This copy of the document is intended for use by ashkan babazadeh only. Downloaded 2024-04-29.

No further distribution shall be made.

SECTION 3 FATIGUE STRENGTH REPRESENTATION

Section 3
1 Introduction
The basis for the fatigue assessment is an estimate of the stress range (hot spot, nominal or local at free
plate edge) which is described in [2]. When assessing the fatigue strength the time in corrosive and in-
air environment, the time in different loading conditions and the time in port need to be accounted for as
described in [3]. The results of the fatigue assessment can be presented in different ways depending on what
is found convenient:
— Calculated fatigue damage versus allowable fatigue damage, [3]
— Calculated fatigue life versus design fatigue life, [3]
— Actual stress range (calculated) versus permissible peak stress range (calculated), [4]
— Given stress concentration factor versus permissible stress concentration factor (calculated), [5]
— Given FAT class versus required FAT class (calculated), [5].
The scope and extent of the fatigue assessment, the loading conditions and related parameters to consider
and the time at sea are given in the rules for the standard ship types.

2 Fatigue stress range

2.1 General
The calculated stress range should follow the net scantling approach, tn50, for prescriptive fatigue assessment
and net or gross, t, based on FE analysis, where net or gross scantlings is defined in the rules for the specific
ship types. Since there is a difference on how the corrosion affects the global and local stress, a correction
factor fc needs to be included. In addition, the calculated stress depends on the wave environment which is
represented by the environmental factor fe.
Before performing fatigue strength assessment, the calculated stress entered into the S-N curve needs to be
corrected for other fatigue capacity effects such as:
— Mean stress
— Thickness
— Material
— Post-weld treatment.
Results from performed fatigue analyses are presented in App.C in terms of allowable stress ranges as
function of the Weibull shape parameter. The basis for the allowable stress ranges is that long term stress
ranges can be described by a two parameter Weibull distribution.

2.2 Scantlings approach factor, fc

The global stress is less affected by the corrosion than the local stress due to the probabilistic nature of
corrosion, which gives less average corrosion on a global level compared to at a local level. Thus, the local
stress should be based on tn50 while the global stress should be based on tn25.
For prescriptive fatigue assessment using a beam model based on tn50, the correction factor fc only applies to
the global stresses. When using a global FE model based on tn50 or tgr, the same correction factor applies to
both the global and local stresses, i.e. directly on the stress.
The correction factor should be taken as:
a) Prescriptive fatigue strength assessment
— fc = 1.0 for local plate and stiffener bending based on tn50
— fc = 0.95 for hull girder stresses based on tn50

Class guideline — DNV-CG-0129. Edition October 2021 Page 43

Fatigue assessment of ship structures

DNV AS
 This copy of the document is intended for use by ashkan babazadeh only. Downloaded 2024-04-29. No further distribution shall be made.
b) FE based fatigue strength assessment

Section 3
— fc = 1.0 for stresses based on tgr
— fc = 0.95 for stresses based on tn50
Since FE models may be based on either tn50 or tgr, the notation t is used for the thickness in relation to
FE analysis. However, prescriptive assessment should be based on tn50.

2.3 Environmental factor, fe

The environmental factor fe is covering World Wide and North Atlantic wave environment for merchant
vessels. The fe factor for prescriptive approach is listed in DNV Pt.3 Ch.9 Sec.4 [4.2].
In direct hydrodynamic calculations based on a specific scatter diagram, the fe factor is set to 1.0 or can be
omitted.

2.4 Operational factor, fR

While the environmental factor fe refers to the operational area, the operational factor fR relates to how the
vessel operates in the operational area regarding weather routing. The operational factor for prescriptive
loads is defined in DNV Pt.3 Ch.9 Sec.4 [4.3].
In direct hydrodynamic calculations, the operational factor, fR, and possibly also the fatigue coefficient,
ffa, and the vibration factor, fvib, depending on the scatter diagram used, are multiplied with the directly
calculated stress range, , as explained in DNV DNV-CG-0130 Sec.2 [7] Wave loads to find the fatigue
stress range:

For site specific operation the fvib is reduced to 1.05. For site specific scatter diagram and trade specific
scatter diagram with full effect of routing the ffa and fR are omitted.

2.5 Post-weld treatment

Reference is made to the rules for the application of the post-weld treatment. There are several limitations
for the benefit of post-weld treatment. This factor should preferably not be considered in the initial design
phase. For burr grinding the stress range can be multiplied with the following factor:

where:

fT = Fatigue life factor as given in Sec.7 [2.3].

2.6 Fatigue stress range

The fatigue stress range ΔσFS in N/mm to be used in the assessment is based on the calculated stress range
2

Δσ times the corrections for mean stress effect, thickness effect, material factor, scantlings approach factor,
the environmental factor and the post-weld treatment factor:

Class guideline — DNV-CG-0129. Edition October 2021 Page 44

Fatigue assessment of ship structures

DNV AS
 This copy of the document is intended for use by ashkan babazadeh only. Downloaded 2024-04-29. No further distribution shall be made.
Section 3
-2
The stress range should be estimated at a 10 probability level of exceedance.
The straight line spectrum, i.e. 2-parameter Weibull distribution with Weibull slope ξ = 1.0, is displayed
in Figure 1 where it can also be seen that in this case the permissible stress range at a probability level of
-2 -8
10 is 25% of the stress range at a probability level of 10 . In Figure 2 the distribution of the damage is
illustrated at different probability levels based on the straight line spectrum, suggesting that most of the
-2
fatigue damage comes from smaller stress ranges close to a probability level of 10 .

Figure 1 Straight line spectrum giving a damage of 1.0 based on S-N curve D/FAT 90.

Figure 2 Damage distribution for a straight line spectrum based on S-N curve D/FAT 90.

Class guideline — DNV-CG-0129. Edition October 2021 Page 45

Fatigue assessment of ship structures

DNV AS
 This copy of the document is intended for use by ashkan babazadeh only. Downloaded 2024-04-29. No further distribution shall be made.
Section 3
3 Fatigue damage and fatigue life calculation

3.1 Fatigue damage accumulation with Palmgren - Miner’s rule

Fatigue assessment may be carried out by estimating the linear cumulative fatigue damage D by using
Palmgren-Miner’s rule formulated as:

where:

ni = Number of cycles at stress range Δσi

Ni = Number of cycles to failure at stress range Δσi
ntot = Total number of stress range blocks
i = Stress range block index.

The number of load cycles, ni, is determined by the long term stress distribution, e.g. represented by the
Weibull distribution, while the number of cycles to failure, Ni, is represented by the S-N curve.

The acceptance criteria to the damage D is given by the rules.

3.2 Time in air and corrosive environment

The design fatigue life is divided into two time periods due to limitation of the corrosion protection. It is
assumed that the corrosion protection (i.e. coating system) is only effective for a limited number of years
during which the structural details are exposed to in-air environment. During the remaining part of the
design life, TC, as specified in the rules, the structural details are unprotected i.e. exposed to corrosive
environment.

3.3 Annual fatigue damage

The annual fatigue damage is the damage in-air accumulated during a specific loading condition (j) assumed
100% time at sea. The annual fatigue damage, can be based on the fatigue stress range obtained from the
predominant load case (EDW giving the highest and thereby the most representative fatigue stress range,
ΔσFS) in the prescriptive fatigue assessment. This can be calculated based on a closed form formulation as:

Class guideline — DNV-CG-0129. Edition October 2021 Page 46

Fatigue assessment of ship structures

DNV AS
 This copy of the document is intended for use by ashkan babazadeh only. Downloaded 2024-04-29. No further distribution shall be made.
Section 3
where:

Total number of stress cycles possible per year, where 1/(4logL) is assumed as the zero
ND =
up-crossing frequency
Fatigue stress range from the predominate load case at the reference probability level of
ΔσFS,(j) = -2 2
exceedance of 10 , in N/mm
-2
Number of cycles corresponding to the reference probability of exceedance of 10 . NR =
NR =
100.

ξ = Weibull shape parameter, ξ=1

Γ(ξ) = Complete Gamma function

K2 = Constant of the design S-N curve, as given in Table 1 for in-air environment

μ(j) = Coefficient taking into account the change of inverse slope of the S-N curve, m

γ(a,ξ) = Incomplete Gamma function

2
Stress range, in N/mm , corresponding to the intersection of the two segments of design
Δσq = 7
S-N curve at N = 10 cycles, as given in Table 1
7
Δm = Change in inverse slope of S-N curve at N = 10 cycles.
Δm = 2

3.4 The combined fatigue damage

The combined fatigue damage is the damage accumulated in both in-air and in corrosive environment for a
specific loading condition (j) during a specified design fatigue life, TDF, with 100% time at sea. The combined
fatigue damage for the design fatigue life, TDF, for each loading condition (j) can be calculated as:

where:

TC,25 = Time in corrosive environment, in years, within the duration of the design life, TD

Class guideline — DNV-CG-0129. Edition October 2021 Page 47

Fatigue assessment of ship structures

DNV AS
 This copy of the document is intended for use by ashkan babazadeh only. Downloaded 2024-04-29. No further distribution shall be made.
TD = Design life, in years, taken as TD = 25 years.

Section 3
DE,(j) = The annual fatigue damage for in-air environment for loading condition (j).

3.5 The total fatigue damage from multiple loading conditions

Total fatigue damage within the design fatigue life, TDF, is the sum of the combined fatigue damages obtained
for all loading conditions and accounting for the fraction of the time at sea and the fraction in each loading
condition. The total fatigue damage for all applicable loading conditions is calculated as:

where:

f0 = Factor taking into account time in seagoing operations given in the rules

αj = Fraction of time in each loading condition given in the rules

nLC = Total number of loading conditions given in the rules

D(j) = Combined fatigue damage for each applicable loading condition.

3.6 Fatigue life

The predicted fatigue life, TF, is based on different part times in-air and corrosive environment. For short
fatigue lives less than TD − TC,25 years, the predicted fatigue life is based on in-air environment. For
predicted fatigue lives between TD and TC,25, the time after TD − TC,25 is based on corrosive environment.
For predicted fatigue lives above TD, it is assumed that the predicted fatigue life is based on a regular
maintenance intervals from the delivery of the vessel. This may be approximated by a constant annual
damage rate which also reflects a relative amount of corrosive environment. The predicted fatigue life above
TD = 25 years is independent on design fatigue life TDF (which is equal or longer than TD). If the design
fatigue life is changed at a later time due to the need for life time extension, the predicted fatigue life for the
same operation area does not change. The formulation is illustrated in Figure 3.The predicted fatigue life is
calculated as:

for

for

for

where:

Class guideline — DNV-CG-0129. Edition October 2021 Page 48

Fatigue assessment of ship structures

DNV AS
 This copy of the document is intended for use by ashkan babazadeh only. Downloaded 2024-04-29. No further distribution shall be made.
Section 3
and where:

D = Total damage given by [3.5]

DE,(j) = The annual fatigue damage for in-air environment for loading condition (j).

Figure 3 Illustration of fatigue damage accumulation for different assumptions of time in

corrosive environment

4 Permissible stress range

The stress range should be less or equal to the permissible stress range:

In general, the permissible stress range, Δσperm, can be assessed iteratively as the stress range which yields
a total damage D = 1. However, the assessment of Δσperm, is limited to cases where there is only one loading
condition applicable for the whole design fatigue life. However, it can be used on several loading conditions,
-2
to see which is worst. The permissible stress range at 10 probability level can be calculated as:

Class guideline — DNV-CG-0129. Edition October 2021 Page 49

Fatigue assessment of ship structures

DNV AS
 This copy of the document is intended for use by ashkan babazadeh only. Downloaded 2024-04-29. No further distribution shall be made.
2 6 2
FAT = Reference stress, in N/mm , range of the S-N curve at 2·10 load cycles (FAT class), in N/mm

Section 3
Factor for the spectrum’s shape (straight-line) and the number of load cycles
= 0.786 log(ND · f0 · TDF-C) − 4.966 for welded details with S-N curves D, E, F and FAT according
= to Sec.2 Table 1
fN 0.681 log(ND · f0 · TDF-C) − 4.188 for free plate edges with S-N curves C, C1 and C2, Sec.2
=
Table 1
=
0.6007 log(ND · f0 · TDF-C) − 3.578 for base material with S-N curve B1 and free plate edges
with S-N curves B and B2, Sec.2 Table 1.

To reflect the part time in the different environments (protected or corrosive), for the determination of fN, the
design fatigue life should be taken as:

5 Permissible stress concentration factor and required FAT class

In general, the permissible stress concentration factor or required FAT class can be estimated iteratively,
yielding a total damage D = 1. In cases where there is only one loading condition applicable for the whole
design fatigue life, the permissible stress concentration factor, Kperm, can be estimated as:

where:
2 2
Δσn = Nominal stress range of a straight-line spectrum at 10 load cycles, in N/mm .

6
The required reference stress range of the S-N curve at 2·10 load cycles, FATreq, can be estimated as:

for nominal stress, where:

2
Nominal stress range or local stress range at free plate edge of a straight-line spectrum at 10
Δσ = 2
load cycles, in N/mm .

The estimation of Kperm or FATreq should be limited to cases where all load components refer to the same
stress concentration factor. As the stress concentration factors K for bending and axial loading of longitudinal
end connections may differ, this method is not regarded applicable, or too conservative for these details. If
different loading conditions should be assessed for fractions of the time of the design fatigue life, Kperm or
FATreq can only be estimated iteratively, yielding a total damage D = 1.

Class guideline — DNV-CG-0129. Edition October 2021 Page 50

Fatigue assessment of ship structures

DNV AS