International Journal of Fatigue 59 (2014) 137144

Contents lists available at ScienceDirect

International Journal of Fatigue

journal homepage: www.elsevier.com/locate/ijfatigue

Fatigue strength experiments of corroded small scale steel specimens

Y. Garbatov a, C. Guedes Soares a,, J. Parunov b
a
Centre for Marine Technology and Engineering (CENTEC), Instituto Superior Tcnico, Technical University of Lisbon, Av. Rovisco Pais, 1049-001 Lisboa, Portugal
b
University of Zagreb, Faculty of Mechanical Engineering and Naval Architecture, Croatia

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

Article history: The objective of this work is to analyze the fatigue strength of small scale corroded steel specimens. The
Received 10 June 2013 specimens were cut from a box girder, which was initially corroded in real sea water conditions. The sur-
Received in revised form 8 September 2013 face of 11 corroded specimens was analyzed applying photogrammetry techniques and a description of
Accepted 10 September 2013
an idealized corroded surface was established. The non-corroded specimens are identied as FAT 86 cat-
Available online 27 September 2013
egory but the fatigue test demonstrated that due to the severe corrosion degradation the experimental
fatigue results of the corroded specimens are located above the fatigue design category FAT 50 and below
Keywords:
FAT 100 referring to the nominal stress approach. The regression analysis of fatigue test results leads to
Fatigue life
Fatigue tests
m = 3.094 and Dr97:7%;m3 64:95 MPa. Fatigue assessment of crack propagation on a pit like crack ow
Corrosion based on a failure assessment diagram was performed. The admissible initial idealized ow defect has
been dened, which matches the fatigue life achieved by the fatigue test for different load categories
and corrosion degradation level.
2013 Elsevier Ltd. All rights reserved.

1. Introduction Since the phenomena of corrosion deterioration of structural

Corrosion and fatigue cracking may be the two most important
types of damage in aging ship structures, which lead to surface
roughness, reduction of the plate thickness and strength and even-
tually to leakage.
Corrosion in ship structures has an important role in the long-
term structural integrity. Under conditions of high temperature,
inappropriate ventilation, high stress concentration, high stress
cycling, very high rates of corrosion can be achieved in spaces such
as ballast tanks and at specic structural details such as horizontal
stringers or longitudinals and web frames. This has been a source
of concern by ship operators and Classication Societies that have
collected much service data [14]. Different models have been pro-
posed to explain the growth of corrosion wastage in ship structures
combining in general some considerations about the physics of
corrosion growth with the tting of service data [37]. Depending
on the location of the ship structural elements the corrosion rate
characteristics are different and models have also been proposed
that account for the effect of environmental factors on the growth
of corrosion [810] .
Corrosion has clear consequences in degrading the ultimate
strength of ship structures [1113] and also affecting the fatigue
strength by the resulting increased level of stresses [1418] and
also by the direct degradation of fatigue strength as studied in this
paper.
Corresponding author.
E-mail address: guedess@mar.ist.utl.pt (C. Guedes Soares).
members are the consequences of extremely complex phenomena
governed by many factors, it is necessary to establish corrosion
margins and permissible corrosion levels by taking into account
past records. An average annual corrosion rate obtained by divid-
ing the thickness reduction of an aged member by a ships age at
a given time has conventionally been used as the basic criteria
[1,2], due to easy of assessing and handling, but more rational cri-
teria, assessing it with a probabilistic model is needed.
Fatigue is an important design criterion for welded components
and global structures. The fatigue damage may even further reduce
the structural strength due to the presence of different kind of
imperfections, which can lead to local increase of the stresses and
to an acceleration of fatigue crack initiation and propagation. For fa-
tigue life assessments different procedures have been developed
based on databases of the fatigue behavior of welded structural com-
ponents as a result of both tests and theoretical investigations [19].
The main steps in fatigue analysis are based on direct calcula-
tions that involve the description of the wave induced loading
[20] the stress distribution in the structure [21], the model of fatigue
damage (SN approach) or fracture mechanics approach [22] and
the probabilistic evaluation of the different steps to arrive at a safety
index or time dependent reliability as has been developed in [23].
The analysis of stresses is a complex task due to the complexity
of a ship structure. Nowadays the method that is mostly acceptable
and spread for analysing a complex welded structure is the hot
spot stress approach [24] based on the effective notch stress
approach [25,26]. Recommendations for fatigue stress assessment
can be found in guidelines [2729].

0142-1123/$ - see front matter 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.ijfatigue.2013.09.005
138 Y. Garbatov et al. / International Journal of Fatigue 59 (2014) 137144
However, ship welded structures are not perfect and their
behavior depends on a variety of inuential factors, namely geo-
metric [30] and material properties [31], loadings, initial or post
built imperfections, deterioration, crack propagation denting etc.
The imperfections change permanently the structural capacity of
welded structures that initially have been designed to resist load-
ings, keeping a certain level of safety.
In a recent work box girders have been corroded in sea environ-
ment [32] and then have been tested up to ultimate strength,
showing important reduction of strength [33]. The analysis of the
results suggested that this might have been done by changes in
mechanical properties of the corroded steel [34], which has sug-
gested that this may also occur with the fatigue strength. Thus,
the objective of this work is to analyze the fatigue strength of cor-
roded small scale steel specimens that have been cut from those
tested box girder specimens.
Corrosion environment drastically reduces fatigue strength
with respect to dry air conditions. There are important differences
in the results of separate and simultaneous action of corrosion
environment and fatigue load. Fatigue strength can be drastically
reduced if the steel detail is rst corroded and later tested in dry
air condition and in this case the fatigue SN curve may have a hor-
izontal sector in the high frequency regime. In the simultaneous
application of corrosion environment and fatigue cyclic load, the
fatigue strength is drastically reduced and the fatigue SN curve
is lower and parallel to the non-corroded one.
Fatigue damage in corrosion environment is accompanied by
the appearance of many cracks, but only one of them will reach
the size of failure. In aged structures, the corrosion degradation ini-
tially is smoothing the existing sharp corners on hot spots and lat-
ter due the penetration effect of corrosion, new hot spots are
created leading to faster fatigue damage.
The work presented here is analysing a small scale specimen of
a transversely stiffened welded plate, which was rstly corroded
and then fatigue tested. The surface of 11 corroded specimens
was analyzed applying photogrammetry techniques to obtain the
description of the corroded surfaces. The non-corroded specimens
are identied as FAT 86 fatigue design category. The admissible ini-
tial ow defect has been dened, which matches the fatigue life
achieved by the fatigue test for different load categories and corro-
sion degradation.
2. Corroded test specimen
As indicated, the fatigue specimens have been cut from box
girders that have been subjected to corrosion, in a real corrosive
environment in direct contact with sea water.
Fig. 1. Box girder after the test in hot sea water with anodic polarization [32].
Table 1
Chemical composition of steel.
Element Concentration (wt.%) Element Concentration (wt.%)
C 0.079 P <0.001
Mn 0.612 S 0.00133
Si 0.017 Cr 0.0115
Cu 0.0474 Ni <0.001
Fe Remainder
The dimensions of the box girder specimen were 1400  800 
600 mm. The deck panel was stiffened with ve longitudinal at
bars with a spacing of 150 mm. The side panel was stiffened with
two stiffeners on a distance of 300 and 500 mm respectively and
the bottom panel was stiffened with one stiffener in the middle
as may be seen in Fig. 1. The box girder was made of normal ship-
building steel with yield stress of 235 MPa. The chemical compo-
nents of the steel used for constructing the box girders as weight
percentage is presented in Table 1.
The box-girder specimen was exposed to the Baltic seawater
and tested in hot water. The box girder was placed in a large tank
and seawater was pumped into the tank continuously. The temper-
ature of sea water was increased and additionally oxygen depolar-
ization sub process rate was increased by the agitation of seawater,
which resulted in a corrosion rate increase.
To model corrosion degradation, acceleration anodic polariza-
tion of the metal surface was used. Anodic electric current was
supplied by an external source. The test duration was 90 days
without polarization. More detailed information about the corro-
sion set up may be seen in [32]. The total weight loss observed
was 56 kg (23% of initial weight). The average value of the electro-
lyte ow rate was 308 dm3/h, water temperature is 48 C and
pH = 7.93.
After the test was completed the box girder was covered with
iron corrosion products. This box girder has been subjected to ulti-
mate strength tests as described in [3335]. Eleven fatigue test
specimens were cut from the box girder from the side shell panel
around the neutral axis of the box girder (see the locations marked
in Fig. 1). The shape of the specimens can be seen in Fig. 2.
3. Photogrammetry analysis of corroded surfaces
To properly characterize the test specimens an analysis of their
thickness has been made employing photogrammetry techniques.
The photogrammetry techniques have already been used in various
cases for close-range measurements, including for structural test
specimen. The principles of photogrammetry are known since the
 Y. Garbatov et al. / International Journal of Fatigue 59 (2014) 137144
Fig. 2. Fatigue test specimen conguration.
end of 19th century, but their real application in the industry only
became possible with the development of digital cameras and
powerful computers. They provide an automatic image analysis
and fast and complex calculations. The position of a point in the
3D space can be determined by triangulating multiple bundles of
observation rays. If the spatial orientation of each bundle is known
in the object coordinate system, the intersection of the rays
delivers the desired 3D object coordinates. Such sophisticated
technique of estimating the three-dimensional coordinates of
points on an object by measurements made in two or more
photographic images taken from different positions , is known as
stereophotogrammetry.
Examples are available of the use of photogrammetry in ships.
Koelman [36] presented an industrial application of CAD that con-
cerned the measurements and re-engineering of the shape of a
complete ship hull and ship parts by using photogrammetry. In
his study, he considered separately the 3D model measurements
and the topological proprieties. Furthermore, he went through a
comparison between laser and photogrammetric techniques con-
cluding that the latter is more appropriate for ship hull inverse
engineering measurements. Ljubenkov et al. [37] used photogram-
metry methods in measurements of the structural deection in the
machinery space and displacements of the main engine were taken
before and after the launching. After the launching, they succeed in
obtaining structural deections up to 3 mm at the narrowest parts
of the hull with an accuracy of 0.2 mm. The results of a digital pho-
togrammetric survey, performed on the 81-foot Italian Navy motor
yacht Argo, were reported by Menna et al. [38].
A photogrammetric approach has also been used for laboratory
specimens, in this case for measuring weld-induced initial distor-
tions in plated structures [3941].
Using photogrammetry, a fast and detailed acquisition of a cor-
roded surface shape is established by combining a stereo setup
with a special projector unit, which projects different fringe pat-
terns onto the object to be measured [42] (Fig. 3). These patterns
Fig. 3. 3D scanner ATOS [40].
139
are then recorded by two calibrated cameras, so the computer
automatically calculates the 3D coordinates for each of the up to
4 million camera pixels with high precision within just a few sec-
onds [42].
In order to completely digitize a corroded surface, several indi-
vidual measurements from various views were performed. While
measuring the surface, the ATOS system automatically identies
the reference markers that are placed directly on the surface or
in surrounding xtures. All measurements are automatically
merged together into the global coordinate system. The measuring
accuracy depends on the object size. Objects of 4 m are typically
measured with a 3D accuracy of approximately 0.1 mm. The accu-
racy for small objects is better than 0.01 mm.
After scanning, the ATOS software calculates a complete high-
resolution polygon mesh of the object surface, creating small trian-
gles in curved zones and large triangles in atter areas without
diminishing the meshs accuracy. Polygon meshes consist of sev-
eral millions of measuring points and describe the object in a very
detailed way.
The detailed polygon meshes captured by ATOS enable ap-
proaches like rapid surface reconstruction or class-A surface mod-
eling. The surfaces of two areas on both sides of 3 cm2 for any
corroded specimen have been analyzed here (Fig. 4). First, the
upper and lower surface topology has been established and based
on them the corroded thickness has been evaluated. The non-cor-
roded thickness of the specimen was 4.5 mm. As can be seen from
Fig. 4, the corrosion distribution is non-uniform and severe corro-
sion deterioration is observed in many locations of the specimen.
4. Corrosion degradation analysis
It is considered that the corrosion data from all specimens fol-
low the same time dependency for a mean value and standard
deviation. The minimum, maximum, mean values and standard
deviation of measured corroded specimen thickness are given in
Table 2. However, if a different corrosion environment is present,
the time-dependent model will be different, but the corrosion
wastage dependency will remain the same. Therefore, a degree of
degradation (DOD) of corrosion wastage is presented here to ana-
lyze the effect of corrosion deterioration on the tested fatigue
specimens.
The degree of degradation indicates the plate deterioration,
which also may identify the degree of indentations, which is
dened as:
140 Y. Garbatov et al. / International Journal of Fatigue 59 (2014) 137144

(a)

Upper surface
(b)

Lower surface

(c) Thickness

Fig. 4. Thickness measurements.

Table 2
Principal characteristics of corroded specimen thickness.

A6 A7 A8 B9 A10 B14 B17 C4 C11 C12 C13

Maximum (mm) 3.08 3.07 3.23 3.22 3.07 3.23 3.14 3.28 3.29 3.09 2.98
Minimum (mm) 1.50 1.52 1.62 2.13 1.48 2.10 2.09 2.04 1.83 1.96 2.02
Mean value (mm) 2.44 2.39 2.53 2.72 2.40 2.69 2.60 2.53 2.57 2.61 2.52
Standard deviation (mm) 0.39 0.28 0.32 0.19 0.36 0.11 0.17 0.19 0.23 0.22 0.16

V0  Vc tion of the indented surface. The degree of degradation is the true

DOD%   100% 1
V0 measure of wastage of the material. In practice the DOD is calcu-
lated by using the mass of the plate instead of its volume.
where V0 is the intact plate volume, L  b  ho and Vc denotes the Another common way to assess the material loss caused by cor-
corroded volume, dened as: rosion degradation is the concept of net thickness. This concept
Z Z measures the corrosion degradation by considering an equivalent
L b
thickness, heq of the structural component thickness as:
Vc hx; ydxdy 2
0 0
Vc
heq  h0 1  DOD 3
The corroded or indented plate surface is discretized, which im- bL
plies smooth discrete values, leading to a multi-linear approxima-
or by considering the average thickness, hav, which may be ex-
pressed as:

hav  h 4
where h denotes the mean value of the corroded plate surface. For a
relatively uniform surface it can be proven that Eqs. (3) and (4) are
approximately identical. The relation between the mean corrosion
depth and the degree of degradation, DOD for the tested specimen
is shown in Fig. 5.

5. Fatigue test setup

Fig. 2 shows a steel corroded specimen that is fatigue tested to

Fig. 5. Mean corrosion depth of specimens as a function of DOD.
axial cyclic load. A transverse stiffener is welded to one side of the
specimen. The specimen are not symmetrical about the load axis.
Eleven specimen with different corrosion degradation levels are
used to analyze the effects of the degradation and to evaluate
their fatigue performance. The specimen are identied as
 Y. Garbatov et al. / International Journal of Fatigue 59 (2014) 137144 141

Table 3
Principal characteristics of corroded specimens.

Load group Specimen Cross section, (mm2) R Dr (MPa) Frequency (Hz) Number of cycles
A 6 52.74 0.05 120 10 552,360
7 51.48 0.05 120 10 169,023
8 53.86 0.05 120 10 915,974
10 53.37 0.05 120 10 212,719
B 9 55.67 0.05 155 10 368,200
14 51.70 0.05 155 10 842,368
17 53.33 0.05 155 10 315,611
C 4 50.12 0.05 200 10 411,43
11 51.63 0.05 200 10 69,252
12 52.72 0.05 200 10 98,715
13 51.99 0.05 200 10 119,801

non-load-carrying transverse stiffener with llet weld of a FAT 100

fatigue design category based on the hot spot stress approach and a
FAT 86 category based on the nominal stress approach. The FAT
categories determine the fatigue life of a given detail as a function
of fatigue detail geometry and orientation with respect to the
direction of loading [43,44].
The design of the specimen assumes that if the specimen is in-
tact, without corrosion, the failure will occur along the toe of the
transverse lled weld. Fig. 2 shows the dimensions of specimens.
All specimens were 180 mm in length and had a plate thickness
of intact non-corroded specimen of 4.5 mm. The stiffener of intact
non-corroded specimen was also 4.5 mm thick. The stiffener width
was in average for all specimens about 17.6 mm. The specimen
were welded based on normal shipyard practice and were ma-
chined to smooth the edges.
All specimen were tested under cyclic axial loading at a nominal
stress range, Dr as is shown in Table 3. The stress range is the dif-
ference between the minimum and maximum nominal stresses.
The stress range across the net section of the plate, which varies
from specimen to specimen as can be seen in Table 3, is modeled
as 120, 155 and 200 MPa, respectively and were calculated by
dividing the applied axial loading by the mean cross-sectional area
Fig. 7. Fatigue test ruptured specimen.
of the plate. The minimum nominal stress corresponds to 6, 8 and
11 MPa respectively, whereas the maximum stress was 126, 163
and 210 MPa respectively. It should be noted that the three groups
of loading have been simulated A, B and C. The rst group consists
of specimens 6, 7, 8 and 10, the second group covers the specimens
9, 14 and 17 and the third one includes 4, 11, 12 and 13.
A servo hydraulic testing machine w + b LFV50-HH/EDC120
with a control unit DIGWIN 2000 of manufacturer Bai [45], shown
in Fig. 6, with a 63 kN capacity was used to apply the test load.
All specimen were tested under cyclic loading at a nominal
stress range indicated in Table 3 and a stress ratio, R = 0.05. The
stress ratio is dened as the minimum nominal stress divided by
the maximum nominal stress. All specimen were tested at 10 Hz.
Fig. 6 shows a specimen installed in the test machine prior to
loading.
The loads and operating frequency were recorded into the
computer via Dion-Pro software [46]. The loads and operating
frequency were achieved through small increments, bringing the
machine up to speed at a controlled rate.
Specimen were considered failed if either an error detector was
triggered by an increase in specimen displacement due to a crack
opening or rupture occurred due to rapid crack propagation (see
Fig. 7). When triggered, the detectors were set to immediately stop
all testing by turning off the hydraulics. The specimens were then
ungripped and removed from the machine.

6. Fatigue strength assessment

The fatigue life, or number of loading cycles to failure, was re-

Fig. 6. Uniaxial testing machine (left) and with mounted specimen (right).
corded and presented for each specimen in Table 3. A very large
scatter of fatigue life can be observed. Fatigue life data exhibit
widely scattered results because of inherent microstructural inho-
mogeneity in the material properties, differences in the surface and
material properties due to corrosion. It has been observed that
once pits nucleate cracks in a material at high-stress levels, these
cracks have a better chance of overcoming the surrounding
microstructure.
The fatigue test results for the different corrosion degradation
level are given in Fig. 8. The statistical evaluation of the results
was performed assuming a slope exponent m = 3 of the SN curve
being typical for welded joints and close to that which resulted
from a regression analysis. The resulting fatigue strength at two
142 Y. Garbatov et al. / International Journal of Fatigue 59 (2014) 137144
Fig. 8. Estimated 50% and 97.7% survival probabilities of fatigue test results.
Fig. 9. Comparison of fatigue test results with Category 50 and 100.
million cycles was determined for probabilities of survival
Ps = 97.7%, 90%, 50% and 10%. The ratio between the fatigue lives
for Ps = 90% and 10% yields the scatter ratio TN. The regression anal-
ysis leads to m = 3.094 with Dr97:7%; m3 64:945 MPa and TN = 1.3.
Specimen A8 had the longest fatigue life, failing at 915,974 cy-
cles of loading at a measured stress range of 120 MPa. Specimen
C11 had the shortest fatigue life, failing at 69,252 cycles. A nominal
stress range of 200 MPa was assumed for specimen C11. The
average fatigue life of all specimens tested is 336,833 cycles. The
average measured stress range is 158 MPa.
Fig. 9 shows the fatigue test results plotted together with the
FAT design categories 50 and 100. Ten out of eleven observed fail-
ures occurred between FAT 50 and 100 categories. The corrosion
degradation of the specimens was the cause of the shorter fatigue
life relative to the FAT design category 100. All of the failures initi-
ated at a specic corrosion pit at the narrowed plate away from the
lled weld due to severe corrosion in the specimens.
7. Initial corrosion dent identication
To identify the initial pit size and shape, the corrosion pits are
assumed as a aw allowing existing or postulated pit geometry
to be modeled by a geometrically simple crack amenable to a frac-
ture mechanics analysis. The characterized crack-like corrosion
pits are intended to lead to idealized crack geometries that are
more severe than the actual crack geometry they represent, but
will arrive to the fatigue life as it was dened by the fatigue test.
Nakai et al. [12,47,48] performed extensive studies on the effect
of pit corrosion occurring to bulk carriers and oil tanker. They mea-
sured the pit corrosion geometry on both types of ships and found
that the ratio of 2c/a is between two and ve. It is assumed here
that this ratio is 3. The idealized shape used to evaluate crack like
aws is shown in Fig. 10.
Fig. 10. Corrosion pit characterization.
The aw length is not difcult to determine for surface breaking
aws. If the aw is oriented perpendicular to the plane of the max-
imum principal tensile stress, then the aw length to be used in
calculations is merely the measured length 2c. If the aw is not ori-
ented in a principal plane, then an equivalent aw dimension with
a Mode I [49] orientation has to be identied. An initially mixed-
mode crack will tend to become a mode I crack when it propagates,
assuming the material is relatively homogeneous. If there are no
planes of weakness in the material, the propagating crack will fol-
low the path where the driving force is highest, which turns out to
be approximately normal to the maximum principal stress in most
cases.
The Failure Assessment Diagram (FAD) Level 2 assessment,
based on BS7910 [50] Level 2 approach is used here for estimating
the initial corrosion pit resulting in unstable brittle fracture. The
p
applied relative Crack Tip Opening Displacement (CTOD), dr is
used here, which has been shown to have potential in quantifying
crack tip deformations during stable fatigue crack growth.
p p
The FAD curve Level 2, dr FAD is calculated as follows. When dr
p
exceeds dr FAD , the driving force for fatigue crack propagation will
lead to unstable fracture.
p 6 ry ru
dr FAD 1  0:14  L2r 0:3 0:7  e0:65Lr for Lr 6 5
2ry
where ry is the yield stress of material, ru is the ultimate stress of
material, and Lr is the relative load ratio, which is given by:
rref
Lr 6
ry
where rref is the resultant stress remote from the crack. In the DAF
procedure Level 2 of BS 7910 crack assessment, Lr represents the
stress level in the plate relative to yield, which is relative to the
plastic collapse of the plate, without the crack. According to the
BS7910 standard, the residual weld stress is assumed to be at yield
level and the effect of through thickness variation is negligible. The
p
applied relative CTOD, dr [50] is dened by using CTOD, d accord-
ing to the BS7910:
s
p d
dr 7
dmat
where dmat is the parameter characterizing material resistance
against unstable crack propagation. For most marine structural
steels and weldments, the fracture toughness is most commonly
measured as the critical CTOD value dmat that corresponds to unsta-
ble fracture initiation in the specimen ASTM [51]. dmat equals to 0.67
mm is used here.
A relationship between the stress intensity factor and CTOD
have been established for the case of small scale yielding around
the crack tip and for the plan stress case, which on a thin plate is
given as:
K2
d  8
ry  E
where the applied crack
p
tip stress intensity factor is dened by
p
K Y2c=Wr p2c and dr is calculated for the incremental
p
fatigue crack length until it equals to dr FAD , where a critical crack
size, 2ccrit W, of 17.6 mm is reached. The dimensionless geometry
function Y2c=W accounts for loading mode, crack and joint
geometry. The geometrical function is based on the work of Gurney
 Y. Garbatov et al. / International Journal of Fatigue 59 (2014) 137144 143

Table 4 Acknowledgements
Initial defect size.

Specimen group Dr (MPa) ao (mm) 2co (mm) 2ccrit = W (mm) The experimental work on the box girders fom where the pres-
A 120 0.000104 0.000311 17.6 ent specimens were extracted has been performed within the pro-
B 155 0.000651 0.001953 17.6 ject MARSTRUCT Network of Excellence on Marine Structures,
C 200 0.008199 0.024597 17.6 (www.mar.ist.utl.pt/marstruct/), which has been funded by the
European Union through the Growth program under contract
TNE3-CT-2003-506141. The present work has been funded by the
Table 5
Portuguese Foundation for Science and Technology through its an-
Initial defect SCF. nual funding to CENTEC, the Centre for Marine Technology and
Engineering.
Specimen Dr (MPa) Number of cycles SCF
4 200 41,143 1.829
References
11 200 69,252 1.538
12 200 98,715 1.366
[1] TSCF. Condition Evaluation and Maintenance of Tanker Structures (Tanker
13 200 119,801 1.281
Structure Co-operative Forum). Witherby & Co., London; 1992.
7 120 169,023 1.899
[2] TSCF, Guidance Manual for Tanker Structures (Tanker Structure Co-operative
10 120 212,719 1.759
Forum), Tanker Structure Cooperative Forum; 1997.
17 155 315,611 1.196 [3] Yamamoto N, Ikagaki K. A study on the degradation of coating and corrosion on
9 155 368,200 1.136 ships hull based on the probabilistic approach. J Offshore Mech Arct
6 120 552,360 1.280 1998;120:1218.
14 155 842,368 0.862 [4] Wang G, Spencer J, Elsayed T. Estimation of corrosion rates of oil tankers. In:
8 120 915,974 1.081 Proceedings of the 22nd International Conference on Offshore Mechanics and
Arctic Engineering, ASME. OMAE 2003; 2003. p. 37361.
[5] Paik JK, Kim S, Lee S, Park Y. A probabilistic corrosion rate estimation model for
[52], who derived numerical values for Y2c=W for a transversely longitudinal strength members of bulk carriers. J Ship Ocean Technol
1998;2:5870.
stiffened welded plate with a crack developed on the plate with
[6] Guedes Soares C, Garbatov Y. Reliability of maintained, corrosion protected
an average weld toe prole. The initial crack depth, ao and length, plates subjected to non-linear corrosion and compressive loads. Marine Struct
2co , where 2co =3ao are assumed, according to the initial pit size def- 1999:42545.
[7] Garbatov Y, Guedes Soares C, Wang G. Nonlinear time dependent corrosion
inition as it is given in [12,47,48]. The initial pit like aw defect, for
wastage of deck plates of ballast and cargo tanks of tankers. J Offshore Mech
all specimens, are calculated and presented in Table 4. Arct Eng Trans ASME 2007;129:4855.
To identify how sharp is the initial corrosion dent or what is the [8] Guedes Soares C, Garbatov Y, Zayed A, Wang G. Corrosion wastage model for
SCF of the dent, which may initiate crack propagation, the fatigue ship crude oil tanks. Corros Sci 2008;50:3095106.
[9] Guedes Soares C, Garbatov Y, Zayed A, Wang G. Inuence of environmental
test results are analyzed again. The SCF for any individual speci- factors on corrosion of ship structures in marine atmosphere. Corros Sci
men is calculated based on the applied load for fatigue testing, 2009;51:201426.
achieved fatigue life and using a FAT 100 fatigue design category, [10] Guedes Soares C, Garbatov Y, Zayed A. Effect of environmental factors on steel
plate corrosion under marine immersion conditions. Corros Eng Sci Technol
Dr97:7%; m3 100 MPa using the following equation: 2011;46:52441.
[11] Paik JK, Lee JM, Ko MJ. Ultimate strength of plate elements with pit corrosion
10LogaLogN=m wastage. J Eng Maritime Environ 2003;217:185200.
SCF
Dr [12] Nakai T, Matsushita H, Yamamoto N. Effect of pitting corrosion on the ultimate
strength of steel plates subjected to in-plane compression and bending. J Mar
where Loga Log2:106 mLogDr97:7%; m3 . Sci Technol 2006;11:5264.
[13] Jiang X, Guedes Soares C. A closed form formula to predict the ultimate
The calculated SCFs are shown in Table 5. It can be seen that the
capacity of pitted mild steel plate under biaxial compression. Thin Walled
calculated SCF and initial defect size are in synchrony with the Struct 2012;59:2734.
achieved fatigue life. However, in the case of transversely stiffened [14] Guedes Soares C, Garbatov Y. Reliability of corrosion protected and maintained
ship hulls subjected to corrosion and fatigue. J Ship Res 1999;43:6578.
welded plate, without corrosion degradation, for a plate of less
[15] Garbatov Y, Rudan S, Guedes Soares C. Fatigue damage of structural joints
than 25 mm, SCF = 1.13 according to DnV [53]. accounting for nonlinear corrosion. J Ship Res 2002;46:28998.
[16] Moan T, Ayala-Uraga E. Reliability-based assessment of deteriorating ship
structures operating in multiple sea loading climates. Reliab Eng Syst Safe
8. Conclusion 2008;93:43346.
[17] Tran Nguyen K, Garbatov Y, Guedes Soares C. Fatigue damage assessment
The work presented here analyzes the effect of corrosion dete- based on global and local stress approaches accounting for corrosion. Int J
Fatigue 2012;43:197206.
rioration on a transversely stiffened welded joint subjected to cyc- [18] Tran Nguyen K, Garbatov Y, Guedes Soares C. Spectral fatigue damage
lic loading. The fatigue behavior is analyzed based on the fatigue assessment of tanker deck structural detail subjected to time-dependent
test results. The specimens were cut from a box girder that was ini- corrosion. Int J Fatigue 2013;48:14755.
[19] Fricke W, Petershagen H. Detail design of welded ship structures based on hot-
tially corroded in real sea water conditions. In order to transfer the spot stresses. In: Caldwell JB, Ward G, editors. Proceedings of the practical
results to other geometries loads, corrosion degradation and initial design of ships and mobile units. Elsevier Science Limited; 1992. p. 1087100.
surface roughness of 11 corroded specimens were analyzed apply- [20] Guedes Soares C, Moan T. Model uncertainty in the long-term distribution of
wave-induced bending moment for fatigue design of ship structure. Marine
ing photogrammetry techniques and a description of the idealized Struct 1991;4:295315.
corroded surface was established. The non-corroded specimens are [21] Guedes Soares C, Garbatov Y, von Selle H. Fatigue damage assessment of ship
identied as FAT 86 category but the fatigue test demonstrated structural components based on the long-term distribution of local stresses.
Int Shipbuild Progr 2003;50:3556.
that due to the severe corrosion degradation the experimental [22] Paris P, Erdogan F. A critical analysis of crack propagation laws. J Basic Eng
fatigue results are located above the fatigue design category FAT 1963;85:52834.
50 and below FAT 100 referring to the nominal stress approach. [23] Guedes Soares C, Garbatov Y. Reliability of maintained ship hull subjected to
corrosion and fatigue. Struct Safe 1998;20:20119.
The regression analysis of fatigue test results leads to m = 3.094
[24] Petershagen H, Fricke W, Massel T. Application of the local approach to the
and Dr97:7%;m3 64:95 MPa. Fatigue assessment of crack propaga- fatigue strength assessment of welded structures in ships. In: International
tion on pit like crack ow based on a failure assessment diagram Institute of Welding; 1991.
was performed demonstrating a direct relation between the initial [25] Radaj D. Design and analysis of fatigue-resistant welded
structures. Cambridge: Abington Publishing; 1990.
corrosion surface roughness and pit depth to the fatigue life [26] Radaj D, Sonsino CM, Fricke W. Recent developments in local concepts of
estimation. fatigue assessment of welded joints. Int J Fatigue 2009;31:211.
144 Y. Garbatov et al. / International Journal of Fatigue 59 (2014) 137144
[27] Niemi E. Determination of stresses for fatigue analysis of welded components.
In: Proceedings of the IIW, Conference on Design in Welded, Construction;
1992. p. 5774.
[28] Niemi E. Recommendation concerning stress determination for fatigue
analysis of weld components. Abington Publishing; 1995.
[29] Niemi E, Fricke W, Maddox S. Structural stress approach to fatigue analysis of
welded components designers guide. In: IIW Doc. XIII-1819-00/XV-1090-01;
2004.
[30] Chakarov K, Garbatov Y, Guedes Soares C. Fatigue analysis of ship deck
structure accounting for imperfections. Int J Fatigue 2008;30:188197.
[31] Garbatov Y, Guedes Soares C. Inuence of steel strength on the fatigue
reliability of welded structural components. Int J Fatigue 2004;26:75362.
[32] Domzalicki P, Skalski I, Guedes Soares C, Garbatov Y. Large scale corrosion
tests. In: Guedes Soares C, Das PK, editors. Marstruct Bk Series. London,
UK: Taylor & Francis Group; 2009. p. 1938.
[33] Saad-Eldeen S, Garbatov Y, Guedes Soares C. Corrosion-dependent ultimate
strength assessment of aged box girders based on experimental results. J Ship
Res 2011;55:289300.
[34] Saad-Eldeen S, Garbatov Y, Guedes Soares C. Effect of corrosion degradation on
the ultimate strength of steel box girders. Corros Eng Sci Technol
2012;47:27283.
[35] Saad-Eldeen S, Garbatov Y, Guedes Soares C. Ultimate strength assessment of
corroded box girders. Ocean Eng 2013;58:3547.
[36] Koelman H. Application of a photogrammetry-based system to measure and
reengineer ship hulls and ship parts: An industrial practices-based report.
Comput Aid Des 2010;42:73143.
[37] Ljubenkov B, Zaplatic T, Ziha K. Measurements for reliability of ship design,
Production and operation interacion; 2008.
[38] Menna F, Ackermann S, Scamardella A, Troisi S. Digital photogrammetry: a
useful tool for shipbuilding applications. In: Proceedings of the 13th Congress
of International Maritime Association of Mediterranean, IMAM, Istanbul,
Turkey; 2009. p. 60714.
[39] Chen BQ, Garbatov Y, Guedes Soares C. Displacement Measurement of
Box Girder Based on Photogrammetry. In: Spaier S, Estefen S, Pasqualino I,
editors. Proceedings of the 11th international symposium on practical design
of ships and other oating structures (PRADS 2010), paper PRADS2010-
20083. Rio de Janeiro, Brasil: Aped - Apoio e Produo Ltda; 2010.
[40] Chen BQ, Garbatov Y, Guedes Soares C. Measurement of weld induced
deformations in three-dimensional structures based on photogrammetry
technique. J Ship Product Des 2011;27:5162.
[41] Chen BQ, Garbatov Y, Guedes Soares C. Automatic approach for measuring
deformations in complex structures using photogrammetry technique. In:
Proceedings of the XXII Pan American Conference of Naval Engineering,
Maritime Transportation & Ports Engineering, Buenos Aires, Argentina; 2011.
[42] ATOS, 3D scanner ATOS, <http://www.gom.com/metrology-systems/system-
overview/atos-triple-scan.html>; 2013.
[43] Eurocode. Design of steel structures. Brssel: European Committee for
Standardisation; 1992.
[44] Hobbacher A. Recommendations for fatigue design of welded joints and
components, IIW doc.1823-07, Welding Research Council Bulletin 520,
International Institute of Welding, New York; 2009.
[45] Bai W. Servohydraulic testing machine w+b LFV50-HH/EDC120 with control
unit DIGWIN 2000 of manufacturer Walter Bai, <http://www.walterbai.com/>;
2013.
[46] Dion-Pro-software, <http://software.informer.com/>; 2013.
[47] Nakai T, Matsushita H, Yamamoto N. Effect of pitting corrosion on local
strength of hold frames of bulk carriers (2nd report) - Lateral-distortional
buckling and local face buckling. Mar Struct 2004;17:61241.
[48] Nakai T, Matsushita H, Yamamoto N, Arai H. Effect of pitting corrosion on local
strength of hold frames of bulk carriers. Marine Struct 2004;17:40332. 1st
report.
[49] Hellan T. Introduction to fracture mechanics. McGraw-Hill Book Company;
1988.
[50] BS7910. British Standard BS7910, Guide to methods for assessing the
acceptability of aws in metallic structures, BSI, London; 2005.
[51] ASTM. Standard E 1290. Standard test method for crack-tip opening
displacement (CTOD) fracture toughness measurement, American Society of
Testing and Materials, Philadelphia; 1995.
[52] Gurney TR. Finite element analysis of some joints with the welds transverse to
the direction of stress. Weld Res Int 1976;6:4072.
[53] DnV, Fatigue assessment of ship structures, Classication notes 30.7, Det
Norske Veritas, Hovik, Norway, 2003.