You are on page 1of 6

Ocean Engineering 35 (2008) 281286

Methods for ultimate limit state assessment of ships and ship-shaped


offshore structures: Part III hull girders
Jeom Kee Paik

, Bong Ju Kim, Jung Kwan Seo


Department of Naval Architecture and Ocean Engineering, Pusan National University, 30 Jangjeon-Dong, Geumjeong-Gu, Busan 609-735, Republic of Korea
Received 6 December 2006; accepted 6 August 2007
Available online 24 October 2007
Abstract
The present paper is Part III of a series of three papers on the methods useful for ultimate limit state assessment of ships
and ship-shaped offshore structures. It is focused on the methods for the progressive collapse analysis of hull girders under bending
moments, in contrast to the previous two papers (Parts I and II), respectively, dealing with methods for the ultimate limit
state assessment of unstiffened plates and stiffened panels. An AFRAMAX-class hypothetical double hull oil tanker structure designed
by IACS common structural rules (CSR) method is studied as an illustrative example. The ultimate vertical bending moment capacity of
the hull structure is then analyzed by ANSYS FEA, ALPS/HULL, and IACS CSR methods, and their resulting computations are
compared.
r 2007 Elsevier Ltd. All rights reserved.
Keywords: Ultimate limit states; Limit state design; Ships; Ship-shaped offshore structures; IACS common structural rules; ANSYS FEA; DNV PULS;
ALPS/ULSAP; ALPS/HULL
1. Introduction
In maritime industry, the ultimate limit state is now
applied as a basis of structural design and strength assess-
ment (ISO, 2006, 2007; IMO, 2006; IACS, 2006a, b). This
means that the ultimate limit state assessment is now a
mandatory task for structural design of ships and ship-
shaped offshore structures.
For the ultimate limit state assessment of hull girders
that are composed of a number of structural components,
it is desirable to perform the progressive collapse analysis.
This is because the progressive collapse analysis can reect
the local failures of individual structural components and
their interacting effects in the best way.
The aim of the present series of the study has been to
evaluate methods useful for the ultimate limit state
assessment of ships and ship-shaped offshore structures.
Part I (Paik et al., 2007a) and Part II (Paik et al., 2007b)
deal with methods for the ultimate limit state assessment of
plates and stiffened panels, respectively, using ANSYS
nonlinear nite element analysis (FEA) (ANSYS, 2006),
DNV PULS (DNV, 2006), and ALPS/ULSAP (2006). The
present paper (Part III) is focused on methods for the
progressive hull girder collapse analysis.
As an illustrative example, an AFRAMAX-class hypo-
thetical double hull oil tanker structure designed by IACS
CSR method is studied. The ultimate vertical bending
moments of the hull girder in sagging and hogging
condition are analyzed using ANSYS FEA, ALPS/HULL
(2006), and IACS common structural rules (CSR) (IACS,
2006a), and their computational results are compared.
2. Candidate methods
For ultimate vertical bending moment calculations of
hull girders, the following three methods are used, namely
ANSYS nonlinear FEA (ANSYS, 2006);
ALPS/HULL (2006); and
IACS CSR method (IACS, 2006a).
For the ultimate bending moment calculations, IACS
CSR method provides two approaches, namely single-step
ARTICLE IN PRESS
www.elsevier.com/locate/oceaneng
0029-8018/$ - see front matter r 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.oceaneng.2007.08.008

Corresponding author. Tel.: +82 51 510 2429; fax: +82 51 512 8836.
E-mail address: jeompaik@pusan.ac.kr (J.K. Paik).
method and incremental-iterative (or multi-step) method.
The single-step method is applicable only for ultimate
sagging moment capacity prediction, while the multi-step
method can be applied for both sagging and hogging
conditions. It is noted that the IACS CSR method is
more likely to be closed-form expressions that cannot
take into account the progressive failures of individual
components and their interacting effects. In the present
benchmark study, both single- and multi-step methods
are considered, and their resulting computations are
compared.
For nonlinear analysis of a system structure that is
composed of a number of individual structural compo-
nents, the progressive collapse analysis is highly desirable.
This is because the progressive failures of individual
components and their interacting effects cannot be
accounted for unless otherwise. Another common issue
arose in the nonlinear computations of large complex
system structures is computational efforts. It is now well
recognized that idealized structural unit method (ISUM)
can resolve the two issues above (Paik and Thayamballi,
2003, 2007).
ALPS/HULL method is a special program developed for
the progressive hull collapse analysis of ships and ship-
shaped offshore structures (e.g., FPSOs) using ISUM,
while ALPS/GENERAL program can deal with general
types of steel or aluminum plated structures. The latest
version of ALPS/HULL (version 2006.3) uses advanced
ISUM elements in which the number of degree of
freedom at each nodal point of individual ISUM elements
is six, i.e., three translational degrees of freedom and
three rotational degrees of freedom, although the old
version used the elements with only three translational
degrees of freedom at each nodal point. The benet of the
use of the advanced ISUM elements with six degrees of
freedom at each nodal point is that the method can now be
applied for three-dimensional structures under more
general condition of hull girder actions. For ultimate limit
state assessment of individual structural components with-
in ALPS/HULL, ALPS/ULSAP is employed as a major
module.
For the present benchmark study purpose using the
candidate methods, the initial deection of plating
(between support members) and stiffener web are assumed
to be the following:
w
opl

b
200
, (1.a)
w
ow

h
w
200
, (1.b)
where w
opl
is the maximum plate initial deection, w
ow
the
maximum stiffener web initial deection, b the breadth of
plating between longitudinal stiffeners, and h
w
is the
stiffener web height (following the denition of ALPS/
ULSAP, i.e., excluding stiffener ange thickness).
It is basically assumed that the pattern of the plate initial
deection is equivalent to the plate-buckling mode which
may give the lowest resistance against the actions, but the
effect of some different shapes of plate initial deection
is also studied in the present study. No residual
stress is supposed to exist for the present benchmark study
purpose, although residual stress can signicantly affect the
ultimate strength characteristics, and all the candidate
methods can deal with the residual stress as a parameter of
inuence.
For the present study, the fabrication related initial
distortions of stiffeners are assumed as follows
w
oc
w
os

a
1000
, (2)
where w
oc
is the column type initial deection of stiffeners
in the vertical direction, w
os
the sideways initial deection
of stiffeners in the horizontal direction, and a is the
stiffener length.
The pattern of both column type initial deection
and sideways initial deection of stiffeners is supposed
to be buckling mode that results in the minimum buck-
ling strength of stiffeners. The consideration of stiffener
initial distortions in Eq. (2) is important, because the
stiffener column type initial distortion can signi-
cantly affect the beam-column type collapse mode,
while the stiffener sideways initial distortions can signi-
cantly affect the exural-torsional buckling (or tripping)
mode.
3. The object ship and its structural characteristics
The object ship is an AFRAMAX-class hypothetical
double hull oil tanker designed by CSR methods as
indicated in Table 1. While some details of the object ship
are presented in Parts I and II, they are briey described in
the present paper for convenience. Fig. 1 shows the
mid-ship section of the ship. Figs. 2 and 3 show the
stiffened plate structures of the object ship at deck and
outer bottom, respectively. The entire ship structure is
made of 32AH high tensile steel with yield stress s
Y
of
315 MPa, Youngs modulus E of 205.8 GPa, and Poissons
ratio n of 0.3.
ARTICLE IN PRESS
Table 1
Principal particulars of an AFRAMAX-class hypothetical double hull oil
tanker designed by IACS CSR method
Length O.A. (m) 250.000
Length B.P. (m) 239.000
Length scantling (m) 236.292
Breadth (m) 43.800
Depth (m) 21.000
Designed draught (m) 13.600
Scantling draught (m) 14.900
Block coefcient 0.87
J.K. Paik et al. / Ocean Engineering 35 (2008) 281286 282
4. Progressive hull collapse analysis
The ultimate vertical bending moment of the hull
structure is now studied using ANSYS FEA, IACS CSR,
and ALPS/HULL methods. It is noted that the hull
structural dimensions applied for the present analysis were
dened by excluding 50% corrosion margin values of
individual structural components as specied by IACS
(2006a), because the hull structure was designed to meet
IACS CSR requirements.
Figs. 4 and 5 show ANSYS FEA and ALPS/HULL
models employed for the progressive hull collapse analysis
under vertical bending, respectively. A sliced single hull
cross-section model between two adjacent transverse
frames at mid-ship is adopted as the extent of the analysis.
For ALPS/HULL method modeling, plate elements
between support members are idealized by ISUM plate
elements, and support members (excluding attached plat-
ing) are idealized by ISUM beam-column elements,
although very ne mesh modeling with the similar size of
nite elements applied for the deck and bottom stiffened
panels discussed above is adopted for nonlinear FEA.
It is important to realize that the buckling collapse may
take place in vertical members of the hull structure until the
hull girder reaches the ultimate limit state. This implies that
such ne mesh modeling should also be applied for vertical
members as well as deck or bottom plate structures for
nonlinear FEA. Further, the neutral axis of the hull cross-
section can be changed as the collapse of individual
structural components progressively occurs. Therefore,
both ANSYS and ALPS/HULL take into account
the change of the neutral axis position as the vertical
bending moment is increasingly applied. The initial
distortions of plating and stiffeners are supposed as dened
in Eqs. (1.a)(2). No residual stresses exist.
ARTICLE IN PRESS
T
r
a
n
s
. fr
a
m
e
B
a
/2
a
/2
a
X Y
Z
1
5
m
m
3
3
7

9
6

8
/1
3
1
8
m
m
1
9
m
m
96 mm
13 mm
337 mm
8 mm
Fig. 2. Deck stiffened plate structure of an AFRAMAX size double hull
oil tanker designed by CSR method (IACS, 2006a).
172 mm 172 mm
17 mm
19 mm
463 mm
475 mm
8 mm 8 mm
T
r
a
n
s
. flo
o
r
B
a
/2
a
/2
a
X Y
Z
1
9
m
m
1
6
.5
m
m
4
6
3

8
+
1
7
2

1
7
(
T
)
4
7
5

8
+
1
7
2

1
9
(
T
)
1
8
.5
m
m
1
6
.5
m
m
Fig. 3. Outer bottom stiffened plate structure of an AFRAMAX size
double hull oil tanker designed by CSR method (IACS, 2006a).
Fig. 4. ANSYS nonlinear FEA model used for the progressive hull
collapse analysis under vertical bending.
Deck stiffened panel
Bottom stiffened panel
Fig. 1. Midship section of the object ship.
J.K. Paik et al. / Ocean Engineering 35 (2008) 281286 283
The cross-sectional data of the hull structure computed
by ALPS/HULL is as follows:
Cross-sectional area=5.4982 m
2
.
Moment of inertia at vertical direction=378.4787 m
4
.
Section modulus=40.7442 m
3
at bottom, 29.9172 m
3
at
deck.
Full plastic bending moment=12742.76 MNm.
Fig. 6 shows the deformed shape of the hull structure at
the ultimate limit state under sagging bending moment,
obtained by nonlinear FEA. It is evident from Fig. 6 that
deck part and some large areas of vertical members have
failed until the hull structure entirely reached the ultimate
limit state. Fig. 7 represents the von Mises stress distri-
bution over the hull cross-section at the ultimate limit state
under sagging.
Fig. 8 shows the vertical bending moment versus
curvature curves under both sagging and hogging, ob-
tained by ANSYS FEA and ALPS/HULL. The gure also
plots the IACS CSR predictions: ultimate sagging
moments M
usag_CSR_single-step
by single-step method and
M
usag_CSR_multi-step
by multi-step method, and ultimate
hogging moment M
uhog_CSR_multi-step
by multi-step method,
where the single-step method is not applicable for the
ultimate hogging moment prediction. The IACS CSR
ARTICLE IN PRESS
Beam-column
unit
Plate unit
X
X
Y
Z
Z
Y
Fig. 5. ALPS/HULL model used for the progressive hull collapse analysis under vertical bending.
Fig. 6. Deformed shape of the hull at the ultimate limit state under
sagging moment, obtained by ANSYS nonlinear FEA method.
a
b
Fig. 7. von Mises stress distribution over the hull section at the ultimate
limit state under sagging moment obtained by (a) nonlinear FEA method
and (b) ALPS/HULL method.
J.K. Paik et al. / Ocean Engineering 35 (2008) 281286 284
ARTICLE IN PRESS
4.0E004 2.0E004 0.0E+000 2.0E004 4.0E004
Curvature (1/m)
2.0E+010
1.0E+010
0.0E+000
1.0E+010
V
e
r
t
i
c
a
l

b
e
n
d
i
n
g

m
o
m
e
n
t

(
N
m
)
Hog
Sag
M
uhog_req.
=5.768 GNm
w
opl
=b/200
1)
, w
oc
=w
os
=0.001a
M
usag_HULL
/ M
usag_req.
=1.083
M
uhog_HULL
/ M
uhog_req.
=1.640
M
usag_CSR_multi step
/ M
usag_req.
=1.111
Buckling mode shape
w
opl
=b/200
1)
, w
oc
=w
os
=0.001a
M
usag_HULL
/ M
usag_FEA
=1.027
M
uhog_CSR_multi step
/ M
uhog_req.
=1.915
M
usag_CSR_multi step
/ M
usag_FEA
=1.053
M
uhog_CSR_multi step
/ M
uhog_FEA
=1.156
M
uhog_HULL
/ M
uhog_FEA
=0.990
w
opl
=b/200, w
oc
=w
os
=0.001a
M
usag_CSR_single step
/ M
usag_FEA
=1.029
M
usag_CSR_single step
/ M
usag_req.
=1.085
M
uhog_CSR_multi step
=11.049 GNm
M
uhog_FEA
=9.554 GNm
M
uhog_HULL
=9.465 GNm
M
usag_req.
=7.686 GNm
M
usag_FEA
=8.107 GNm
M
usag_CSR_single step
=8.342 GNm
Fig. 8. Vertical bending moment versus curvature curves of the hull structure obtained by the candidate methods, IACS CSR bending capacity
requirements.
Table 2
Ultimate hull girder strengths predicted by ANSYS FEA, ALPS/HULL, and IACS CSR methods, together with IACS CSR ultimate bending capacity
requirements
Required and predicted ultimate hull girder strengths Initial imperfections
w
opl
b/200, w
oc
w
os
0.001a
(buckling mode shape)
IACS CSR requirements M
uhog_req.
Hog 5.768 GNm
M
usag_req.
Sag 7.686 GNm
IACS CSR prediction by multi-step
method
M
uhog_CSR_multi-step
Hog 11.049 GNm
M
usag_CSR_multi-step
Sag 8.540 GNm
M
uhog_CSR_multi-step/
M
uhog_req.
Hog 1.915
M
usag_CSR_multi-step/
M
usag_req.
Sag 1.111
IACS CSR prediction by single-step
method
M
usag_CSR_single-step
Sag 8.342 GNm
M
usag_CSR_single-step/
M
usag_req.
Sag 1.085
ALPS/HULL prediction M
uhog_HULL
Hog 9.465 GNm
M
usag_HULL
Sag 8.329 GNm
M
uhog_HULL
/M
uhog_req.
Hog 1.640
M
usag_HULL
/M
usag_req.
Sag 1.083
ANSYS nonlinear FEA prediction M
uhog_FEA
Hog 9.554 GNm
M
usag_FEA
Sag 8.107 GNm
M
uhog_FEA
/M
uhog_req.
Hog 1.656
M
usag_FEA
/M
usag_req.
Sag 1.054
Comparisons of the ultimate strengths
among FEA, ALPS, and CSR methods
M
uhog_CSR_multi-step
/M
uhog_FEA
Hog 1.156
M
uhog_CSR_multi-step
/M
usag_FEA
Sag 1.053
M
uhog_CSR_single-step
/M
usag_FEA
Sag 1.028
M
uhog_HULL
/M
uhog_FEA
Hog 0.991
M
usag_HULL
/M
usag_FEA
Sag 1.027
J.K. Paik et al. / Ocean Engineering 35 (2008) 281286 285
bending capacity requirements denoted by M
uhog_req.
for
hogging and M
usag_req.
for sagging are also plotted. Table 2
summarizes the resulting computations.
It is observed that IACS CSR predictions of ultimate
sagging moment correspond well with more rened ANSYS
FEA and ALPS/HULL progressive analysis results, but the
IACS CSR method overestimates the ultimate hogging
bending capacity by some 15.6% compared to the nonlinear
FEA, for the present specic ship hull structure. The
difference between ANSYS nonlinear FEA and ALPS/
HULL is 2.7% for sagging and 0.9% for hogging, showing
that ALPS/HULL method solutions are in very good
agreement with more rened nonlinear FEA.
In terms of safety margin against hull girder collapse, it
is found that the double hull oil tanker structure considered
in the present study has a safety margin greater than the
required capacity specied by IACS CSR, at least by some
5.4% for sagging and by 54% for hogging.
5. Concluding remarks
The aim of the present study has been to identify the
accuracy and applicability of some candidate methods,
which are considered to be useful for ultimate limit state
assessment of ships and offshore structures.
For ultimate hull girder strength predictions, it is found
that the IACS CSR method by either single- or multi-step
approach gives accurate results compared with more
rened ANSYS FEA and ALPS/HULL progressive
hull collapse analysis method solutions for sagging.
However, the ultimate hogging moment predicted by the
application of the IACS CSR method by multi-step
approach has a large discrepancy by overestimating the
ultimate strength to some large extent, while the IACS
CSR method by single-step approach is not applicable for
hogging. On the other hand, ALPS/HULL method
solutions are in very good agreement with more rened
nonlinear FEA results in both sagging and hogging
conditions.
Acknowledgments
The present study was undertaken at the Ship and
Offshore Structural Mechanics Laboratory (http://alps.ac),
Pusan National University, Korea, which is a National
Research Laboratory funded by the Korea Science and
Engineering Foundation (Grant no. ROA-2006-000-
10239-0). The authors are pleased to acknowledge the
support of Samsung Heavy Industries. Part of the present
paper was presented at MARSTRUCT 2007 Conference,
Glasgow, UK.
References
ALPS/HULL, 2006. A Computer Program for Progressive
Collapse Analysis of Ship Hulls (Version 2006.3). Ship and Offshore
Structural Mechanics Laboratory, Pusan National University,
Korea.
ALPS/ULSAP, 2006. A Computer Program for Ultimate Limit State
Assessment for Stiffened Panels (Version 2006.3). Ship and Offshore
Structural Mechanics Laboratory, Pusan National University,
Korea.
ANSYS, 2006. Users Manual (Version 10.0). Swanson Analysis Systems
Inc., Houston.
DNV PULS, 2006. Users manual (version 2.05). Technical Report
No. 2004-0406, Det Norske Veritas, Oslo.
IACS, 2006a. Common Structural Rules for Double Hull Oil Tankers.
International Association of Classication Societies, London http://
www.iacs.org.
IACS, 2006b. Common Structural Rules for Bulk Carriers. International
Association of Classication Societies, London.
IMO, 2006. Goal-Based Standards. International Maritime Organization,
London November.
ISO, 2006. ISO/CD 18072-2: Ships and Marine TechnologyShip
StructuresPart 2: Requirements for Their Ultimate Limit State
Assessment. International Organization for Standardization, Geneva
November.
ISO, 2007. ISO 18072-1: Ships and Marine TechnologyShip Struc-
turesPart 1: General Requirements for Their Limit State Assess-
ment. International Organization for Standardization, Geneva
November.
Paik, J.K., Thayamballi, A.K., 2003. Ultimate Limit State Design of Steel-
Plated Structures. Wiley, Chichester, UK.
Paik, J.K., Thayamballi, A.K., 2007. Ship-Shaped Offshore Installations:
Design, Building, and Operation. Cambridge University Press,
Cambridge, UK.
Paik, J.K., Kim, B.J., Seo, J.K., 2007a. Methods for ultimate limit state
assessment of ships and ship-shaped offshore structures: Part I
unstiffened panels. Ocean Engineering, doi:10.1016/j.oceaneng.
2007.08.004.
Paik, J.K., Kim, B.J., Seo, J.K., 2007b. Methods for ultimate limit
state assessment of ships and ship-shaped offshore structures:
Part II stiffened panels. Ocean Engineering, doi:10.1016/j.oceaneng.
2007.08.007.
ARTICLE IN PRESS
J.K. Paik et al. / Ocean Engineering 35 (2008) 281286 286

You might also like