Original Article

Proc IMechE Part M:

J Engineering for the Maritime Environment
2021, Vol. 235(1) 110–126
Efficient fatigue assessment of the Ó IMechE 2020
Article reuse guidelines:
upper and lower hopper knuckle sagepub.com/journals-permissions
DOI: 10.1177/1475090220945460

connections of an oil tanker journals.sagepub.com/home/pim

Ozgur Ozguc

Abstract
The fatigue of structural details of ships is of great importance in the maritime industry as they can lead to cracks which
can jeopardize structural integrity. The hopper knuckle is among the most vulnerable areas in a ship with respect to fati-
gue damage. Analysis of the hopper knuckle has become mandatory for tankers in the Class Ship Rules and in the
International Association of Classification Societies Common Structural Rules. Along with the fatigue damage record of
the vessels, the fatigue analysis of the hopper knuckles is essential to ensure that the vessels have sufficient fatigue
strength. In the current work, fatigue calculations are performed of the upper and lower hopper knuckle connections
within midship of the oil tanker using simplified fatigue calculations based on Det Norske Veritas (DNVGL) Classification
Note No. 30.7. The fatigue analysis is based on 25 years of operation in worldwide wave environment. A cargo hold
model (½ + 1 + ½) amidships and a local finite element model of the hopper knuckle are generated. The local model
provides relevant hotspot stress for fatigue life calculations. The results from the hopper knuckle fatigue analysis show
that the vessels may expect fatigue cracks at lower hopper knuckle before the vessel reaches its design life of 25 years.
Based on the findings, it is recommended to fit soft brackets and close scallops at the lower hopper knuckle on selected
common frames in all cargo holds.

Keywords
Upper hopper knuckle, lower hopper knuckle, finite element analysis, fatigue assessment, structural integrity

Date received: 1 May 2020; accepted: 24 June 2020

Introduction All ships are designed to avoid dangerous fatigue

damage. Some ships that will trade in particular severe
A ship is made up of welded steel plates with numerous wave areas may be designed with higher fatigue stan-
of built-in microcracks and weld defects. These weld dards on request from the ship owner. Typically, these
defects that are present from the first day of the ship’s are ships trading in the North Atlantic, in areas with
life together with the dynamic sea loads that the ship is possible icy conditions or if the owner would like to
subjected to will, at some higher stressed locations in avoid the cost of future repair work and unscheduled
the ship structure, eventually develop into larger fatigue docking. The probability of fatigue damages in ships
cracks. How long it will take depends on the stress level
will be reduced by designing ships with a higher fatigue
in the ship structure. The stress level depends on the
standard.
detailed structural configuration, plate thicknesses and
The complexity of all the uncertainties involved in
the wave environment the ship is trading in. In a ship
fatigue calculations means that even ships designed with
structure, there are thousands of potential fatigue crack
the most advanced fatigue methods may experience
locations (hot spots) and it would be practically impos-
some fatigue cracks during the design life. These cracks
sible to design a ship with a strength level such that dur-
ing the entire design life absolutely no fatigue cracks
will occur. Even a ship that is designed for the highest Department of Naval Architecture and Ocean Engineering, Istanbul
fatigue standard with assumed no uncertainty in wave Technical University, Istanbul, Turkey
load level, stress level and manufacturing (which in real-
Corresponding author:
ity is impossible to obtain) should theoretically experi- Ozgur Ozguc, Department of Naval Architecture and Ocean Engineering,
ence fatigue cracking at 3 out of 100 structural details Istanbul Technical University, Ayazaga Campus, 34469 Istanbul, Turkey.
having a design fatigue damage of 1.0. Email: ozguco@itu.edu.tr
Ozguc 111

are normally detected by visible inspection long before S–N curves and the calculation of hotspot and notch
the cracks becomes dangerous. Repair is easy but may spot stress concentration factors.6
be time consuming and costly for the ship owner. Most The simplified procedures of the Classification
ships will have some fatigue damage repair through Societies are sufficient for fatigue screening of the
their lifetime. details of a ship structure for pointing out the poten-
Fatigue damage assessment is an important issue for tially fatigue critical ones, but they cannot be realisti-
ship structures, and special attention to this failure cally applied to the design of new types of structures
mode should be paid. For structural applications such such as catamarans and fast ships. For practical eva-
as rule-based and direct calculation approaches, two luation of fatigue life, these techniques must be further
basic approaches have been developed. adapted for various ship types, and the ship structural
The rules-based approach is based on previous expe- detail database must also be complemented and incor-
rience and has the advantage of its simplicity, but the porated with the findings of both experiments and the-
specific features of an individual ship are not fully con- oretical investigations.7
sidered. Fricke et al.1 made a comparative assessment Choung and Yoon8 investigated to identify how the
of fatigue damage of the hatch cover bearing pad using items could influence on the stochastic fatigue assess-
different approaches proposed from different classifica- ment (SFA) and the discussion items related to wave
tion societies showing a large scatter, between 2 and conditions, wave spectrum combinations, screening
21 years in predicting fatigue lives. procedures, fabrication tolerances, welding types, cor-
Current practice to fatigue analyses of weld compo- rosion allowance and Morison loads. The theoretical
nents includes different approaches. The main steps in approach was also addressed in the analyses of stochas-
fatigue analysis based on direct calculations involve the tic fatigue damage calculation, in which hotspot stress
description of the wave-induced loading, the stress dis- computation was in detail discussed. Furthermore, a
tribution in the structure, the model of fatigue damage full stochastic spectral fatigue analysis software was
or fatigue crack growth and the probabilistic evaluation developed and outstanding features of the software
of the different steps to arrive at a safety index or time- were provided.
dependent reliability. Ships are susceptible to fatigue damage mainly com-
The analysis of stresses is a complex task due to the ing from waves and changing dynamic loading condi-
complexity of a typical ship structure. Fatigue analysis tions because of high cyclic load. Fatigue is therefore an
must be based on local stress response to the global and important design criterion. Fatigue damage decreases
local loads. The global analysis may be performed using the load-carrying capacity of the structure, which may
a finite element model (FEM) with a relatively coarse lead, in extreme circumstances, to leakage resulting in
mesh. cargo mixing and gas accumulating in confined areas.9
Sophisticated models of finite elements can approxi- Ozguc10 addressed the long-term loads and fatigue
mate local stress responses, but traditional multilevel damage accumulation for different trading routes rela-
hierarchical modeling is used to determine loading and tive to North Atlantic operations based on
boundary conditions of a local high-level model with a International Association of Classification Societies
lower-level global model. Not only is the mesh density (IACS) scatter diagram for a 216,000 m3 LNGC (lique-
of the local FEM but also the accuracy of the load used fied natural gas carrier). The evaluation was based on
for the local model responsible for the precision of cal-
direct wave load analysis using WASIM (time domain
culated local stress.2
analysis of ships with forward speed, with conversion
The target of the local finite element analysis is nor-
to frequency domain) and scatter diagrams created in
mally not to calculate directly the stress at a detail, but
accordance with DNV Code 30.5.11
to compute the stress distribution in the hotspot area.
Kapnopoulou and Caridis12 investigated the fatigue
The second step is to employ these stresses as a basis
life of a vessel detail that was considered as hotspot
for evaluation of the geometric stress concentration fac-
approach, using the finite element method by means of
tors. The literature in this regard has a broad variety of
Abaqus and the Common Structural Rules (CSR). A
techniques for modeling finite elements.3–5
vessel of 181k-ton deadweight (DWT) bulk carrier was
For ship structures, the fatigue analysis is usually
used, and the detail that is sub-modeled was the lower
developed accordingly to procedures with predefined
hopper knuckle connection at the midship location.
long-term stress range distribution and assuming a
The findings of the stress distribution and fatigue life
design life corresponding to the service life typically for
of the hopper knuckle connection were addressed. It
25 years, which means that fatigue cracking during ser-
was discovered that alternate condition drew the most
vice life is accepted and the fail-safe principle is applied.
The approaches for the fatigue strength assessment damage compared with all the other conditions.
utilized the S–N curves for classified joints being a Ozguc13 described fatigue analysis procedures that
common basis, with normal, hotspot and notch stress were supported by a developed tool to be used in the
approach being generally accepted and Palmgren– calculations. Three details of local fine mesh models
Miner linear damage summation being applied. such as deck erection butt weld, longitudinal stiffener
Differences remain on the definition of wave loads, the through web-frame and bottom erection butt weld have
112 Proc IMechE Part M: J Engineering for the Maritime Environment 235(1)

been analyzed. The results have been compared with ship and offshore structures in harsh environments like
the component-based approach. the North Sea during the 1970s. The need for fatigue
In the current work, fatigue calculations are per- design of ship structures became increased as more high
formed of upper and lower hopper knuckle connections strength steel was being used in these structures during
within midship of the oil tanker using simplified fatigue the 1970s.
calculations based on DNVGL Classification Note No. Fischer at al.17 addressed the fatigue assessment of
30.7.14 The fatigue analysis is based on 25 years of oper- three ship structural details. While local approaches,
ation in worldwide wave environment. A cargo hold which were proposed in relevant literature and which
model (½ + 1 + ½) amidships and a local FEM of nowadays were also widely applied in design practice,
the hopper knuckle are generated. The local model pro- can successfully assess the fatigue strength of several
vides relevant hotspot stress for fatigue life calculations. welded joints, web-stiffened complex geometries
The results from the hopper knuckle fatigue analysis showed peculiar features, needing to bias the results of
shows that the vessels may expect fatigue cracks at local approaches. The work identified effects, which
lower hopper knuckle before the vessel reaches its altered particularly the crack propagation behavior.
design life of 25 years. Based on the findings, it is rec- Some test cases were re-evaluated using approaches
ommended to fit soft brackets and close scallops at the referring to the effective notch stress and to the strain
lower hopper knuckle on selected common frames in all energy density. Results were compared with the test
cargo holds. results in order to provide explanations to the above-
mentioned biased assessments.
Gaidai et al.18 contributed on development of novel
Literature review fatigue estimation techniques that utilized available
limited data set more efficiently. The continuous stress
Fatigue damage is the most common damage to ship time series covering nearly 2 years duration was ana-
structures due to cyclic wave loads. Fluctuating stresses lyzed. Rainflow counting method was subsequently
arising from wave loads can initiate fatigue cracks in used for the accumulated fatigue damage assessment.
the vicinity of joints which are inadequately designed, An efficient distribution tail extrapolation technique
constructed and maintained. Although the load is not was proposed to accurately predict fatigue damage.
large enough to cause immediate failure, microscopic The proposed technique utilized available data more
cracks gradually increase in size. Failure occurs after a efficiently, than direct fatigue estimation. Fatigue and
certain number of cycles when the accumulated damage corrosion are two major deterioration mechanisms of
reaches threshold limit. ship structures. These mechanisms can lead to a reduc-
Crack propagation may lead to the failure of pri- tion of structural safety.
mary members. The cost of inspection and repair of Han et al.19 proposed a probabilistic management
joints affected by fatigue cracks are high. Hence, the framework for ship structures under coupled
fatigue design should be addressed properly at the early corrosion–fatigue deterioration processes. Various
design stage. The cumulative damage due to fluctuating uncertainties arising from material properties, coupled
loads leads to fatigue fracture which is the main cause corrosion–fatigue modeling, loading conditions and
of the fracture of offshore vessels and structures. inspection techniques were considered in the proposed
Knuckle joints are the most critical area due to its sus- framework. The detrimental effects of corrosion on
ceptibility to fatigue failure. This is mainly due to a fatigue crack growth were taken into account in life-
high-stress concentration at knuckle joints. cycle optimization of inspection/repair actions. It was
Fatigue damage computations involve design vari- found that the effects of corrosion on fatigue can signif-
ables such as S–N curve data, wave scatter data and icantly affect the results of the optimal life-cycle man-
wave spectrum. Current rules of classification societies agement strategy and, therefore, must be properly
are used to evaluate the fatigue damage of knuckle incorporated in the planning of life-cycle maintenance
joint. In general, the spectral fatigue calculation is cum- actions.
bersome due to time-consuming calculation process. Garbatov and Huang20 addressed the reliability-
Naik15 performed the spectral fatigue analysis for based design and optimization of ship structural com-
148k, Moss type spherical tank LNGC. The study was ponents subjected to stochastic loads and accounting
focused on the fatigue damage evaluation of hopper for the local fatigue damage and buckling and ultimate
knuckle joint details by full spectral fatigue analysis. It global strength of the ship hull. The multi-objective
was also drawn that the weld preparation quality and structural optimization was performed in minimizing
inspection category had a direct consequence on the the component net-section area, lateral deflection and
fatigue life. Higher weld quality and stringent inspec- fatigue damage, avoiding local buckling.
tion type caused increased fatigue life. Gledić et al.21 investigated the possibility of crack
Lotsberg16 introduced an overview of the develop- initiation in damaged ship structure during salvage
ment of fatigue design standards for marine structures period. Hypothesis was that low-cycle wave loading
over the last 40 years. Fatigue design standards for off- could cause high-amplitude fluctuating stresses at the
shore structures became needed with development of edge of damage opening. FEM of oil tanker was created
Ozguc 113

for intact ship and 50 random damage scenarios based leading to fatigue accumulation in ship structures. For
on accident statistics. Low-cycle fatigue damage was the fatigue life prediction of ship structures, it is impor-
estimated by the strain-life method using Monte Carlo tant to obtain both the long-term distribution and the
simulation of wave load amplitudes. Parametric study time history of wave-induced loads. An essential step is
investigated sensitivity of results to material properties, to get reliable wave statistics and accurate description
wave environment and salvage duration. of the stochastic nature of sea state along a ship’s sail-
Garbatov et al.22 addressed crack initiation and ing routes during her service time.27
growth under cyclic loading as well as unstable crack Fischer and Fricke28 simulated the numerical crack
propagation and tearing in the ship and offshore struc- propagation which was performed for geometrically
tures. Due attention shall be paid to the suitability and different variants. The variants differed in global geo-
uncertainty of physical models and testing. metry, boundary conditions and weld shape. The analy-
Consideration was given to practical application, statis- ses investigated how the crack propagation was altered
tical description and fracture control methods in if the structural configuration became more complex. It
design, fabrication and service. was drawn that the load-carrying grade of the weld, the
De Gracia et al.23 presented a comparative study weld flank angle and the geometrical configuration also
based on the S–N fatigue assessments for a welded joint have an impact on both the notch effect and the local
in a container vessel following two different shipping stress concentration.
routes such as North Atlantic and North Pacific Fricke29 reviewed the different approaches, high-
Oceans. Fatigue damages accumulated by the ship fol- lighting their advantages and limitations in fatigue
lowing a ship weather route were calculated. Short-sea strength assessment of welded joints. In this connec-
sequences were generated by a wave load sequence tion, the problematic distinction between crack initia-
model. Based on these results, the effect of different tion and propagation phases was discussed, followed
shipping routes on the cumulative fatigue damage was by considerations about some parameters which had
discussed. large influence on the fatigue behavior of welded joints
Nair et al.24 presented criteria for assessment of but are considered differently in the approaches, such
cracks based on inducement factors which influenced as plate thickness and stress gradient effects, multiaxial
crack initiation in the hull structure even during the stress states, welding-induced distortions and residual
early stages of operation. Detectable cracks identified stresses.
during inspection were analyzed to determine the indu- Park et al.30 performed the spectral fatigue analysis
cement factor and the underlying causes for such crack through the process of hydrodynamic response analysis,
initiation corresponding to the ship operational life. A global structural analysis, local structural analysis and
relationship was established between these inducement calculation of fatigue damage. In these processes, fati-
factors and the ship operational life to simplify the gue damage was affected by many variables. The repre-
inspection and crack assessment process. The circular sentative variables are S–N curve data, wave scatter
data visualization technique was adopted to represent data, wave spectrum and bandwidth effect. The effects
this relationship, and a procedure was developed to of these variables to the fatigue damage were analyzed
demonstrate the use of the proposed crack assessment through the spectral fatigue analysis for 170k LNGC.
criteria. Kim and Kim31 introduced a fatigue assessment pro-
Ozguc25 calculated the fatigue life of longitudinal cedure using mesh-insensitive structural stress method
members of an aging LPG carrier within amidships. based on the CSR for bulk carriers by considering
These fatigue calculations were theoretical calculations important factors, such as mean stress and thickness
that would be used as guidance for close-up inspections effects. The fatigue assessment result of mesh-
when ships are surveyed periodically to verify that they insensitive structural stress method were compared with
are maintained in an acceptable condition in accor- CSR procedure based on equivalent notch stress at
dance with international conventions, the rules of clas- major hotspot points in the area near the ballast hold
sification societies and so on. for a 180k bulk carrier. The possibility of implementing
Fatigue analysis was influenced by the uncertainty mesh-insensitive structural stress method in the fatigue
in the input parameters and modeling procedure. assessment procedure for ship structures was discussed.
Magoga26 presented validation of SFA against trials Garbatov32 performed fatigue strength and reliabil-
data, followed by a study on the sensitivity of fatigue ity assessment of complex double hull oil tanker struc-
damage incurred in a naval high-speed light craft tures, based on different local structural finite element
(HSLC) to various parameters. It was found that the approaches accounting for the uncertainties originating
fatigue damage was most sensitive to the significant from load, nominal stresses, hotspot stress calculations,
wave height, though when operational conditions were weld quality estimations and misalignments and fatigue
taken into account the relative importance of the speed S–N parameters including the correlation between load
and heading increased. cases and the coating life and corrosion degradation.
Ocean crossing ship structures are continuously suf- Bernatowska et al.33 presented a research on com-
fering from wave loads when sailing at sea. The wave parison of fatigue life predictions based on nominal
loads cause large variation of structural stresses, stress method and on structural stress method, made
114 Proc IMechE Part M: J Engineering for the Maritime Environment 235(1)

for a couple of constructional details. The work investi- Table 1. Principle particulars for the vessel.
gated influence of geometrical parameters on stress con-
centration factors and highlights a wider flexibility of Length between perpendiculars, LPP 260.00 m
Breadth molded, B 44.20 m
structural stress method. Depth molded, D 24.20 m
Pedersen34 conducted a review on the fatigue Draught, T (scantling) 17.52 m
strength of thick butt joints in steel. The causes of the Displacement 149,775 tons
thickness effect were discussed with emphasis on the
peculiarities of axially loaded butt joints. The available
experimental investigations in the open literature were
Two loading conditions for the fatigue analysis are
reviewed and discussed. Two specific investigations
defined as follows:
were highlighted, in which testing was conducted under
particularly representative conditions, which both
showed limited/non-existing thickness dependency. It LC1—full-load condition;
LC2—ballast condition.
was concluded that the reduction of fatigue strength in
current codes seem too conservative for the case of axi-
ally loaded butt joints. Main characteristics of these loading conditions are
Cai et al.35 discussed some key issues including the shown in Table 2, and fraction of time at sea is given in
Table 3.
rationalities of stress read out method and stress mag-
Worldwide wave environment is assumed in the fati-
nification factors, of screening fatigue assessment.
gue calculations, that is, in the simplified fatigue
Screening fatigue assessment, a fatigue evaluation
approach is an environmental reduction factor fe = 0.8
approach based on fine mesh FEM with elements size
applied on the stress range. Cargo oil tanks are consid-
of 50 mm 3 50 mm, was introduced into CSR for bulk
ered to be non-corrosive environment and S–N curves
carriers and oil tankers according to the requirement of
in air are employed for the whole specified design life.
Goal-Based Standards (GBS). Calculations were car-
ried out for relevant details of some typical oil tankers
and bulk carriers. Software tools used
Bigot and Derbanne36 presented the validation of
fatigue evaluation methods accounting for the intermit- In order to execute the set of fatigue analysis, SESAM
tent wetting effect. Spectral analysis using a stochastic software package developed by DNVGL has been
linearization was compared with the nonlinear time used. SESAM is a complete strength assessment pro-
domain simulation, considered as the reference model. gram for engineering of ships, offshore structures and
In order to compare the different intermittent wetting risers based on the finite element methodology. This
software package has various modules which can be
models and assess their accuracy versus the reference
used depending on the type numerical simulation that
model, numerical applications were done for 18 trans-
is required to be carried out. For this study of tanker
verse sections of six container carriers, taking into
carrier the following modules used has been selected.
account the actual relative influence of the different
types of loads. The fatigue damage predicted by the
current models was compared with the fatigue damage Sesam Genie
provided by the reference model.
GeniE38 is the design analysis tool in SESAM used for
Wang37 calculated the fatigue life of a ship structural
designing and analyzing the offshore and maritime
detail using a spectral approach. The wave-induced ver-
structures made of beams/shells. GeniE can be used as
tical and horizontal bending moments, two base vessel
a stand-alone tool and user where the user can
loading conditions and the non-operating time have all
been taken into account in the spectral fatigue damage
 Model structure, equipment, environment and
calculation. The predicted fatigue life value using the
other loads;
spectral approach was compared with the one calcu-
 Calculate hydrodynamic loads and run static struc-
lated using the IACS R 56, and various factors indu-
tural analyses including non-linear pile soil analysis;
cing uncertainties in the spectral method are further  Visualize and post-process results;
identified and investigated. Finally, recommendations  Perform code checking based on recognized
on how to choose the specific parameters and how to
standards.
model the random wave environment in the spectral
approach were outlined.
Sestra
Sestra is a program for linear static and dynamic struc-
Basis for the analysis
tural analysis within the SESAM program system. It
Table 1 presents the principle dimensions of the vessel uses a displacement-based finite element method. Sestra
studied. computes the local element matrices and load vectors,
Ozguc 115

Table 2. Loading conditions.

Loading condition Draught AP (m) Draught FP (m) Displacement (tons) Cb (m) GM (m)

LC1 (full) 17.624 17.415 174,245.7 0.865 3.472

LC2 (ballast) 10.043 7.473 81,503.6 0.812 12.356

AP: after perpendicular; FP: forward perpendicular; Cb: block coefficient; GM: the metacentric height.

Table 3. Fraction of time at sea. Structural FEM

Operation Load condition
In the shipbuilding, fatigue life is an analysis that was
often overlooked which caused the unknown of the
Ballast Full load operating life of a ship. Ningrum and Ari42 discussed
Tanker for oil 0.425 0.425 the calculation of the tanker fatigue life. This calcula-
tion was to determine the age of a tanker construction
so that it could be a reference to determine the age of
and assembles them into global matrices and load vec- the ship in the future. The procedure used was by mod-
tors. The global matrices are used by algebraic numeri- eling (½ + 1 + ½) cargo oil tank construction using
cal algorithms to do the requested static, dynamic, or finite element method modeling. The method of fatigue
linearized buckling analysis. It has interface with other life calculation was simplified fatigue assessment
program modules of SESAM. method. The rules used was CSR 2012 standard, and
the ship subject of this research was 17,500 light dead-
weight (LTDW) Crude Oil Tanker Ship. The results
Xtract showed that the greatest stress occurred at the meeting
Xtract software39 is post-processing tool having fea- between hopper plates and double-hull plates.
tures for selecting, further processing, displaying, tabu- Niemi at al.43 introduced several hints regarding
lating and animating results from static and dynamic stress determination and interpretation in FEMs and
structural analysis as well as results from various types discussed the choice of suitable finite elements with
of hydrodynamic analysis. With its high-performance respect to structural hotspot stress analyses, covering
three-dimensional (3D) graphics enables easy and effi- shell as well as solid elements. Three methods of
cient interactive rotation, zooming and panning of the deriving structural hotspot stresses were explained
model for viewing and animation. Based on the finite and discussed in detail, such as the through-thickness
element (FE) analysis results with Xtract, it is possible stress linearization at the hot spot, the surface stress
to present the decomposed stress components into extrapolation to the hot spot and the determination
membrane and bending parts, principal stresses and of the structural hotspot stress at a certain point in
von Mises stresses. Furthermore, with Xtract, it is also front of or below the hot spot. The detailed hints
possible to present deformed model, contour curves of were given regarding the choice of the element type
stresses and displacement-Y graphs and tabulated data. and size as well as stress evaluation in case of coarse
In addition, the motion of vessel can be animated. and fine finite element meshes, supplemented by
some remarks on the weld modeling when using shell
elements.
Submod Linear stress analysis is carried out on a cargo hold
The submod programme40 allows a part of a global model of the mid-ship area and on a local model of the
model to be re-analyzed to produce more accurate hopper knuckle. The models are subjected to external
results. For the local model separated from the global and internal pressure loads on 10–4 probability level
model, the displacements from global model analysis are according to DNVGL Classification Note No. 30.7.
applied as prescribed displacements at the boundary. Pressures corresponding to the load conditions
described are applied.

Stofat
Global FEM
The Stofat41 is software tool for fatigue design. Stofat
performs stochastic fatigue analysis on structures mod- The global FEM which is a cargo hold model
eled by 3D shell and solid elements and assesses (½ + 1 + ½) amidships is shown in Figure 1. The
whether the structure is likely to suffer failure due to middle hold of the model represents Tank no. 4. For
the action of repeated loading. Stresses from the global symmetry purpose with respect to plan (0, x, z), only
or local FE analysis serves as stress transfer function half part of the ship is modeled.44,45
for fatigue calculation. Accumulated damage is calcu- Structures as shell, deck, bulkheads, girders, strin-
lated using stress transfer function, S–N curve data, gers, plating and web frames are represented by three-
wave spectrum, weighted over sea states and wave. and four-node shell elements. Stiffeners are represented
116 Proc IMechE Part M: J Engineering for the Maritime Environment 235(1)

Figure 1. FE model of the cargo hold in midship area.

Figure 3. Location of the local model into the global model.

three-node beam elements. At critical hotspot locations,

the mesh size is equal to the plate thickness. Two hot-
spot locations are considered at midship area such as
the lower and the upper hopper knuckle as seen in
Figure 2. The mesh in these hotspot locations is shown
in Figures 4 and 5, respectively, for the upper and lower
hopper knuckle.
The plate thickness used in both models is in accor-
Figure 2. Local FEM of Hopper knuckle in midship area.
dance with net scantlings, that is, as built scantlings
minus a corrosion reduction according to DNVGL CN
by two-node beam elements. The mesh density of the
30.7. Figure 6 shows the plating thickness used at the
model is typically four elements over the web frame
hotspot points.
spacing in longitudinal direction. Between longitudinal
stiffeners, the mesh density is one element over the stif-
fener spacing in transverse and vertical direction. Loads and boundary conditions
Rules pressure loads based on DNVGL Classification
Note No. 30.7 are applied to both structural models.
Local FEM—hopper knuckle Six load cases are taken into account:
A local stress concentration model of hopper knuckle is
created. A fine mesh FE model with element size in the  External dynamic pressure in full-load condition
order of the plate thickness at hot spots is required to (LC1);
obtain the relevant fatigue stress, that is, hotspot stress  Internal dynamic pressure in full-load condition
range. (LC2);
Deformations from the cargo hold model are trans-  External dynamic pressure in ballast condition
ferred to the local model as prescribed displacement (LC3);
using of sub-modeling technique. In addition, the local  Internal dynamic pressure in ballast condition
model is also subjected to the same pressures as the (LC4);
cargo hold model. The local model is shown in Figure  Static pressure (external and internal) in full-load
2. The location of the local model with respect to the condition (LC5);
global model can be seen in Figure 3. The local model  Static pressure (external and internal) in ballast
is made by use of six- and eight-node shell elements and condition (LC6).
Ozguc 117

Figure 4. Mesh in the upper hopper knuckle.

Figure 7. Dynamic pressure loads (four load cases).

Figure 5. Mesh in the lower hopper knuckle.

Figure 8. Static pressure loads (two load cases).

Boundary conditions for cargo hold model

The boundary conditions applied to the cargo hold model
are according to DNVGL Classification Note No. 31.3.46
The boundary conditions are shown in Figure 9 and
applied to the FE model visualized in Figure 10.

Local model of hopper knuckle

Boundary conditions in the local model are set according
to sub-modeling technique. Subnodes are applied to the
boundary of the local model. These subnodes will obtain
Figure 6. Thickness of the plating in the cargo holds at the deflections according to the results of the cargo hold
hopper knuckle model. model in order to account for the effect of global defec-
tions in the stress result of local model. The sub nodes on
the boundary of the local model are shown in Figure 11.
Dynamic and static pressure loads
The dynamic pressure fields for the first four load cases
Structural analysis results
are shown in Figure 7. The static pressure fields for the The global cargo hold model displacements that can be
two static load cases are shown in Figure 8. seen in Figures 12 and 13 illustrates the sub-modeling
118 Proc IMechE Part M: J Engineering for the Maritime Environment 235(1)

Figure 9. Boundary conditions applied according to DNVGL CN31.3.

Figure 10. Boundary conditions applied on the global model.

technique and shows the continuity of the displacement

magnitude in between the two models.
Figure 14 shows the local model stress results for the
two hot spots in the hopper knuckle such as upper and
lower hopper knuckle. Figure 11. Boundary conditions applied on the local model.

damage is calculated based on extreme stress level at

Fatigue analysis of hopper knuckles 10–4 probability level assuming a long-term Weibull dis-
tribution of stress ranges.47–49
A simplified fatigue analysis is performed based on the
The fatigue design is based on use of S–N curves that
stress results of the local hopper knuckle model. The
are obtained from fatigue tests. The design S–N curves
local model subjected to pressure load on 10–4 prob-
that follow are in accordance with mean-minus-two-
ability level and with the very fine mesh at critical hot
standard-deviation curves for relevant experimental
spots provides relevant hotspot stress for fatigue analy-
data hence associated with a 97.6% probability of sur-
sis. The fatigue calculations are performed according to
vival. The S–N parameters are provided in Table 4.
DNVGL Classification Note 30.7.

Fatigue basis S–N curves—cumulative Fatigue results of hopper knuckles

damage Fatigue results from local FE model stresses
Fatigue damage is computed on basis of the Palmgren– For each hot spot, there are two possible crack loca-
Miner rule, assuming linear cumulative damage. The tions. These possible crack locations can be seen in
Ozguc 119

Table 4. S–N parameters (DNVGL CN 30.7).

S–N curve Material Environment N 4 107 cycles N . 107 cycles

Log (a) m Log (a) m

I Welded joint Air or with cathodic protection 12.65 3 16.42 5

II Welded joint Corrosive 12.38 3 12.38 3
III Base material Air or with cathodic protection 12.89 3 16.81 5
IV Base material Corrosive 12.62 3 12.62 3

CN: Classification Note.

Figure 14. Stress results for hopper knuckle, two hotspot

points (LC1).

Figure 12. Displacements magnitude under external pressure

in full-load condition (LC1).

Figure 13. Displacements magnitude for cargo hold model and

hopper knuckle model (LC1).

Figure 15. For each of them, the fatigue life is com- Figure 15. Possible crack locations.
puted and presented in Table 5.
The results show that the existing design of the upper Corrective actions
hopper knuckle have a fatigue life expectancy well
above the assumed design life of 25 years operation in Based on the results obtained that it is recommended
the world wide, whereas the existing design of the lower that brackets are fitted at the lower hopper knuckle at
hopper knuckle does not comply with a fatigue design first possible occasion. Present work shows that how
life of 25 years in the world wide. and where these brackets shall be fitted, as well as the
120 Proc IMechE Part M: J Engineering for the Maritime Environment 235(1)

Table 5. Fatigue life results in worldwide environment.

Stress read out position Damage Fatigue life (years)

Upper knuckle side plate 0.00157 . 100

Upper knuckle hopper plate 0.00112 . 100
Lower knuckle hopper plate 0.96 26.0
Lower knuckle bottom plate 3.69 7.0

Figure 17. Mesh of the bracket (inner shell).

Figure 16. Reinforcement of the lower hopper knuckle with

bracket.

fatigue life of the lower hopper knuckle with bracket

inserted.
Until brackets are fitted, it is recommended to moni-
tor the situation at lower hopper knuckle carefully by

 Inspection from ballast tank side for leakage during

cargo voyages;
 Inspection from cargo tank side when possible.

Inspection is the best way of controlling this situa-

tion, as a possible crack will be visible long before it
develops to become critical.

Figure 18. Filling up of the scallop.

Reinforced hopper knuckle
Bracket insertion at the lower hopper knuckle
To reinforce the lower hopper knuckle, a bracket is
inserted in the local FEM. Figure 16 shows the bracket
location in the hopper knuckle model. The mesh is
refined with an element size equal to the plate thickness
(18.5 mm) at the expected hotspot locations. The
expected hotspot locations are at the bracket toes and
in the middle of the bracket edge. Figure 17 shows the
mesh of the bracket from inner side. Moreover, the
scallop at midship has been removed as it is shown in
Figure 18. Figure 19 depicts the mesh of bracket at
midship. The thickness of the bracket is 25 mm.

Reinforcement at lower hopper knuckle

The strengthening bracket should have the main dimen-
sions shown in Figure 20 with a thickness of 25 mm.
The bracket should be made of high tensile steel. In Figure 19. Mesh of the bracket.
Ozguc 121

Closing of scallop
The scallops in transverse web plating at lower hopper
knuckle should be closed at the frames and results are
provided in Figure 21.
Fatigue calculations are performed for following
hotspot locations:

 The bracket toe at hopper plate;

 The bracket toe at bottom plate—the middle of the
bracket edge;
 The critical hotspot point of the hopper knuckle as
built;
 The critical hotspot point in inner-shell of the rein-
forced lower hopper knuckle.
Figure 20. Dimensions of bracket in way of lower hopper
knuckle.
The hotspot locations are indicated in Figure 22.
order to obtain the improved fatigue results, the brack-
ets should be fitted at the lower hopper knuckle.
Before brackets are fitted to the frames, it is impor-
Fatigue results
tant to do the following at the knuckles: Stress results from hopper knuckle model (design as
built) are presented in Table 6.
 NDT should be performed to reveal potential Stress results from hopper knuckle model with rein-
cracks. forcement at the lower hopper knuckle are demon-
 If cracks are found the weld should be gauged and strated in Table 7, and Table 8 presents the stress
re-welded. results for middle of the bracket edge.

Figure 21. Stress comparison between the lower hopper knuckle as built and the reinforced lower hopper knuckle.
122 Proc IMechE Part M: J Engineering for the Maritime Environment 235(1)

Table 6. Stress results from hopper knuckle model as built condition.

Stress (MPa) Point 1 (t1/2) 9.25 mm Principal Point 2 (t1) 18.5 mm s (HS)
stress direction
Lower knuckle bottom plate s surface s membrane s surface s membrane

LC1 full-load ext. pressure –253.2 –120.3 P2 –221.3 –119.8 –198.9

LC2 full-load int. pressure 110.6 47.9 P1 95.9 48.1 84.6
LC3 ballast ext. pressure –188.0 –85.7 P2 –163.4 –85.7 –145.7
LC4 ballast int. pressure 11.0 4.5 P1 9.0 4.5 8.0
LC5 full-load static 348.3 159.5 P1 305.4 159.5 271.8
LC6 ballast static –154.8 –57.1 P2 –133.6 –58.5 –114.1

Lower knuckle hopper plate

LC1 full-load ext. pressure –170.2 –78.6 P2 –148.7 –76.9 –132.3
LC2 full-load int. pressure 76.7 38,7 P1 67.9 37.7 61.5
LC3 ballast ext. pressure –128.3 –62.3 P2 –112.7 –60.8 –101,4
LC4 ballast int. pressure 8.1 5.0 P1 7.0 4.8 6.8
LC5 full-load static 235.3 109.5 P1 207.7 107.2 184.5
LC6 ballast static –111.4 –61.0 P2 –101.2 –59.4 –92.8

Upper knuckle side plate

LC1 full-load ext. pressure 0.1 –7.6 P2 0.1 –11.9 –4.7
LC2 full-load int. pressure 34.3 18.8 P1 29.3 19.7 27.1
LC3 ballast ext. pressure –17.8 –12.6 P2 –15.3 –14.0 –15.7
LC4 ballast int. pressure 3.9 2.0 P1 2.9 1.8 2.7
LC5 full-load static 61.9 40.4 P1 55.2 46.2 54.5
LC6 ballast static –134.3 –65.3 P2 –115.4 –64.6 –101.4

Upper knuckle hopper plate

LC1 full-load ext. pressure 28.8 –9.2 P1 15.2 –6.3 9.3
LC2 full-load int. pressure 19.7 16.1 P1 18.7 15.3 18.5
LC3 ballast ext. pressure –3.5 –11.3 P2 –6.4 –10.1 –7.8
LC4 ballast int. pressure 3.5 1.5 P1 2.5 1.5 2.4
LC5 full-load static 10.3 37.1 P1 22.2 32.9 25.8
LC6 ballast static –104.2 –53.6 P2 –88.1 –53.4 –80.9

LC: loaded condition.

The results show that the reinforced lower hopper

knuckle complies with fatigue requirement of 25 years
of fatigue life in the worldwide wave environment.

Concluding remarks
The fatigue damage is a potential for the structural
safety of ship structures and accurate prediction of fati-
gue crack propagation, and its impact on the ultimate
strength of the ships is an issue to be solved. Present
work performs this duty and achieves ensuring the
safety of ship structures with fatigue life extension of
the upper and lower hopper knuckle connection at an
oil tanker by modifying properly the structural arrange-
ment of the configuration.
It is known across literature that low fatigue life is
Figure 22. Calculated hotspot locations on reinforced design. mainly developed at this location. First, the fatigue life
of the original configuration is identified by a detailed
fatigue analysis. To extend the fatigue life of the detail,
The fatigue calculations are performed for the possi- minimization of the stress values that induce fatigue is
ble crack locations. The results in term of fatigue life attained by the process of structural optimization. To
are computed and are presented in Table 9. this end, the geometric parameters of the detail that
Ozguc 123

Table 7. Stress results from hopper knuckle model with reinforcement at the lower hopper knuckle.

Stress (MPa) Point 1 (t1/2) 9.25 mm Principal stress Point 2 (t1) 18.5 mm s (HS)
direction
Lower knuckle bottom plate frame #156 s surface s membrane s surface s membrane

LC1 full-load ext. pressure –57.4 –44.3 P2 –40.7 –45.0 –66.2

LC2 full-load int. pressure 23.2 15.9 P1 14.8 16.4 26.8
LC3 ballast ext. pressure –41.6 –30.5 P2 –28.1 –31.2 –48.0
LC4 ballast int. pressure 2.6 1.9 P1 1.4 2.0 3.2
LC5 full-load static 75.2 55.1 P1 52.4 56.2 86.0
LC6 ballast static –26.6 –13.6 P2 –13.9 –14.4 –30.4

Lower knuckle hopper plate frame #156

LC1 full-load ext. pressure 35.8 –24.3 P2 –27.0 –25.3 49.7
LC2 full-load int. pressure 19.0 14.8 P1 15.4 14.8 20.9
LC3 ballast ext. pressure –30.1 –22.3 P2 –23.7 –22.7 –32.9
LC4 ballast int. pressure 2.7 2.6 P1 2.2 2.6 3.2
LC5 full-load static 49.4 33.1 P1 38.4 34.0 52.6
LC6 ballast static –30.9 –26.0 P2 –27.0 –25.0 –34.1

Lower knuckle bottom plate critical hot spot

LC1 full-load ext. pressure –79.0 –51.9 P2 –67.2 –51.4 –81.2
LC2 full-load int. pressure 32.5 18.3 P1 27.1 18.2 32.4
LC3 ballast ext. pressure –56.8 –35.3 P2 –47.7 –35.0 –57.9
LC4 ballast int. pressure 3.0 2.3 P1 2.2 2.2 3.4
LC5 full-load static 107.4 63.8 P1 92.4 63.3 106.8
LC6 ballast static –40.9 –15.1 P2 –33.8 –15.1 –37.5

Bracket toe bottom plate

LC1 full-load ext. pressure –92.2 P2 –103.3
LC2 full-load int. pressure 34.0 P1 38.1
LC3 ballast ext. pressure –62.3 P2 –69.8
LC4 ballast int. pressure 2.6 P1 2.9
LC5 full-load static 122.2 P1 136.9
LC6 ballast static –34.6 P1 –38.8

Bracket toe hopper plate

LC1 full-load ext. pressure –68.5 P2 –76.7
LC2 full-load int. pressure 31.3 P1 35.1
LC3 ballast ext. pressure –52.4 P2 –58.7
LC4 ballast int. pressure 6.2 P1 7.0
LC5 full-load static 81.8 P1 91.7
LC6 ballast static –35.7 P2 –40.0

LC: loaded condition.

Table 8. Stress results for middle of the bracket edge. It is not unusual that some fatigue cracks occurs in
ships after about 10 years. The main reason to this is
Stress (MPa) Maximum stress along the edge normally weld defects or unfavorable design at difficult
Middle of the bracket edge s surface Principal s (HS) connections in the structure. The fatigue crack that is
stress often found in way of upper hopper knuckle on ships
direction initiated due to unfavorable design at a difficult transi-
LC1 full-load ext. pressure –150.6 P2 –150.6 tion in hopper knuckle due to change in hull shape.
LC2 full-load int. pressure 66.9 P1 66.9 The crack in the way of bottom bilge plating was initi-
LC3 ballast ext. pressure 109.2 P2 109.2 ated most likely due to weld defects and/or large plat-
LC4 ballast int. pressure 3.2 P1 3.2 ing span adjacent the longitudinal fillet weld.
LC5 full-load static 225.2 P1 225.2 The following conclusions are drawn based on an
LC6 ballast static –103.2 P2 –103.2
assumed vessel design life of 25 years of operation in
LC: loaded condition. world wave environment:

 Upper hopper knuckle: The design as built proves

affect these stress values are identified and introduced to satisfy a design fatigue life of 25 years of opera-
as the optimization problem design variables. tion in worldwide wave environment.
124 Proc IMechE Part M: J Engineering for the Maritime Environment 235(1)

Table 9. Fatigue results of reinforced lower hopper knuckle in The following points are recommended for the future
worldwide. work:
Stress read out position Damage Fatigue life (years)
of reinforced lower 1. The fatigue damage assessment by stochastic
hopper knuckle spectral fatigue yields reliable results, but involves
time-consuming computational procedures. The
Lower hopper knuckle 0.066 377 equivalent design wave method may be used to
bottom plate
Lower hopper knuckle 0.084 299 assess the fatigue damage at the preliminary design
hopper plate stages. The comparison of damages against the
Critical hotspot location 0.154 163 spectral fatigue gives the factor of safety embedded
in the reinforced lower in the equivalent design approach. However, the
hopper knuckle DNV GL classification society requires fatigue
Bracket toe at bottom 0.389 64
plate limit state on a critical area such as hopper knuckle
Bracket toe at hopper 0.116 215 joints that should be assessed by directly calculated
plate loads. The equivalent design wave approach at the
Middle of the bracket 0.169 148 preliminary design stage of the study can be used
edge to identify the critical fatigue locations in the
vessel.
2. Welds are ignored in the present fatigue analysis
 Lower hopper knuckle: The excepted fatigue life of and the effect of the weld can be considered to
the as built design is below 25 years in worldwide account for tensile residual stresses from welding
wave environment, and it is recommended to add which will further decrease the fatigue life.
soft brackets at the lower hopper knuckle at 3. The variables involved in the fatigue design are sub-
selected frame positions to increase fatigue strength. jected to significant uncertainty, and hence the result-
ing fatigue damage assessment is also associated with
In order to document fatigue strength of the rein- large uncertainties. The calculated loads on the vessel
forced hopper knuckle, the bracket included in the are uncertain due to the stochastic nature of the
FEM and a reanalysis is performed. The following con- wave loads and uncertainties in the sailing route.
clusion is drawn for the reinforced design: Furthermore, uncertainties associated in FE model-
ing, the evaluation of stress concentration, local weld
 Reinforced lower hopper knuckle: The design of geometry and initial imperfections will attribute to
reinforced lower hopper knuckle proves to satisfy a the uncertainties in the calculated fatigue life of s
design fatigue life of 25 years of operation in world- structural detail. A reliability-based fatigue assess-
wide wave environment. ment should be performed directly through a sto-
chastic modeling of these uncertain variables.
The fatigue crack that was found in way of upper
hopper knuckle on oil tanker was initiated due to unfa- Declaration of conflicting interests
vorable design at a difficult transition in hopper
The author(s) declared no potential conflicts of interest
knuckle due to change in hull shape. The crack in the
with respect to the research, authorship, and/or publi-
way of bottom bilge plating on oil tankers was initiated
cation of this article.
most likely due to weld defects and large plating span
adjacent to the longitudinal fillet weld. Such cracks are
not unusual on ships with an age of about 10 years. Funding
The modifications and repairs are recommended to
avoid future fatigue problems. The results from the The author(s) received no financial support for the
hopper knuckle fatigue analysis show that the vessels research, authorship, and/or publication of this article.
may expect fatigue cracks at lower hopper knuckle
before the vessel reaches its design life of 25 years.
ORCID iD
It is recommended that the brackets should be fitted
at first possible occasion. In the meantime, it is recom- Ozgur Ozguc https://orcid.org/0000-0002-7238-0642
mended to perform corrective actions that could relieve
the stress level at lower hopper knuckle. Inspection for References
leakage during cargo voyages and stress relieve by hav-
1. Fricke W, Cui W, Kierkegaard H, et al. Comparative
ing partly filled ballast tanks during cargo voyage fatigue strength assessment of a structural detail in a con-
should be considered. When soft brackets have been tainership using various approaches of classification soci-
fitted and scallops are closed at lower hopper knuckle eties. Mar Struct 2002; 15: 1–13.
at all recommended positions, the vessels are consid- 2. Garbatov Y, Rudan S and Guedes Soares C. Fatigue
ered to have generally good fatigue strength. assessment of welded trapezoidal joints of a very fast
Ozguc 125

ferry subjected to combined load. Eng Struct 2010; 32: 21. Gledić I, Parunov J, Prebeg P, et al. Low-cycle fatigue of
800–807. ship hull damaged in collision. Eng Failure Anal 2019; 96:
3. Niemi E, Fricke W and Maddox S. Structural stress 436–454.
approach to fatigue analysis of welded components— 22. Garbatov Y, Ås SK, Branner K, et al. Committee III. 2:
designer’s guide. IIW Doc. XIII-181900/XV-1090-01, 2004. fatigue and fracture. In: 20th international ship and off-
4. Fricke W and Kahl A. Comparison of different struc- shore structures congress (ISSC 2018), Liege - Belgium &
tural stress approaches for fatigue assessment of welded Amsterdam - The Netherlands, 9–13 September 2018,
ship structures. Mar Struct 2005; 18: 473–488. pp.441–547. IOS Press.
5. Chakarov K, Garbatov Y and Guedes Soares C. Fatigue 23. De Gracia LC, Tamaru H, Osawa N, et al. A study on
analysis of ship deck structure accounting for imperfec- the effect on fatigue damage due to different shipping
tions. Int J Fatig 2008; 30: 1881–1897. routes. In: Proceeding of the VI international ship design &
6. Fricke W, Bollero A, Chirica I, et al. Round Robin study Naval Engineering Congress (CIDIN) and XXVI pan-
on structural hot-spot and effective notch stress analysis. American Congress of Naval Engineering, Maritime Trans-
Ship Offshore Struct 2008; 30: 335–345. portation and Port Engineering (COPINAVAL), Carta-
7. Guedes Soares C, Garbatov Y and von Selle H. Fatigue gena, Colombia, 13–15 March 2019, pp.175–184. Cham:
damage assessment of ship structural components based Springer.
on the long-term distribution of local stresses. Int Ship- 24. Nair A, Sivaprasad K and Nandakumar CG. Crack
build Prog 2003; 50: 35–56. assessment criteria for ship hull structure based on ship
8. Choung J and Yoon KY. Fully stochastic fatigue analy- operational life. Cogent Eng 2017; 4(1): 1345044.
sis for FPSO based on shipyard practices. Int J Offshore 25. Ozguc O. Simplified fatigue analysis of structural details
Polar Eng 2008; 18(2): 142–148. of an ageing LPG carrier. J Marine Eng Technol 2018;
9. Ozguc O. Fatigue assessment of longitudinal stiffener 17(1): 33–42.
end connections for ageing bulk carriers. J Marine Sci 26. Magoga T. Fatigue damage sensitivity analysis of a naval
Technol 2017; 25(5): 543–551. high-speed light craft via spectral fatigue analysis. Ships
10. Ozguc O. Evaluation of different trading routes on fati- Offshore Struct 2020; 15(3): 236–248.
gue damage for a 216K m3 LNG carrier. J Marine Sci 27. De Gracia L, Wang H, Mao W, et al. Comparison of two
Technol 2017; 25(4): 458–463. statistical wave models for fatigue and fracture analysis
11. DNVGL classification notes no. 30.5: environmental of ship structures. Ocean Eng 2019; 187: 106161.
conditions and environmental loads, 2012, https://infost 28. Fischer C and Fricke W. Effect of the stress distribution
ore.saiglobal.com/en-us/Standards/DNV-CLASSIFICAT in simple welded specimens and complex components on
ION-NOTE-30-5-LATEST-458283_SAIG_DNV_DNV_ the crack propagation life. Int J Fatig 2016; 92: 488–498.
1033408/ 29. Fricke W. Recent developments and future challenges in
12. Kapnopoulou VK and Caridis PA. Fatigue life extension fatigue strength assessment of welded joints. Proc
of the lower hopper knuckle connection of a carrier by IMechE, Part C: J Mechanical Engineering Science 2015;
optimization of the structural arrangement. Mar Syst 229(7): 1224–1239.
Ocean Technol 2019; 14(4): 192–209. 30. Park TY, Jang CD, Suh YS, et al. A parametric study
13. Ozguc O. Procedures of fatigue analysis by supporting based on spectral fatigue analysis for 170k LNGC. Int J
direct load application on midship sections. Trans Mari- Naval Arch Ocean Eng 2011; 3(2): 116–121.
time Sci 2020; 9(1): 6–22. 31. Kim SM and Kim MH. Incorporating mesh-insensitive
14. DNVGL. Fatigue assessment of ship structures, DNV structural stress into the fatigue assessment procedure of
classification note 30.7, 2015. common structural rules for bulk carriers. Int J Naval
15. Naik S. Fatigue strength comparative study of knuckle Arch Ocean Eng 2015; 7(1): 10–24.
joints in LNG carrier by different approaches of classifica- 32. Garbatov Y. Fatigue strength assessment of ship struc-
tion society’s rules. MSc Thesis, Université de Liège, tures accounting for a coating life and corrosion degrada-
Liège, 2017. tion. Int J Struct Integr 2016; 7(2). DOI: 10.1108/IJSI-04-
16. Lotsberg I. Development of fatigue design standards for 2014-0017.
marine structures. J Offshore Mech Arctic Eng 2019; 33. Bernatowska E, Leń D and Śle˛czka L. Comparative study
141(3): 031301. of fatigue life assessment made by different approaches.
17. Fischer C, Fricke W and Rizzo CM. Fatigue assessment In: Blikharskyy Z, Koszelnik P and Mesaros P (eds) Pro-
of web-stiffened corners in plated structures by local ceedings of CEE 2019. CEE 2019. Lecture notes in civil
approaches. Ship Tech Res 2018; 65(2): 69–78. engineering, vol. 47, pp.428–435. Cham: Springer, 2020.
18. Gaidai O, Storhaug G, Naess A, et al. Efficient fatigue 34. Pedersen MM. Thickness effect in fatigue of welded butt
assessment of ship structural details. Ship Offshore Struct. joints: a review of experimental works. Int J Steel Struct
Epub ahead of print 5 September 2019. DOI: 10.1080/ 2019; 19(6): 1930–1938.
17445302.2019.1661623. 35. Cai S, Wu J, Wang Y, et al. Research on screening fati-
19. Han X, Yang DY and Frangopol DM. Probabilistic life- gue assessment method based on new common structural
cycle management framework for ship structures sub- rules. In The 27th international ocean and polar engineer-
jected to coupled corrosion–fatigue deterioration pro- ing conference, San Francisco, CA, 25–30 June 2017.
cesses. J Struct Eng 2019; 145(10): 04019116. Mountain View, CA: International Society of Offshore
20. Garbatov Y and Huang YC. Multiobjective reliability- and Polar Engineers.
based design of ship structures subjected to fatigue dam- 36. Bigot F, Mahérault-Mougin S and Derbanne Q. Compar-
age and compressive collapse. J Offshore Mech Arctic ison of different models for the fatigue analysis of details
Eng 2020; 142(5): OMAE2019-96666. subject to side shell intermittent wetting effect. In:
126 Proc IMechE Part M: J Engineering for the Maritime Environment 235(1)

Proceedings of PRADS2016, Copenhagen, 4–8 September (iCAST), Manado, Indonesia, 26–27 October 2018,
2016. Technical University of Denmark. pp.205–209. New York: IEEE.
37. Wang Y. Spectral fatigue analysis of a ship structural 43. Niemi E, Fricke W and Maddox SJ. Structural hot-spot
detail—a practical case study. Int J Fatigue 2010; 32(2): stress determination using finite element analysis. In: Niemi
310–317. E, Fricke W and Maddox SJ (eds) Structural hot-spot stress
38. DNV. SESAM user manual genie: concept design and approach to fatigue analysis of welded components. IIW col-
analysis of offshore structures, version 6.9, 2014. lection. Singapore: Springer, 2018, pp.17–31.
39. DNV. SESAM Xtract user manual: postprocessor for 44. DNVGL-CG-0127. Class guideline—finite element analy-
presentation, animation and reporting of results, version sis, 2016.
3.0, 2011. 45. DNVGL-CG-0129. Class guideline—fatigue assessment
40. DNV GL. SESAM Submod user manual: transfer displa- of ship structures, 2015.
cements from global model to sub-model, version 3.0, 46. DNVGL: strength analysis of hull structures in tankers,
2016. DNV Classification Note 31.3, 2009.
41. DNV GL. SESAM Stofat user manual: fatigue damage 47. DNVGL rules for classification of ships, part 3, chapter
calculation of welded plates and shells, version 3.7-05, 1, 2019.
2016. 48. DNV classification notes—no. 34.1. CSA—Direct Analy-
42. Ningrum M and Ari M. Fatigue life ship construction sis of Ship Structures, 2013.
analysis of 17500 LTDW crude oil tanker. In: 2018 inter- 49. DNVGL-RP-0005. Recommended practice, fatigue
national conference on applied science and technology design of offshore steel structures, 2014.