Guide for Spectral-Based Fatigue Analysis for Vessels

GUIDE FOR

SPECTRAL-BASED FATIGUE ANALYSIS FOR VESSELS

(FOR THE ‘SFA (years)’ CLASSIFICATION NOTATION)

JUNE 2016 (Updated October 2017 – see next page)

American Bureau of Shipping

Incorporated by Act of Legislature of
the State of New York 1862

 2016 American Bureau of Shipping. All rights reserved.

ABS Plaza
16855 Northchase Drive
Houston, TX 77060 USA
Updates

October 2017 consolidation includes:

• February 2017 version plus Notice No. 2

February 2017 consolidation includes:

• June 2016 version plus Notice No. 1 and Corrigenda/Editorials
Foreword

Foreword (1 October 2017)

This Guide provides information about the optional classification notations, ‘Spectral Fatigue Analysis’ –
SFA (years) and SFA (years, WWT) – which are available to qualifying vessels as described in 1-1-3/21
of the ABS Rules for Conditions of Classification (Part 1).
The source document for this Guide is the Guidance Notes on Spectral-Based Fatigue Analysis for Vessels
(2004). Promoting the Guidance Notes to a Guide adheres to ABS standard practice as the source document
was published more than ten years ago. Furthermore, classification notations are customarily provided by
only Rules or Guides. Promotion to a Guide alleviates any confusion caused by deviation from standard
practice.
The technical content in the Guide has not been significantly changed from that in the Guidance Notes.
However, it has been significantly reorganized such that it more closely aligns with the procedure and
steps commonly used to conduct spectral-based fatigue analyses for ship structures. Nomenclature has
also been modified such that it is more consistent throughout the document. The organization of the Guide
is detailed in Section 2. A flowchart depicting the spectral-based fatigue procedure is given by Section 2,
Figure 1.
The table below shows how the existing Guidance Notes have been reorganized in the creation of this
Guide.
Section in SFA Guide Section in SFA Guidance Notes Remarks
1 Introduction 1 Introduction
2 Spectral-based Fatigue Analysis 1 Introduction SFA procedure overview was
Procedure expanded and moved to its own
section.
3 Seakeeping Analysis 2 Establishing Fatigue Demand Seakeeping analysis was addressed in
4 Motion Analysis and Wave- one section instead of across three
induced Load Components sections. The content in the
seakeeping analysis section is in
6 Loading for Global Finite agreement with the DLA/SFA
Element Method (FEM) System.
Structural Analysis Model
4 Structural Analysis 5 Wave-induced Load Components Structural analysis was addressed in
6 Structural Modeling and one section instead of across two
Analysis sections. The content in the structural
analysis section is in agreement with
the DLA/SFA System.
5 Spectral Analysis 9 Fatigue Life (Damage) The discussion of spectral analysis
was moved to its own section and
expanded.
6 Fatigue Strength Assessment 8 Fatigue Strength The fatigue strength assessment
9 Fatigue Life (Damage) section in the Guide contains most of
the content in the fatigue strength and
fatigue life sections from the
Guidance Notes.

This Guide becomes effective on the first day of the month of publication.
Users are advised to check periodically on the ABS website www.eagle.org to verify that this version of
this Guide is the most current.
We welcome your feedback. Comments or suggestions can be sent electronically by email to [email protected].

ABS GUIDE FOR SPECTRAL-BASED FATIGUE ANALYSIS FOR VESSELS . 2016 iii
 Table of Contents

GUIDE FOR

SPECTRAL-BASED FATIGUE ANALYSIS FOR VESSELS

CONTENTS
SECTION 1 Introduction ............................................................................................ 1
1 Purpose and Applicability....................................................................1
3 Background .........................................................................................2
5 Areas for Fatigue Strength Evaluation ................................................2
7 Methodology and Assumptions ...........................................................2

SECTION 2 Spectral-based Fatigue Analysis Procedure Overview ....................... 4

1 General ...............................................................................................4
1.1 Seakeeping Analysis (Section 3) ..................................................... 4
1.3 Structural Analysis (Section 4) ......................................................... 4
1.5 Spectral Analysis (Section 5) ........................................................... 5
1.7 Fatigue Strength Assessment (Section 6) ....................................... 5

FIGURE 1 Schematic Spectral-based Fatigue Analysis Procedure ...........6

SECTION 3 Seakeeping Analysis .............................................................................. 7

1 Scope of Seakeeping Analysis ...........................................................7
3 Vessel Loading Conditions .................................................................7
5 Hydrodynamic Models ........................................................................7
7 Initial Balance Check ..........................................................................8
9 Seakeeping Analysis Methods ............................................................8
11 Effects of Ship Motions and Wave Pressures.....................................8
11.1 Inertial and Gravitational Loads due to Ship Motions....................... 8
11.3 Real and Imaginary Parts ................................................................ 8
11.5 Pressure Mapping............................................................................ 9
11.7 Intermittent Wetting.......................................................................... 9

SECTION 4 Structural Analysis ............................................................................... 10

1 General .............................................................................................10
3 Internal Load .....................................................................................11
3.1 Liquid Tank Pressures ................................................................... 11
3.3 Bulk Cargo Pressure...................................................................... 12
3.5 Container Loads ............................................................................ 14
3.7 Loads on Lightship Structure and Equipment ................................ 15

iv ABS GUIDE FOR SPECTRAL-BASED FATIGUE ANALYSIS FOR VESSELS . 2016

 5 Load Cases and Model Equilibrium .................................................. 15
5.1 Number of Load Cases.................................................................. 15
5.3 Equilibrium Check ......................................................................... 16
7 Structural Modeling and Analysis ..................................................... 16
7.1 3-D Global Analysis Modeling ....................................................... 16
7.3 Analyses of Local Structure ........................................................... 16
7.5 Hot Spot Stress Concentration ...................................................... 16

FIGURE 1 Ship Coordinate System ......................................................... 10

FIGURE 2 Liquid Pressure on a Completely Filled Tank ......................... 12
FIGURE 3 Liquid Pressure on a Partially Filled Tank .............................. 12

FIGURE 4 Depiction of Surface Normal Vector, n , Acceleration
Vector, a , and α ..................................................................... 13
FIGURE 5 Vertical and Transverse Container Load Components ........... 15
FIGURE 6 Definition of Hot Spot Stress ................................................... 17

SECTION 5 Spectral Analysis ................................................................................. 18

1 General ............................................................................................. 18
3 Wave Spectra and Wave Scatter Diagram ....................................... 18
5 Stress Range Distributions ............................................................... 18

TABLE 1 ABS Wave Scatter Diagram for Unrestricted Service

Classification (Representing the North Atlantic) ..................... 19
TABLE 2 ABS Wave Scatter Diagram for World-Wide Trading Service
Classification ........................................................................... 19

SECTION 6 Fatigue Strength Assessment ............................................................. 20

1 General ............................................................................................. 20
3 S-N Data ........................................................................................... 20
5 Fatigue Strength ............................................................................... 21
7 Combined Fatigue Life from Multiple Loading Conditions ................ 21
9 Acceptance Criteria........................................................................... 22

APPENDIX 1 Basic Design S-N Curves .................................................................... 23

TABLE 1 Parameters for Basic S-N Design Curves .............................. 24

FIGURE 1 S-N Curves.............................................................................. 23

APPENDIX 2 Outline of a Closed Form Spectral-based Fatigue Analysis

Procedure ............................................................................................. 25
1 General ............................................................................................. 25
3 Key Steps in Closed Form Damage Calculation .............................. 25
5 Closed Form Damage Expression .................................................... 28

FIGURE 1 Spreading Angles Definition.................................................... 26

ABS GUIDE FOR SPECTRAL-BASED FATIGUE ANALYSIS FOR VESSELS . 2016 v

This Page Intentionally Left Blank
 Section 1: Introduction

SECTION 1 Introduction

1 Purpose and Applicability (1 October 2017)

Part 5C of the ABS Rules for Building and Classing Steel Vessels (Steel Vessel Rules) includes the fatigue
assessment criteria for the classification of various types of specialized vessels covered by the Rules. Part
5A and 5B of the ABS Rules for Building and Classing Steel Vessels (Steel Vessel Rules) contains fatigue
assessment guidance for vessels subject to the “Common Structural Rules for Bulk Carriers and Oil Tankers”.
A brief description of the background and objectives of these fatigue criteria is given in Subsection 1/3.
In addition to the simplified fatigue strength criteria required for classification by ABS, the Owner may
wish to apply more extensive Spectral-based Fatigue Analysis (SFA) techniques to the vessel’s structural
systems. It may be an added objective of these Spectral-based Fatigue Analyses to demonstrate a longer
design fatigue life than that required for classification.
Spectral-based Fatigue Analysis techniques are used in addition to the SafeHull Fatigue Assessment technique,
a Permissible Stress Range method (discussed in Subsection 1/3). The fatigue life of each critical location
in the structural system is assessed for adequacy. The critical locations are to be selected using the results
of the SafeHull Fatigue Assessment technique which is to be employed in the overall structural design and
analysis effort.
The list of critical structural locations which are to be subjected to Spectral-based Fatigue Analysis is to be
submitted to ABS for approval.
Provided that Spectral-based Fatigue Analysis is conducted in accordance with the procedures included in
this Guide or equivalent, ABS will grant the optional classification notation, SFA (years) or SFA
(years, WWT). In either case, as a minimum, a vessel is to meet the fatigue strength criteria in the Steel
Vessel Rules as described above. The SFA (years) notation is granted if the design fatigue life value is
equal to 20 years or greater under the wave conditions of unrestricted service defined in Section 5, Table 1.
The SFA (years, WWT) notation is granted if the design fatigue life value is equal to 40 years or greater
under the wave conditions of worldwide trading service defined in Section 5, Table 2.
If a vessel is primarily operating in seas that exceed the worldwide trading pattern (i.e., a vessel operating
half or more than half of its life in onerous seas including North Atlantic or North Pacific), SFA (years,
WWT) is not applicable. It should be noted that the calculated fatigue lives for different design wave
conditions may produce significantly different fatigue lives.
The value in parentheses is the design fatigue life in years specified by the applicant in 5-year increments
starting at 20 for SFA (years) and 40 for SFA (years, WWT). The structural system as a whole is analyzed
to verify that the calculated fatigue life values for the entire system meet or exceed the design fatigue life.
The actual service life of a vessel is dependent on many factors such as initial design, operational and
maintenance schemes. The SFA (years) or SFA (years, WWT) notation denotes the design fatigue life
of a vessel and is not a guarantee that the vessel or structure will achieve the design fatigue life.
In order to obtain the SFA (years) or SFA (years, WWT) notations, SFA methods must be applied to
the entire vessel.
For vessels complying with Part 5A and 5B “Common Structural Rules for Bulk Carriers and Oil Tankers”
of the Steel Vessel Rules, the design fatigue life for Spectral-based Fatigue Analysis is equal to 25 years or
greater in 5-year increments.

ABS GUIDE FOR SPECTRAL-BASED FATIGUE ANALYSIS FOR VESSELS . 2016 1

Section 1 Introduction

3 Background (1 October 2017)

In the application of Part 5C of the Steel Vessel Rules, the SafeHull Fatigue Assessment technique is typically
used to evaluate fatigue strength. The SafeHull Fatigue Assessment technique is a permissible stress range
approach that is readily applied to large portions of a vessel’s hull structure. The technique is required for
certain vessels such as an oil tanker with a Rule length greater than 150 m for which the technique is detailed
in Appendix 5C-1-A1 of the Steel Vessel Rules. The technique was derived considering unrestricted ocean
service wave loads representing the North Atlantic and a design fatigue life of 20 years.
Part 5A and 5B “Common Structural Rules for Bulk Carriers and Oil Tankers” of the Steel Vessel Rules
requires a design fatigue life of 25 years with wave loads representing the North Atlantic and multiple
fatigue assessment techniques are employed. A simplified stress analysis method based on beam theory is
presented in Part 5A, Pt 1, Ch 9, Sec 4 of the Steel Vessel Rules, a finite element stress analysis approach is
described in Part 5A, Pt 1, Ch 9, Sec 5 of the Steel Vessel Rules and structural design following the detailed
design standards is discussed in Part 5A, Pt 1, Ch 9, Sec 6 of the Steel Vessel Rules.
Supplementary to the SafeHull Fatigue Assessment technique or methods employed for vessels complying
with Part 5C of the Steel Vessel Rules, ABS may require the use of additional or alternative techniques to
demonstrate the fatigue strength adequacy of structural components. These techniques may include
Spectral-based Fatigue Analysis methods. In many instances the structural details cannot be adequately
analyzed via the permissible stress range fatigue assessment approach. An optional classification notation,
FL (years), may be requested in cases in which the owner or designer increases the target fatigue lives of
some or all of the structural components above the 20 year minimum value. This notation is discussed in the
Steel Vessel Rules and does not require the use of SFA methods.

5 Areas for Fatigue Strength Evaluation

The Steel Vessel Rules provide guidance on which locations should be included in the fatigue assessment.
Various appendices provide guidance on which locations should be included in the fatigue assessment for
different vessel types. For example, 5C-1-A1/3.3 of the Steel Vessel Rules provides guidance for location
selection for oil tankers with a Rule length greater than or equal to 150 m and 5C-3-A1/3.3 of the Steel
Vessel Rules provides guidance for location selection for bulk carriers with a Rule length greater than or
equal to 150 m.

7 Methodology and Assumptions

Spectral-based Fatigue Analysis is a complex and numerically-intensive technique and there are multiple
valid implementations of the method. ABS does not wish to eliminate the use of any valid approach by
over-specifying the SFA technique. However, it is necessary to be clear about the basic assumptions that
form the basis of a valid method and highlight key details that are to be incorporated in the method to
produce acceptable results. The remainder of this Guide is devoted in large part to the presentation of these
topics.
A typical spectral fatigue analysis for a structural location is to evaluate its fatigue strength by comparing
its stress range distribution against its fatigue capacity. The following definitions are used in the context of
this Guide:
• Stress Range Distribution: Stress range probability density functions calculated per this Guide.
• Fatigue Capacity: S-N data (S-N curves) representing the number of stress cycles at fatigue failure.
• Fatigue Strength: Fatigue life (or damage) calculated per this Guide.
• Fatigue Demand: Design fatigue life.

2 ABS GUIDE FOR SPECTRAL-BASED FATIGUE ANALYSIS FOR VESSELS . 2016

Section 1 Introduction

The main underlying assumptions of the Spectral-based Fatigue Analysis method are:
i) Ocean waves are the source of the fatigue stress range acting on the structural system.
ii) The load and structural analyses are assumed to be linear as required for the frequency domain
formulation and the associated probabilistic analysis to be valid. As such, scaling and the
superposition of stress transfer functions from unit amplitude waves are considered valid.
iii) Non-linearities due to non-linear roll motion and intermittent loads, such as wetting of the side
shell in the splash zone, can be effectively accounted for using correction factors.
iv) Due to their insignificant contributions in typical load cases it is appropriate to disregard structural
dynamic amplification, transient loads and effects such as springing. This allows for the use of
quasi-static finite element analysis.
For the specific SFA method presented in Appendix 2, it is assumed that the short-term stress variation for
a given sea-state is a random, narrow-banded, stationary process. Therefore a Rayleigh distribution can be
used to represent the short-term stress range distribution.
The Spectral-based Fatigue Analysis method is applied to each of the selected structural locations by
implementing the following process:
• Determine the stress range distributions.
• Determine fatigue capacity (S-N data).
• Calculate fatigue strength (fatigue life or damage).
Section 2 provides an overview of the spectral-based fatigue analysis procedure. A schematic representation
of the SFA procedure can be found in Section 2, Figure 1.
An effort is made in this Guide to avoid the discussion of complicated formulae and instead to focus on the
concepts being presented. It is necessary to present the complex formulae used in the calculation of fatigue
damage resulting from the predicted stress range distributions, which are presented in Appendix 2. It should
be noted that the contents of Appendix 2 are intended to serve as an example of a valid SFA method. It is
not necessary that the process be followed exactly; however, any method used should adhere to the same
principles.

ABS GUIDE FOR SPECTRAL-BASED FATIGUE ANALYSIS FOR VESSELS . 2016 3

 Section 2: Spectral-based Fatigue Analysis Procedure Overview

SECTION 2 Spectral-based Fatigue Analysis Procedure

Overview

1 General
This Section provides an overview of the spectral-based fatigue analysis procedure for sea-going vessels.
The section in which a detailed discussion of each step can be found is included in parentheses.
The spectral fatigue analysis for ocean-going vessels covers a wide scope of technical disciplines ranging
from the analysis of ship’s Response Amplitude Operators (RAOs) to the assessment of fatigue damages of
structural details. Four main steps are typically involved:
• Seakeeping Analysis: to determine ship motion and wave pressure RAOs.
• Structural Analysis: to determine stress transfer functions.
• Spectral Analysis: to determine stress range distributions.
• Fatigue Strength Assessment: to determine fatigue life or damage.
The tasks that comprise each step are described below.

1.1 Seakeeping Analysis (Section 3)

1.1.1 Selection of Loading Conditions (Subsection 3/3)
Vessel loading conditions that a vessel is expected to experience during its service life are to be
reviewed and selected. At least two loading conditions are to be analyzed – one representative of
the most probable deepest draft and one representative of the most probable shallowest draft.
1.1.2 Development of Hydrodynamic Models (Subsection 3/5)
A hydrodynamic model is to be created for each loading condition to be evaluated. This Guide is
written with the assumption that the seakeeping analysis will be conducted using a three dimensional
panel method. As such, all discussion of a seakeeping model in this Guide will be with respect to
a three dimensional panel model.
1.1.3 Determination of Ship Motion and Wave Pressure RAOs (Subsection 3/11)
The primary objective of seakeeping analysis is the determination of Response Amplitude Operators,
which are mathematical representations of the vessel responses and load effects to unit amplitude
sinusoidal waves. The ship motion and wave pressure RAOs are to be calculated for each vessel
loading condition.

1.3 Structural Analysis (Section 4)

1.3.1 Development of Structural Models (Subsection 4/7)
A model representing the entire hull structure is to be created in order to determine the stress
transfer functions. Typically, a three dimensional finite element model is used. However, a more
simplified approach such as beam theory may be used instead. This Guide is written with the
assumption that a finite element model will be created.
The global structural finite element model should adequately represent the stiffness and inertial
properties of the vessel. In order to analyze fatigue critical details, the creation of fine mesh models
is typically necessary.

4 ABS GUIDE FOR SPECTRAL-BASED FATIGUE ANALYSIS FOR VESSELS . 2016

Section 2 Spectral-based Fatigue Analysis Procedure Overview

1.3.2 Determination of Tank and Inertial Load RAOs (Subsection 4/3)

RAOs should be calculated for the following motion-induced load effects in addition to ship motions
and wave pressures:
• Internal tank pressures and cargo hold loads due to fluid and cargo inclinations and accelerations,
• Inertial forces on the masses of structural components and, as applicable, significant equipment
items.
1.3.3 Determination of Stress Transfer Functions (Subsection 4/1)
The stress transfer function for each location where the fatigue strength is to be evaluated should
be determined via a suitable structural analysis approach. Finite element analysis (FEA) is a commonly
used approach.
The stress transfer functions are required in order to derive the stress range distributions. Therefore,
structural analysis is a fundamental task of spectral-based fatigue analysis.

1.5 Spectral Analysis (Section 5)

1.5.1 Selection of Wave Spectra and Wave Scatter Diagrams (Subsection 5/3)
Wave spectra and wave scatter diagrams (wave scatter data) are environmental data and need to be
decided prior to performing Spectral-based Fatigue Analysis. They are to be used in conjunction
with the stress transfer functions to determine the stress spectra.
1.5.2 Determination of Stress Range Probability Density Functions (Subsection 5/5)
The stress spectra and spectral moments are calculated according to the selected wave spectra and
scatter diagram. It is customary to apply a wave spreading function to the spectral moment to account
for the fact that most fatigue damage is due to low or moderate seas which tend to be short-crested
and result in kinetic energy spread. The stress zero up-crossing frequencies, bandwidth parameters
and stress range probability density functions are calculated using the stress spectral moments.

1.7 Fatigue Strength Assessment (Section 6)

1.7.1 Selection of S-N Data (Subsection 6/3)
The capacity of a location to resist fatigue damage is characterized by the application of S-N data.
The selection of S-N data depends on the structural details and stresses used.
1.7.2 Calculation of Fatigue Damage and Life (Subsections 6/5 and 6/7)
For a given time span, the short term fatigue damage corresponding to each individual sea state
can be calculated from the stress range distributions, the selected S-N data and the probability of the
sea sate. The long term fatigue damage can be calculated by applying a linear damage summation
technique such as Miner’s Rule to the short term fatigue damages.
In order to determine the suitability of the structure for a design fatigue life, it is necessary to calculate
the total long term fatigue damage for the design fatigue life. One can use the total fatigue damage
and the design fatigue life to calculate the fatigue life of a detail.
1.7.3 Acceptance Criteria (Subsection 6/9) (1 October 2017)
To meet the requirements of the SFA (years) or SFA (years, WWT) notations, the calculated
fatigue life is to be equal to or greater than the design fatigue life.

ABS GUIDE FOR SPECTRAL-BASED FATIGUE ANALYSIS FOR VESSELS . 2016 5

Section 2 Spectral-based Fatigue Analysis Procedure Overview

FIGURE 1
Schematic Spectral-based Fatigue Analysis Procedure

Select Loading Conditions

Subsection 3/3

SEAKEEPING ANALYSIS
(Section 3)
Develop Hydrodynamic Model
Subsection 3/5

Determine Ship Motion and

Wave Pressure RAOs
Subsection 3/11

Develop Structural Model

Subsection 4/7
STRUCTURAL ANALYSIS
For Each Loading Condition

(Section 3)

Determine Tank and

Inertial Load RAOs
Subsection 4/3

Determine Stress
Transfer Functions
Subsection 4/1

SPECTRAL ANALYSIS (Section 5)

For Each Structural Detail

Determine Stress Range Select Wave Spectra and

Probability Density Functions Wave Scatter Diagrams
Subsection 5/5 Subsection 5/3

Calculate Fatigue Damage

Select S-N Data
and Life
Subsection 6/3
Subsections 6/5 and 6/7

Calculate Combined Fatigue Life

Subsection 6/7

Compare Combined Fatigue Life

to Acceptance Criteria
Subsection 6/9

FATIGUE STRENGTH ASSESSMENT (Section 6)

6 ABS GUIDE FOR SPECTRAL-BASED FATIGUE ANALYSIS FOR VESSELS . 2016

 Section 3: Seakeeping Analysis

SECTION 3 Seakeeping Analysis

1 Scope of Seakeeping Analysis

Response Amplitude Operators (RAOs) are mathematical representations of the vessel responses and load
effects to unit amplitude sinusoidal waves. A seakeeping analysis is to be conducted primarily to determine
the ship motion and wave pressure RAOs. This section is referring to motion and load RAOs when
discussing wave induced motions and load components.
A seakeeping analysis is to be performed at 75% of the vessel’s design speed for each wave frequency,
vessel heading angle and selected loading condition to be used in the spectral analysis. Normally, the frequency
range to be used is 0.1 rad/sec to 1.80 rad/sec in increments not larger than 0.1 rad/sec. However, depending
on the characteristics of the response, it may be necessary to consider a different frequency range. The
wave heading range is to be 0 to 360 degrees in increments not larger than 30 degrees.

3 Vessel Loading Conditions

The loading conditions defined in the loading manual relate to the probable variations in loading that the
vessel will experience during its service life. Prior to conducting a seakeeping analysis, appropriate vessel
cargo loading conditions need to be selected. The main parameters for selecting loading conditions are tank
or hold loading and ballast arrangements, draft and trim. These parameters have a direct influence on the
static stress components of the hull’s response, and they also affect the wave-induced variable stress range
experienced at a structural location.
Because of the variability of the loading conditions and their impact on fatigue strength predictions, it is
necessary to consider more than one loading condition in the fatigue analysis. At a minimum, two cases,
the most probable deepest and shallowest drafts that the vessel is expected to experience during its service
life, should be represented and used in the Spectral-based Fatigue Analysis process.

5 Hydrodynamic Models
Boundary element methods generally require that the wetted surface of the vessel be discretized into many
three dimensional panels. The panel mesh should be fine enough to resolve the radiation and diffraction
waves with reasonable accuracy. A general rule is that there are at least four panels within the shortest
wave length being analyzed. The plots of the RAOs versus frequency should be reviewed for any abrupt
changes; they are indicative of an insufficient number of panels to fully capture the wave characteristics.
The hull offset data can be used to facilitate the generation of the panel model. For each loading condition,
the draft at the forward and aft perpendiculars, the location of the center of gravity, the radii of gyration
and the sectional mass distribution along the ship length are to be in accordance with the Loading Manual.
In addition, there should be sufficient compatibility between the hydrodynamic and structural models such
that the mapping of fluid pressures onto the structural model can be done appropriately.

ABS GUIDE FOR SPECTRAL-BASED FATIGUE ANALYSIS FOR VESSELS . 2016 7

Section 3 Seakeeping Analysis

7 Initial Balance Check

The seakeeping analysis should be performed on the hydrodynamic model with respect to static initial
conditions representing the vessel geometry and loading conditions (see Subsection 3/3). The still-water
bending moment (SWBM) and still water shear force (SWSF) distributions should be calculated at a
sufficient number of transverse sections along the hull length such that weight distribution discontinuities
are accounted for accurately. A recognized hydrostatic analysis program should be used to perform these
calculations. The convergence of the displacement, Longitudinal Center of Buoyancy (LCB) and trim of
the hydrodynamic model should be checked to meet the following tolerances:
• Displacement: ±1%
• Trim: ±0.5 degrees
• Draft:
- Forward ±1 cm
- Mean ±1 cm
- Aft ±1 cm
• LCB: ±0.1% of length
• SWBM: ±5%
Additionally, the longitudinal locations of the maximum and the minimum SWBMs and SWSFs and, if
appropriate, those of zero SWBM and zero SWSF should be checked to ensure proper distribution of the
SWBM along the vessel’s length.

9 Seakeeping Analysis Methods

Computations of the wave-induced motions and loads should be carried out using appropriate, proven methods.
Preference should be given to the application of seakeeping analysis codes utilizing three-dimensional,
potential flow-based diffraction-radiation theory. These codes, based on linear wave and motion assumptions,
make use of boundary element methods with constant or higher order sink-source panels over the entire wetted
surface of the hull on which the hydrodynamic pressures are computed. The analysis should account for
rigid body motions in all six degrees-of-freedom.
For the load component types and structural responses of primary interest, analysis software formulations
derived from linear idealizations are deemed to be sufficient.

11 Effects of Ship Motions and Wave Pressures

11.1 Inertial and Gravitational Loads due to Ship Motions

Wave-induced loads on a buoyant structure are complex. In addition to direct forces (e.g., wave pressures
on the external surface of the hull), there are indirect force components produced by the rigid body motions
of the vessel. The motions result in both inertial forces and gravitational loads due to the vessel’s inclinations
(the rotational components of motion). The inclination induced gravitational loads are in addition to the
pure static (still water) gravitational loads on the hull structure and are referred to as quasi-static loads in
this Guide. Therefore, the total loads in general are composed of static and dynamic components, and the
dynamic component can be further decomposed into quasi-static and inertial components. These load
components are discussed in detail in Section 4.

11.3 Real and Imaginary Parts

The treatment of ship motions and various loads is typically done through the use of their real and imaginary
parts which are employed separately in structural analyses. In a physical sense, the real and imaginary
parts correspond to two wave systems that are 90 degrees out of phase.

8 ABS GUIDE FOR SPECTRAL-BASED FATIGUE ANALYSIS FOR VESSELS . 2016

Section 3 Seakeeping Analysis

11.5 Pressure Mapping

The wave pressures on a hydrodynamic panel model need to be mapped to its corresponding structural model
for structural analysis. The mesh densities of the hydrodynamic panel model and the structural model are
typically different. Consequently, it is usually necessary to interpolate the pressure distribution (via three
dimensional linear interpolation) when transferring it from the coarser hydrodynamic model to the finer
structural model.

11.7 Intermittent Wetting

Seakeeping analysis based on linear theory will not predict the non-linear effects due to intermittent wetting
near the mean waterline. In actual service, this phenomenon is manifested by a reduction in the number of
fatigue cracks at side shell plating stiffeners located near the mean waterline compared to those about four
or five bays below. To take into account the pressure reduction near the mean waterline due to this non-
linearity, the following reduction factor can be applied to the pressure RAOs:
RF = 0.5[1.0 + tanh(0.35d)]
where d is depth, in meters, of the field point below the still-water line.
In order to correctly account for the intermittent wetting effects, it is recommended that the vertical height
of the panels near the waterline be no greater than two times the longitudinal spacing. If the panels are too
large, the pressure reduction factor can be applied to the structural model after the wave pressures are
mapped from the hydrodynamic model.

ABS GUIDE FOR SPECTRAL-BASED FATIGUE ANALYSIS FOR VESSELS . 2016 9

 Section 4: Structural Analysis

SECTION 4 Structural Analysis

1 General (1 October 2017)

The structural analysis procedure described in this Section is based on the assumption that the structural
analysis is performed through a typical finite element analysis. Subsections 4/3 and 4/5 address the loading of
the structural analysis finite element model. Finite element modeling guidance is provided in Subsection 4/7.
Motion and load components are presented in ship fixed coordinate system in this section. The origin is
located at the intersection of the aft perpendicular, centerline and baseline. The positive sign convention is
longitudinal forward, transverse port and vertical upwards (See Section 4, Figure 1).

FIGURE 1
Ship Coordinate System

A structural analysis is to be conducted to calculate the stress transfer functions necessary for spectral-
based fatigue analysis. The stress transfer functions are used in the spectral response analysis to calculate
the stress range distributions as discussed in Section 5.
The wave induced motion and load components described in this Section are RAOs, and the stress responses
are stress transfer functions. The following loads are to be calculated via a structural analysis: liquid tank
pressures (4/3.1), dry bulk cargo loads (4/3.3) and container loads (4/3.5). In each Paragraph, the static,
quasi-static, and inertial loads are discussed. In addition, the inertial loads of discrete masses, which are
applicable to significant items of equipment, are discussed (4/3.7).
The effect of sloshing loads due to fluid in partially filled tanks is not within the scope of the SFA (years)
or SFA (years, WWT) classification notations. However, the designer is encouraged to perform and
submit such calculations, if deemed important.
There may be situations where partial models of the structural system are used. In such instances, hull
girder shear and axial forces, and bending and torsional moments should be determined to appropriately
represent the boundary conditions at the ends of the partial model.
The general approach used in the calculation methods described below is to calculate the total stress response
considering both static and dynamic loads. Subsequently, the static stress is deducted from the total,
leaving the pure wave-induced stress response. Alternative methods and formulations that directly produce
the dynamic stress response may also be used.

10 ABS GUIDE FOR SPECTRAL-BASED FATIGUE ANALYSIS FOR VESSELS . 2016

Section 4 Structural Analysis

3 Internal Load
The procedures to calculate liquid tank pressure, bulk cargo pressure, container loads, and loads on discrete
masses are presented in this subsection. Each of these loads is composed of static and dynamic components.
The static load component results from gravity. The dynamic load component can be further decomposed
into quasi-static and inertial components. The quasi-static load component results from gravity, considering
the roll and pitch inclinations of the vessel. The direction of gravitational forces in the ship-fixed coordinate
system varies with roll and pitch motions, resulting in the quasi-static loads. The inertial load component
results from the local accelerations caused by the ship motions in six degrees-of-freedom. It is assumed
that the ship motions are calculated in the ship-fixed coordinate system. Therefore, coordinate transformations
are not needed for the local accelerations.
The following definitions are used in this Subsection:
θ = roll angle, positive starboard down
φ = pitch angle, positive bow down
gV = acceleration of gravity. It can be considered as the vertical component of gravitational
acceleration relative to the ship-fixed coordinate system to the accuracy of the first
order.
gV ≈ g
gT, gL = transverse and longitudinal components of gravitational acceleration relative to the
ship-fixed coordinate system due to roll and pitch inclinations
= (–g sin φ, g sin θ)
aV, aT, aL = vertical, transverse and longitudinal components of local accelerations caused by ship
motions relative to the ship-fixed coordinate system at a calculation point
  
= A + Θ× R

A = surge, sway and heave acceleration vector at vessel’s center of gravity

Θ = roll, pitch and yaw acceleration vector

R = distance vector from the vessel’s center of gravity to the calculation point

3.1 Liquid Tank Pressures

As appropriate, the liquid pressure on liquid tank (or liquid cargo hold) boundaries should be considered in
the analysis. The pressures acting on the internal surfaces of liquid tanks are to be calculated and applied
to the structural model. Static and dynamic pressures on completely filled and/or partially filled tanks are
to be considered in the analysis. Tank sloshing loads are not included in the analysis. These sloshing loads
are to be treated in accordance with the current Rule requirements.
The liquid tank pressure for each of the tank boundary points can be calculated from the following equation,
which is expressed as a combination of the static and dynamic pressure components:
p = po + ρ hi [(g + aV)2 + (gT + aT)2 + (gL + aL)2 ]1/2
where
p = liquid tank pressure at a tank boundary point
po = the vapor pressure or the pressure setting on the pressure/vacuum relief valve

ρ = density of tank liquid

ABS GUIDE FOR SPECTRAL-BASED FATIGUE ANALYSIS FOR VESSELS . 2016 11

Section 4 Structural Analysis

hi = internal pressure head defined by the height of projected liquid column in the direction
of a resultant acceleration vector. For a completely filled tank, the pressure head is to
be measured from the highest point of the tank to the load point (see Section 4, Figure 2).
For a partially filled tank, the pressure head is to be measured from the free surface
level to the load point (see Section 4, Figure 3). The free surface is defined as the liquid
surface normal to the resultant acceleration vector. In the aforementioned figures,
only vertical and transverse accelerations are indicated.
aV, aT, aL = local accelerations at the center of gravity (CG) of tank contents
The liquid tank pressure p in the above formula becomes the static pressure if aV, aT, aL, gT and gL are all
set to zero, and hi is calculated in the direction of g.

FIGURE 2
Liquid Pressure on a Completely Filled Tank

z z
y aV y
θe θe

aT
hi
hi
hi
hi

FIGURE 3
Liquid Pressure on a Partially Filled Tank

z aV z
θe
y y
aT

hi
hi
hi hi hi

3.3 Bulk Cargo Pressure

A loaded bulk cargo hold can be partially or completely filled. For light bulk cargo, the bulk cargo may be
loaded up to the top of the hatch coaming. For heavy bulk cargo, unless otherwise specified, the top surface
of bulk cargo may be considered to have a slope equal to half the angle of repose at sides and have a flat
surface with half the cargo hold beam width.
The bulk cargo pressures acting on the internal surfaces of cargo holds are to be calculated and applied to
the structural model for analysis. Static and dynamic bulk cargo pressures should be included in the analysis
assuming there is no relative motion between the cargo hold and contained bulk cargo.

12 ABS GUIDE FOR SPECTRAL-BASED FATIGUE ANALYSIS FOR VESSELS . 2016

Section 4 Structural Analysis

The following definitions are used in the formulas for the static and dynamic load calculations in this Paragraph.
αo = angle of repose for the bulk cargo considered
= 30 deg. in general, 35 deg. for iron ore, 25 deg. for cement
ρ = density of the bulk cargo
aV, aT, aL = local accelerations at the CG of cargo contents

n = surface unit normal vector in the ship coordinate system, towards the inside of the
cargo hold
  
= nx i + ny j + nz k
   
a = (aL + gL) i +(aT + gT) j + (aV + g) k

a = magnitude of a

= |a |
 
α = the angle between the surface unit normal vector, n , and the acceleration vector, a
(see Section 4, Figure 4)
 
 a⋅n 
= cos −1    
 | a || n | 
Both static and dynamic bulk cargo pressures can be decomposed into normal and tangential components
relative to the surface of cargo hold. The positive normal component is defined in the opposite direction of
   
n , and the positive tangential component is defined in the direction of n × ( n × a ).

FIGURE 4
 
Depiction of Surface Normal Vector, n , Acceleration Vector, a , and α

n
α

The normal and tangential components of bulk cargo pressure can be calculated using the following equations,
each of which is expressed as a combination of the static and dynamic pressure components:
pn = ρ hTa{cos2α + (1 − sin αo) sin2α} 0° ≤ α ≤ 90°
pn = 0 otherwise

pt = ρ hTa{sin αo sin α cos α} 0° ≤ α ≤ 90°

pt = 0 otherwise

ABS GUIDE FOR SPECTRAL-BASED FATIGUE ANALYSIS FOR VESSELS . 2016 13

Section 4 Structural Analysis

where
pn = pressure component normal to the hold surface
pt = pressure component tangential to the hold surface
hT = bulk cargo pressure head defined by the height of projected bulk cargo column in the
direction of the acceleration vector. The pressure head is to be measured from the top
of the cargo surface to the load point.
The total pressure components pn and pt in the above formulae become the static pressure components if
aV, aT , aL , gT and gL are all set to zero, and hT is calculated in the direction of g.

3.5 Container Loads

The container loads acting on the cargo holds and deck are to be calculated and applied to the structural
model for analysis. Static and dynamic container loads should be included in the analysis assuming that
there is no relative motion between the hull and the containers.
3.5.1 Static Load
The static container load due to gravity acting on the cargo hold bottom or on deck can be expressed as:
FS = mg
where
m = mass of the container.
The static load due to a stack of containers may be summed and applied to appropriate nodes on
the cargo hold bottom plate. Total vertical load due to the containers on the deck may be applied
to the appropriate nodes on the hatch coaming top plates.
3.5.2 Dynamic Load
The vertical component (see Section 4, Figure 5) of dynamic container load can be calculated using
the following equation:
FV = m a V
where
m = mass of the container.
aV = local vertical acceleration at the CG of a container.
The transverse component (see Section 4, Figure 5) of dynamic container load can be calculated
using the following equation which is expressed as a combination of the quasi-static and inertial
load components:
FT = m (gT + aT)
where
aT = local transverse acceleration at the CG of a container.
The transverse load due to containers may be distributed to appropriate nodes on the bulkhead
structure via the container cell guide. The total transverse load due to the containers on deck may be
applied to the appropriate nodes on the hatch coaming top plates via the container lashing system.
The longitudinal component of dynamic container load can be calculated using the following
equation which is expressed as a combination of the quasi-static and inertial load components:
FL = m (gL + aL)
where
aL = local longitudinal acceleration at the CG of a container.

14 ABS GUIDE FOR SPECTRAL-BASED FATIGUE ANALYSIS FOR VESSELS . 2016

Section 4 Structural Analysis

FIGURE 5
Vertical and Transverse Container Load Components

∇ FT ∇

FV

aV

aT

3.7 Loads on Lightship Structure and Equipment

The static and dynamic loads acting on the lightship structure and equipment are to be calculated and applied
to the structural model for the structural analysis.
The static load acting on a node of the lightship structure and equipment can be expressed as:
FS = mg
where
m = point mass of the structural member or equipment
The dynamic components of the load acting on the node in vertical, transverse and longitudinal directions can
be calculated from the following equations:
FV = m aV
FT = m (gT + aT)
FL = m (gL + aL)
where
aV, aT, aL = local accelerations at the mass point
The formulas for FT and FL above represent a combination of the quasi-static and inertial load components.

5 Load Cases and Model Equilibrium

5.1 Number of Load Cases

A load case in the spectral-based fatigue analysis corresponds to a combination of a loading condition, a vessel
heading angle, a wave frequency, and the real or imaginary component of the wave. The structural analysis is to
be carried out at each wave frequency and vessel heading angle for each considered loading condition.
The stress transfer function for each location of fatigue strength evaluation needs to be calculated for each
load case. It is likely that the number of analysis load cases will be very large. The recommended parameters
for a spectral-based fatigue analysis for each considered loading condition are 18 wave frequencies (0.1 to
1.80 rad/s in increments of 0.10 rad/s), 12 wave headings (0 to 360 degrees in 30 degree increments) and
the real and imaginary wave components. In the case of a standard analysis, the total number of load cases
per loading condition will be (18 · 12 · 2) = 432. In addition, a static load case may be needed. Provided
that the shape of each RAO and its local maxima are adequately captured, larger frequency steps may be
used to reduce the number of analysis load cases.

ABS GUIDE FOR SPECTRAL-BASED FATIGUE ANALYSIS FOR VESSELS . 2016 15

Section 4 Structural Analysis

5.3 Equilibrium Check

The applied hydrodynamic external pressure should be in equilibrium with the other applied loads on a full
length ship structural model. For each load case, the forces and moments in each global direction should
be summed to calculate the force and moment imbalances. A suitable load balancing scheme should be
applied to the structural model to balance the unbalanced forces and moments prior to conducting the
structural analysis.

7 Structural Modeling and Analysis

A three dimensional (3-D) global FE model representing the entire hull structure is to be created in order to
determine the stress transfer functions. While the global FE model analysis may produce results of
sufficient accuracy, it is typically necessary to perform fine mesh FE analyses of local areas. The FE
models to be used for the local fine mesh analyses can be created either by refining a region of the global
FE model or by creating a separate fine mesh FE model of the local area and applying boundary conditions
determined from the global FE analysis. The load cases discussed in Subsection 4/5 are to be used for both
global and fine mesh FE analysis.
Reference should be made to additional ABS guidance on modeling and analysis techniques for vessel
structure (e.g., the ABS Guidance Notes on SafeHull Finite Element Analysis for Hull Structures). While
the partial hull model used for a SafeHull analysis is different from the full ship model used for an SFA
analysis, the modeling practices and considerations presented are applicable. Detailed modeling guidance
relating to items such as element types, mesh sizes, dependence between local and global models, etc., are
also presented in the ABS Guidance Notes on SafeHull Finite Element Analysis for Hull Structures.

7.1 3-D Global Analysis Modeling

The global structural finite element model should be as detailed and complete as practicable. Gross
scantlings are ordinarily used for the Spectral-based Fatigue Analysis of a new-build vessel.
While creating the model, a sufficient number of nodes, elements and degrees of freedom are to be used to
adequately represent the stiffness and inertial properties of the vessel.
Three types of elements are typically used to approximate the geometry, configuration, and stiffness of the
vessel structure:
i) Truss or rod elements with axial stiffness only.
ii) Beam elements with axial, shear and bending stiffness.
iii) Membrane and bending plate elements, either triangular or quadrilateral.

7.3 Analyses of Local Structure

Fine mesh finite element models should be created in order to obtain more refined stress distributions of
local structure. When creating the fine mesh models, care is to be taken to ensure that the geometry and
stiffness of the structure is accurately represented. If a separate model is created, displacements extracted
from the global 3-D FE analysis are to be used as boundary conditions. In addition to the boundary
constraints, all pertinent local loads should be applied to the fine mesh models.

7.5 Hot Spot Stress Concentration

The Hot Spot Stress Approach, which is often used to characterize the fatigue strength of details such as
the toe of a fillet weld, requires the use of a regimented procedure. The two major parts of the procedure
are:
• The selection of an S-N data class (see Subsection 6/3 and Appendix 1) that applies for each instance.
• Creating the fine mesh FEM adjacent to the detail and calculating the stress at the hot spot location via
extrapolation of the stress distribution.

16 ABS GUIDE FOR SPECTRAL-BASED FATIGUE ANALYSIS FOR VESSELS . 2016

Section 4 Structural Analysis

Section 4, Figure 6 depicts an acceptable method that can be used to calculate the hot spot stress for a weld
toe. Element sizes near the detail of interest are to be approximately equal to the plating thickness. The
hot spot stress is found via linear extrapolation of the calculated stress results at distances of t/2 and 3t/2
from the detail. For stresses obtained using this approach, the use of E class S-N data (see Appendix 1) is
considered to be most appropriate. The surface stresses (considering a “bending plate” element type) should
be used to determine the hot spot stress. A detailed description of the numerical extrapolation procedure
can be found in 5C-1-A1/13.7 of the Steel Vessel Rules.

FIGURE 6
Definition of Hot Spot Stress
Peak Stress

Weld Toe Hot Spot Stress

t Weld Toe

≈t
Weld Toe Location

t/2

3t/2

ABS GUIDE FOR SPECTRAL-BASED FATIGUE ANALYSIS FOR VESSELS . 2016 17

 Section 5: Spectral Analysis

SECTION 5 Spectral Analysis

1 General
The spectral analysis is used to calculate the stress range distributions using the stress transfer functions
calculated via structural analysis and the selected wave spectra and a wave scatter diagram. Mathematically,
Spectral-based Fatigue Analysis begins after the determination of the stress transfer functions. A detailed
presentation of an example method is presented in Appendix 2.

3 Wave Spectra and Wave Scatter Diagram (1 October 2017)

Wave data are represented in terms of wave energy spectra and wave scatter diagrams.
The Bretschneider or two parameter Pierson-Moskowitz wave spectrum is recommended for the North Atlantic,
described by the following expression:

H2  2
4
 5  1  2  4 
4

S() = s    exp     
4  Tz     Tz  
 

where
Hs = significant wave height, in m

 = wave frequency, in rad/s

Tz = average zero up-crossing wave period, in seconds
Spectral-based fatigue analysis typically uses environmental data for ocean waves that are given in a wave
scatter diagram format. A wave scatter diagram consists of cells, or sea-states, which are short-term descriptions
of the sea in terms of joint probability of occurrence of a significant wave height, Hs, and a characteristic
period. Each cell contains three data items:
i) The significant wave height, Hs, (typically in meters),
ii) The characteristic wave period (in seconds),
iii) The number of occurrences for the sea state.
Section 5, Tables 1 and 2 present the wave scatter diagram data that is to be used in the spectral-based
fatigue analysis of a vessel classed for unrestricted service and for world-wide trading service, respectively. It
can be assumed that there is an equal probability of each vessel heading relative to the direction of the waves.

5 Stress Range Distributions

The calculation of the stress transfer functions is done via a structural analysis as presented in Section 4.
The stress amplitude can be determined by selecting two points on the sinusoidal stress time series of
frequency  with a 90-degree phase shift.
Once the stress transfer functions are obtained via a structural analysis and the wave spectra are selected
for each of the cells in the wave scatter diagram, the stress spectra and spectral moments can be calculated.
With the available stress spectra, the Rayleigh probability density function (pdf) describing the short term
stress-range distribution, the zero up-crossing frequency of the stress response, and the stress bandwidth
parameter used in calculating rainflow correction can be calculated for each sea state in the wave scatter
diagram. The aforementioned terms are defined in Appendix 2 which presents a detailed example of a closed
form spectral-based fatigue analysis procedure. The stress range is considered to be twice the stress amplitude.

18 ABS GUIDE FOR SPECTRAL-BASED FATIGUE ANALYSIS FOR VESSELS . 2016

Section 5 Spectral Analysis

TABLE 1
ABS Wave Scatter Diagram for Unrestricted Service Classification
(Representing the North Atlantic)
* Wave heights taken as significant values, Hs
** Wave periods taken as zero crossing values, Tz
Wave Period (sec)** Sum Over
3.50 4.50 5.50 6.50 7.50 8.50 9.50 10.50 11.50 12.50 13.50 All Periods
0.5 8 260 1344 2149 1349 413 76 10 1 5610
1.5 55 1223 5349 7569 4788 1698 397 69 9 1 21158
2.5 9 406 3245 7844 7977 4305 1458 351 65 10 25670
3.5 2 113 1332 4599 6488 4716 2092 642 149 28 20161
4.5 30 469 2101 3779 3439 1876 696 192 43 12625
Wave Height (m)*

5.5 8 156 858 1867 2030 1307 564 180 46 7016

6.5 2 52 336 856 1077 795 390 140 40 3688
7.5 1 18 132 383 545 452 247 98 30 1906
8.5 6 53 172 272 250 150 65 22 990
9.5 2 22 78 136 137 90 42 15 522
10.5 1 9 37 70 76 53 26 10 282
11.5 4 18 36 42 32 17 7 156
12.5 2 9 19 24 19 11 4 88
13.5 1 4 10 14 12 7 3 51
>14.5 1 5 13 19 19 13 7 77
Sum over 8 326 3127 12779 24880 26874 18442 8949 3335 1014 266 100000
All Heights

TABLE 2
ABS Wave Scatter Diagram for World-Wide Trading Service
Classification (1 October 2017)
* Wave heights taken as significant values, Hs
** Wave periods taken as zero crossing values, Tz
Wave Period (sec)** Sum Over
3.50 4.50 5.50 6.50 7.50 8.50 9.50 10.50 11.50 12.50 13.50 All Periods
0.5 1888 5417 5836 3392 1260 313 55 8 1 18170
1.5 493 3739 8374 9339 6517 2985 940 219 41 6 1 32654
2.5 96 1128 3682 5934 6159 4216 1930 629 158 33 6 23971
3.5 18 294 1217 2439 3269 2959 1769 734 229 57 12 12997
4.5 4 79 388 893 1417 1568 1142 568 208 60 15 6342
Wave Height (m)*

5.5 1 23 130 328 581 745 631 361 150 49 13 3012

6.5 8 47 126 239 341 325 208 96 34 10 1434
7.5 2 18 51 101 156 163 115 58 22 7 693
8.5 1 7 21 44 72 82 63 34 14 5 343
9.5 3 10 20 34 42 34 20 9 3 175
10.5 1 5 10 17 22 19 11 5 2 92
11.5 1 2 5 9 12 10 7 3 1 50
12.5 1 3 5 6 6 4 2 1 28
13.5 1 1 2 3 4 3 1 1 16
>14.5 1 2 3 5 5 4 2 1 23
Sum over 2500 10691 19704 22543 19628 13425 7127 2983 1024 297 78 100000
All Heights

ABS GUIDE FOR SPECTRAL-BASED FATIGUE ANALYSIS FOR VESSELS . 2016 19

 Section 6: Fatigue Strength Assessment

SECTION 6 Fatigue Strength Assessment

1 General
Section 4 of this Guide addressed calculating the stress transfer functions for locations in the structure for
which the adequacy of fatigue strength is to be evaluated. Appendix 2 presents a spectral-based analysis
procedure that may be used for the calculation of the stress range distributions. The capacity of a location
to resist fatigue damage is characterized by the use of S-N Data, which are described below. Refer to
Appendix 1 of this Guide and Part 5C of the Steel Vessel Rules concerning the S-N Data recommended by
ABS.
S-N data is typically used in two ways. In the first, the nominal stress approach, it is assumed that the acting
variable stress range can be adequately determined from the nominal stress distribution (which may include
geometric stress concentration effects) in the area surrounding the location for which the fatigue life is to
be evaluated. The second method, the hot spot approach (see 4/7.3), is used for locations in which complicated
geometry or relatively steep local stress gradients invalidate the use of the nominal stress approach.
Reference should be made to Part 5C of the Steel Vessel Rules for further explanation and application of
these two approaches and for guidance on the categorization of structural details into the various S-N data
classes.

3 S-N Data (15 February 2017)

To provide a ready reference, the S-N Data recommended by ABS are given in Appendix 1 of this Guide.
(Note: source United Kingdom’s Dept of Energy (HSE) Guidance Notes, 4th Edition.)
It may be necessary to make adjustments to the S-N data to reduce the fatigue capacity of the detail in
some instances. Conversely, if the detail has characteristics beneficial to fatigue performance, the S-N data
may be adjusted to increase its fatigue capacity.
A reduction in fatigue capacity may be required to account for factors such as the absence of corrosion
protection (coating) for structural steel or relatively large plate thickness. The reduction is to be done in
accordance with standard ABS practice. The fatigue capacity of a detail may be increased to account for
conditions such as compressive mean stress effects – a high compressive portion of the acting variable
stress range – or the use of weld improvement techniques.
The use of a weld improvement technique such as weld toe grinding or peening to relieve ambient residual
stress can be effective in increasing fatigue life. However, such effects should not be considered in the
design of the structure. Typically, an increase in fatigue capacity through weld improvement techniques
will be considered only for situations arising during construction, operation or future reconditioning of the
structure. If the design fatigue life cannot be obtained using preferred design measures such as layout
refinement, geometric changes, scantling adjustment or weld profile modification to reduce fatigue damage,
an exception may be granted.
Weld improvement techniques such as weld toe grinding or ultrasonic peening may be used to improve
fatigue life only if the calculated fatigue life is sufficient when the weld improvement techniques are not
considered. The required calculated fatigue life without considering weld improvements for various design
fatigue lives are listed in the table below:

20 ABS GUIDE FOR SPECTRAL-BASED FATIGUE ANALYSIS FOR VESSELS . 2016

Section 6 Fatigue Strength Assessment

Calculated Fatigue Life without Considering

Design Fatigue Life
Weld Improvements
20 years 15 years
25 years 17 years
Greater than 25 years Greater than 2/3 design fatigue life

Where an improvement technique is applied, full details of the technique are to be submitted for review
along with supporting calculations indicating the proposed fatigue life improvement factor. If grinding is
used, the full details of the grinding standard, including the extents, profile smoothness particulars, final
weld profile, improved workmanship technique and quality acceptance criteria are to be clearly indicated
on the applicable drawings.
Grinding is to produce a smooth concave profile at the weld toe with a penetration depth of at least 0.5 mm
into the plate surface below the bottom of any visible undercut. It is preferable that a rotary burr be used
and that the grinding extend below the plate surface in order to remove toe defects. The ground area is to
have effective corrosion protection. The groove depth is to be minimized and generally should not exceed
1 mm. In no circumstances may the grinding depth exceed the lesser of 2 mm or 7% of the plate gross
thickness. The grinding area is to extend well beyond the high stress region.
A weld surface treated using ultrasonic peening should have a smooth finished shape and all traces of the
weld toe are to be removed. The minimum peening depth is to be 0.2 mm below the original surface.
Generally, the maximum peening depth is 0.5 mm.
A maximum fatigue life improvement of 2 times may be granted provided the above recommendations are
followed.

5 Fatigue Strength
An appropriate method is to be employed to calculate the fatigue damage resulting from each considered
sea state. The damage resulting from individual sea states is referred to as short-term. The total fatigue
damage resulting from combining the damage from each of the short-term conditions can be calculated
using a weighted linear summation technique (e.g., Miner’s Rule).
Appendix 2 contains a detailed description of the steps involved in a suggested Spectral-based Fatigue Analysis
method that follows the basic elements mentioned above. ABS should be provided with background and
verification information that demonstrates the suitability of the analytical method employed.

7 Combined Fatigue Life from Multiple Loading Conditions

To calculate combined fatigue life from multiple loading conditions, exposure time factors need to be specified
for each individual loading condition. In addition, for conventional trading vessels a factor of 0.85 is to be
applied to account for non-sailing time for operations such as loading and unloading, repairs, etc. For
vessels engaged in a service with less sailing time, special consideration may be given to accepting a lower
factor subject to Owner agreement.
The fatigue life for each loading condition is to be calculated separately. If the exposure time ratios have
been considered in the fatigue life calculations for each loading condition, the combined fatigue life is
given by:
1
LC =
1 1 1 
αs  + ++ 
 L1 L2 Ln 

where
LC = combined fatigue life

αs = factor of 0.85 to account for non-sailing time

ABS GUIDE FOR SPECTRAL-BASED FATIGUE ANALYSIS FOR VESSELS . 2016 21

Section 6 Fatigue Strength Assessment

Li = fatigue life for the ith loading condition (i = 1 to n) prior to accounting for non-sailing
time
n = number of loading conditions.
If the exposure time ratios have not been considered in the fatigue life calculations for each loading condition,
the combined fatigue life is given by:
1
LC =
 p1 p 2 p 
αs  + ++ n 
 L1 L2 Ln 

where
pi = exposure time ratio of the ith loading condition (i =1 to n) satisfying:
n

∑p
i =1
i = 1.0

9 Acceptance Criteria (1 October 2017)

The minimum required fatigue life for a spectral-based approach is customarily stated as either a damage
ratio (D) or a design fatigue life (L). The latter is employed in this Guide. To meet the requirements of the
SFA (years) or SFA (years, WWT) notations, the calculated fatigue life is to be equal to or greater than
the design fatigue life.

22 ABS GUIDE FOR SPECTRAL-BASED FATIGUE ANALYSIS FOR VESSELS . 2016

 Appendix 1: Basic Design S-N Curves

APPENDIX 1 Basic Design S-N Curves

FIGURE 1
S-N Curves

The S-N Curves are represented by the following equation:

Sm N = A
where
S = stress range
N = number of cycles to failure
A, m = parameters representing the intercept and inverse slope of the upper (left) portion of
the S-N Curve. These change at N = 107 cycles to C and r, respectively. Values of
these parameters are given in the following table.

ABS GUIDE FOR SPECTRAL-BASED FATIGUE ANALYSIS FOR VESSELS . 2016 23

Appendix 1 Basic Design S-N Curves

TABLE 1
Parameters for Basic S-N Design Curves
N ≤ 107 N > 107
A m C r
Class (For MPa units) (For MPa units)
B 1.013 × 1015 4 1.020 × 1019 6
C 4.227 × 1013 3.5 2.584 × 1017 5.5
D 1.519 × 1012 3 4.239 × 1015 5
E 1.035 × 1012 3 2.300 × 1015 5
F 6.315 × 1011 3 9.975 × 1014 5
F2 4.307 × 1011 3 5.278 × 1014 5
G 2.477 × 1011 3 2.138 × 1014 5
W 1.574 × 1011 3 1.016 × 1014 5

Refer to Part 5C of the Steel Vessel Rules for guidance on the categorization of structural details
into the indicated classes.
Notes for Application of Classes:
Class B: Parent material with automatic flame-cut edges ground to remove flame cutting drag line.
Class C: Parent material with automatic flame-cut edges and full penetration butt welds ground flush in way of
hatch corners in container carriers or similar deck areas in other vessel types.
Class D: Full penetration butt welds in way of hatch corners in container carriers or similar deck areas in other
vessel types.

24 ABS GUIDE FOR SPECTRAL-BASED FATIGUE ANALYSIS FOR VESSELS . 2016

 App en di x 2 : Outlin e of a Clo s ed F or m S pe c tr al- bas ed F atig ue An al ysi s Pr oce du r e

APPENDIX 2 Outline of a Closed Form Spectral-based Fatigue

Analysis Procedure
Notes:
(1) (1 October 2017) This Appendix is referred to in Section 6. It is provided to describe the formulations comprising
a Spectral-based Fatigue Analysis approach, which can be employed to satisfy the criteria to obtain the SFA (years) or
SFA (years, WWT) Classification notations. However, it is at this formulation level that there are multiple valid
methods that may be selected. For that reason, it is emphasized that the contents of this Appendix are provided
primarily to illustrate principle rather than to give mandatory steps for the Spectral-based Fatigue method.
(2) The procedure described below considers the use of a wave scatter diagram (e.g., Section 5, Table 1). Where a
different base period for the wave scatter diagram is employed, the procedure must be suitably modified.

1 General
In the “short-term closed form” approach described below, the stress range is normally expressed in terms
of probability density functions for different short-term intervals corresponding to the individual cells or
bins of the wave scatter diagram. These short-term probability density functions are derived by a spectral
approach based on the Rayleigh distribution method, whereby, it is assumed that the variation of stress is a
narrow-banded random Gaussian process. To take into account effects of swell, which are not accounted
for when the wave environment is represented by the scatter diagram, Wirsching’s “rainflow correction”
factor is applied in the calculation of short-term fatigue damage. Having calculated the short-term damage,
the total fatigue damage is calculated through their weighted linear summation (using Miner’s rule).
Mathematical representations of the steps of the Spectral-based Fatigue Analysis approach just described
are given next.

3 Key Steps in Closed Form Damage Calculation

1. Determine the complex stress transfer function, H (|), at a structural location of interest for a
particular loading condition. This is done in a direct manner where structural analyses are performed
for the specified ranges of wave frequencies and headings, and the resulting stresses are used to
explicitly generate the stress transfer function.
2. Generate a stress energy spectrum, S (|Hs, Tz, ), by scaling the wave energy spectrum S (|Hs,
Tz) in the following manner:
S(|Hs, Tz, ) = |H (|)|2 S(|Hs, Tz).......................................................................... (1)
3. Calculate the spectral moments at the same ship forward speed, V, used in the seakeeping analysis.
The nth spectral moment, mn, is calculated as follows:



mn  (  V 2 cos θ / g ) n S ( | H s , Tz , θ)d .............................................................. (2)
0

Most fatigue damage is associated with low or moderate seas, hence, confused short-crested sea
conditions must be allowed. Confused short-crested seas result in a kinetic energy spread, which
is modeled using the cosine-squared approach, (2/) cos2. Generally, cosine-squared spreading is
assumed from +90 to –90 degrees on either side of the selected wave heading (refer to Appendix 2,
Figure 1). Applying the wave spreading function modifies the spectral moment as follows:
θ  90
2  
  cos   θ    (  V cos  / g ) S ( | H s , Tz ,  )d d .................. (3)
 
2 2 n
mn 
   
θ 90 0 

ABS GUIDE FOR SPECTRAL-BASED FATIGUE ANALYSIS FOR VESSELS . 2016 25

Appendix 2 Outline of a Closed-form Spectral-based Fatigue Analysis Procedure

FIGURE 1
Spreading Angles Definition
0

ve
Wa
θ
α

θ - 90°

θ + 90°

4. Using the spectral moments, the Rayleigh probability density function (pdf) describing the short term
stress-range distribution, the zero up-crossing frequency of the stress response and the bandwidth
parameter used in calculating Wirsching’s “rainflow correction” are calculated as follows:
Rayleigh pdf:

s   s 
2
g (s) = exp −    ......................................................................................... (4)
4σ 2   2 2σ  
Zero-up crossing frequency, in Hz:

1 m2
f = ................................................................................................................... (5)
2π m0

Bandwidth Parameter:

m 22
ε = 1− ................................................................................................................ (6)
m0 m 4

where
s = stress range (twice the stress amplitude)

σ = m0

m0, m2, m4 = spectral moments

26 ABS GUIDE FOR SPECTRAL-BASED FATIGUE ANALYSIS FOR VESSELS . 2016

Appendix 2 Outline of a Closed-form Spectral-based Fatigue Analysis Procedure

5. Calculate cumulative fatigue damage based on Palmgren-Miner’s rule, which assumes that the
cumulative fatigue damage (D) inflicted by a group of variable amplitude stress cycles is the sum
of the damage inflicted by each stress range (di), independent of the sequence in which the stress
cycles occur:
J J

∑ ∑N
ni
D= di = ......................................................................................................... (7)
i =1 i =1 i

where
ni = number of stress cycles of a particular stress range
Ni = average number of loading cycles to failure under constant amplitude loading
at that stress range according to the relevant S-N curve
J = number of considered stress range intervals
Failure is predicted to occur when the cumulative damage (D) over J exceeds a critical value equal
to unity.
The short term damage incurred in the i-th sea-state, assuming an S-N curve of the form N = AS-m,
is given by:
∞
T 
Di =   ∫ (kt kms kh s ) m f 0i pi g ( s )i ds ........................................................................... (8)
 A 0
where
Di = damage incurred in the i-th sea-state
kh = a factor for high tensile steel, applicable to parent material only
= 1.000 for mild steel or welded connections
= 0.926 for H32 steel
= 0.885 for H36 steel
= 0.870 for H40 steel
kt = a factor for thickness effect, which is not applicable to longitudinal stiffeners
which are flat bars or bulb flats
n
 t 
=   for t ≥ 22 mm
 22 
= 1.0 for t < 22 mm
n = 0.25 for cruciform joints, transverse T joints and plates with
transverse attachments
= 0.20 for transverse butt welds
= 0.10 for butt welds ground flush, base metal, longitudinal welds or
attachments
If it can be conclusively established that the detail under consideration is always subject
to a mean stress of σms, D is to be adjusted by a factor kms.
kms = a factor for mean stress effect, which is
= 1.0 for σms > s4/2
= 0.85 + 0.3σms/s4 for –s4/2 ≤ σms ≤ s4/2
= 0.7 for σms < –s4/2

ABS GUIDE FOR SPECTRAL-BASED FATIGUE ANALYSIS FOR VESSELS . 2016 27

Appendix 2 Outline of a Closed-form Spectral-based Fatigue Analysis Procedure

σms = mean stress

s4 = long-term stress range corresponding to the representative probability level
of 10-4
m, A = physical parameters describing the S-N curve
T = design fatigue life, in seconds
f0i = zero-up-crossing frequency of the stress response, Hz
pi = joint probability of Hs and Tz
gi = probability density function governing s in the i-th sea state
s = specific value of stress range
Summing Di over all of the sea-states in the wave scatter diagram leads to the total cumulative
damage, D. Therefore:
∞
m f 0T  m  M 
D = (k h k t k ms )   s  ∫ ∑
 A  0  i =1
f 0i pi g i / f 0  ds ................................................................ (9)


where
D = total cumulative damage
f0 = “average” frequency of s over the design fatigue life

= Σi pif0i (where the summation is done from i = 1 to M, the number of

considered sea-states)
Introducing long-term probability density function, g(s) of the stress range as:

∑f pg 0i i i
g (s) = i
......................................................................................................... (10)
∑f p i
0i i

and
NT = total number of cycles in design fatigue life = f0T
the expression for total cumulative damage, D, can be rewritten as:
∞
D = (k h k t k ms )
NT
∫s
m m
g ( s )ds ..................................................................................... (11)
A 0

6. If the total number of cycles, NT, corresponds to a required minimum design fatigue life of 20
years, the calculated fatigue life would then be equal to 20/D. Increasing the design fatigue life to
higher values can be done accordingly. The fatigue safety check is to be done in accordance with
the applicable Rules where factors of safety (or Fatigue Design Factors) are specified.

5 Closed Form Damage Expression

The closed form expression of damage can be obtained using equation 11 and assuming an analytical
distribution for the long term probability density function of the stress range g(s) (e.g., Weibull probability
density function). For all one-segment linear S-N curves, the closed form expression of damage, D, as
given by equation 9, is as follows:
M

∑ λ(m, ε ) f
T
D= (2 2 ) m Γ(m/2 + 1) i 0i pi (k h k t k msσ i ) m ........................................................ (12)
A i =1

28 ABS GUIDE FOR SPECTRAL-BASED FATIGUE ANALYSIS FOR VESSELS . 2016

Appendix 2 Outline of a Closed-form Spectral-based Fatigue Analysis Procedure

where
Γ = complete gamma function with the argument (m/2 + 1)

σi = m0 for the i-th considered sea state

λ = rainflow factor of Wirsching and is defined as:

λ(m, εi) = a(m) + [1 – a(m)][1 – εi]b(m) .................................................................... (13)
where
a(m) = 0.926 – 0.033m
b(m) = 1.587m – 2.323
εi = Spectral Bandwidth (equation 6)
For bi-linear S-N curves (see Appendix 1) where the negative slope changes at point Q = (NQ, SQ) from m
to r = m + ∆m (∆m > 0) and the constant A changes to C, the expression for damage, as given in equation
12, is as follows:
M

∑ λ(m, ε )µ f
T
D= (2 2 ) m Γ(m/2 + 1) i i 0i pi (k h k t k msσ i ) m ................................................... (14)
A i =1

where µi is the endurance factor having its value between 0 and 1 and measuring the contribution of the
lower branch to the damage. It is defined as:

[ ]∫ s
SQ SQ
 A
∫ s g i ds −   (k h k t k ms )
m ∆m m + ∆m
g i ds
C 
µi = 1 − 0
∞
0
.................................................................... (15)

∫s
m
g i ds
0

If g(s) is a Rayleigh distribution, then µi is:

µi = 1 −
[ ]
Γ0 (m / 2 + 1, vi ) − (1 / vi ) ∆m / 2 (k h k t k ms )∆m Γ0 (r / 2 + 1, vi )
................................................ (16)
Γ(m / 2 + 1)
where
2
 SQ 
vi =  
 2 2σ 
 i 
Γ0 = incomplete gamma function and is
x

∫
Γ0 (a, x) = u a −1 exp(−u )du
0

See steps 2-6 above, regarding the fatigue safety check.

ABS GUIDE FOR SPECTRAL-BASED FATIGUE ANALYSIS FOR VESSELS . 2016 29