22 and 23 represent the results of the nonlinear
2-mm interval. As the slenderness ratio is determined by a given finite element analysis when α ¼1.0, 1.33, 1.5 and 2.0 for material
aspect ratio and thickness, it is likely to vary between cases. The types and boundary conditions, respectively. Generally, it is
types of steel most commonly used in the shipbuilding industry observed that the larger the material's yield stress and the thicker
are mild steel, AH32 high-tensile steel and AH36 high-tensile the bracket, the higher the ultimate bending moment of the
steel, for which σ Y ¼235 MPa, 313.6 MPa and 352.8 MPa, bracket. This result is unsurprising.
 S.E. Lee et al. / Ocean Engineering 101 (2015) 182–200 195

Simply supported Fixed

1000 2000

800 1600
Mu (kNm)

Mu (kNm)
600 1200

400 800

200 400

0
10 20 30 40 50 60 10 20 30 40 50 60
β1 β1

Simply supported Fixed

1000 2000

800 1600
Mu (kNm)

Mu (kNm)
600 1200

400 800

200 400

0 0
10 20 30 40 50 60 10 20 30 40 50 60
β1 β1

Simply supported Fixed

1000 2000

800 1600
Mu (kNm)

Mu (kNm)

600 1200

400 800

200 400

0 0
10 20 30 40 50 60 10 20 30 40 50 60
β1 β1
Fig. 21. Ultimate strength of steel brackets for α¼ 1.33 (h1:b1 ¼3.0:4.0): (a) simply supported and (b) fixed.

Fig. 24 illustrates the sensitivity of the bracket's ultimate 5. Empirical formulations of the bracket ultimate strength
strength to the design variables. First, boundary-condition type
has little effect on its sensitivity. Second, thickness is found to be The results of the parametric analysis described in Section 4.3
the most significant determinant of ultimate strength. Third, are used to derive empirical formulations of predicting the ultimate
ultimate strength is less sensitive to material type than to the strength of a steel bracket, as follows:
other three variables. These findings suggest that the most
effective design strategy for preventing local buckling is to control ðM u =M P ÞS ¼ γ i ðh1 =tÞ2 þ ηi ðh1 =tÞ þ κ i ðsimply supportedÞ ð6Þ
the thickness of the steel bracket. When the variation in thickness
is limited, increasing the breadth is a more efficient means of
enhancing the load-carrying capacity than increasing the height. ðM u =M P ÞF ¼ ςi ðh1 =tÞ3 þ ξi ðh1 =tÞ2 þ ψ i ðh1 =tÞ þ ζ i ðfixedÞ ð7Þ
196 S.E. Lee et al. / Ocean Engineering 101 (2015) 182–200

1000 2000

800 1600
Mu (kNm)

Mu (kNm)
600 1200

400 800

200 400

0 0
10 20 30 40 50 60 10 20 30 40 50 60
β1 β1

Simply supported Fixed

1000 2000

800 1600
Mu (kNm)

Mu (kNm)
600 1200

400 800

200 400

0 0
10 20 30 40 50 60 10 20 30 40 50 60
β1 β1

Simply supported Fixed

1000 2000

800 1600
Mu (kNm)

Mu (kNm)

600 1200

400 800

200 400

0 0
10 20 30 40 50 60 10 20 30 40 50 60
β1 β1
Fig. 22. Ultimate strength of steel brackets for α¼ 1.5 (h1:b1 ¼ 1.0:1.5): (a) simply supported and (b) fixed.

In the above, ðM u =M P ÞS and ðM u =M P ÞF are the non-dimen- structural design, the formula for the simply supported condition
sionalised ultimate bending moment for the simply supported is recommended.
and fixed boundary conditions, respectively, i is the number of
fitting equations.
Fig. 25 reveals the non-dimentionalised ultimate strength of 6. Discussion and usage of the design formulations
steel brackets and their fitting curves for α ¼1.0, 1.33, 1.5 and 2.0.
The coefficients of the design formula for different material types In the present study, the ultimate strength of steel brackets in
and aspect ratios are indicated in Table 7. For use in conservative ship structures is numerically investigated with varying parameters
 S.E. Lee et al. / Ocean Engineering 101 (2015) 182–200 197

Simply supported Fixed

1000 2000

800 1600
Mu (kNm)

Mu (kNm)
600 1200

400 800

200 400

0 0
10 20 30 40 50 60 10 20 30 40 50 60
β1 β1

Simply supported Fixed

1000 2000

800 1600
Mu (kNm)

600 Mu (kNm) 1200

400 800

200 400

0 0
10 20 30 40 50 60 10 20 30 40 50 60
β1 β1

Simply supported Fixed

1000 2000

800 1600
Mu (kNm)

Mu (kNm)

600 1200

400 800

200 400

0 0
10 20 30 40 50 60 10 20 30 40 50 60
β β
1 1

Fig. 23. Ultimate strength of steel brackets for α ¼ 2.0 (h1 : b1 ¼1.0:2.0): (a) simply supported and (b) fixed.

of influence such as radius ratio, aspect ratio and combination of health monitoring; for basic design (Varela et al., 2011; Beghin, 2013;
them. Based on the computations, two sets of empirical formulae for Payer and Schellin, 2013; Zanic, 2013; Zanic et al., 2013; Na and Karr,
predicting the ultimate strength are proposed for simply supported 2013), for detailed design (Hwang et al., 2010), and for least cost
and fixed boundary conditions. optimisation (Richir et al., 2007). They will also be useful for fatigue
It is realized that the design formulations developed can be used design (Lotsberg, 2007; Petinov et al., 2006; Lotsberg et al., 2008;
for various purposes within the framework of multi-criteria synthesis Gayton and Lemaire, 2009; Yu et al., 2009; Lotsberg, 2009, Gaspar
of ship and offshore structures in terms of design, construction and et al., 2011), for structural modifications (Rizzo, 2006; Henriksen
198 S.E. Lee et al. / Ocean Engineering 101 (2015) 182–200

Simply Supported Fixed

1.0 1.0

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0.0 0.0
h1 b1 t σY h1 b1 t σY
(mm) (mm) (mm) (MPa) (mm) (mm) (mm) (MPa)

Fig. 24. Sensitivity of height, breadth, thickness and yield stress on the ultimate strength.

α= 1.00 α= 1.33
1.20 1.20
FEM Fitting FEM Fitting
Simply supported Simply supported
1.00 Fixed 1.00 Fixed
Mu / MP

Mu / MP
0.80 0.80

0.60 0.60

0.40 0.40

0.20 0.20

0.00 0.00
0 20 40 60 80 100 0 20 40 60 80 100
β1 β1

α= 1.50 α= 2.00
1.20 1.20
FEM Fitting FEM Fitting
Simply supported Simply supported
1.00 Fixed 1.00 Fixed
Mu / MP

Mu / MP

0.80 0.80

0.60 0.60

0.40 0.40

0.20 0.20

0.00 0.00
0 20 40 60 80 100 0 20 40 60 80 100
β1 β1
Fig. 25. Nondimensionalised ultimate strength and design formula of each material for α ¼1.0, 1.33, 1.5 and 2.0.

et al., 2008), for inspections (Anastasopoulos et al., 2009), and for Table 7
maintenance and repair (Turan et al., 2009). Brackets also play a key Coefficients of design formulae for different material types and aspect ratios.
role in accidental conditions (Haris and Amdahl, 2012). It has been
α Fixed Simply Supported
recognised that failure of brackets may accelerate the hull collapse
and fracture of ship structures (Drouin, 2006). ς (10 7
) ξ (10 4
) ψ (10 3
) ζ γ (10  5) η (10  2) κ
However, the computations showed a lot of uncertainties due
to geometrical and boundary conditions, among others. While the 1.00 7.506  1.461 2.578 0.953 10.337  1.576 0.849
1.33 7.581  1.502 2.017 0.886 13.096  1.748 0.808
design formulae developed in the present study cover an extensive
1.50 9.536  1.635 2.140 0.821 1.412  1.749 0.762
range of possible geometrical variations in steel brackets and they 2.00 6.265  1.076 0.349 0.657 12.628  1.427 0.586
must be a good guidance for steel bracket design, it should be
 S.E. Lee et al. / Ocean Engineering 101 (2015) 182–200 199

cautioned that they may need to be validated by comparison with Conway, H.D., Leissa, A.W., 1960. A method for investigating certain eigenvalue
more specific computations and experiments when one may aim problems of the buckling and vibration of plates. J. Appl. Mech. 27 (3), 557–558.
Cox, H.L., Klein, B., 1955. The buckling of isosceles triangular plates. J. Aeronaut. Sci.
at using them for some special cases of geometric and boundary 22 (5), 321–325.
conditions. DNVGL, 2012. Rules for classification of ships, Part 3.. Det Norske Veritas, Oslo,
Norway (Chapter 1).
Drouin, P., 2006. Brittle fracture in ships – a lingering problem. Ships Offshore
Struct. 1 (3), 229–233.
7. Conclusion Gaspar, B., Garbatov, Y., Guedes Soares, C., 2011. Effect of weld shape imperfections
on the structural hot-spot stress distribution. Ships Offshore Struct. 6 (1–2),
The aims of this study were to numerically examine the effects 145–159.
Gayton, N., Lemaire, M., 2009. Reliability assessment of structures subjected to
of various design variables on the ultimate-strength characteristics
fatigue failure. Ships Offshore Struct. 4 (3), 229–239.
of steel brackets, and to propose a design formula for the ultimate Han, L.S., 1960. The buckling and deflection of isosceles triangular plates. ASME J.
strength of a steel bracket. A series of nonlinear finite element Appl. Mech. 27 (1), 207–208.
computations were undertaken to achieve these objectives. Sev- Haris, S., Amdahl, J., 2012. An analytical model to assess a ship side during a
collision. Ships Offshore Struct. 7 (4), 431–448.
eral conclusions can be drawn from the results, as outlined below. Henriksen, L., Kaminshohara, A., Hosokawa, H., 2008. Structural analysis and
modifications – 2 tankers for offshore FPSO and FSO service. Ships Offshore
(1) First, the steel brackets used in real ship structures were investi- Struct. 3 (2), 127–134.
Hughes, O.F., Paik, J.K., 2013. Ship Structural Analysis and Design. The Society of
gated with reference to six types of vessel: three oil tankers, two
Naval Architects and Marine Engineers, Washington, DC, USA.
bulk carriers and one LNGC. The actual dimensional characteristics Hwang, J.K., Roh, M.I., Lee, K.Y., 2010. Detailed design and construction of the hull of
of these brackets were identified from the data collected and a floating, production, storage and off-loading (FPSO) unit. Ships Offshore
analysed. Struct. 5 (2), 93–104.
IACS, 2006a. Common Structural Rules for Double Hull Oil Tankers. International
(2) As radiused steel brackets were found to have significantly less Association of Classification Societies, London, UK.
ultimate strength than triangular brackets, the latter were recom- IACS, 2006b. Common Structural Rules for Bulk Carriers. International Association
mended for general use. of Classification Societies, London, UK.
Jaunky, N., Knight, N.F., Ambur, D.R., 1995. Buckling analysis of general triangular
(3) To enhance the structural reliability and integrity of ship anisotropic plates using polynomials. AIAA J. 33 (12), 2647–2654.
structures, steel brackets with aspect ratios either equal to or Krishnakumar, S., 1988. Large deflection of simply supported isosceles triangular
greater than 1.0 should be used. plates subjected to edge shear and compressive loading. Int. J. Eng. Sci. 26 (7),
673–685.
(4) The results of the sensitivity analysis indicate that increasing
Li, Y.C., 1959. Buckling problem of right-angled isosceles triangular plates. Publ. de
the thickness of a steel bracket is the most efficient way to l'Inst. Math. 13 (19), 85–94.
increase its ultimate strength. If this is impossible, it is Lotsberg, I., 2007. Recent advances on fatigue limit state design for FPSOs. Ships
advisable to increase the breadth of the bracket (rather than Offshore Struct. 2 (1), 49–68.
Lotsberg, I., 2009. Assessment of design criteria for fatigue cracking from weld toes
its height) to enhance the load-carrying capacity. subjected to proportional loading. Ships Offshore Struct. 4 (2), 175–187.
(5) Based on the results of the parametric analysis, a design Lotsberg, I., Rundhaug, T.A., Thorkildsen, H., Boe, A., Lindemark, T., 2008. A
formula accommodating a whole range of actual dimensional procedure for fatigue design of web-stiffened cruciform connections. Ships
Offshore Struct. 3 (2), 113–126.
characteristics is proposed. This formula has the potential to
LR, 2012. Rules and Regulations for the Classification of Ships, Part 3. Ship
improve the design and safety assessment of steel brackets. Structures. Lloyd's Register, London, UK.
(6) In conservative structural design point of view, the design formula Na, S.S., Karr, D.G., 2013. An efficient stiffness method for the optimum design of
of the simply supported boundary condition is recommended to ship structures based on common structural rules. Ships Offshore Struct. 8 (1),
29–44.
use in design of steel brackets. Paik, J.K., Kim, D.K., Park, D.H., Kim, H.B., Mansour, A.E., Caldwell, J.B., 2012.
(7) Considering that a lot of uncertainties are involved due to Modified Paik–Mansour formula for ultimate strength calculations of ship
geometrical and boundary conditions, among others, further hulls. Ships Offshore Struct. 8 (3–4), 245–260.
Paik, J.K., Seo, J.K., 2009a. Nonlinear finite element method models for ultimate
researches are recommended to conduct experiments on full-
strength analysis of steel stiffened-plate structures under combined biaxial
scale or large-scale bracket models which will be used to compression and lateral pressure actions – Part I: plate elements. Thin-Walled
examine the collapse mechanism more realistically and also to Struct. 47 (8–9), 1008–1017.
validate the computations and design formulae. Paik, J.K., Seo, J.K., 2009b. Nonlinear finite element method models for ultimate
strength analysis of steel stiffened-plate structures under combined biaxial
compression and lateral pressure actions – Part II: stiffened panels. Thin-
Walled Struct. 47 (8–9), 998–1017.
Paik, J.K., Thayamballi, A.K., 2003. Ultimate Limit State Design of Steel-Plated
Structures. Wiley, Chichester, UK.
Acknowledgements
Payer, H.G., Schellin, T.E., 2013. A class society's view on rationally based ship
structural design. Ships Offshore Struct. 8 (3–4), 319–336.
This work was supported by a 2-Year Research Grant of Pusan Petinov, S.V., Kim, W.S., Paik, Y.M., 2006. Assessment of fatigue strength of weld
root in ship structure: an approximate procedure. Ships Offshore Struct. 1 (1),
National University. This study was undertaken at The Lloyd's Register
55–60.
Foundation Research Centre of Excellence at Pusan National University. Richir, T., Caprace, J.D., Losseau, N., Pircalabu, E., Toderan, C., Rigo, P., 2007. Least
Lloyd's Register Foundation (LRF), a UK registered charity and sole cost optimization of large passenger vessels. Ships Offshore Struct. 2 (4),
shareholder of Lloyd's Register Group Ltd., invests in science, engineer- 339–345.
Rizzo, C.M., 2006. The lengthening of a small chemical tanker: a comparison of class
ing and technology for public benefit, worldwide. scantling rules. Ships Offshore Struct. 1 (2), 71–81.
Safar, S.S., Machaly, E.B., 2005. Finite element analysis of unstiffened triangular
bracket plates. In: Proceedings of the 11th International Colloquium on
References
Structural and Geotechnical Engineering. May 17–19, Cairo, Egypt.
Salmon, C.G., 1962. Analysis of triangular bracket-type plates. J. Eng. Mech. Div. 88
Anastasopoulos, A., Kourousis, D., Botten, S., Wang, G., 2009. Acoustic emission (EM6), 41–87.
monitoring for detecting structural defects in vessels and offshore structures. Salmon, C.G., Buettner, D.R., O’Sheridan, T.C., 1964. Laboratory investigation of
Ships Offshore Struct. 4 (4), 363–372. unstiffened triangular bracket plates. J. Struct. Div. 90 (ST2), 257–278.
ANSYS, 2012. User's Manual (Version 14.0). ANSYS Inc., Canonsburg, PA, USA. Shi, G.J., Wang, D.Y., 2012. Ultimate strength model experiment regarding a
Aung, T.M., 2006. Plastic Buckling of Mindlin Plates PhD thesis. National University container ship's hull structures. Ships Offshore Struct. 7 (2), 165–184.
of Singapore, Singapore. Tan, H.K.V., Bettess, P., Bettess, J.A., 1983. Elastic buckling of isotropic triangular flat
Beghin, D., 2013. Technical evolution in the ship structural analysis over the last 50 plates by finite elements. Appl. Math. Model. 7 (5), 311–316.
years. Ships Offshore Struct. 8 (3–4), 337–345. Turan, O., Oicer, A.I., Rigo, P., Caprace, J.D., 2009. Maintenance/repair and
Burchard, W., 1937. Beulspannungen der quadratischen Platte mit Schragsteife production-oriented life cycle cost/earning model for ship structural optimisa-
unter Druck bzw. Schub. Ing. Archiv. 21 (107), 332–348. tion during conceptual design stage. Ships Offshore Struct. 4 (2), 107–125.
200 S.E. Lee et al. / Ocean Engineering 101 (2015) 182–200

Ueda, Y., Kuramoto, Y., Yao, T., 1976. Effects of initial imperfections due to welding Wang, C.M., Xiang, Y., Kitipornchai, S., Liew, K.M., 1994. Buckling solutions for
on strength and rigidity of triangular corner bracket. Trans. JWRI 6 (1), 39–45. Mindlin plates of various shapes. Eng. Struct. 16 (2), 119–127.
Ueda, Y., Kuramoto, Y., Yao, T., 1977. Strength and rigidity of triangular corner Wittrick, W.H., 1954. Symmetrical buckling of right-angled isosceles triangular
bracket. J. Soc. Nav. Archit. Jpn. 139, 122–128. plates. Aero. Quart. 5, 131–143.
Ueda, Y., Yao, T., 1987. Determination of optimum thickness of corner bracket Woinowsky-Krieger, S., 1933. Berechnung der ringsum frei aufliegenden gleichzei-
considering buckling and/or plastic collapse. Trans. JWRI 16 (2), 383–392. tigen dreiecksplatte. Ing. Archiv. 4, 254–262.
Vaaraniemi, K., Pramila, A., Enlund, H., Olknuora, P., 2003. Ultimate load carrying Xiang, Y., 2002. Buckling of triangular plates with elastic edge constraints. Acta
capacity of brackets. Intern. Shipbuild. Prog. 50 (1–2), 89–108. Mech. 156 (1–2), 63–77.
Varela, J.M., Ventura, M., Guedes Soares, C., 2011. Product data model of hull Xiang, Y., Wang, C.M., Kitipornchai, S., Liew, K.M., 1994. Buckling of triangular
structures and digital prototyping system for basic structural design. Ships Mindlin plates under isotropic in-plane compression. Acta Mech. 102 (1–4),
Offshore Struct. 6 (1–2), 3–14. 123–135.
Vhanmane, S., Bhattacharya, B., 2008. Estimation of ultimate hull girder strength Yu, L., Das, P.K., Zheng, Y., 2009. A response surface approach to fatigue reliability of
with initial imperfections. Ships Offshore Struct. 3 (3), 149–158.
ship structures. Ships Offshore Struct. 4 (3), 253–259.
Wakasugi, S., 1960a. Buckling of a right-angled isosceles triangular plate under
Zanic, V., 2013. Methods and concepts for the multi-criteria synthesis of ship
edge thrusts. Trans. Japan Soc. Mech. Eng. 21, 474–480.
structures. Ships Offshore Struct. 8 (3–4), 225–244.
Wakasugi, S., 1960b. Buckling of a simply supported equilateral triangular plate.
Zanic, V., Andric, J., Prebeg, P., 2013. Design synthesis of complex structures. Ships
Trans. Jpn. Soc. Mech. Eng. 26 (164), 538–544.
Offshore Struct. 8 (3–4), 383–403.
Wakasugi, S., 1961. Buckling of clamped equilateral triangular plate under uniform
Zhang, S., Kumar, P., Rutherford, S.E., 2008. Ultimate shear strength of plates and
27 (179), 1010–1017Trans. Jpn. Soc. Mech. Eng. 27 (179), 1010–1017.
Wang, C.M., Liew, K.M., 1994. Buckling of triangular plates under uniform stiffened panels. Ships Offshore Struct. 3 (2), 105–112.
compression. Eng. Struct. 16 (1), 43–50.